8 *(Released Fri, 31 Jan 2014)*
10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
14 default. Specify --no-wait-for-sync to override this behavior.
15 - The Ganeti python code now adheres to a private-module layout. In particular,
16 the module 'ganeti' is no longer in the python search path.
17 - On instance allocation, the iallocator now considers non-LVM storage
18 properly. In particular, actual file storage space information is used
19 when allocating space for a file/sharedfile instance.
20 - When disabling disk templates cluster-wide, the cluster now first
21 checks whether there are instances still using those templates.
22 - 'gnt-node list-storage' now also reports storage information about
23 file-based storage types.
24 - In case of non drbd instances, export \*_SECONDARY environment variables
25 as empty strings (and not "None") during 'instance-migrate' related hooks.
30 - KVM hypervisors can now access RBD storage directly without having to
31 go through a block device.
32 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
33 procedure between two Ganeti versions that are both 2.10 or higher.
34 - The move-instance command can now change disk templates when moving
35 instances, and does not require any node placement options to be
36 specified if the destination cluster has a default iallocator.
37 - Users can now change the soundhw and cpuid settings for XEN hypervisors.
38 - Hail and hbal now have the (optional) capability of accessing average CPU
39 load information through the monitoring deamon, and to use it to dynamically
40 adapt the allocation of instances.
41 - Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
42 so that disk and NIC modifications take effect without the need of actual
43 reboot. There are a couple of constrains currently for this feature:
45 - only KVM hypervisor (versions >= 1.0) supports it,
46 - one can not (yet) hotplug a disk using userspace access mode for RBD
47 - in case of a downgrade instances should suffer a reboot in order to
48 be migratable (due to core change of runtime files)
53 - A new test framework for logical units was introduced and the test
54 coverage for logical units was improved significantly.
55 - Opcodes are entirely generated from Haskell using the tool 'hs2py' and
56 the module 'src/Ganeti/OpCodes.hs'.
57 - Constants are also generated from Haskell using the tool
58 'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
59 exception of socket related constants, which require changing the
60 cluster configuration file, and HVS related constants, because they
61 are part of a port of instance queries to Haskell. As a result, these
62 changes will be part of the next release of Ganeti.
67 The following new dependencies have been added/updated.
71 - The version requirements for ``python-mock`` have increased to at least
72 version 1.0.1. It is still used for testing only.
77 - Documentation improvements
78 - Run drbdsetup syncer only on network attach
79 - Include target node in hooks nodes for migration
81 - Support post-upgrade hooks during cluster upgrades
83 Inherited from the 2.9 branch:
85 - Ensure that all the hypervisors exist in the config file (Issue 640)
86 - Correctly recognise the role as master node (Issue 687)
87 - configure: allow detection of Sphinx 1.2+ (Issue 502)
88 - gnt-instance now honors the KVM path correctly (Issue 691)
90 Inherited from the 2.8 branch:
92 - Change the list separator for the usb_devices parameter from comma to space.
93 Commas could not work because they are already the hypervisor option
95 - Add support for blktap2 file-driver (Issue 638)
96 - Add network tag definitions to the haskell codebase (Issue 641)
97 - Fix RAPI network tag handling
98 - Add the network tags to the tags searched by gnt-cluster search-tags
99 - Fix caching bug preventing jobs from being cancelled
100 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
105 - All known issues in 2.10.0 beta1 have been resolved (see changes from
107 - Improve handling of KVM runtime files from earlier Ganeti versions
108 - Documentation fixes
110 Inherited from the 2.9 branch:
112 - use custom KVM path if set for version checking
113 - SingleNotifyPipeCondition: don't share pollers
115 Inherited from the 2.8 branch:
117 - Fixed Luxi daemon socket permissions after master-failover
118 - Improve IP version detection code directly checking for colons rather than
119 passing the family from the cluster object
120 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
121 opportunistically anymore (Issue 622)
122 - Allow link local IPv6 gateways (Issue 624)
123 - Fix error printing (Issue 616)
124 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
125 disk modifications, keep the old one. If name=none then set disk name to
127 - Update build_chroot script to work with the latest hackage packages
128 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
129 - Fix evacuation out of drained node (Issue 615)
130 - Add default file_driver if missing (Issue 571)
131 - Fix job error message after unclean master shutdown (Issue 618)
132 - Lock group(s) when creating instances (Issue 621)
133 - SetDiskID() before accepting an instance (Issue 633)
134 - Allow the ext template disks to receive arbitrary parameters, both at creation
135 time and while being modified
136 - Xen handle domain shutdown (future proofing cherry-pick)
137 - Refactor reading live data in htools (future proofing cherry-pick)
143 *(Released Tue, 17 Dec 2013)*
145 This was the first RC release of the 2.10 series. All important changes
146 are listed in the latest 2.10 entry.
152 *(Released Wed, 27 Nov 2013)*
154 This was the first beta release of the 2.10 series. All important changes
155 are listed in the latest 2.10 entry.
160 The following issues are known to be present in the beta and will be fixed
163 - Issue 477: Wrong permissions for confd LUXI socket
164 - Issue 621: Instance related opcodes do not aquire network/group locks
165 - Issue 622: Assertion Error: Node locks differ from node resource locks
166 - Issue 623: IPv6 Masterd <-> Luxid communication error
172 *(Released Mon, 10 Feb 2014)*
174 - Fix the RAPI instances-multi-alloc call
175 - assign unique filenames to file-based disks
176 - gracefully handle degraded non-diskless instances with 0 disks (issue 697)
177 - noded now runs with its specified group, which is the default group,
178 defaulting to root (issue 707)
179 - make using UUIDs to identify nodes in gnt-node consistently possible
186 *(Released Mon, 27 Jan 2014)*
188 - Ensure that all the hypervisors exist in the config file (Issue 640)
189 - Correctly recognise the role as master node (Issue 687)
190 - configure: allow detection of Sphinx 1.2+ (Issue 502)
191 - gnt-instance now honors the KVM path correctly (Issue 691)
193 Inherited from the 2.8 branch:
195 - Change the list separator for the usb_devices parameter from comma to space.
196 Commas could not work because they are already the hypervisor option
197 separator (Issue 649)
198 - Add support for blktap2 file-driver (Issue 638)
199 - Add network tag definitions to the haskell codebase (Issue 641)
200 - Fix RAPI network tag handling
201 - Add the network tags to the tags searched by gnt-cluster search-tags
202 - Fix caching bug preventing jobs from being cancelled
203 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
209 *(Released Fri, 13 Dec 2013)*
211 - use custom KVM path if set for version checking
212 - SingleNotifyPipeCondition: don't share pollers
214 Inherited from the 2.8 branch:
216 - Fixed Luxi daemon socket permissions after master-failover
217 - Improve IP version detection code directly checking for colons rather than
218 passing the family from the cluster object
219 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
220 opportunistically anymore (Issue 622)
221 - Allow link local IPv6 gateways (Issue 624)
222 - Fix error printing (Issue 616)
223 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
224 disk modifications, keep the old one. If name=none then set disk name to
226 - Update build_chroot script to work with the latest hackage packages
227 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
228 - Fix evacuation out of drained node (Issue 615)
229 - Add default file_driver if missing (Issue 571)
230 - Fix job error message after unclean master shutdown (Issue 618)
231 - Lock group(s) when creating instances (Issue 621)
232 - SetDiskID() before accepting an instance (Issue 633)
233 - Allow the ext template disks to receive arbitrary parameters, both at creation
234 time and while being modified
235 - Xen handle domain shutdown (future proofing cherry-pick)
236 - Refactor reading live data in htools (future proofing cherry-pick)
242 *(Released Wed, 13 Nov 2013)*
244 - fix bug, that kept nodes offline when readding
245 - when verifying DRBD versions, ignore unavailable nodes
246 - fix bug that made the console unavailable on kvm in split-user
248 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
254 *(Released Tue, 5 Nov 2013)*
256 Incompatible/important changes
257 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
259 - hroller now also plans for capacity to move non-redundant instances off
260 any node to be rebooted; the old behavior of completely ignoring any
261 non-redundant instances can be restored by adding the --ignore-non-redundant
263 - The cluster option '--no-lvm-storage' was removed in favor of the new option
264 '--enabled-disk-templates'.
265 - On instance creation, disk templates no longer need to be specified
266 with '-t'. The default disk template will be taken from the list of
267 enabled disk templates.
268 - The monitoring daemon is now running as root, in order to be able to collect
269 information only available to root (such as the state of Xen instances).
270 - The ConfD client is now IPv6 compatible.
271 - File and shared file storage is no longer dis/enabled at configure time,
272 but using the option '--enabled-disk-templates' at cluster initialization and
274 - The default directories for file and shared file storage are not anymore
275 specified at configure time, but taken from the cluster's configuration.
276 They can be set at cluster initialization and modification with
277 '--file-storage-dir' and '--shared-file-storage-dir'.
278 - Cluster verification now includes stricter checks regarding the
279 default file and shared file storage directories. It now checks that
280 the directories are explicitely allowed in the 'file-storage-paths' file and
281 that the directories exist on all nodes.
282 - The list of allowed disk templates in the instance policy and the list
283 of cluster-wide enabled disk templates is now checked for consistency
284 on cluster or group modification. On cluster initialization, the ipolicy
285 disk templates are ensured to be a subset of the cluster-wide enabled
291 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
292 the correct command syntax. It is possible to use different DRBD versions
293 on different nodes as long as they are compatible to each other. This
294 enables rolling upgrades of DRBD with no downtime. As permanent operation
295 of different DRBD versions within a node group is discouraged,
296 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
297 - New "inst-status-xen" data collector for the monitoring daemon, providing
298 information about the state of the xen instances on the nodes.
299 - New "lv" data collector for the monitoring daemon, collecting data about the
300 logical volumes on the nodes, and pairing them with the name of the instances
302 - New "diskstats" data collector, collecting the data from /proc/diskstats and
303 presenting them over the monitoring daemon interface.
304 - The ConfD client is now IPv6 compatible.
308 The following new dependencies have been added.
312 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
313 for the unit tests (and only used for testing).
317 - ``hslogger`` (http://software.complete.org/hslogger) is now always
318 required, even if confd is not enabled.
323 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
325 - Improved error messsages (inherited from stable-2.8)
331 *(Released Tue, 15 Oct 2013)*
333 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
335 - in implicit configuration upgrade, match ipolicy with enabled disk templates
336 - improved harep documentation (inherited from stable-2.8)
342 *(Released Wed, 9 Oct 2013)*
344 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
346 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
347 at least one DRBD instance.
348 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
355 *(Released Tue, 1 Oct 2013)*
357 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
360 - update of the documentation, in particular installation instructions
361 - merging of LD_* constants into DT_* constants
362 - python style changes to be compatible with newer versions of pylint
368 *(Released Thu, 29 Aug 2013)*
370 This was the first beta release of the 2.9 series. All important changes
371 are listed in the latest 2.9 entry.
377 *(Released Thu, 23 Jan 2014)*
379 - Change the list separator for the usb_devices parameter from comma to space.
380 Commas could not work because they are already the hypervisor option
381 separator (Issue 649)
382 - Add support for blktap2 file-driver (Issue 638)
383 - Add network tag definitions to the haskell codebase (Issue 641)
384 - Fix RAPI network tag handling
385 - Add the network tags to the tags searched by gnt-cluster search-tags
386 - Fix caching bug preventing jobs from being cancelled
387 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
393 *(Released Thu, 12 Dec 2013)*
395 - Fixed Luxi daemon socket permissions after master-failover
396 - Improve IP version detection code directly checking for colons rather than
397 passing the family from the cluster object
398 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
399 opportunistically anymore (Issue 622)
400 - Allow link local IPv6 gateways (Issue 624)
401 - Fix error printing (Issue 616)
402 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
403 disk modifications, keep the old one. If name=none then set disk name to
405 - Update build_chroot script to work with the latest hackage packages
406 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
407 - Fix evacuation out of drained node (Issue 615)
408 - Add default file_driver if missing (Issue 571)
409 - Fix job error message after unclean master shutdown (Issue 618)
410 - Lock group(s) when creating instances (Issue 621)
411 - SetDiskID() before accepting an instance (Issue 633)
412 - Allow the ext template disks to receive arbitrary parameters, both at creation
413 time and while being modified
414 - Xen handle domain shutdown (future proofing cherry-pick)
415 - Refactor reading live data in htools (future proofing cherry-pick)
421 *(Released Thu, 07 Nov 2013)*
423 - DRBD: ensure peers are UpToDate for dual-primary
424 - Improve error message for replace-disks
425 - More dependency checks at configure time
426 - Placate warnings on ganeti.outils_unittest.py
432 *(Released Thu, 17 Oct 2013)*
434 - Correctly start/stop luxid during gnt-cluster master-failover
435 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
436 - Fix path for the job queue serial file
437 - Improved harep man page
438 - Minor documentation improvements
444 *(Released Mon, 30 Sep 2013)*
446 Incompatible/important changes
447 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
449 - Instance policy can contain multiple instance specs, as described in
450 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
451 <design-partitioned>`. As a consequence, it's not possible to partially change
452 or override instance specs. Bounding specs (min and max) can be specified as a
453 whole using the new option ``--ipolicy-bounds-specs``, while standard
454 specs use the new option ``--ipolicy-std-specs``.
455 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
456 gnt-instance is a valid YAML object.
457 - hail now honors network restrictions when allocating nodes. This led to an
458 update of the IAllocator protocol. See the IAllocator documentation for
460 - confd now only answers static configuration request over the network. luxid
461 was extracted, listens on the local LUXI socket and responds to live queries.
462 This allows finer grained permissions if using separate users.
467 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
468 to always require authentication, ``--require-authentication``. It can
469 be specified in ``$sysconfdir/default/ganeti``.
470 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
471 be used to manage the disk templates to be used by instances in the cluster.
472 Initially, it will be set to a list that includes plain, drbd, if they were
473 enabled by specifying a volume group name, and file and sharedfile, if those
474 were enabled at configure time. Additionally, it will include all disk
475 templates that are currently used by instances. The order of disk templates
476 will be based on Ganeti's history of supporting them. In the future, the
477 first entry of the list will be used as a default disk template on instance
479 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
480 configuration back to the previous stable version.
481 - Disk templates in group ipolicy can be restored to the default value.
482 - Initial support for diskless instances and virtual clusters in QA.
483 - More QA and unit tests for instance policies.
484 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
485 describing why the opcode itself was executed.
486 - The monitoring daemon is now available. It allows users to query the cluster
487 for obtaining information about the status of the system. The daemon is only
488 responsible for providing the information over the network: the actual data
489 gathering is performed by data collectors (currently, only the DRBD status
490 collector is available).
491 - In order to help developers work on Ganeti, a new script
492 (``devel/build_chroot``) is provided, for building a chroot that contains all
493 the required development libraries and tools for compiling Ganeti on a Debian
495 - A new tool, ``harep``, for performing self-repair and recreation of instances
496 in Ganeti has been added.
497 - Split queries are enabled for tags, network, exports, cluster info, groups,
499 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
500 It prints the command line to set the current policies, to ease
502 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
503 devices for KVM virtio-net interfaces will get created with VNET_HDR
504 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
505 virtio-net interfaces, which prevents host kernel tainting and log
506 flooding, when dealing with broken or malicious virtio-net drivers.
507 It's set to true by default.
508 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
510 - Support 'viridian' parameter in Xen HVM
511 - Support DSA SSH keys in bootstrap
512 - To simplify the work of packaging frameworks that want to add the needed users
513 and groups in a split-user setup themselves, at build time three files in
514 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
515 the groups to be generated, the ``users`` file contains, one per line, the
516 users to be generated, optionally followed by their primary group, where
517 important. The ``groupmemberships`` file contains, one per line, additional
518 user-group membership relations that need to be established. The syntax of
519 these files will remain stable in all future versions.
524 The following new dependencies have been added:
527 - The ``curl`` library is not optional anymore for compiling the Haskell code.
528 - ``snap-server`` library (if monitoring is enabled).
531 - The minimum Python version needed to run Ganeti is now 2.6.
532 - ``yaml`` library (only for running the QA).
536 - Perform proper cleanup on termination of Haskell daemons
537 - Fix corner-case in handling of remaining retry time
543 *(Released Tue, 17 Sep 2013)*
545 - To simplify the work of packaging frameworks that want to add the needed users
546 and groups in a split-user setup themselves, at build time three files in
547 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
548 the groups to be generated, the ``users`` file contains, one per line, the
549 users to be generated, optionally followed by their primary group, where
550 important. The ``groupmemberships`` file contains, one per line, additional
551 user-group membership relations that need to be established. The syntax of
552 these files will remain stable in all future versions.
553 - Add a default to file-driver when unspecified over RAPI (Issue 571)
554 - Mark the DSA host pubkey as optional, and remove it during config downgrade
556 - Some documentation fixes
562 *(Released Tue, 27 Aug 2013)*
564 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
566 - Support 'viridian' parameter in Xen HVM (Issue 233)
567 - Include VCS version in ``gnt-cluster version``
568 - Support DSA SSH keys in bootstrap (Issue 338)
569 - Fix batch creation of instances
570 - Use FQDN to check master node status (Issue 551)
571 - Make the DRBD collector more failure-resilient
577 *(Released Fri, 2 Aug 2013)*
579 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
581 - Fix upgrading/downgrading from 2.7
582 - Increase maximum RAPI message size
583 - Documentation updates
584 - Split ``confd`` between ``luxid`` and ``confd``
585 - Merge 2.7 series up to the 2.7.1 release
586 - Allow the ``modify_etc_hosts`` option to be changed
587 - Add better debugging for ``luxid`` queries
588 - Expose bulk parameter for GetJobs in RAPI client
589 - Expose missing ``network`` fields in RAPI
590 - Add some ``cluster verify`` tests
591 - Some unittest fixes
592 - Fix a malfunction in ``hspace``'s tiered allocation
593 - Fix query compatibility between haskell and python implementations
594 - Add the ``vnet_hdr`` HV parameter for KVM
595 - Add ``--cleanup`` to instance failover
596 - Change the connected groups format in ``gnt-network info`` output; it
597 was previously displayed as a raw list by mistake. (Merged from 2.7)
603 *(Released Mon, 24 Jun 2013)*
605 This was the first beta release of the 2.8 series. All important changes
606 are listed in the latest 2.8 entry.
612 *(Released Thu, 26 Sep 2013)*
614 - Change the connected groups format in ``gnt-network info`` output; it
615 was previously displayed as a raw list by mistake
616 - Check disk template in right dict when copying
617 - Support multi-instance allocs without iallocator
618 - Fix some errors in the documentation
619 - Fix formatting of tuple in an error message
625 *(Released Thu, 25 Jul 2013)*
627 - Add logrotate functionality in daemon-util
628 - Add logrotate example file
629 - Add missing fields to network queries over rapi
630 - Fix network object timestamps
631 - Add support for querying network timestamps
632 - Fix a typo in the example crontab
633 - Fix a documentation typo
639 *(Released Thu, 04 Jul 2013)*
641 Incompatible/important changes
642 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
644 - Instance policies for disk size were documented to be on a per-disk
645 basis, but hail applied them to the sum of all disks. This has been
647 - ``hbal`` will now exit with status 0 if, during job execution over
648 LUXI, early exit has been requested and all jobs are successful;
649 before, exit status 1 was used, which cannot be differentiated from
651 - Compatibility with newer versions of rbd has been fixed
652 - ``gnt-instance batch-create`` has been changed to use the bulk create
653 opcode from Ganeti. This lead to incompatible changes in the format of
654 the JSON file. It's now not a custom dict anymore but a dict
655 compatible with the ``OpInstanceCreate`` opcode.
656 - Parent directories for file storage need to be listed in
657 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
658 write the file automatically based on old configuration values, but it
659 can not distribute it across all nodes and the file contents should be
660 verified. Use ``gnt-cluster copyfile
661 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
662 upgraded. The reason for requiring this list of paths now is that
663 before it would have been possible to inject new paths via RPC,
664 allowing files to be created in arbitrary locations. The RPC protocol
665 is protected using SSL/X.509 certificates, but as a design principle
666 Ganeti does not permit arbitrary paths to be passed.
667 - The parsing of the variants file for OSes (see
668 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
669 lines and comment lines (starting with ``#``) are ignored for better
671 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
672 longer available. ``gnt-node add`` now invokes a new tool on the
673 destination node, named ``prepare-node-join``, to configure the SSH
674 daemon. Paramiko is no longer necessary to configure nodes' SSH
675 daemons via ``gnt-node add``.
676 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
677 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
678 the flag is not reset after a master failover.
679 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
680 is recommended. 2.6 will be mandatory from the 2.8 series.
686 - New network management functionality to support automatic allocation
687 of IP addresses and managing of network parameters. See
688 :manpage:`gnt-network(8)` for more details.
689 - New external storage backend, to allow managing arbitrary storage
690 systems external to the cluster. See
691 :manpage:`ganeti-extstorage-interface(7)`.
692 - New ``exclusive-storage`` node parameter added, restricted to
693 nodegroup level. When it's set to true, physical disks are assigned in
694 an exclusive fashion to instances, as documented in :doc:`Partitioned
695 Ganeti <design-partitioned>`. Currently, only instances using the
696 ``plain`` disk template are supported.
697 - The KVM hypervisor has been updated with many new hypervisor
698 parameters, including a generic one for passing arbitrary command line
699 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
700 compatible up to qemu 1.4.
701 - A new tool, called ``mon-collector``, is the stand-alone executor of
702 the data collectors for a monitoring system. As of this version, it
703 just includes the DRBD data collector, that can be executed by calling
704 ``mon-collector`` using the ``drbd`` parameter. See
705 :manpage:`mon-collector(7)`.
706 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
707 for RAPI users. It allows granting permissions to query for
708 information to a specific user without giving
709 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
710 - A new tool named ``node-cleanup`` has been added. It cleans remains of
711 a cluster from a machine by stopping all daemons, removing
712 certificates and ssconf files. Unless the ``--no-backup`` option is
713 given, copies of the certificates are made.
714 - Instance creations now support the use of opportunistic locking,
715 potentially speeding up the (parallel) creation of multiple instances.
716 This feature is currently only available via the :doc:`RAPI
717 <rapi>` interface and when an instance allocator is used. If the
718 ``opportunistic_locking`` parameter is set the opcode will try to
719 acquire as many locks as possible, but will not wait for any locks
720 held by other opcodes. If not enough resources can be found to
721 allocate the instance, the temporary error code
722 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
723 retried thereafter, with or without opportunistic locking.
724 - New experimental linux-ha resource scripts.
725 - Restricted-commands support: ganeti can now be asked (via command line
726 or rapi) to perform commands on a node. These are passed via ganeti
727 RPC rather than ssh. This functionality is restricted to commands
728 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
729 reasons. The file is not copied automatically.
735 - Diskless instances are now externally mirrored (Issue 237). This for
736 now has only been tested in conjunction with explicit target nodes for
738 - Queries not needing locks or RPC access to the node can now be
739 performed by the confd daemon, making them independent from jobs, and
740 thus faster to execute. This is selectable at configure time.
741 - The functionality for allocating multiple instances at once has been
742 overhauled and is now also available through :doc:`RAPI <rapi>`.
744 There are no significant changes from version 2.7.0~rc3.
750 *(Released Tue, 25 Jun 2013)*
752 - Fix permissions on the confd query socket (Issue 477)
753 - Fix permissions on the job archive dir (Issue 498)
754 - Fix handling of an internal exception in replace-disks (Issue 472)
755 - Fix gnt-node info handling of shortened names (Issue 497)
756 - Fix gnt-instance grow-disk when wiping is enabled
757 - Documentation improvements, and support for newer pandoc
758 - Fix hspace honoring ipolicy for disks (Issue 484)
759 - Improve handling of the ``kvm_extra`` HV parameter
765 *(Released Fri, 24 May 2013)*
767 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
769 - Disks added through ``gnt-instance modify`` or created through
770 ``gnt-instance recreate-disks`` are wiped, if the
771 ``prealloc_wipe_disks`` flag is set.
772 - If wiping newly created disks fails, the disks are removed. Also,
773 partial failures in creating disks through ``gnt-instance modify``
774 triggers a cleanup of the partially-created disks.
775 - Removing the master IP address doesn't fail if the address has been
777 - Fix ownership of the OS log dir
778 - Workaround missing SO_PEERCRED constant (Issue 191)
784 *(Released Fri, 3 May 2013)*
786 This was the first release candidate of the 2.7 series. Since beta3:
788 - Fix kvm compatibility with qemu 1.4 (Issue 389)
789 - Documentation updates (admin guide, upgrade notes, install
790 instructions) (Issue 372)
791 - Fix gnt-group list nodes and instances count (Issue 436)
792 - Fix compilation without non-mandatory libraries (Issue 441)
793 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
794 - Make confd logging more verbose at INFO level (Issue 435)
795 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
796 - Fix failure path for instance storage type conversion (Issue 229)
797 - Update htools text backend documentation
798 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
799 - Disable inter-cluster instance move for file-based instances, because
800 it is dependant on instance export, which is not supported for
801 file-based instances. (Issue 414)
802 - Fix gnt-job crashes on non-ascii characters (Issue 427)
803 - Fix volume group checks on non-vm-capable nodes (Issue 432)
809 *(Released Mon, 22 Apr 2013)*
811 This was the third beta release of the 2.7 series. Since beta2:
813 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
814 - Fix data loss on wrong usage of ``gnt-instance move``
815 - Properly export errors in confd-based job queries
816 - Add ``users-setup`` tool
817 - Fix iallocator protocol to report 0 as a disk size for diskless
818 instances. This avoids hail breaking when a diskless instance is
820 - Fix job queue directory permission problem that made confd job queries
821 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
822 for access to archived jobs (Issue 406).
823 - Limit the sizes of networks supported by ``gnt-network`` to something
824 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
825 - Fix bugs in instance disk template conversion
826 - Fix GHC 7 compatibility
827 - Fix ``burnin`` install path (Issue 426).
828 - Allow very small disk grows (Issue 347).
829 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
830 that noded doesn't import masterd code (Issue 419).
831 - Make sure the default metavg at cluster init is the same as the vg, if
832 unspecified (Issue 358).
833 - Fix cleanup of partially created disks (part of Issue 416)
839 *(Released Tue, 2 Apr 2013)*
841 This was the second beta release of the 2.7 series. Since beta1:
843 - Networks no longer have a "type" slot, since this information was
844 unused in Ganeti: instead of it tags should be used.
845 - The rapi client now has a ``target_node`` option to MigrateInstance.
846 - Fix early exit return code for hbal (Issue 386).
847 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
848 - Fix ``rbd showmapped`` output parsing (Issue 312).
849 - Networks are now referenced indexed by UUID, rather than name. This
850 will require running cfgupgrade, from 2.7.0beta1, if networks are in
852 - The OS environment now includes network information.
853 - Deleting of a network is now disallowed if any instance nic is using
854 it, to prevent dangling references.
855 - External storage is now documented in man pages.
856 - The exclusive_storage flag can now only be set at nodegroup level.
857 - Hbal can now submit an explicit priority with its jobs.
858 - Many network related locking fixes.
859 - Bump up the required pylint version to 0.25.1.
860 - Fix the ``no_remember`` option in RAPI client.
861 - Many ipolicy related tests, qa, and fixes.
862 - Many documentation improvements and fixes.
863 - Fix building with ``--disable-file-storage``.
864 - Fix ``-q`` option in htools, which was broken if passed more than
866 - Some haskell/python interaction improvements and fixes.
867 - Fix iallocator in case of missing LVM storage.
868 - Fix confd config load in case of ``--no-lvm-storage``.
869 - The confd/query functionality is now mentioned in the security
876 *(Released Wed, 6 Feb 2013)*
878 This was the first beta release of the 2.7 series. All important changes
879 are listed in the latest 2.7 entry.
885 *(Released Fri, 21 Dec 2012)*
887 Important behaviour change: hbal won't rebalance anymore instances which
888 have the ``auto_balance`` attribute set to false. This was the intention
889 all along, but until now it only skipped those from the N+1 memory
890 reservation (DRBD-specific).
892 A significant number of bug fixes in this release:
894 - Fixed disk adoption interaction with ipolicy checks.
895 - Fixed networking issues when instances are started, stopped or
896 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
897 - Fixed the warning in cluster verify for shared storage instances not
899 - Fixed removal of storage directory on shared file storage (issue 262).
900 - Fixed validation of LVM volume group name in OpClusterSetParams
901 (``gnt-cluster modify``) (issue 285).
902 - Fixed runtime memory increases (``gnt-instance modify -m``).
903 - Fixed live migration under Xen's ``xl`` mode.
904 - Fixed ``gnt-instance console`` with ``xl``.
905 - Fixed building with newer Haskell compiler/libraries.
906 - Fixed PID file writing in Haskell daemons (confd); this prevents
907 restart issues if confd was launched manually (outside of
908 ``daemon-util``) while another copy of it was running
909 - Fixed a type error when doing live migrations with KVM (issue 297) and
910 the error messages for failing migrations have been improved.
911 - Fixed opcode validation for the out-of-band commands (``gnt-node
913 - Fixed a type error when unsetting OS hypervisor parameters (issue
914 311); now it's possible to unset all OS-specific hypervisor
916 - Fixed the ``dry-run`` mode for many operations: verification of
917 results was over-zealous but didn't take into account the ``dry-run``
918 operation, resulting in "wrong" failures.
919 - Fixed bash completion in ``gnt-job list`` when the job queue has
920 hundreds of entries; especially with older ``bash`` versions, this
921 results in significant CPU usage.
923 And lastly, a few other improvements have been made:
925 - Added option to force master-failover without voting (issue 282).
926 - Clarified error message on lock conflict (issue 287).
927 - Logging of newly submitted jobs has been improved (issue 290).
928 - Hostname checks have been made uniform between instance rename and
930 - The ``--submit`` option is now supported by ``gnt-debug delay``.
931 - Shutting down the master daemon by sending SIGTERM now stops it from
932 processing jobs waiting for locks; instead, those jobs will be started
933 once again after the master daemon is started the next time (issue
935 - Support for Xen's ``xl`` program has been improved (besides the fixes
937 - Reduced logging noise in the Haskell confd daemon (only show one log
938 entry for each config reload, instead of two).
939 - Several man page updates and typo fixes.
945 *(Released Fri, 12 Oct 2012)*
947 A small bugfix release. Among the bugs fixed:
949 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
950 made the command unusable.
951 - Commands that issue many jobs don't fail anymore just because some jobs
952 take so long that other jobs are archived.
953 - Failures during ``gnt-instance reinstall`` are reflected by the exit
955 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
957 - When ``always_failover`` is set, ``--allow-failover`` is not required
958 in migrate commands anymore.
959 - ``bash_completion`` works even if extglob is disabled.
960 - Fixed bug with locks that made failover for RDB-based instances fail.
961 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
962 a random node instead of one based on the allocator metric.
963 - Support for newer versions of pylint and pep8.
964 - Hail doesn't fail anymore when trying to add an instance of type
965 ``file``, ``sharedfile`` or ``rbd``.
966 - Added new Makefile target to rebuild the whole distribution, so that
967 all files are included.
973 *(Released Fri, 27 Jul 2012)*
976 .. attention:: The ``LUXI`` protocol has been made more consistent
977 regarding its handling of command arguments. This, however, leads to
978 incompatibility issues with previous versions. Please ensure that you
979 restart Ganeti daemons soon after the upgrade, otherwise most
980 ``LUXI`` calls (job submission, setting/resetting the drain flag,
981 pausing/resuming the watcher, cancelling and archiving jobs, querying
982 the cluster configuration) will fail.
991 The current ``admin_up`` field, which used to denote whether an instance
992 should be running or not, has been removed. Instead, ``admin_state`` is
993 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
995 The rational behind this is that an instance being “down” can have
998 - it could be down during a reboot
999 - it could be temporarily be down for a reinstall
1000 - or it could be down because it is deprecated and kept just for its
1003 The previous Boolean state was making it difficult to do capacity
1004 calculations: should Ganeti reserve memory for a down instance? Now, the
1005 tri-state field makes it clear:
1007 - in ``up`` and ``down`` state, all resources are reserved for the
1008 instance, and it can be at any time brought up if it is down
1009 - in ``offline`` state, only disk space is reserved for it, but not
1012 The field can have an extra use: since the transition between ``up`` and
1013 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
1014 transition between ``offline`` and ``down`` is done via ``gnt-instance
1015 modify``, it is possible to given different rights to users. For
1016 example, owners of an instance could be allowed to start/stop it, but
1017 not transition it out of the offline state.
1019 Instance policies and specs
1020 +++++++++++++++++++++++++++
1022 In previous Ganeti versions, an instance creation request was not
1023 limited on the minimum size and on the maximum size just by the cluster
1024 resources. As such, any policy could be implemented only in third-party
1025 clients (RAPI clients, or shell wrappers over ``gnt-*``
1026 tools). Furthermore, calculating cluster capacity via ``hspace`` again
1027 required external input with regards to instance sizes.
1029 In order to improve these workflows and to allow for example better
1030 per-node group differentiation, we introduced instance specs, which
1033 - minimum instance disk size, disk count, memory size, cpu count
1034 - maximum values for the above metrics
1035 - and “standard” values (used in ``hspace`` to calculate the standard
1038 The minimum/maximum values can be also customised at node-group level,
1039 for example allowing more powerful hardware to support bigger instance
1042 Beside the instance specs, there are a few other settings belonging to
1043 the instance policy framework. It is possible now to customise, per
1044 cluster and node-group:
1046 - the list of allowed disk templates
1047 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
1048 - the maximum ratio of instance to spindles (see below for more
1049 information) for local storage
1051 All these together should allow all tools that talk to Ganeti to know
1052 what are the ranges of allowed values for instances and the
1053 over-subscription that is allowed.
1055 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
1056 instance configuration, and the physical CPU configuration from the
1057 node. For the spindle ratios however, we didn't track before these
1058 values, so new parameters have been added:
1060 - a new node parameter ``spindle_count``, defaults to 1, customisable at
1061 node group or node level
1062 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
1064 Note that spindles in this context doesn't need to mean actual
1065 mechanical hard-drives; it's just a relative number for both the node
1066 I/O capacity and instance I/O consumption.
1068 Instance migration behaviour
1069 ++++++++++++++++++++++++++++
1071 While live-migration is in general desirable over failover, it is
1072 possible that for some workloads it is actually worse, due to the
1073 variable time of the “suspend” phase during live migration.
1075 To allow the tools to work consistently over such instances (without
1076 having to hard-code instance names), a new backend parameter
1077 ``always_failover`` has been added to control the migration/failover
1078 behaviour. When set to True, all migration requests for an instance will
1079 instead fall-back to failover.
1081 Instance memory ballooning
1082 ++++++++++++++++++++++++++
1084 Initial support for memory ballooning has been added. The memory for an
1085 instance is no longer fixed (backend parameter ``memory``), but instead
1086 can vary between minimum and maximum values (backend parameters
1087 ``minmem`` and ``maxmem``). Currently we only change an instance's
1090 - live migrating or failing over and instance and the target node
1091 doesn't have enough memory
1092 - user requests changing the memory via ``gnt-instance modify
1095 Instance CPU pinning
1096 ++++++++++++++++++++
1098 In order to control the use of specific CPUs by instance, support for
1099 controlling CPU pinning has been added for the Xen, HVM and LXC
1100 hypervisors. This is controlled by a new hypervisor parameter
1101 ``cpu_mask``; details about possible values for this are in the
1102 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
1103 VCPU-to-CPU mapping) form will work well only when all nodes in your
1104 cluster have the same amount of CPUs.
1109 Another area in which Ganeti was not customisable were the parameters
1110 used for storage configuration, e.g. how many stripes to use for LVM,
1111 DRBD resync configuration, etc.
1113 To improve this area, we've added disks parameters, which are
1114 customisable at cluster and node group level, and which allow to
1115 specify various parameters for disks (DRBD has the most parameters
1116 currently), for example:
1118 - DRBD resync algorithm and parameters (e.g. speed)
1119 - the default VG for meta-data volumes for DRBD
1120 - number of stripes for LVM (plain disk template)
1123 These parameters can be modified via ``gnt-cluster modify -D …`` and
1124 ``gnt-group modify -D …``, and are used at either instance creation (in
1125 case of LVM stripes, for example) or at disk “activation” time
1126 (e.g. resync speed).
1128 Rados block device support
1129 ++++++++++++++++++++++++++
1131 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
1132 denoted by the ``rbd`` disk template type. This is considered
1133 experimental, feedback is welcome. For details on configuring it, see
1134 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
1139 The existing master IP functionality works well only in simple setups (a
1140 single network shared by all nodes); however, if nodes belong to
1141 different networks, then the ``/32`` setup and lack of routing
1142 information is not enough.
1144 To allow the master IP to function well in more complex cases, the
1145 system was reworked as follows:
1147 - a master IP netmask setting has been added
1148 - the master IP activation/turn-down code was moved from the node daemon
1149 to a separate script
1150 - whether to run the Ganeti-supplied master IP script or a user-supplied
1151 on is a ``gnt-cluster init`` setting
1153 Details about the location of the standard and custom setup scripts are
1154 in the man page :manpage:`gnt-cluster(8)`; for information about the
1155 setup script protocol, look at the Ganeti-supplied script.
1160 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
1163 It is now possible to use TLS-protected connections, and when renewing
1164 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
1165 it is now possible to specify spice or spice CA certificates. Also, it
1166 is possible to configure a password for SPICE sessions via the
1167 hypervisor parameter ``spice_password_file``.
1169 There are also new parameters to control the compression and streaming
1170 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1171 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1172 for the spice parameters.
1174 Lastly, it is now possible to see the SPICE connection information via
1175 ``gnt-instance console``.
1180 A new tool (``tools/ovfconverter``) has been added that supports
1181 conversion between Ganeti and the `Open Virtualization Format
1182 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1185 This relies on the ``qemu-img`` tool to convert the disk formats, so the
1186 actual compatibility with other virtualization solutions depends on it.
1188 Confd daemon changes
1189 ++++++++++++++++++++
1191 The configuration query daemon (``ganeti-confd``) is now optional, and
1192 has been rewritten in Haskell; whether to use the daemon at all, use the
1193 Python (default) or the Haskell version is selectable at configure time
1194 via the ``--enable-confd`` parameter, which can take one of the
1195 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1196 daemon will result in a smaller footprint; for larger systems, we
1197 welcome feedback on the Haskell version which might become the default
1200 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1201 daemon running. The Python version doesn't implement the new call.
1204 User interface changes
1205 ~~~~~~~~~~~~~~~~~~~~~~
1207 We have replaced the ``--disks`` option of ``gnt-instance
1208 replace-disks`` with a more flexible ``--disk`` option, which allows
1209 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1210 disk size and mode can be changed upon recreation (via ``gnt-instance
1211 recreate-disks``, which accepts the same ``--disk`` option).
1213 As many people are used to a ``show`` command, we have added that as an
1214 alias to ``info`` on all ``gnt-*`` commands.
1216 The ``gnt-instance grow-disk`` command has a new mode in which it can
1217 accept the target size of the disk, instead of the delta; this can be
1218 more safe since two runs in absolute mode will be idempotent, and
1219 sometimes it's also easier to specify the desired size directly.
1221 Also the handling of instances with regard to offline secondaries has
1222 been improved. Instance operations should not fail because one of it's
1223 secondary nodes is offline, even though it's safe to proceed.
1225 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1226 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1227 minors to instance name.
1232 RAPI coverage has improved, with (for example) new resources for
1233 recreate-disks, node power-cycle, etc.
1238 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1241 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1242 whether to still support Python 2.4 or move to Python 2.6 as minimum
1245 Support for Fedora has been slightly improved; the provided example
1246 init.d script should work better on it and the INSTALL file should
1247 document the needed dependencies.
1252 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1253 ``Query(what=QR_LOCK, ...)`` instead.
1255 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1256 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1257 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1258 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1259 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1261 RAPI client: ``CertificateError`` now derives from
1262 ``GanetiApiError``. This should make it more easy to handle Ganeti
1265 Deprecation warnings due to PyCrypto/paramiko import in
1266 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1267 make sure to run an up-to-date paramiko version, if you use this tool.
1269 The QA scripts now depend on Python 2.5 or above (the main code base
1270 still works with Python 2.4).
1272 The configuration file (``config.data``) is now written without
1273 indentation for performance reasons; if you want to edit it, it can be
1274 re-formatted via ``tools/fmtjson``.
1276 A number of bugs has been fixed in the cluster merge tool.
1278 ``x509`` certification verification (used in import-export) has been
1279 changed to allow the same clock skew as permitted by the cluster
1280 verification. This will remove some rare but hard to diagnose errors in
1287 *(Released Thu, 19 Jul 2012)*
1289 Very few changes from rc4 to the final release, only bugfixes:
1291 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1292 instance, fix CDROM booting for KVM instances)
1293 - fixed node group modification of node parameters
1294 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1295 - fixed generation of bash completion to ensure a stable ordering
1302 *(Released Fri, 13 Jul 2012)*
1304 Third release candidate for 2.6. The following changes were done from
1307 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1308 - Fixed an inconsistency in the LUXI protocol with the provided
1309 arguments (NOT backwards compatible)
1310 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1311 the cluster ``std`` value
1312 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1313 easier instance debugging on nodes (requires ``hconfd`` to work)
1319 *(Released Tue, 03 Jul 2012)*
1321 Second release candidate for 2.6. The following changes were done from
1324 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1326 - Fixed a RAPI regression on missing beparams/memory
1327 - Fixed redistribution of files on offline nodes
1328 - Added possibility to run activate-disks even though secondaries are
1329 offline. With this change it relaxes also the strictness on some other
1330 commands which use activate disks internally:
1331 * ``gnt-instance start|reboot|rename|backup|export``
1332 - Made it possible to remove safely an instance if its secondaries are
1334 - Made it possible to reinstall even though secondaries are offline
1340 *(Released Mon, 25 Jun 2012)*
1342 First release candidate for 2.6. The following changes were done from
1345 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1347 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1348 behaviour. It supports now the modify operation
1349 - ``hcheck`` implemented to analyze cluster health and possibility of
1350 improving health by rebalance
1351 - ``hbal`` has been improved in dealing with split instances
1357 *(Released Mon, 11 Jun 2012)*
1359 Second beta release of 2.6. The following changes were done from beta2
1362 - Fixed ``daemon-util`` with non-root user models
1363 - Fixed creation of plain instances with ``--no-wait-for-sync``
1364 - Fix wrong iv_names when running ``cfgupgrade``
1365 - Export more information in RAPI group queries
1366 - Fixed bug when changing instance network interfaces
1367 - Extended burnin to do NIC changes
1368 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1369 - Changed default for DRBD barriers
1370 - Fixed DRBD error reporting for syncer rate
1371 - Verify the options on disk parameters
1373 And of course various fixes to documentation and improved unittests and
1380 *(Released Wed, 23 May 2012)*
1382 First beta release of 2.6. The following changes were done from beta1 to
1385 - integrated patch for distributions without ``start-stop-daemon``
1386 - adapted example init.d script to work on Fedora
1387 - fixed log handling in Haskell daemons
1388 - adapted checks in the watcher for pycurl linked against libnss
1389 - add partial support for ``xl`` instead of ``xm`` for Xen
1390 - fixed a type issue in cluster verification
1391 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1394 Plus integrated fixes from the 2.5 branch:
1396 - fixed ``kvm-ifup`` to use ``/bin/bash``
1397 - fixed parallel build failures
1398 - KVM live migration when using a custom keymap
1404 *(Released Tue, 24 Jul 2012)*
1406 A small bugfix release, with no new features:
1408 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1409 different default shell (e.g. Debian, Ubuntu)
1410 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1411 243 and Debian bug #650664)
1412 - fixed compatibility with KVM versions that don't support multiple boot
1413 devices (fixes Issue 230 and Debian bug #624256)
1415 Additionally, a few fixes were done to the build system (fixed parallel
1416 build failures) and to the unittests (fixed race condition in test for
1417 FileID functions, and the default enable/disable mode for QA test is now
1424 *(Released Fri, 11 May 2012)*
1426 A small bugfix release.
1428 The main issues solved are on the topic of compatibility with newer LVM
1431 - fixed parsing of ``lv_attr`` field
1432 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1433 the ``--force`` flag is needed
1435 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1436 to accept kernel 3.x too (was hardcoded to 2.6.*).
1438 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1439 of hook results and that also made display of other errors suboptimal
1440 was fixed; the code behaves now like 2.4 and earlier.
1442 Another change in 2.5, the cleanup of the OS scripts environment, is too
1443 aggressive: it removed even the ``PATH`` variable, which requires the OS
1444 scripts to *always* need to export it. Since this is a bit too strict,
1445 we now export a minimal PATH, the same that we export for hooks.
1447 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1448 integrated into this release.
1450 Finally, a few other miscellaneous changes were done (no new features,
1451 just small improvements):
1453 - Fix ``gnt-group --help`` display
1454 - Fix hardcoded Xen kernel path
1455 - Fix grow-disk handling of invalid units
1456 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1457 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1463 *(Released Thu, 12 Apr 2012)*
1465 Incompatible/important changes and bugfixes
1466 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1468 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1469 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1470 to match the underlying LUXI interface.
1471 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1472 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1473 not maintain backwards-compatibility as the underlying operation
1474 changed in an incompatible way. The RAPI client can talk to old
1475 servers, but it needs to be told so as the return value changed.
1476 - When creating file-based instances via RAPI, the ``file_driver``
1477 parameter no longer defaults to ``loop`` and must be specified.
1478 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1480 - Support for the undocumented and deprecated RAPI instance creation
1481 request format version 0 has been dropped. Use version 1, supported
1482 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1483 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1484 documentation <install>`.
1485 - The "cluster-verify" hooks are now executed per group by the
1486 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1487 if you just run ``gnt-cluster verify``, which generates one opcode per
1489 - The environment as passed to the OS scripts is cleared, and thus no
1490 environment variables defined in the node daemon's environment will be
1491 inherited by the scripts.
1492 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1494 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1495 support operations involving multiple node groups.
1496 - Offline nodes are ignored when failing over an instance.
1497 - Support for KVM version 1.0, which changed the version reporting format
1499 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1501 - ``Makefile`` is now compatible with Automake 1.11.2
1502 - Includes all bugfixes made in the 2.4 series
1507 - The ganeti-htools project has been merged into the ganeti-core source
1508 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1509 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1510 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1511 GPT-style partition tables (requires `parted
1512 <http://www.gnu.org/s/parted/>`_).
1513 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1514 - Allowed adding tags on instance creation.
1515 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1516 - Allow instances to be started in a paused state, enabling the user to
1517 see the complete console output on boot using the console.
1518 - Added new hypervisor flag to control default reboot behaviour
1519 (``reboot_behavior``).
1520 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1521 - Improved out-of-band management support:
1523 - Added ``gnt-node health`` command reporting the health status of
1525 - Added ``gnt-node power`` command to manage power status of nodes.
1526 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1528 - Instance migration can fall back to failover if instance is not
1530 - Filters can be used when listing nodes, instances, groups and locks;
1531 see :manpage:`ganeti(7)` manpage.
1532 - Added post-execution status as variables to :doc:`hooks <hooks>`
1534 - Instance tags are exported/imported together with the instance.
1535 - When given an explicit job ID, ``gnt-job info`` will work for archived
1537 - Jobs can define dependencies on other jobs (not yet supported via
1538 RAPI or command line, but used by internal commands and usable via
1541 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1544 - Instance failover is now available as a RAPI resource
1545 (``/2/instances/[instance_name]/failover``).
1546 - ``gnt-instance info`` defaults to static information if primary node
1548 - Opcodes have a new ``comment`` attribute.
1549 - Added basic SPICE support to KVM hypervisor.
1550 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1552 Node group improvements
1553 ~~~~~~~~~~~~~~~~~~~~~~~
1555 - ``gnt-cluster verify`` has been modified to check groups separately,
1556 thereby improving performance.
1557 - Node group support has been added to ``gnt-cluster verify-disks``,
1558 which now operates per node group.
1559 - Watcher has been changed to work better with node groups.
1561 - One process and state file per node group.
1562 - Slow watcher in one group doesn't block other group's watcher.
1564 - Added new command, ``gnt-group evacuate``, to move all instances in a
1565 node group to other groups.
1566 - Added ``gnt-instance change-group`` to move an instance to another
1568 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1569 per-group operations.
1570 - Node groups can be tagged.
1571 - Some operations switch from an exclusive to a shared lock as soon as
1573 - Instance's primary and secondary nodes' groups are now available as
1574 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1575 and ``snodes.group.uuid``).
1580 - Numerous updates to documentation and manpages.
1582 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1584 - Some opcode/job results are now also documented, see :doc:`RAPI
1587 - A lockset's internal lock is now also visible in lock monitor.
1588 - Log messages from job queue workers now contain information about the
1589 opcode they're processing.
1590 - ``gnt-instance console`` no longer requires the instance lock.
1591 - A short delay when waiting for job changes reduces the number of LUXI
1592 requests significantly.
1593 - DRBD metadata volumes are overwritten with zeros during disk creation.
1594 - Out-of-band commands no longer acquire the cluster lock in exclusive
1596 - ``devel/upload`` now uses correct permissions for directories.
1602 *(Released Fri, 23 Mar 2012)*
1604 This was the sixth release candidate of the 2.5 series.
1610 *(Released Mon, 9 Jan 2012)*
1612 This was the fifth release candidate of the 2.5 series.
1618 *(Released Thu, 27 Oct 2011)*
1620 This was the fourth release candidate of the 2.5 series.
1626 *(Released Wed, 26 Oct 2011)*
1628 This was the third release candidate of the 2.5 series.
1634 *(Released Tue, 18 Oct 2011)*
1636 This was the second release candidate of the 2.5 series.
1642 *(Released Tue, 4 Oct 2011)*
1644 This was the first release candidate of the 2.5 series.
1650 *(Released Wed, 31 Aug 2011)*
1652 This was the third beta release of the 2.5 series.
1658 *(Released Mon, 22 Aug 2011)*
1660 This was the second beta release of the 2.5 series.
1666 *(Released Fri, 12 Aug 2011)*
1668 This was the first beta release of the 2.5 series.
1674 *(Released Thu, 27 Oct 2011)*
1676 - Fixed bug when parsing command line parameter values ending in
1678 - Fixed assertion error after unclean master shutdown
1679 - Disable HTTP client pool for RPC, significantly reducing memory usage
1681 - Fixed queue archive creation with wrong permissions
1687 *(Released Tue, 23 Aug 2011)*
1691 - Fixed documentation for importing with ``--src-dir`` option
1692 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1693 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1699 *(Released Fri, 5 Aug 2011)*
1701 Many bug-fixes and a few small features:
1703 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1704 issues when you tried to add an instance with two MAC addresses in one
1706 - KVM: fixed per-instance stored UID value
1707 - KVM: configure bridged NICs at migration start
1708 - KVM: Fix a bug where instance will not start with never KVM versions
1710 - Added OS search path to ``gnt-cluster info``
1711 - Fixed an issue with ``file_storage_dir`` where you were forced to
1712 provide an absolute path, but the documentation states it is a
1713 relative path, the documentation was right
1714 - Added a new parameter to instance stop/start called ``--no-remember``
1715 that will make the state change to not be remembered
1716 - Implemented ``no_remember`` at RAPI level
1717 - Improved the documentation
1718 - Node evacuation: don't call IAllocator if node is already empty
1719 - Fixed bug in DRBD8 replace disks on current nodes
1720 - Fixed bug in recreate-disks for DRBD instances
1721 - Moved assertion checking locks in ``gnt-instance replace-disks``
1722 causing it to abort with not owning the right locks for some situation
1723 - Job queue: Fixed potential race condition when cancelling queued jobs
1724 - Fixed off-by-one bug in job serial generation
1725 - ``gnt-node volumes``: Fix instance names
1726 - Fixed aliases in bash completion
1727 - Fixed a bug in reopening log files after being sent a SIGHUP
1728 - Added a flag to burnin to allow specifying VCPU count
1729 - Bugfixes to non-root Ganeti configuration
1735 *(Released Thu, 12 May 2011)*
1737 Many bug-fixes and a few new small features:
1739 - Fixed a bug related to log opening failures
1740 - Fixed a bug in instance listing with orphan instances
1741 - Fixed a bug which prevented resetting the cluster-level node parameter
1742 ``oob_program`` to the default
1743 - Many fixes related to the ``cluster-merge`` tool
1744 - Fixed a race condition in the lock monitor, which caused failures
1745 during (at least) creation of many instances in parallel
1746 - Improved output for gnt-job info
1747 - Removed the quiet flag on some ssh calls which prevented debugging
1749 - Improved the N+1 failure messages in cluster verify by actually
1750 showing the memory values (needed and available)
1751 - Increased lock attempt timeouts so that when executing long operations
1752 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1753 too early and thus prevent the use of the 'fair' mechanism
1754 - Changed instance query data (``gnt-instance info``) to not acquire
1755 locks unless needed, thus allowing its use on locked instance if only
1756 static information is asked for
1757 - Improved behaviour with filesystems that do not support rename on an
1759 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1760 kept locks on all nodes during the wipe, which is unneeded
1761 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1762 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1763 10GiB) which caused the wipe to fail right at the end in some cases
1764 - Fixed master IP activation when doing master failover with no-voting
1765 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1766 the master node itself
1767 - Fixed potential data-loss in under disk full conditions, where Ganeti
1768 wouldn't check correctly the return code and would consider
1769 partially-written files 'correct'
1770 - Fixed bug related to multiple VGs and DRBD disk replacing
1771 - Added new disk parameter ``metavg`` that allows placement of the meta
1772 device for DRBD in a different volume group
1773 - Fixed error handling in the node daemon when the system libc doesn't
1774 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1775 - Fixed lock release during replace-disks, which kept cluster-wide locks
1776 when doing disk replaces with an iallocator script
1777 - Added check for missing bridges in cluster verify
1778 - Handle EPIPE errors while writing to the terminal better, so that
1779 piping the output to e.g. ``less`` doesn't cause a backtrace
1780 - Fixed rare case where a ^C during Luxi calls could have been
1781 interpreted as server errors, instead of simply terminating
1782 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1784 - Added a few more parameters to the KVM hypervisor, allowing a second
1785 CDROM, custom disk type for CDROMs and a floppy image
1786 - Removed redundant message in instance rename when the name is given
1788 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1789 disks on new nodes, allowing recreation when the original instance
1790 nodes are completely gone
1791 - Added option when converting disk templates to DRBD to skip waiting
1792 for the resync, in order to make the instance available sooner
1793 - Added two new variables to the OS scripts environment (containing the
1795 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1796 to allow use of ``pvgrub`` as bootloader
1797 - Changed the instance memory modifications to only check out-of-memory
1798 conditions on memory increases, and turned the secondary node warnings
1799 into errors (they can still be overridden via ``--force``)
1800 - Fixed the handling of a corner case when the Python installation gets
1801 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1802 to execute a command that doesn't exist
1803 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1804 node of the instance returned failures during instance shutdown; this
1805 adds the option ``--ignore-consistency`` to gnt-instance move
1807 And as usual, various improvements to the error messages, documentation
1814 *(Released Wed, 09 Mar 2011)*
1816 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1817 the RAPI users file if run twice (even with ``--dry-run``).
1819 The release fixes that bug (nothing else changed).
1825 *(Released Mon, 07 Mar 2011)*
1827 Final 2.4.0 release. Just a few small fixes:
1829 - Fixed RAPI node evacuate
1830 - Fixed the kvm-ifup script
1831 - Fixed internal error handling for special job cases
1832 - Updated man page to specify the escaping feature for options
1838 *(Released Mon, 28 Feb 2011)*
1840 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1841 that this feature wiped the disks of the wrong instance, leading to loss
1846 - Fixed title of query field containing instance name
1847 - Expanded the glossary in the documentation
1848 - Fixed one unittest (internal issue)
1854 *(Released Mon, 21 Feb 2011)*
1856 A number of bug fixes plus just a couple functionality changes.
1858 On the user-visible side, the ``gnt-* list`` command output has changed
1859 with respect to "special" field states. The current rc1 style of display
1860 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1861 the default output mode special fields states are displayed as follows:
1863 - Offline resource: ``*``
1864 - Unavailable/not applicable: ``-``
1865 - Data missing (RPC failure): ``?``
1866 - Unknown field: ``??``
1868 Another user-visible change is the addition of ``--force-join`` to
1873 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1874 - Fixed regression in RAPI/instance reinstall where all parameters were
1875 required (instead of optional)
1876 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1877 - Fixed iallocator usage (offline nodes were not considered offline)
1878 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1879 - Fixed hypervisor and OS parameter validation with respect to
1880 non-vm_capable nodes
1881 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1883 - Fixed ``tools/listrunner`` with respect to agent-based usage
1889 *(Released Fri, 4 Feb 2011)*
1891 Many changes and fixes since the beta1 release. While there were some
1892 internal changes, the code has been mostly stabilised for the RC
1895 Note: the dumb allocator was removed in this release, as it was not kept
1896 up-to-date with the IAllocator protocol changes. It is recommended to
1897 use the ``hail`` command from the ganeti-htools package.
1899 Note: the 2.4 and up versions of Ganeti are not compatible with the
1900 0.2.x branch of ganeti-htools. You need to upgrade to
1901 ganeti-htools-0.3.0 (or later).
1903 Regressions fixed from 2.3
1904 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1906 - Fixed the ``gnt-cluster verify-disks`` command
1907 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1909 - Fixed disk adoption breakage
1910 - Fixed wrong headers in instance listing for field aliases
1915 - Fixed corner case in KVM handling of NICs
1916 - Fixed many cases of wrong handling of non-vm_capable nodes
1917 - Fixed a bug where a missing instance symlink was not possible to
1918 recreate with any ``gnt-*`` command (now ``gnt-instance
1919 activate-disks`` does it)
1920 - Fixed the volume group name as reported by ``gnt-cluster
1922 - Increased timeouts for the import-export code, hopefully leading to
1923 fewer aborts due network or instance timeouts
1924 - Fixed bug in ``gnt-node list-storage``
1925 - Fixed bug where not all daemons were started on cluster
1926 initialisation, but only at the first watcher run
1927 - Fixed many bugs in the OOB implementation
1928 - Fixed watcher behaviour in presence of instances with offline
1930 - Fixed instance list output for instances running on the wrong node
1931 - a few fixes to the cluster-merge tool, but it still cannot merge
1932 multi-node groups (currently it is not recommended to use this tool)
1938 - Improved network configuration for the KVM hypervisor
1939 - Added e1000 as a supported NIC for Xen-HVM
1940 - Improved the lvmstrap tool to also be able to use partitions, as
1941 opposed to full disks
1942 - Improved speed of disk wiping (the cluster parameter
1943 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1944 of instance creations
1945 - Added documentation for the OS parameters
1946 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1947 hypervisor is not responding
1948 - Added display of blacklisted and hidden OS information in
1949 ``gnt-cluster info``
1950 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1951 NIC and node parameters, which might create problems with currently
1952 invalid (but undetected) configuration files, but prevents validation
1953 failures when unrelated parameters are modified
1954 - Changed cluster initialisation to wait for the master daemon to become
1956 - Expanded the RAPI interface:
1958 - Added config redistribution resource
1959 - Added activation/deactivation of instance disks
1960 - Added export of console information
1962 - Implemented log file reopening on SIGHUP, which allows using
1963 logrotate(8) for the Ganeti log files
1964 - Added a basic OOB helper script as an example
1970 *(Released Fri, 14 Jan 2011)*
1975 - Fixed timezone issues when formatting timestamps
1976 - Added support for node groups, available via ``gnt-group`` and other
1978 - Added out-of-band framework and management, see :doc:`design
1979 document <design-oob>`
1980 - Removed support for roman numbers from ``gnt-node list`` and
1981 ``gnt-instance list``.
1982 - Allowed modification of master network interface via ``gnt-cluster
1983 modify --master-netdev``
1984 - Accept offline secondaries while shutting down instance disks
1985 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1986 - Added support for multiple LVM volume groups
1987 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1989 - Added commands to list available fields:
1991 - ``gnt-node list-fields``
1992 - ``gnt-group list-fields``
1993 - ``gnt-instance list-fields``
1995 - Updated documentation and man pages
2000 - Moved ``rapi_users`` file into separate directory, now named
2001 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
2003 - Added new tool for running commands on many machines,
2004 ``tools/ganeti-listrunner``
2005 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
2006 improving the ``gnt-instance console`` output
2007 - Allowed customisation of disk index separator at ``configure`` time
2008 - Export node group allocation policy to :doc:`iallocator <iallocator>`
2009 - Added support for non-partitioned md disks in ``lvmstrap``
2010 - Added script to gracefully power off KVM instances
2011 - Split ``utils`` module into smaller parts
2012 - Changed query operations to return more detailed information, e.g.
2013 whether an information is unavailable due to an offline node. To use
2014 this new functionality, the LUXI call ``Query`` must be used. Field
2015 information is now stored by the master daemon and can be retrieved
2016 using ``QueryFields``. Instances, nodes and groups can also be queried
2017 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
2018 exposed via RAPI). The following commands make use of this
2019 infrastructure change:
2021 - ``gnt-group list``
2022 - ``gnt-group list-fields``
2024 - ``gnt-node list-fields``
2025 - ``gnt-instance list``
2026 - ``gnt-instance list-fields``
2027 - ``gnt-debug locks``
2032 - New RAPI resources (see :doc:`rapi`):
2036 - ``/2/groups/[group_name]``
2037 - ``/2/groups/[group_name]/assign-nodes``
2038 - ``/2/groups/[group_name]/modify``
2039 - ``/2/groups/[group_name]/rename``
2040 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
2044 - Implemented ``no_install`` for instance creation
2045 - Implemented OS parameters for instance reinstallation, allowing
2046 use of special settings on reinstallation (e.g. for preserving data)
2051 - Added IPv6 support in import/export
2052 - Pause DRBD synchronization while wiping disks on instance creation
2053 - Updated unittests and QA scripts
2054 - Improved network parameters passed to KVM
2055 - Converted man pages from docbook to reStructuredText
2061 *(Released Mon, 20 Dec 2010)*
2063 Released version 2.3.1~rc1 without any changes.
2069 *(Released Wed, 1 Dec 2010)*
2071 - impexpd: Disable OpenSSL compression in socat if possible (backport
2072 from master, commit e90739d625b, see :doc:`installation guide
2073 <install-quick>` for details)
2074 - Changed unittest coverage report to exclude test scripts
2075 - Added script to check version format
2081 *(Released Wed, 1 Dec 2010)*
2083 Released version 2.3.0~rc1 without any changes.
2089 *(Released Fri, 19 Nov 2010)*
2091 A number of bugfixes and documentation updates:
2093 - Update ganeti-os-interface documentation
2094 - Fixed a bug related to duplicate MACs or similar items which should be
2096 - Fix breakage in OS state modify
2097 - Reinstall instance: disallow offline secondaries (fixes bug related to
2098 OS changing but reinstall failing)
2099 - plus all the other fixes between 2.2.1 and 2.2.2
2105 *(Released Tue, 2 Nov 2010)*
2107 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
2108 - Fixed master failover race with watcher
2109 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
2111 - Accept previously stopped instance for export with instance removal
2112 - Simplify and extend the environment variables for instance OS scripts
2113 - Added new node flags, ``master_capable`` and ``vm_capable``
2114 - Added optional instance disk wiping prior during allocation. This is a
2115 cluster-wide option and can be set/modified using
2116 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
2117 - Added IPv6 support, see :doc:`design document <design-2.3>` and
2118 :doc:`install-quick`
2119 - Added a new watcher option (``--ignore-pause``)
2120 - Added option to ignore offline node on instance start/stop
2121 (``--ignore-offline``)
2122 - Allow overriding OS parameters with ``gnt-instance reinstall``
2123 - Added ability to change node's secondary IP address using ``gnt-node
2125 - Implemented privilege separation for all daemons except
2126 ``ganeti-noded``, see ``configure`` options
2127 - Complain if an instance's disk is marked faulty in ``gnt-cluster
2129 - Implemented job priorities (see ``ganeti(7)`` manpage)
2130 - Ignore failures while shutting down instances during failover from
2132 - Exit daemon's bootstrap process only once daemon is ready
2133 - Export more information via ``LUInstanceQuery``/remote API
2134 - Improved documentation, QA and unittests
2135 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
2136 restart if the file was created or changed
2137 - Added LUXI protocol version sent with each request and response,
2138 allowing detection of server/client mismatches
2139 - Moved the Python scripts among gnt-* and ganeti-* into modules
2140 - Moved all code related to setting up SSH to an external script,
2142 - Infrastructure changes for node group support in future versions
2148 *(Released Fri, 19 Nov 2010)*
2150 A few small bugs fixed, and some improvements to the build system:
2152 - Fix documentation regarding conversion to drbd
2153 - Fix validation of parameters in cluster modify (``gnt-cluster modify
2155 - Fix error handling in node modify with multiple changes
2156 - Allow remote imports without checked names
2162 *(Released Tue, 19 Oct 2010)*
2164 - Disable SSL session ID cache in RPC client
2170 *(Released Thu, 14 Oct 2010)*
2172 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
2173 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2175 - Fix problems with interaction between Curl and Python's HTTP server,
2176 resulting in increased speed in many RPC calls
2177 - Improve our release script to prevent breakage with older aclocal and
2184 *(Released Thu, 7 Oct 2010)*
2186 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2187 value retrieved from master
2188 - Added support for blacklisted or hidden OS definitions
2189 - Added simple lock monitor (accessible via (``gnt-debug locks``)
2190 - Added support for -mem-path in KVM hypervisor abstraction layer
2191 - Allow overriding instance parameters in tool for inter-cluster
2192 instance moves (``tools/move-instance``)
2193 - Improved opcode summaries (e.g. in ``gnt-job list``)
2194 - Improve consistency of OS listing by sorting it
2195 - Documentation updates
2201 *(Released Fri, 8 Oct 2010)*
2203 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2209 *(Released Mon, 4 Oct 2010)*
2211 - Fixed regression in ``gnt-instance rename``
2217 *(Released Wed, 22 Sep 2010)*
2219 - Fixed OS_VARIANT variable for OS scripts
2220 - Fixed cluster tag operations via RAPI
2221 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2222 - Disabled RAPI CA checks in watcher
2228 *(Released Mon, 23 Aug 2010)*
2230 - Support DRBD versions of the format "a.b.c.d"
2232 - Re-introduce support for usage from multiple threads in RAPI client
2233 - Instance renames and modify via RAPI
2234 - Work around race condition between processing and archival in job
2236 - Mark opcodes following failed one as failed, too
2237 - Job field ``lock_status`` was removed due to difficulties making it
2238 work with the changed job queue in Ganeti 2.2; a better way to monitor
2239 locks is expected for a later 2.2.x release
2240 - Fixed dry-run behaviour with many commands
2241 - Support ``ssh-agent`` again when adding nodes
2242 - Many additional bugfixes
2248 *(Released Fri, 30 Jul 2010)*
2250 Important change: the internal RPC mechanism between Ganeti nodes has
2251 changed from using a home-grown http library (based on the Python base
2252 libraries) to use the PycURL library. This requires that PycURL is
2253 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2254 against GnuTLS by default. cURL's support for GnuTLS had known issues
2255 before cURL 7.21.0 and we recommend using the latest cURL release or
2256 linking against OpenSSL. Most other distributions already link PycURL
2257 and cURL against OpenSSL. The command::
2259 python -c 'import pycurl; print pycurl.version'
2261 can be used to determine the libraries PycURL and cURL are linked
2264 Other significant changes:
2266 - Rewrote much of the internals of the job queue, in order to achieve
2267 better parallelism; this decouples job query operations from the job
2268 processing, and it should allow much nicer behaviour of the master
2269 daemon under load, and it also has uncovered some long-standing bugs
2270 related to the job serialisation (now fixed)
2271 - Added a default iallocator setting to the cluster parameters,
2272 eliminating the need to always pass nodes or an iallocator for
2273 operations that require selection of new node(s)
2274 - Added experimental support for the LXC virtualization method
2275 - Added support for OS parameters, which allows the installation of
2276 instances to pass parameter to OS scripts in order to customise the
2278 - Added a hypervisor parameter controlling the migration type (live or
2279 non-live), since hypervisors have various levels of reliability; this
2280 has renamed the 'live' parameter to 'mode'
2281 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2282 logical volumes, meaning that cluster verify will ignore them and not
2283 flag their presence as errors
2284 - The watcher will now reset the error count for failed instances after
2285 8 hours, thus allowing self-healing if the problem that caused the
2286 instances to be down/fail to start has cleared in the meantime
2287 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2288 check for, and warn, if the drbd module parameter ``usermode_helper``
2289 is not consistent with the cluster-wide setting; this is needed to
2290 make diagnose easier of failed drbd creations
2291 - Started adding base IPv6 support, but this is not yet
2292 enabled/available for use
2293 - Rename operations (cluster, instance) will now return the new name,
2294 which is especially useful if a short name was passed in
2295 - Added support for instance migration in RAPI
2296 - Added a tool to pre-configure nodes for the SSH setup, before joining
2297 them to the cluster; this will allow in the future a simplified model
2298 for node joining (but not yet fully enabled in 2.2); this needs the
2299 paramiko python library
2300 - Fixed handling of name-resolving errors
2301 - Fixed consistency of job results on the error path
2302 - Fixed master-failover race condition when executed multiple times in
2304 - Fixed many bugs related to the job queue (mostly introduced during the
2305 2.2 development cycle, so not all are impacting 2.1)
2306 - Fixed instance migration with missing disk symlinks
2307 - Fixed handling of unknown jobs in ``gnt-job archive``
2308 - And many other small fixes/improvements
2312 - Enhanced both the unittest and the QA coverage
2313 - Switched the opcode validation to a generic model, and extended the
2314 validation to all opcode parameters
2315 - Changed more parts of the code that write shell scripts to use the
2317 - Switched the master daemon to use the asyncore library for the Luxi
2324 *(Released Thu, 17 Jun 2010)*
2326 - Added tool (``move-instance``) and infrastructure to move instances
2327 between separate clusters (see :doc:`separate documentation
2328 <move-instance>` and :doc:`design document <design-2.2>`)
2329 - Added per-request RPC timeout
2330 - RAPI now requires a Content-Type header for requests with a body (e.g.
2331 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2332 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2333 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2335 - Implemented initial support for running Ganeti daemons as separate
2336 users, see configure-time flags ``--with-user-prefix`` and
2337 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2339 - Instances can be removed after export (``gnt-backup export
2340 --remove-instance``)
2341 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2342 key (instead of 1024 bit)
2343 - Added new cluster configuration file for cluster domain secret
2344 - Import/export now use SSL instead of SSH
2345 - Added support for showing estimated time when exporting an instance,
2346 see the ``ganeti-os-interface(7)`` manpage and look for
2353 *(Released Tue, 16 Nov 2010)*
2355 Some more bugfixes. Unless critical bugs occur, this will be the last
2358 - Fix case of MAC special-values
2359 - Fix mac checker regex
2360 - backend: Fix typo causing "out of range" error
2361 - Add missing --units in gnt-instance list man page
2367 *(Released Tue, 24 Aug 2010)*
2370 - Don't ignore secondary node silently on non-mirrored disk templates
2372 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2373 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2374 - Properly document the usb_mouse parameter
2375 - Fix path in ganeti-rapi(8) (issue 116)
2376 - Adjust error message when the ganeti user's .ssh directory is
2378 - Add same-node-check when changing the disk template to drbd
2384 *(Released Fri, 16 Jul 2010)*
2387 - Add an option to only select some reboot types during qa/burnin.
2388 (on some hypervisors consequent reboots are not supported)
2389 - Fix infrequent race condition in master failover. Sometimes the old
2390 master ip address would be still detected as up for a short time
2391 after it was removed, causing failover to fail.
2392 - Decrease mlockall warnings when the ctypes module is missing. On
2393 Python 2.4 we support running even if no ctypes module is installed,
2394 but we were too verbose about this issue.
2395 - Fix building on old distributions, on which man doesn't have a
2397 - Fix RAPI not to ignore the MAC address on instance creation
2398 - Implement the old instance creation format in the RAPI client.
2404 *(Released Thu, 01 Jul 2010)*
2406 A small bugfix release:
2407 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2408 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2409 a non-existing option
2410 - Fix instance create: the --force-variant option was ignored
2411 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2412 - Fix modify node storage with non-FQDN arguments
2413 - Fix RAPI client to authenticate under Python 2.6 when used
2414 for more than 5 requests needing authentication
2415 - Fix gnt-instance modify -t (storage) giving a wrong error message
2416 when converting a non-shutdown drbd instance to plain
2422 *(Released Fri, 18 Jun 2010)*
2424 A small bugfix release:
2426 - Fix live migration of KVM instances started with older Ganeti
2427 versions which had fewer hypervisor parameters
2428 - Fix gnt-instance grow-disk on down instances
2429 - Fix an error-reporting bug during instance migration
2430 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2431 silently ignoring broken ones
2432 - Fix an RPC error reporting bug affecting, for example, RAPI client
2434 - Fix bug triggered by different API version os-es on different nodes
2435 - Fix a bug in instance startup with custom hvparams: OS level
2436 parameters would fail to be applied.
2437 - Fix the RAPI client under Python 2.6 (but more work is needed to
2438 make it work completely well with OpenSSL)
2439 - Fix handling of errors when resolving names from DNS
2445 *(Released Thu, 3 Jun 2010)*
2447 A medium sized development cycle. Some new features, and some
2448 fixes/small improvements/cleanups.
2450 Significant features
2451 ~~~~~~~~~~~~~~~~~~~~
2453 The node deamon now tries to mlock itself into memory, unless the
2454 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2455 its logs, and falls back to console logging. This allows emergency
2456 features such as ``gnt-node powercycle`` to work even in the event of a
2457 broken node disk (tested offlining the disk hosting the node's
2458 filesystem and dropping its memory caches; don't try this at home)
2460 KVM: add vhost-net acceleration support. It can be tested with a new
2461 enough version of the kernel and of qemu-kvm.
2463 KVM: Add instance chrooting feature. If you use privilege dropping for
2464 your VMs you can also now force them to chroot to an empty directory,
2465 before starting the emulated guest.
2467 KVM: Add maximum migration bandwith and maximum downtime tweaking
2468 support (requires a new-enough version of qemu-kvm).
2470 Cluster verify will now warn if the master node doesn't have the master
2471 ip configured on it.
2473 Add a new (incompatible) instance creation request format to RAPI which
2474 supports all parameters (previously only a subset was supported, and it
2475 wasn't possible to extend the old format to accomodate all the new
2476 features. The old format is still supported, and a client can check for
2477 this feature, before using it, by checking for its presence in the
2478 ``features`` RAPI resource.
2480 Now with ancient latin support. Try it passing the ``--roman`` option to
2481 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2482 (requires the python-roman module to be installed, in order to work).
2487 As usual many internal code refactorings, documentation updates, and
2490 - Lots of improvements and cleanups to the experimental Remote API
2491 (RAPI) client library.
2492 - A new unit test suite for the core daemon libraries.
2493 - A fix to creating missing directories makes sure the umask is not
2494 applied anymore. This enforces the same directory permissions
2496 - Better handling terminating daemons with ctrl+c (used when running
2497 them in debugging mode).
2498 - Fix a race condition in live migrating a KVM instance, when stat()
2499 on the old proc status file returned EINVAL, which is an unexpected
2501 - Fixed manpage checking with newer man and utf-8 charachters. But now
2502 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2508 *(Released Fri, 7 May 2010)*
2510 Fix a bug which prevented untagged KVM instances from starting.
2516 *(Released Fri, 7 May 2010)*
2518 Another release with a long development cycle, during which many
2519 different features were added.
2521 Significant features
2522 ~~~~~~~~~~~~~~~~~~~~
2524 The KVM hypervisor now can run the individual instances as non-root, to
2525 reduce the impact of a VM being hijacked due to bugs in the
2526 hypervisor. It is possible to run all instances as a single (non-root)
2527 user, to manually specify a user for each instance, or to dynamically
2528 allocate a user out of a cluster-wide pool to each instance, with the
2529 guarantee that no two instances will run under the same user ID on any
2532 An experimental RAPI client library, that can be used standalone
2533 (without the other Ganeti libraries), is provided in the source tree as
2534 ``lib/rapi/client.py``. Note this client might change its interface in
2535 the future, as we iterate on its capabilities.
2537 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2538 replace the cluster's certificates and crypto keys. This might help in
2539 case they have been compromised, or have simply expired.
2541 A new disk option for instance creation has been added that allows one
2542 to "adopt" currently existing logical volumes, with data
2543 preservation. This should allow easier migration to Ganeti from
2544 unmanaged (or managed via other software) instances.
2546 Another disk improvement is the possibility to convert between redundant
2547 (DRBD) and plain (LVM) disk configuration for an instance. This should
2548 allow better scalability (starting with one node and growing the
2549 cluster, or shrinking a two-node cluster to one node).
2551 A new feature that could help with automated node failovers has been
2552 implemented: if a node sees itself as offline (by querying the master
2553 candidates), it will try to shutdown (hard) all instances and any active
2554 DRBD devices. This reduces the risk of duplicate instances if an
2555 external script automatically failovers the instances on such nodes. To
2556 enable this, the cluster parameter ``maintain_node_health`` should be
2557 enabled; in the future this option (per the name) will enable other
2558 automatic maintenance features.
2560 Instance export/import now will reuse the original instance
2561 specifications for all parameters; that means exporting an instance,
2562 deleting it and the importing it back should give an almost identical
2563 instance. Note that the default import behaviour has changed from
2564 before, where it created only one NIC; now it recreates the original
2567 Cluster verify has added a few new checks: SSL certificates validity,
2568 /etc/hosts consistency across the cluster, etc.
2573 As usual, many internal changes were done, documentation fixes,
2576 - Fixed cluster initialization with disabled cluster storage (regression
2577 introduced in 2.1.1)
2578 - File-based storage supports growing the disks
2579 - Fixed behaviour of node role changes
2580 - Fixed cluster verify for some corner cases, plus a general rewrite of
2581 cluster verify to allow future extension with more checks
2582 - Fixed log spamming by watcher and node daemon (regression introduced
2584 - Fixed possible validation issues when changing the list of enabled
2586 - Fixed cleanup of /etc/hosts during node removal
2587 - Fixed RAPI response for invalid methods
2588 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2589 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2590 from ide disks, etc.)
2591 - Allow OS changes without re-installation (to record a changed OS
2592 outside of Ganeti, or to allow OS renames)
2593 - Allow instance creation without OS installation (useful for example if
2594 the OS will be installed manually, or restored from a backup not in
2596 - Implemented option to make cluster ``copyfile`` use the replication
2598 - Added list of enabled hypervisors to ssconf (possibly useful for
2600 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2602 - A partial form of node re-IP is possible via node readd, which now
2603 allows changed node primary IP
2604 - Command line utilities now show an informational message if the job is
2606 - The logs of the master daemon now show the PID/UID/GID of the
2613 *(Released Fri, 12 Mar 2010)*
2615 During the 2.1.0 long release candidate cycle, a lot of improvements and
2616 changes have accumulated with were released later as 2.1.1.
2621 The node evacuate command (``gnt-node evacuate``) was significantly
2622 rewritten, and as such the IAllocator protocol was changed - a new
2623 request type has been added. This unfortunate change during a stable
2624 series is designed to improve performance of node evacuations; on
2625 clusters with more than about five nodes and which are well-balanced,
2626 evacuation should proceed in parallel for all instances of the node
2627 being evacuated. As such, any existing IAllocator scripts need to be
2628 updated, otherwise the above command will fail due to the unknown
2629 request. The provided "dumb" allocator has not been updated; but the
2630 ganeti-htools package supports the new protocol since version 0.2.4.
2632 Another important change is increased validation of node and instance
2633 names. This might create problems in special cases, if invalid host
2634 names are being used.
2636 Also, a new layer of hypervisor parameters has been added, that sits at
2637 OS level between the cluster defaults and the instance ones. This allows
2638 customisation of virtualization parameters depending on the installed
2639 OS. For example instances with OS 'X' may have a different KVM kernel
2640 (or any other parameter) than the cluster defaults. This is intended to
2641 help managing a multiple OSes on the same cluster, without manual
2642 modification of each instance's parameters.
2644 A tool for merging clusters, ``cluster-merge``, has been added in the
2645 tools sub-directory.
2650 - Improved the int/float conversions that should make the code more
2651 robust in face of errors from the node daemons
2652 - Fixed the remove node code in case of internal configuration errors
2653 - Fixed the node daemon behaviour in face of inconsistent queue
2654 directory (e.g. read-only file-system where we can't open the files
2656 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2657 now it either aborts cleanly or, if given the new "auto_promote"
2658 parameter, will automatically promote other nodes as needed
2659 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2660 completely prevent Ganeti from working
2661 - Fixed bug for instance export when not all disks were successfully
2663 - Fixed behaviour of node add when the new node is slow in starting up
2665 - Fixed handling of signals in the LUXI client, which should improve
2666 behaviour of command-line scripts
2667 - Added checks for invalid node/instance names in the configuration (now
2668 flagged during cluster verify)
2669 - Fixed watcher behaviour for disk activation errors
2670 - Fixed two potentially endless loops in http library, which led to the
2671 RAPI daemon hanging and consuming 100% CPU in some cases
2672 - Fixed bug in RAPI daemon related to hashed passwords
2673 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2675 - Enhanced compatibility with non-Debian OSes, but not using absolute
2676 path in some commands and allowing customisation of the ssh
2677 configuration directory
2678 - Fixed possible future issue with new Python versions by abiding to the
2679 proper use of ``__slots__`` attribute on classes
2680 - Added checks that should prevent directory traversal attacks
2681 - Many documentation fixes based on feedback from users
2686 - Added an "early_release" more for instance replace disks and node
2687 evacuate, where we release locks earlier and thus allow higher
2688 parallelism within the cluster
2689 - Added watcher hooks, intended to allow the watcher to restart other
2690 daemons (e.g. from the ganeti-nbma project), but they can be used of
2691 course for any other purpose
2692 - Added a compile-time disable for DRBD barriers, to increase
2693 performance if the administrator trusts the power supply or the
2694 storage system to not lose writes
2695 - Added the option of using syslog for logging instead of, or in
2696 addition to, Ganeti's own log files
2697 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2698 can indeed boot from a paravirtual NIC
2699 - Added a generic debug level for many operations; while this is not
2700 used widely yet, it allows one to pass the debug value all the way to
2702 - Enhanced the hooks environment for instance moves (failovers,
2703 migrations) where the primary/secondary nodes changed during the
2704 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2705 - Enhanced data validations for many user-supplied values; one important
2706 item is the restrictions imposed on instance and node names, which
2707 might reject some (invalid) host names
2708 - Add a configure-time option to disable file-based storage, if it's not
2709 needed; this allows greater security separation between the master
2710 node and the other nodes from the point of view of the inter-node RPC
2712 - Added user notification in interactive tools if job is waiting in the
2713 job queue or trying to acquire locks
2714 - Added log messages when a job is waiting for locks
2715 - Added filtering by node tags in instance operations which admit
2716 multiple instances (start, stop, reboot, reinstall)
2717 - Added a new tool for cluster mergers, ``cluster-merge``
2718 - Parameters from command line which are of the form ``a=b,c=d`` can now
2719 use backslash escapes to pass in values which contain commas,
2720 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2722 - For KVM, the instance name is the first parameter passed to KVM, so
2723 that it's more visible in the process list
2729 *(Released Tue, 2 Mar 2010)*
2731 Ganeti 2.1 brings many improvements with it. Major changes:
2733 - Added infrastructure to ease automated disk repairs
2734 - Added new daemon to export configuration data in a cheaper way than
2735 using the remote API
2736 - Instance NICs can now be routed instead of being associated with a
2738 - Improved job locking logic to reduce impact of jobs acquiring multiple
2739 locks waiting for other long-running jobs
2741 In-depth implementation details can be found in the Ganeti 2.1 design
2747 - Added chroot hypervisor
2748 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2750 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2751 ``bootloader_path`` and ``bootloader_args``)
2752 - Added the ``use_localtime`` option for the xen-hvm and kvm
2753 hypervisors, and the default value for this has changed to false (in
2754 2.0 xen-hvm always enabled it)
2755 - Added luxi call to submit multiple jobs in one go
2756 - Added cluster initialization option to not modify ``/etc/hosts``
2758 - Added network interface parameters
2759 - Added dry run mode to some LUs
2760 - Added RAPI resources:
2762 - ``/2/instances/[instance_name]/info``
2763 - ``/2/instances/[instance_name]/replace-disks``
2764 - ``/2/nodes/[node_name]/evacuate``
2765 - ``/2/nodes/[node_name]/migrate``
2766 - ``/2/nodes/[node_name]/role``
2767 - ``/2/nodes/[node_name]/storage``
2768 - ``/2/nodes/[node_name]/storage/modify``
2769 - ``/2/nodes/[node_name]/storage/repair``
2771 - Added OpCodes to evacuate or migrate all instances on a node
2772 - Added new command to list storage elements on nodes (``gnt-node
2773 list-storage``) and modify them (``gnt-node modify-storage``)
2774 - Added new ssconf files with master candidate IP address
2775 (``ssconf_master_candidates_ips``), node primary IP address
2776 (``ssconf_node_primary_ips``) and node secondary IP address
2777 (``ssconf_node_secondary_ips``)
2778 - Added ``ganeti-confd`` and a client library to query the Ganeti
2779 configuration via UDP
2780 - Added ability to run hooks after cluster initialization and before
2782 - Added automatic mode for disk replace (``gnt-instance replace-disks
2784 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2785 after catastrophic data-loss
2786 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2788 - Added ``gnt-instance move`` command to move instances
2789 - Added ``gnt-cluster watcher`` command to control watcher
2790 - Added ``gnt-node powercycle`` command to powercycle nodes
2791 - Added new job status field ``lock_status``
2792 - Added parseable error codes to cluster verification (``gnt-cluster
2793 verify --error-codes``) and made output less verbose (use
2794 ``--verbose`` to restore previous behaviour)
2795 - Added UUIDs to the main config entities (cluster, nodes, instances)
2796 - Added support for OS variants
2797 - Added support for hashed passwords in the Ganeti remote API users file
2799 - Added option to specify maximum timeout on instance shutdown
2800 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2801 - Added new helper script to start and stop Ganeti daemons
2802 (``daemon-util``), with the intent to reduce the work necessary to
2803 adjust Ganeti for non-Debian distributions and to start/stop daemons
2805 - Added more unittests
2806 - Fixed critical bug in ganeti-masterd startup
2807 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2808 now customisable at the cluster level for both the KVM and Xen
2809 hypervisors using the new ``migration_port`` parameter
2810 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2811 reinstalling an instance
2812 - Allowed ``@`` in tag names
2813 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2814 - Many documentation updates
2815 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2816 - ``gnt-instance reinstall`` can now reinstall multiple instances
2817 - Updated many command line parameters
2818 - Introduced new OS API version 15
2819 - No longer support a default hypervisor
2820 - Treat virtual LVs as inexistent
2821 - Improved job locking logic to reduce lock contention
2822 - Match instance and node names case insensitively
2823 - Reimplemented bash completion script to be more complete
2830 *(Released Thu, 4 Feb 2010)*
2832 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2833 - Fix a string formatting bug
2834 - Improve safety of the code in some error paths
2835 - Improve data validation in the master of values returned from nodes
2841 *(Released Thu, 17 Dec 2009)*
2843 - Fix security issue due to missing validation of iallocator names; this
2844 allows local and remote execution of arbitrary executables
2845 - Fix failure of gnt-node list during instance removal
2846 - Ship the RAPI documentation in the archive
2852 *(Released Wed, 30 Sep 2009)*
2854 - Fixed many wrong messages
2855 - Fixed a few bugs related to the locking library
2856 - Fixed MAC checking at instance creation time
2857 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2858 - Fixed a few issues related to signal handling in both daemons and
2860 - Fixed the example startup script provided
2861 - Fixed insserv dependencies in the example startup script (patch from
2863 - Fixed handling of drained nodes in the iallocator framework
2864 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2866 - Fixed error related to invalid job IDs in job polling
2867 - Fixed job/opcode persistence on unclean master shutdown
2868 - Fixed handling of partial job processing after unclean master
2870 - Fixed error reporting from LUs, previously all errors were converted
2871 into execution errors
2872 - Fixed error reporting from burnin
2873 - Decreased significantly the memory usage of the job queue
2874 - Optimised slightly multi-job submission
2875 - Optimised slightly opcode loading
2876 - Backported the multi-job submit framework from the development
2877 branch; multi-instance start and stop should be faster
2878 - Added script to clean archived jobs after 21 days; this will reduce
2879 the size of the queue directory
2880 - Added some extra checks in disk size tracking
2881 - Added an example ethers hook script
2882 - Added a cluster parameter that prevents Ganeti from modifying of
2884 - Added more node information to RAPI responses
2885 - Added a ``gnt-job watch`` command that allows following the ouput of a
2887 - Added a bind-address option to ganeti-rapi
2888 - Added more checks to the configuration verify
2889 - Enhanced the burnin script such that some operations can be retried
2891 - Converted instance reinstall to multi-instance model
2897 *(Released Fri, 7 Aug 2009)*
2899 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2900 to allow using the pre-2.0.2 behaviour in activation, if any existing
2901 instances have mismatched disk sizes in the configuration
2902 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2903 any configuration mismatches for disk sizes
2904 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2905 failover to work on two-node clusters
2906 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2908 - Fixed detection of OS script errors in ``gnt-backup export``
2909 - Fixed exit code of ``gnt-backup export``
2915 *(Released Fri, 17 Jul 2009)*
2917 - Added experimental support for stripped logical volumes; this should
2918 enhance performance but comes with a higher complexity in the block
2919 device handling; stripping is only enabled when passing
2920 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2921 affected even in the non-stripped mode
2922 - Improved resiliency against transient failures at the end of DRBD
2923 resyncs, and in general of DRBD resync checks
2924 - Fixed a couple of issues with exports and snapshot errors
2925 - Fixed a couple of issues in instance listing
2926 - Added display of the disk size in ``gnt-instance info``
2927 - Fixed checking for valid OSes in instance creation
2928 - Fixed handling of the "vcpus" parameter in instance listing and in
2929 general of invalid parameters
2930 - Fixed http server library, and thus RAPI, to handle invalid
2931 username/password combinations correctly; this means that now they
2932 report unauthorized for queries too, not only for modifications,
2933 allowing earlier detect of configuration problems
2934 - Added a new "role" node list field, equivalent to the master/master
2935 candidate/drained/offline flags combinations
2936 - Fixed cluster modify and changes of candidate pool size
2937 - Fixed cluster verify error messages for wrong files on regular nodes
2938 - Fixed a couple of issues with node demotion from master candidate role
2939 - Fixed node readd issues
2940 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2941 - Added a new ``--no-voting`` option for masterfailover to fix failover
2942 on two-nodes clusters when the former master node is unreachable
2943 - Added instance reinstall over RAPI
2949 *(Released Tue, 16 Jun 2009)*
2951 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2952 allow re-adding the start in single-user option (regression from 1.2)
2953 - the watcher writes the instance status to a file, to allow monitoring
2954 to report the instance status (from the master) based on cached
2955 results of the watcher's queries; while this can get stale if the
2956 watcher is being locked due to other work on the cluster, this is
2957 still an improvement
2958 - the watcher now also restarts the node daemon and the rapi daemon if
2960 - fixed the watcher to handle full and drained queue cases
2961 - hooks export more instance data in the environment, which helps if
2962 hook scripts need to take action based on the instance's properties
2963 (no longer need to query back into ganeti)
2964 - instance failovers when the instance is stopped do not check for free
2965 RAM, so that failing over a stopped instance is possible in low memory
2967 - rapi uses queries for tags instead of jobs (for less job traffic), and
2968 for cluster tags it won't talk to masterd at all but read them from
2970 - a couple of error handling fixes in RAPI
2971 - drbd handling: improved the error handling of inconsistent disks after
2972 resync to reduce the frequency of "there are some degraded disks for
2973 this instance" messages
2974 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2975 error handling path called a wrong function name)
2981 *(Released Wed, 27 May 2009)*
2983 - no changes from rc5
2989 *(Released Wed, 20 May 2009)*
2991 - fix a couple of bugs (validation, argument checks)
2992 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2993 - some small improvements to RAPI and IAllocator
2994 - make watcher automatically start the master daemon if down
3000 *(Released Mon, 27 Apr 2009)*
3002 - change the OS list to not require locks; this helps with big clusters
3003 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
3004 volume group is broken
3005 - ``gnt-instance info``, without any arguments, doesn't run for all
3006 instances anymore; either pass ``--all`` or pass the desired
3007 instances; this helps against mistakes on big clusters where listing
3008 the information for all instances takes a long time
3009 - miscellaneous doc and man pages fixes
3015 *(Released Wed, 8 Apr 2009)*
3017 - Change the internal locking model of some ``gnt-node`` commands, in
3018 order to reduce contention (and blocking of master daemon) when
3019 batching many creation/reinstall jobs
3020 - Fixes to Xen soft reboot
3021 - No longer build documentation at build time, instead distribute it in
3022 the archive, in order to reduce the need for the whole docbook/rst
3029 *(Released Fri, 27 Mar 2009)*
3031 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
3032 - Fix watcher startup sequence, improves the behaviour of busy clusters
3033 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
3034 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
3035 bind address and other places
3036 - Some documentation fixes and updates
3042 *(Released Mon, 2 Mar 2009)*
3044 - More documentation updates, now all docs should be more-or-less
3046 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
3048 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
3050 - Fix an issue related to $libdir/run/ganeti and cluster creation
3056 *(Released Thu, 19 Feb 2009)*
3058 - Xen PVM and KVM have switched the default value for the instance root
3059 disk to the first partition on the first drive, instead of the whole
3060 drive; this means that the OS installation scripts must be changed
3062 - Man pages have been updated
3063 - RAPI has been switched by default to HTTPS, and the exported functions
3064 should all work correctly
3065 - RAPI v1 has been removed
3066 - Many improvements to the KVM hypervisor
3067 - Block device errors are now better reported
3068 - Many other bugfixes and small improvements
3074 *(Released Mon, 26 Jan 2009)*
3076 - Version 2 is a general rewrite of the code and therefore the
3077 differences are too many to list, see the design document for 2.0 in
3078 the ``doc/`` subdirectory for more details
3079 - In this beta version there is not yet a migration path from 1.2 (there
3080 will be one in the final 2.0 release)
3081 - A few significant changes are:
3083 - all commands are executed by a daemon (``ganeti-masterd``) and the
3084 various ``gnt-*`` commands are just front-ends to it
3085 - all the commands are entered into, and executed from a job queue,
3086 see the ``gnt-job(8)`` manpage
3087 - the RAPI daemon supports read-write operations, secured by basic
3088 HTTP authentication on top of HTTPS
3089 - DRBD version 0.7 support has been removed, DRBD 8 is the only
3090 supported version (when migrating from Ganeti 1.2 to 2.0, you need
3091 to migrate to DRBD 8 first while still running Ganeti 1.2)
3092 - DRBD devices are using statically allocated minor numbers, which
3093 will be assigned to existing instances during the migration process
3094 - there is support for both Xen PVM and Xen HVM instances running on
3096 - KVM virtualization is supported too
3097 - file-based storage has been implemented, which means that it is
3098 possible to run the cluster without LVM and DRBD storage, for
3099 example using a shared filesystem exported from shared storage (and
3100 still have live migration)
3106 *(Released Tue, 13 Jan 2009)*
3108 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
3109 - Reuse the old instance mac address by default on instance import, if
3110 the instance name is the same.
3111 - Handle situations in which the node info rpc returns incomplete
3113 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
3114 - Improved version of batcher:
3116 - state file support
3117 - instance mac address support
3118 - support for HVM clusters/instances
3120 - Add an option to show the number of cpu sockets and nodes in
3122 - Support OSes that handle more than one version of the OS api (but do
3123 not change the current API in any other way)
3124 - Fix ``gnt-node migrate``
3125 - ``gnt-debug`` man page
3126 - Fixes various more typos and small issues
3127 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
3133 *(Released Wed, 24 Sep 2008)*
3135 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
3136 type of disk exported to fully virtualized instances.
3137 - provide access to the serial console of HVM instances
3138 - instance auto_balance flag, set by default. If turned off it will
3139 avoid warnings on cluster verify if there is not enough memory to fail
3140 over an instance. in the future it will prevent automatically failing
3141 it over when we will support that.
3142 - batcher tool for instance creation, see ``tools/README.batcher``
3143 - ``gnt-instance reinstall --select-os`` to interactively select a new
3144 operating system when reinstalling an instance.
3145 - when changing the memory amount on instance modify a check has been
3146 added that the instance will be able to start. also warnings are
3147 emitted if the instance will not be able to fail over, if auto_balance
3149 - documentation fixes
3150 - sync fields between ``gnt-instance list/modify/add/import``
3151 - fix a race condition in drbd when the sync speed was set after giving
3152 the device a remote peer.
3158 *(Released Tue, 22 Jul 2008)*
3160 - note: the allowed size and number of tags per object were reduced
3161 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
3162 - fixed twisted 8.x compatibility
3163 - fixed ``gnt-instance replace-disks`` with iallocator
3164 - add TCP keepalives on twisted connections to detect restarted nodes
3165 - disk increase support, see ``gnt-instance grow-disk``
3166 - implement bulk node/instance query for RAPI
3167 - add tags in node/instance listing (optional)
3168 - experimental migration (and live migration) support, read the man page
3169 for ``gnt-instance migrate``
3170 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3171 has some small improvements in handling its state file
3177 *(Released Fri, 13 Jun 2008)*
3179 - Experimental readonly, REST-based remote API implementation;
3180 automatically started on master node, TCP port 5080, if enabled by
3181 ``--enable-rapi`` parameter to configure script.
3182 - Instance allocator support. Add and import instance accept a
3183 ``--iallocator`` parameter, and call that instance allocator to decide
3184 which node to use for the instance. The iallocator document describes
3185 what's expected from an allocator script.
3186 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3187 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3188 node is lost there is still enough memory to fail over the instances
3190 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3191 ``gnt-cluster verify``, to check for site-specific compliance. All the
3192 hooks will run, and their output, if any, will be displayed. Any
3193 failing hook will make the verification return an error value.
3194 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3195 primary and secondary interfaces
3196 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3197 that is still declared as part of the cluster and has failed.
3198 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3199 specifying output fields, that just adds the chosen fields to the
3201 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3202 export from the filesystem.
3203 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3204 have been added. Using them you can enable/disable acpi and pae
3205 support, and specify a path for a cd image to be exported to the
3206 instance. These parameters as the name suggest only work on HVM
3208 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3209 PAE support will be set to the previously hardcoded values, but the
3210 (previously hardcoded) path to the CDROM ISO image will be unset and
3211 if required, needs to be set manually with ``gnt-instance modify``
3213 - The address to which an instance's VNC console is bound is now
3214 selectable per-instance, rather than being cluster wide. Of course
3215 this only applies to instances controlled via VNC, so currently just
3216 applies to HVM clusters.
3222 *(Released Mon, 18 Feb 2008)*
3224 - more tweaks to the disk activation code (especially helpful for DRBD)
3225 - change the default ``gnt-instance list`` output format, now there is
3226 one combined status field (see the manpage for the exact values this
3228 - some more fixes for the mac export to hooks change
3229 - make Ganeti not break with DRBD 8.2.x (which changed the version
3230 format in ``/proc/drbd``) (issue 24)
3231 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3232 template, allowing migration from DRBD0.7+MD to DRBD8
3238 *(Released Wed, 30 Jan 2008)*
3240 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3242 - add command aliases infrastructure and a few aliases
3243 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3244 man pages and the ``--help`` option of ``gnt-node
3245 list``/``gnt-instance list``
3246 - fix ``gnt-backup list`` with down nodes (issue 21)
3247 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3248 - fix the dist archive and add a check for including svn/git files in
3250 - some developer-related changes: improve the burnin and the QA suite,
3251 add an upload script for testing during development
3257 *(Released Wed, 16 Jan 2008)*
3259 - experimental HVM support, read the install document, section
3260 "Initializing the cluster"
3261 - allow for the PVM hypervisor per-instance kernel and initrd paths
3262 - add a new command ``gnt-cluster verify-disks`` which uses a new
3263 algorithm to improve the reconnection of the DRBD pairs if the device
3264 on the secondary node has gone away
3265 - make logical volume code auto-activate LVs at disk activation time
3266 - slightly improve the speed of activating disks
3267 - allow specification of the MAC address at instance creation time, and
3268 changing it later via ``gnt-instance modify``
3269 - fix handling of external commands that generate lots of output on
3271 - update documentation with regard to minimum version of DRBD8 supported
3277 *(Released Tue, 4 Dec 2007)*
3279 - Log the ``xm create`` output to the node daemon log on failure (to
3280 help diagnosing the error)
3281 - In debug mode, log all external commands output if failed to the logs
3282 - Change parsing of lvm commands to ignore stderr
3288 *(Released Wed, 28 Nov 2007)*
3290 - Another round of updates to the DRBD 8 code to deal with more failures
3291 in the replace secondary node operation
3292 - Some more logging of failures in disk operations (lvm, drbd)
3293 - A few documentation updates
3300 *(Released Tue, 13 Nov 2007)*
3302 - Change configuration file format from Python's Pickle to JSON.
3303 Upgrading is possible using the cfgupgrade utility.
3304 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3305 faster replace disks and is more stable (DRBD 8 has many improvements
3306 compared to DRBD 0.7)
3307 - Added command line tags support (see man pages for ``gnt-instance``,
3308 ``gnt-node``, ``gnt-cluster``)
3309 - Added instance rename support
3310 - Added multi-instance startup/shutdown
3311 - Added cluster rename support
3312 - Added ``gnt-node evacuate`` to simplify some node operations
3313 - Added instance reboot operation that can speedup reboot as compared to
3315 - Soften the requirement that hostnames are in FQDN format
3316 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3318 - Removed dependency on debian's patched fping that uses the
3319 non-standard ``-S`` option
3320 - Now the OS definitions are searched for in multiple, configurable
3321 paths (easier for distros to package)
3322 - Some changes to the hooks infrastructure (especially the new
3323 post-configuration update hook)
3324 - Other small bugfixes
3326 .. vim: set textwidth=72 syntax=rst :