8 *(Released Tue, 17 Dec 2013)*
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 - All known issues in 2.10.0 beta1 have been resolved (see changes from
79 - Improve handling of KVM runtime files from earlier Ganeti versions
82 Inherited from the 2.9 branch:
84 - use custom KVM path if set for version checking
85 - SingleNotifyPipeCondition: don't share pollers
87 Inherited from the 2.8 branch:
89 - Fixed Luxi daemon socket permissions after master-failover
90 - Improve IP version detection code directly checking for colons rather than
91 passing the family from the cluster object
92 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
93 opportunistically anymore (Issue 622)
94 - Allow link local IPv6 gateways (Issue 624)
95 - Fix error printing (Issue 616)
96 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
97 disk modifications, keep the old one. If name=none then set disk name to
99 - Update build_chroot script to work with the latest hackage packages
100 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
101 - Fix evacuation out of drained node (Issue 615)
102 - Add default file_driver if missing (Issue 571)
103 - Fix job error message after unclean master shutdown (Issue 618)
104 - Lock group(s) when creating instances (Issue 621)
105 - SetDiskID() before accepting an instance (Issue 633)
106 - Allow the ext template disks to receive arbitrary parameters, both at creation
107 time and while being modified
108 - Xen handle domain shutdown (future proofing cherry-pick)
109 - Refactor reading live data in htools (future proofing cherry-pick)
115 *(Released Wed, 27 Nov 2013)*
117 This was the first beta release of the 2.10 series. All important changes
118 are listed in the latest 2.10 entry.
123 The following issues are known to be present in the beta and will be fixed
126 - Issue 477: Wrong permissions for confd LUXI socket
127 - Issue 621: Instance related opcodes do not aquire network/group locks
128 - Issue 622: Assertion Error: Node locks differ from node resource locks
129 - Issue 623: IPv6 Masterd <-> Luxid communication error
135 *(Released Mon, 27 Jan 2014)*
137 - Ensure that all the hypervisors exist in the config file (Issue 640)
138 - Correctly recognise the role as master node (Issue 687)
139 - configure: allow detection of Sphinx 1.2+ (Issue 502)
140 - gnt-instance now honors the KVM path correctly (Issue 691)
142 Inherited from the 2.8 branch:
144 - Change the list separator for the usb_devices parameter from comma to space.
145 Commas could not work because they are already the hypervisor option
146 separator (Issue 649)
147 - Add support for blktap2 file-driver (Issue 638)
148 - Add network tag definitions to the haskell codebase (Issue 641)
149 - Fix RAPI network tag handling
150 - Add the network tags to the tags searched by gnt-cluster search-tags
151 - Fix caching bug preventing jobs from being cancelled
152 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
158 *(Released Fri, 13 Dec 2013)*
160 - use custom KVM path if set for version checking
161 - SingleNotifyPipeCondition: don't share pollers
163 Inherited from the 2.8 branch:
165 - Fixed Luxi daemon socket permissions after master-failover
166 - Improve IP version detection code directly checking for colons rather than
167 passing the family from the cluster object
168 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
169 opportunistically anymore (Issue 622)
170 - Allow link local IPv6 gateways (Issue 624)
171 - Fix error printing (Issue 616)
172 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
173 disk modifications, keep the old one. If name=none then set disk name to
175 - Update build_chroot script to work with the latest hackage packages
176 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
177 - Fix evacuation out of drained node (Issue 615)
178 - Add default file_driver if missing (Issue 571)
179 - Fix job error message after unclean master shutdown (Issue 618)
180 - Lock group(s) when creating instances (Issue 621)
181 - SetDiskID() before accepting an instance (Issue 633)
182 - Allow the ext template disks to receive arbitrary parameters, both at creation
183 time and while being modified
184 - Xen handle domain shutdown (future proofing cherry-pick)
185 - Refactor reading live data in htools (future proofing cherry-pick)
191 *(Released Wed, 13 Nov 2013)*
193 - fix bug, that kept nodes offline when readding
194 - when verifying DRBD versions, ignore unavailable nodes
195 - fix bug that made the console unavailable on kvm in split-user
197 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
203 *(Released Tue, 5 Nov 2013)*
205 Incompatible/important changes
206 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
208 - hroller now also plans for capacity to move non-redundant instances off
209 any node to be rebooted; the old behavior of completely ignoring any
210 non-redundant instances can be restored by adding the --ignore-non-redundant
212 - The cluster option '--no-lvm-storage' was removed in favor of the new option
213 '--enabled-disk-templates'.
214 - On instance creation, disk templates no longer need to be specified
215 with '-t'. The default disk template will be taken from the list of
216 enabled disk templates.
217 - The monitoring daemon is now running as root, in order to be able to collect
218 information only available to root (such as the state of Xen instances).
219 - The ConfD client is now IPv6 compatible.
220 - File and shared file storage is no longer dis/enabled at configure time,
221 but using the option '--enabled-disk-templates' at cluster initialization and
223 - The default directories for file and shared file storage are not anymore
224 specified at configure time, but taken from the cluster's configuration.
225 They can be set at cluster initialization and modification with
226 '--file-storage-dir' and '--shared-file-storage-dir'.
227 - Cluster verification now includes stricter checks regarding the
228 default file and shared file storage directories. It now checks that
229 the directories are explicitely allowed in the 'file-storage-paths' file and
230 that the directories exist on all nodes.
231 - The list of allowed disk templates in the instance policy and the list
232 of cluster-wide enabled disk templates is now checked for consistency
233 on cluster or group modification. On cluster initialization, the ipolicy
234 disk templates are ensured to be a subset of the cluster-wide enabled
240 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
241 the correct command syntax. It is possible to use different DRBD versions
242 on different nodes as long as they are compatible to each other. This
243 enables rolling upgrades of DRBD with no downtime. As permanent operation
244 of different DRBD versions within a node group is discouraged,
245 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
246 - New "inst-status-xen" data collector for the monitoring daemon, providing
247 information about the state of the xen instances on the nodes.
248 - New "lv" data collector for the monitoring daemon, collecting data about the
249 logical volumes on the nodes, and pairing them with the name of the instances
251 - New "diskstats" data collector, collecting the data from /proc/diskstats and
252 presenting them over the monitoring daemon interface.
253 - The ConfD client is now IPv6 compatible.
257 The following new dependencies have been added.
261 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
262 for the unit tests (and only used for testing).
266 - ``hslogger`` (http://software.complete.org/hslogger) is now always
267 required, even if confd is not enabled.
272 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
274 - Improved error messsages (inherited from stable-2.8)
280 *(Released Tue, 15 Oct 2013)*
282 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
284 - in implicit configuration upgrade, match ipolicy with enabled disk templates
285 - improved harep documentation (inherited from stable-2.8)
291 *(Released Wed, 9 Oct 2013)*
293 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
295 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
296 at least one DRBD instance.
297 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
304 *(Released Tue, 1 Oct 2013)*
306 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
309 - update of the documentation, in particular installation instructions
310 - merging of LD_* constants into DT_* constants
311 - python style changes to be compatible with newer versions of pylint
317 *(Released Thu, 29 Aug 2013)*
319 This was the first beta release of the 2.9 series. All important changes
320 are listed in the latest 2.9 entry.
326 *(Released Thu, 23 Jan 2014)*
328 - Change the list separator for the usb_devices parameter from comma to space.
329 Commas could not work because they are already the hypervisor option
330 separator (Issue 649)
331 - Add support for blktap2 file-driver (Issue 638)
332 - Add network tag definitions to the haskell codebase (Issue 641)
333 - Fix RAPI network tag handling
334 - Add the network tags to the tags searched by gnt-cluster search-tags
335 - Fix caching bug preventing jobs from being cancelled
336 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
342 *(Released Thu, 12 Dec 2013)*
344 - Fixed Luxi daemon socket permissions after master-failover
345 - Improve IP version detection code directly checking for colons rather than
346 passing the family from the cluster object
347 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
348 opportunistically anymore (Issue 622)
349 - Allow link local IPv6 gateways (Issue 624)
350 - Fix error printing (Issue 616)
351 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
352 disk modifications, keep the old one. If name=none then set disk name to
354 - Update build_chroot script to work with the latest hackage packages
355 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
356 - Fix evacuation out of drained node (Issue 615)
357 - Add default file_driver if missing (Issue 571)
358 - Fix job error message after unclean master shutdown (Issue 618)
359 - Lock group(s) when creating instances (Issue 621)
360 - SetDiskID() before accepting an instance (Issue 633)
361 - Allow the ext template disks to receive arbitrary parameters, both at creation
362 time and while being modified
363 - Xen handle domain shutdown (future proofing cherry-pick)
364 - Refactor reading live data in htools (future proofing cherry-pick)
370 *(Released Thu, 07 Nov 2013)*
372 - DRBD: ensure peers are UpToDate for dual-primary
373 - Improve error message for replace-disks
374 - More dependency checks at configure time
375 - Placate warnings on ganeti.outils_unittest.py
381 *(Released Thu, 17 Oct 2013)*
383 - Correctly start/stop luxid during gnt-cluster master-failover
384 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
385 - Fix path for the job queue serial file
386 - Improved harep man page
387 - Minor documentation improvements
393 *(Released Mon, 30 Sep 2013)*
395 Incompatible/important changes
396 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
398 - Instance policy can contain multiple instance specs, as described in
399 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
400 <design-partitioned>`. As a consequence, it's not possible to partially change
401 or override instance specs. Bounding specs (min and max) can be specified as a
402 whole using the new option ``--ipolicy-bounds-specs``, while standard
403 specs use the new option ``--ipolicy-std-specs``.
404 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
405 gnt-instance is a valid YAML object.
406 - hail now honors network restrictions when allocating nodes. This led to an
407 update of the IAllocator protocol. See the IAllocator documentation for
409 - confd now only answers static configuration request over the network. luxid
410 was extracted, listens on the local LUXI socket and responds to live queries.
411 This allows finer grained permissions if using separate users.
416 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
417 to always require authentication, ``--require-authentication``. It can
418 be specified in ``$sysconfdir/default/ganeti``.
419 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
420 be used to manage the disk templates to be used by instances in the cluster.
421 Initially, it will be set to a list that includes plain, drbd, if they were
422 enabled by specifying a volume group name, and file and sharedfile, if those
423 were enabled at configure time. Additionally, it will include all disk
424 templates that are currently used by instances. The order of disk templates
425 will be based on Ganeti's history of supporting them. In the future, the
426 first entry of the list will be used as a default disk template on instance
428 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
429 configuration back to the previous stable version.
430 - Disk templates in group ipolicy can be restored to the default value.
431 - Initial support for diskless instances and virtual clusters in QA.
432 - More QA and unit tests for instance policies.
433 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
434 describing why the opcode itself was executed.
435 - The monitoring daemon is now available. It allows users to query the cluster
436 for obtaining information about the status of the system. The daemon is only
437 responsible for providing the information over the network: the actual data
438 gathering is performed by data collectors (currently, only the DRBD status
439 collector is available).
440 - In order to help developers work on Ganeti, a new script
441 (``devel/build_chroot``) is provided, for building a chroot that contains all
442 the required development libraries and tools for compiling Ganeti on a Debian
444 - A new tool, ``harep``, for performing self-repair and recreation of instances
445 in Ganeti has been added.
446 - Split queries are enabled for tags, network, exports, cluster info, groups,
448 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
449 It prints the command line to set the current policies, to ease
451 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
452 devices for KVM virtio-net interfaces will get created with VNET_HDR
453 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
454 virtio-net interfaces, which prevents host kernel tainting and log
455 flooding, when dealing with broken or malicious virtio-net drivers.
456 It's set to true by default.
457 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
459 - Support 'viridian' parameter in Xen HVM
460 - Support DSA SSH keys in bootstrap
461 - To simplify the work of packaging frameworks that want to add the needed users
462 and groups in a split-user setup themselves, at build time three files in
463 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
464 the groups to be generated, the ``users`` file contains, one per line, the
465 users to be generated, optionally followed by their primary group, where
466 important. The ``groupmemberships`` file contains, one per line, additional
467 user-group membership relations that need to be established. The syntax of
468 these files will remain stable in all future versions.
473 The following new dependencies have been added:
476 - The ``curl`` library is not optional anymore for compiling the Haskell code.
477 - ``snap-server`` library (if monitoring is enabled).
480 - The minimum Python version needed to run Ganeti is now 2.6.
481 - ``yaml`` library (only for running the QA).
485 - Perform proper cleanup on termination of Haskell daemons
486 - Fix corner-case in handling of remaining retry time
492 *(Released Tue, 17 Sep 2013)*
494 - To simplify the work of packaging frameworks that want to add the needed users
495 and groups in a split-user setup themselves, at build time three files in
496 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
497 the groups to be generated, the ``users`` file contains, one per line, the
498 users to be generated, optionally followed by their primary group, where
499 important. The ``groupmemberships`` file contains, one per line, additional
500 user-group membership relations that need to be established. The syntax of
501 these files will remain stable in all future versions.
502 - Add a default to file-driver when unspecified over RAPI (Issue 571)
503 - Mark the DSA host pubkey as optional, and remove it during config downgrade
505 - Some documentation fixes
511 *(Released Tue, 27 Aug 2013)*
513 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
515 - Support 'viridian' parameter in Xen HVM (Issue 233)
516 - Include VCS version in ``gnt-cluster version``
517 - Support DSA SSH keys in bootstrap (Issue 338)
518 - Fix batch creation of instances
519 - Use FQDN to check master node status (Issue 551)
520 - Make the DRBD collector more failure-resilient
526 *(Released Fri, 2 Aug 2013)*
528 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
530 - Fix upgrading/downgrading from 2.7
531 - Increase maximum RAPI message size
532 - Documentation updates
533 - Split ``confd`` between ``luxid`` and ``confd``
534 - Merge 2.7 series up to the 2.7.1 release
535 - Allow the ``modify_etc_hosts`` option to be changed
536 - Add better debugging for ``luxid`` queries
537 - Expose bulk parameter for GetJobs in RAPI client
538 - Expose missing ``network`` fields in RAPI
539 - Add some ``cluster verify`` tests
540 - Some unittest fixes
541 - Fix a malfunction in ``hspace``'s tiered allocation
542 - Fix query compatibility between haskell and python implementations
543 - Add the ``vnet_hdr`` HV parameter for KVM
544 - Add ``--cleanup`` to instance failover
545 - Change the connected groups format in ``gnt-network info`` output; it
546 was previously displayed as a raw list by mistake. (Merged from 2.7)
552 *(Released Mon, 24 Jun 2013)*
554 This was the first beta release of the 2.8 series. All important changes
555 are listed in the latest 2.8 entry.
561 *(Released Thu, 26 Sep 2013)*
563 - Change the connected groups format in ``gnt-network info`` output; it
564 was previously displayed as a raw list by mistake
565 - Check disk template in right dict when copying
566 - Support multi-instance allocs without iallocator
567 - Fix some errors in the documentation
568 - Fix formatting of tuple in an error message
574 *(Released Thu, 25 Jul 2013)*
576 - Add logrotate functionality in daemon-util
577 - Add logrotate example file
578 - Add missing fields to network queries over rapi
579 - Fix network object timestamps
580 - Add support for querying network timestamps
581 - Fix a typo in the example crontab
582 - Fix a documentation typo
588 *(Released Thu, 04 Jul 2013)*
590 Incompatible/important changes
591 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
593 - Instance policies for disk size were documented to be on a per-disk
594 basis, but hail applied them to the sum of all disks. This has been
596 - ``hbal`` will now exit with status 0 if, during job execution over
597 LUXI, early exit has been requested and all jobs are successful;
598 before, exit status 1 was used, which cannot be differentiated from
600 - Compatibility with newer versions of rbd has been fixed
601 - ``gnt-instance batch-create`` has been changed to use the bulk create
602 opcode from Ganeti. This lead to incompatible changes in the format of
603 the JSON file. It's now not a custom dict anymore but a dict
604 compatible with the ``OpInstanceCreate`` opcode.
605 - Parent directories for file storage need to be listed in
606 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
607 write the file automatically based on old configuration values, but it
608 can not distribute it across all nodes and the file contents should be
609 verified. Use ``gnt-cluster copyfile
610 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
611 upgraded. The reason for requiring this list of paths now is that
612 before it would have been possible to inject new paths via RPC,
613 allowing files to be created in arbitrary locations. The RPC protocol
614 is protected using SSL/X.509 certificates, but as a design principle
615 Ganeti does not permit arbitrary paths to be passed.
616 - The parsing of the variants file for OSes (see
617 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
618 lines and comment lines (starting with ``#``) are ignored for better
620 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
621 longer available. ``gnt-node add`` now invokes a new tool on the
622 destination node, named ``prepare-node-join``, to configure the SSH
623 daemon. Paramiko is no longer necessary to configure nodes' SSH
624 daemons via ``gnt-node add``.
625 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
626 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
627 the flag is not reset after a master failover.
628 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
629 is recommended. 2.6 will be mandatory from the 2.8 series.
635 - New network management functionality to support automatic allocation
636 of IP addresses and managing of network parameters. See
637 :manpage:`gnt-network(8)` for more details.
638 - New external storage backend, to allow managing arbitrary storage
639 systems external to the cluster. See
640 :manpage:`ganeti-extstorage-interface(7)`.
641 - New ``exclusive-storage`` node parameter added, restricted to
642 nodegroup level. When it's set to true, physical disks are assigned in
643 an exclusive fashion to instances, as documented in :doc:`Partitioned
644 Ganeti <design-partitioned>`. Currently, only instances using the
645 ``plain`` disk template are supported.
646 - The KVM hypervisor has been updated with many new hypervisor
647 parameters, including a generic one for passing arbitrary command line
648 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
649 compatible up to qemu 1.4.
650 - A new tool, called ``mon-collector``, is the stand-alone executor of
651 the data collectors for a monitoring system. As of this version, it
652 just includes the DRBD data collector, that can be executed by calling
653 ``mon-collector`` using the ``drbd`` parameter. See
654 :manpage:`mon-collector(7)`.
655 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
656 for RAPI users. It allows granting permissions to query for
657 information to a specific user without giving
658 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
659 - A new tool named ``node-cleanup`` has been added. It cleans remains of
660 a cluster from a machine by stopping all daemons, removing
661 certificates and ssconf files. Unless the ``--no-backup`` option is
662 given, copies of the certificates are made.
663 - Instance creations now support the use of opportunistic locking,
664 potentially speeding up the (parallel) creation of multiple instances.
665 This feature is currently only available via the :doc:`RAPI
666 <rapi>` interface and when an instance allocator is used. If the
667 ``opportunistic_locking`` parameter is set the opcode will try to
668 acquire as many locks as possible, but will not wait for any locks
669 held by other opcodes. If not enough resources can be found to
670 allocate the instance, the temporary error code
671 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
672 retried thereafter, with or without opportunistic locking.
673 - New experimental linux-ha resource scripts.
674 - Restricted-commands support: ganeti can now be asked (via command line
675 or rapi) to perform commands on a node. These are passed via ganeti
676 RPC rather than ssh. This functionality is restricted to commands
677 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
678 reasons. The file is not copied automatically.
684 - Diskless instances are now externally mirrored (Issue 237). This for
685 now has only been tested in conjunction with explicit target nodes for
687 - Queries not needing locks or RPC access to the node can now be
688 performed by the confd daemon, making them independent from jobs, and
689 thus faster to execute. This is selectable at configure time.
690 - The functionality for allocating multiple instances at once has been
691 overhauled and is now also available through :doc:`RAPI <rapi>`.
693 There are no significant changes from version 2.7.0~rc3.
699 *(Released Tue, 25 Jun 2013)*
701 - Fix permissions on the confd query socket (Issue 477)
702 - Fix permissions on the job archive dir (Issue 498)
703 - Fix handling of an internal exception in replace-disks (Issue 472)
704 - Fix gnt-node info handling of shortened names (Issue 497)
705 - Fix gnt-instance grow-disk when wiping is enabled
706 - Documentation improvements, and support for newer pandoc
707 - Fix hspace honoring ipolicy for disks (Issue 484)
708 - Improve handling of the ``kvm_extra`` HV parameter
714 *(Released Fri, 24 May 2013)*
716 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
718 - Disks added through ``gnt-instance modify`` or created through
719 ``gnt-instance recreate-disks`` are wiped, if the
720 ``prealloc_wipe_disks`` flag is set.
721 - If wiping newly created disks fails, the disks are removed. Also,
722 partial failures in creating disks through ``gnt-instance modify``
723 triggers a cleanup of the partially-created disks.
724 - Removing the master IP address doesn't fail if the address has been
726 - Fix ownership of the OS log dir
727 - Workaround missing SO_PEERCRED constant (Issue 191)
733 *(Released Fri, 3 May 2013)*
735 This was the first release candidate of the 2.7 series. Since beta3:
737 - Fix kvm compatibility with qemu 1.4 (Issue 389)
738 - Documentation updates (admin guide, upgrade notes, install
739 instructions) (Issue 372)
740 - Fix gnt-group list nodes and instances count (Issue 436)
741 - Fix compilation without non-mandatory libraries (Issue 441)
742 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
743 - Make confd logging more verbose at INFO level (Issue 435)
744 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
745 - Fix failure path for instance storage type conversion (Issue 229)
746 - Update htools text backend documentation
747 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
748 - Disable inter-cluster instance move for file-based instances, because
749 it is dependant on instance export, which is not supported for
750 file-based instances. (Issue 414)
751 - Fix gnt-job crashes on non-ascii characters (Issue 427)
752 - Fix volume group checks on non-vm-capable nodes (Issue 432)
758 *(Released Mon, 22 Apr 2013)*
760 This was the third beta release of the 2.7 series. Since beta2:
762 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
763 - Fix data loss on wrong usage of ``gnt-instance move``
764 - Properly export errors in confd-based job queries
765 - Add ``users-setup`` tool
766 - Fix iallocator protocol to report 0 as a disk size for diskless
767 instances. This avoids hail breaking when a diskless instance is
769 - Fix job queue directory permission problem that made confd job queries
770 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
771 for access to archived jobs (Issue 406).
772 - Limit the sizes of networks supported by ``gnt-network`` to something
773 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
774 - Fix bugs in instance disk template conversion
775 - Fix GHC 7 compatibility
776 - Fix ``burnin`` install path (Issue 426).
777 - Allow very small disk grows (Issue 347).
778 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
779 that noded doesn't import masterd code (Issue 419).
780 - Make sure the default metavg at cluster init is the same as the vg, if
781 unspecified (Issue 358).
782 - Fix cleanup of partially created disks (part of Issue 416)
788 *(Released Tue, 2 Apr 2013)*
790 This was the second beta release of the 2.7 series. Since beta1:
792 - Networks no longer have a "type" slot, since this information was
793 unused in Ganeti: instead of it tags should be used.
794 - The rapi client now has a ``target_node`` option to MigrateInstance.
795 - Fix early exit return code for hbal (Issue 386).
796 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
797 - Fix ``rbd showmapped`` output parsing (Issue 312).
798 - Networks are now referenced indexed by UUID, rather than name. This
799 will require running cfgupgrade, from 2.7.0beta1, if networks are in
801 - The OS environment now includes network information.
802 - Deleting of a network is now disallowed if any instance nic is using
803 it, to prevent dangling references.
804 - External storage is now documented in man pages.
805 - The exclusive_storage flag can now only be set at nodegroup level.
806 - Hbal can now submit an explicit priority with its jobs.
807 - Many network related locking fixes.
808 - Bump up the required pylint version to 0.25.1.
809 - Fix the ``no_remember`` option in RAPI client.
810 - Many ipolicy related tests, qa, and fixes.
811 - Many documentation improvements and fixes.
812 - Fix building with ``--disable-file-storage``.
813 - Fix ``-q`` option in htools, which was broken if passed more than
815 - Some haskell/python interaction improvements and fixes.
816 - Fix iallocator in case of missing LVM storage.
817 - Fix confd config load in case of ``--no-lvm-storage``.
818 - The confd/query functionality is now mentioned in the security
825 *(Released Wed, 6 Feb 2013)*
827 This was the first beta release of the 2.7 series. All important changes
828 are listed in the latest 2.7 entry.
834 *(Released Fri, 21 Dec 2012)*
836 Important behaviour change: hbal won't rebalance anymore instances which
837 have the ``auto_balance`` attribute set to false. This was the intention
838 all along, but until now it only skipped those from the N+1 memory
839 reservation (DRBD-specific).
841 A significant number of bug fixes in this release:
843 - Fixed disk adoption interaction with ipolicy checks.
844 - Fixed networking issues when instances are started, stopped or
845 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
846 - Fixed the warning in cluster verify for shared storage instances not
848 - Fixed removal of storage directory on shared file storage (issue 262).
849 - Fixed validation of LVM volume group name in OpClusterSetParams
850 (``gnt-cluster modify``) (issue 285).
851 - Fixed runtime memory increases (``gnt-instance modify -m``).
852 - Fixed live migration under Xen's ``xl`` mode.
853 - Fixed ``gnt-instance console`` with ``xl``.
854 - Fixed building with newer Haskell compiler/libraries.
855 - Fixed PID file writing in Haskell daemons (confd); this prevents
856 restart issues if confd was launched manually (outside of
857 ``daemon-util``) while another copy of it was running
858 - Fixed a type error when doing live migrations with KVM (issue 297) and
859 the error messages for failing migrations have been improved.
860 - Fixed opcode validation for the out-of-band commands (``gnt-node
862 - Fixed a type error when unsetting OS hypervisor parameters (issue
863 311); now it's possible to unset all OS-specific hypervisor
865 - Fixed the ``dry-run`` mode for many operations: verification of
866 results was over-zealous but didn't take into account the ``dry-run``
867 operation, resulting in "wrong" failures.
868 - Fixed bash completion in ``gnt-job list`` when the job queue has
869 hundreds of entries; especially with older ``bash`` versions, this
870 results in significant CPU usage.
872 And lastly, a few other improvements have been made:
874 - Added option to force master-failover without voting (issue 282).
875 - Clarified error message on lock conflict (issue 287).
876 - Logging of newly submitted jobs has been improved (issue 290).
877 - Hostname checks have been made uniform between instance rename and
879 - The ``--submit`` option is now supported by ``gnt-debug delay``.
880 - Shutting down the master daemon by sending SIGTERM now stops it from
881 processing jobs waiting for locks; instead, those jobs will be started
882 once again after the master daemon is started the next time (issue
884 - Support for Xen's ``xl`` program has been improved (besides the fixes
886 - Reduced logging noise in the Haskell confd daemon (only show one log
887 entry for each config reload, instead of two).
888 - Several man page updates and typo fixes.
894 *(Released Fri, 12 Oct 2012)*
896 A small bugfix release. Among the bugs fixed:
898 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
899 made the command unusable.
900 - Commands that issue many jobs don't fail anymore just because some jobs
901 take so long that other jobs are archived.
902 - Failures during ``gnt-instance reinstall`` are reflected by the exit
904 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
906 - When ``always_failover`` is set, ``--allow-failover`` is not required
907 in migrate commands anymore.
908 - ``bash_completion`` works even if extglob is disabled.
909 - Fixed bug with locks that made failover for RDB-based instances fail.
910 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
911 a random node instead of one based on the allocator metric.
912 - Support for newer versions of pylint and pep8.
913 - Hail doesn't fail anymore when trying to add an instance of type
914 ``file``, ``sharedfile`` or ``rbd``.
915 - Added new Makefile target to rebuild the whole distribution, so that
916 all files are included.
922 *(Released Fri, 27 Jul 2012)*
925 .. attention:: The ``LUXI`` protocol has been made more consistent
926 regarding its handling of command arguments. This, however, leads to
927 incompatibility issues with previous versions. Please ensure that you
928 restart Ganeti daemons soon after the upgrade, otherwise most
929 ``LUXI`` calls (job submission, setting/resetting the drain flag,
930 pausing/resuming the watcher, cancelling and archiving jobs, querying
931 the cluster configuration) will fail.
940 The current ``admin_up`` field, which used to denote whether an instance
941 should be running or not, has been removed. Instead, ``admin_state`` is
942 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
944 The rational behind this is that an instance being “down” can have
947 - it could be down during a reboot
948 - it could be temporarily be down for a reinstall
949 - or it could be down because it is deprecated and kept just for its
952 The previous Boolean state was making it difficult to do capacity
953 calculations: should Ganeti reserve memory for a down instance? Now, the
954 tri-state field makes it clear:
956 - in ``up`` and ``down`` state, all resources are reserved for the
957 instance, and it can be at any time brought up if it is down
958 - in ``offline`` state, only disk space is reserved for it, but not
961 The field can have an extra use: since the transition between ``up`` and
962 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
963 transition between ``offline`` and ``down`` is done via ``gnt-instance
964 modify``, it is possible to given different rights to users. For
965 example, owners of an instance could be allowed to start/stop it, but
966 not transition it out of the offline state.
968 Instance policies and specs
969 +++++++++++++++++++++++++++
971 In previous Ganeti versions, an instance creation request was not
972 limited on the minimum size and on the maximum size just by the cluster
973 resources. As such, any policy could be implemented only in third-party
974 clients (RAPI clients, or shell wrappers over ``gnt-*``
975 tools). Furthermore, calculating cluster capacity via ``hspace`` again
976 required external input with regards to instance sizes.
978 In order to improve these workflows and to allow for example better
979 per-node group differentiation, we introduced instance specs, which
982 - minimum instance disk size, disk count, memory size, cpu count
983 - maximum values for the above metrics
984 - and “standard” values (used in ``hspace`` to calculate the standard
987 The minimum/maximum values can be also customised at node-group level,
988 for example allowing more powerful hardware to support bigger instance
991 Beside the instance specs, there are a few other settings belonging to
992 the instance policy framework. It is possible now to customise, per
993 cluster and node-group:
995 - the list of allowed disk templates
996 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
997 - the maximum ratio of instance to spindles (see below for more
998 information) for local storage
1000 All these together should allow all tools that talk to Ganeti to know
1001 what are the ranges of allowed values for instances and the
1002 over-subscription that is allowed.
1004 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
1005 instance configuration, and the physical CPU configuration from the
1006 node. For the spindle ratios however, we didn't track before these
1007 values, so new parameters have been added:
1009 - a new node parameter ``spindle_count``, defaults to 1, customisable at
1010 node group or node level
1011 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
1013 Note that spindles in this context doesn't need to mean actual
1014 mechanical hard-drives; it's just a relative number for both the node
1015 I/O capacity and instance I/O consumption.
1017 Instance migration behaviour
1018 ++++++++++++++++++++++++++++
1020 While live-migration is in general desirable over failover, it is
1021 possible that for some workloads it is actually worse, due to the
1022 variable time of the “suspend” phase during live migration.
1024 To allow the tools to work consistently over such instances (without
1025 having to hard-code instance names), a new backend parameter
1026 ``always_failover`` has been added to control the migration/failover
1027 behaviour. When set to True, all migration requests for an instance will
1028 instead fall-back to failover.
1030 Instance memory ballooning
1031 ++++++++++++++++++++++++++
1033 Initial support for memory ballooning has been added. The memory for an
1034 instance is no longer fixed (backend parameter ``memory``), but instead
1035 can vary between minimum and maximum values (backend parameters
1036 ``minmem`` and ``maxmem``). Currently we only change an instance's
1039 - live migrating or failing over and instance and the target node
1040 doesn't have enough memory
1041 - user requests changing the memory via ``gnt-instance modify
1044 Instance CPU pinning
1045 ++++++++++++++++++++
1047 In order to control the use of specific CPUs by instance, support for
1048 controlling CPU pinning has been added for the Xen, HVM and LXC
1049 hypervisors. This is controlled by a new hypervisor parameter
1050 ``cpu_mask``; details about possible values for this are in the
1051 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
1052 VCPU-to-CPU mapping) form will work well only when all nodes in your
1053 cluster have the same amount of CPUs.
1058 Another area in which Ganeti was not customisable were the parameters
1059 used for storage configuration, e.g. how many stripes to use for LVM,
1060 DRBD resync configuration, etc.
1062 To improve this area, we've added disks parameters, which are
1063 customisable at cluster and node group level, and which allow to
1064 specify various parameters for disks (DRBD has the most parameters
1065 currently), for example:
1067 - DRBD resync algorithm and parameters (e.g. speed)
1068 - the default VG for meta-data volumes for DRBD
1069 - number of stripes for LVM (plain disk template)
1072 These parameters can be modified via ``gnt-cluster modify -D …`` and
1073 ``gnt-group modify -D …``, and are used at either instance creation (in
1074 case of LVM stripes, for example) or at disk “activation” time
1075 (e.g. resync speed).
1077 Rados block device support
1078 ++++++++++++++++++++++++++
1080 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
1081 denoted by the ``rbd`` disk template type. This is considered
1082 experimental, feedback is welcome. For details on configuring it, see
1083 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
1088 The existing master IP functionality works well only in simple setups (a
1089 single network shared by all nodes); however, if nodes belong to
1090 different networks, then the ``/32`` setup and lack of routing
1091 information is not enough.
1093 To allow the master IP to function well in more complex cases, the
1094 system was reworked as follows:
1096 - a master IP netmask setting has been added
1097 - the master IP activation/turn-down code was moved from the node daemon
1098 to a separate script
1099 - whether to run the Ganeti-supplied master IP script or a user-supplied
1100 on is a ``gnt-cluster init`` setting
1102 Details about the location of the standard and custom setup scripts are
1103 in the man page :manpage:`gnt-cluster(8)`; for information about the
1104 setup script protocol, look at the Ganeti-supplied script.
1109 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
1112 It is now possible to use TLS-protected connections, and when renewing
1113 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
1114 it is now possible to specify spice or spice CA certificates. Also, it
1115 is possible to configure a password for SPICE sessions via the
1116 hypervisor parameter ``spice_password_file``.
1118 There are also new parameters to control the compression and streaming
1119 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1120 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1121 for the spice parameters.
1123 Lastly, it is now possible to see the SPICE connection information via
1124 ``gnt-instance console``.
1129 A new tool (``tools/ovfconverter``) has been added that supports
1130 conversion between Ganeti and the `Open Virtualization Format
1131 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1134 This relies on the ``qemu-img`` tool to convert the disk formats, so the
1135 actual compatibility with other virtualization solutions depends on it.
1137 Confd daemon changes
1138 ++++++++++++++++++++
1140 The configuration query daemon (``ganeti-confd``) is now optional, and
1141 has been rewritten in Haskell; whether to use the daemon at all, use the
1142 Python (default) or the Haskell version is selectable at configure time
1143 via the ``--enable-confd`` parameter, which can take one of the
1144 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1145 daemon will result in a smaller footprint; for larger systems, we
1146 welcome feedback on the Haskell version which might become the default
1149 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1150 daemon running. The Python version doesn't implement the new call.
1153 User interface changes
1154 ~~~~~~~~~~~~~~~~~~~~~~
1156 We have replaced the ``--disks`` option of ``gnt-instance
1157 replace-disks`` with a more flexible ``--disk`` option, which allows
1158 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1159 disk size and mode can be changed upon recreation (via ``gnt-instance
1160 recreate-disks``, which accepts the same ``--disk`` option).
1162 As many people are used to a ``show`` command, we have added that as an
1163 alias to ``info`` on all ``gnt-*`` commands.
1165 The ``gnt-instance grow-disk`` command has a new mode in which it can
1166 accept the target size of the disk, instead of the delta; this can be
1167 more safe since two runs in absolute mode will be idempotent, and
1168 sometimes it's also easier to specify the desired size directly.
1170 Also the handling of instances with regard to offline secondaries has
1171 been improved. Instance operations should not fail because one of it's
1172 secondary nodes is offline, even though it's safe to proceed.
1174 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1175 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1176 minors to instance name.
1181 RAPI coverage has improved, with (for example) new resources for
1182 recreate-disks, node power-cycle, etc.
1187 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1190 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1191 whether to still support Python 2.4 or move to Python 2.6 as minimum
1194 Support for Fedora has been slightly improved; the provided example
1195 init.d script should work better on it and the INSTALL file should
1196 document the needed dependencies.
1201 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1202 ``Query(what=QR_LOCK, ...)`` instead.
1204 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1205 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1206 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1207 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1208 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1210 RAPI client: ``CertificateError`` now derives from
1211 ``GanetiApiError``. This should make it more easy to handle Ganeti
1214 Deprecation warnings due to PyCrypto/paramiko import in
1215 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1216 make sure to run an up-to-date paramiko version, if you use this tool.
1218 The QA scripts now depend on Python 2.5 or above (the main code base
1219 still works with Python 2.4).
1221 The configuration file (``config.data``) is now written without
1222 indentation for performance reasons; if you want to edit it, it can be
1223 re-formatted via ``tools/fmtjson``.
1225 A number of bugs has been fixed in the cluster merge tool.
1227 ``x509`` certification verification (used in import-export) has been
1228 changed to allow the same clock skew as permitted by the cluster
1229 verification. This will remove some rare but hard to diagnose errors in
1236 *(Released Thu, 19 Jul 2012)*
1238 Very few changes from rc4 to the final release, only bugfixes:
1240 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1241 instance, fix CDROM booting for KVM instances)
1242 - fixed node group modification of node parameters
1243 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1244 - fixed generation of bash completion to ensure a stable ordering
1251 *(Released Fri, 13 Jul 2012)*
1253 Third release candidate for 2.6. The following changes were done from
1256 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1257 - Fixed an inconsistency in the LUXI protocol with the provided
1258 arguments (NOT backwards compatible)
1259 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1260 the cluster ``std`` value
1261 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1262 easier instance debugging on nodes (requires ``hconfd`` to work)
1268 *(Released Tue, 03 Jul 2012)*
1270 Second release candidate for 2.6. The following changes were done from
1273 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1275 - Fixed a RAPI regression on missing beparams/memory
1276 - Fixed redistribution of files on offline nodes
1277 - Added possibility to run activate-disks even though secondaries are
1278 offline. With this change it relaxes also the strictness on some other
1279 commands which use activate disks internally:
1280 * ``gnt-instance start|reboot|rename|backup|export``
1281 - Made it possible to remove safely an instance if its secondaries are
1283 - Made it possible to reinstall even though secondaries are offline
1289 *(Released Mon, 25 Jun 2012)*
1291 First release candidate for 2.6. The following changes were done from
1294 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1296 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1297 behaviour. It supports now the modify operation
1298 - ``hcheck`` implemented to analyze cluster health and possibility of
1299 improving health by rebalance
1300 - ``hbal`` has been improved in dealing with split instances
1306 *(Released Mon, 11 Jun 2012)*
1308 Second beta release of 2.6. The following changes were done from beta2
1311 - Fixed ``daemon-util`` with non-root user models
1312 - Fixed creation of plain instances with ``--no-wait-for-sync``
1313 - Fix wrong iv_names when running ``cfgupgrade``
1314 - Export more information in RAPI group queries
1315 - Fixed bug when changing instance network interfaces
1316 - Extended burnin to do NIC changes
1317 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1318 - Changed default for DRBD barriers
1319 - Fixed DRBD error reporting for syncer rate
1320 - Verify the options on disk parameters
1322 And of course various fixes to documentation and improved unittests and
1329 *(Released Wed, 23 May 2012)*
1331 First beta release of 2.6. The following changes were done from beta1 to
1334 - integrated patch for distributions without ``start-stop-daemon``
1335 - adapted example init.d script to work on Fedora
1336 - fixed log handling in Haskell daemons
1337 - adapted checks in the watcher for pycurl linked against libnss
1338 - add partial support for ``xl`` instead of ``xm`` for Xen
1339 - fixed a type issue in cluster verification
1340 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1343 Plus integrated fixes from the 2.5 branch:
1345 - fixed ``kvm-ifup`` to use ``/bin/bash``
1346 - fixed parallel build failures
1347 - KVM live migration when using a custom keymap
1353 *(Released Tue, 24 Jul 2012)*
1355 A small bugfix release, with no new features:
1357 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1358 different default shell (e.g. Debian, Ubuntu)
1359 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1360 243 and Debian bug #650664)
1361 - fixed compatibility with KVM versions that don't support multiple boot
1362 devices (fixes Issue 230 and Debian bug #624256)
1364 Additionally, a few fixes were done to the build system (fixed parallel
1365 build failures) and to the unittests (fixed race condition in test for
1366 FileID functions, and the default enable/disable mode for QA test is now
1373 *(Released Fri, 11 May 2012)*
1375 A small bugfix release.
1377 The main issues solved are on the topic of compatibility with newer LVM
1380 - fixed parsing of ``lv_attr`` field
1381 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1382 the ``--force`` flag is needed
1384 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1385 to accept kernel 3.x too (was hardcoded to 2.6.*).
1387 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1388 of hook results and that also made display of other errors suboptimal
1389 was fixed; the code behaves now like 2.4 and earlier.
1391 Another change in 2.5, the cleanup of the OS scripts environment, is too
1392 aggressive: it removed even the ``PATH`` variable, which requires the OS
1393 scripts to *always* need to export it. Since this is a bit too strict,
1394 we now export a minimal PATH, the same that we export for hooks.
1396 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1397 integrated into this release.
1399 Finally, a few other miscellaneous changes were done (no new features,
1400 just small improvements):
1402 - Fix ``gnt-group --help`` display
1403 - Fix hardcoded Xen kernel path
1404 - Fix grow-disk handling of invalid units
1405 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1406 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1412 *(Released Thu, 12 Apr 2012)*
1414 Incompatible/important changes and bugfixes
1415 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1417 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1418 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1419 to match the underlying LUXI interface.
1420 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1421 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1422 not maintain backwards-compatibility as the underlying operation
1423 changed in an incompatible way. The RAPI client can talk to old
1424 servers, but it needs to be told so as the return value changed.
1425 - When creating file-based instances via RAPI, the ``file_driver``
1426 parameter no longer defaults to ``loop`` and must be specified.
1427 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1429 - Support for the undocumented and deprecated RAPI instance creation
1430 request format version 0 has been dropped. Use version 1, supported
1431 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1432 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1433 documentation <install>`.
1434 - The "cluster-verify" hooks are now executed per group by the
1435 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1436 if you just run ``gnt-cluster verify``, which generates one opcode per
1438 - The environment as passed to the OS scripts is cleared, and thus no
1439 environment variables defined in the node daemon's environment will be
1440 inherited by the scripts.
1441 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1443 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1444 support operations involving multiple node groups.
1445 - Offline nodes are ignored when failing over an instance.
1446 - Support for KVM version 1.0, which changed the version reporting format
1448 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1450 - ``Makefile`` is now compatible with Automake 1.11.2
1451 - Includes all bugfixes made in the 2.4 series
1456 - The ganeti-htools project has been merged into the ganeti-core source
1457 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1458 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1459 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1460 GPT-style partition tables (requires `parted
1461 <http://www.gnu.org/s/parted/>`_).
1462 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1463 - Allowed adding tags on instance creation.
1464 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1465 - Allow instances to be started in a paused state, enabling the user to
1466 see the complete console output on boot using the console.
1467 - Added new hypervisor flag to control default reboot behaviour
1468 (``reboot_behavior``).
1469 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1470 - Improved out-of-band management support:
1472 - Added ``gnt-node health`` command reporting the health status of
1474 - Added ``gnt-node power`` command to manage power status of nodes.
1475 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1477 - Instance migration can fall back to failover if instance is not
1479 - Filters can be used when listing nodes, instances, groups and locks;
1480 see :manpage:`ganeti(7)` manpage.
1481 - Added post-execution status as variables to :doc:`hooks <hooks>`
1483 - Instance tags are exported/imported together with the instance.
1484 - When given an explicit job ID, ``gnt-job info`` will work for archived
1486 - Jobs can define dependencies on other jobs (not yet supported via
1487 RAPI or command line, but used by internal commands and usable via
1490 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1493 - Instance failover is now available as a RAPI resource
1494 (``/2/instances/[instance_name]/failover``).
1495 - ``gnt-instance info`` defaults to static information if primary node
1497 - Opcodes have a new ``comment`` attribute.
1498 - Added basic SPICE support to KVM hypervisor.
1499 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1501 Node group improvements
1502 ~~~~~~~~~~~~~~~~~~~~~~~
1504 - ``gnt-cluster verify`` has been modified to check groups separately,
1505 thereby improving performance.
1506 - Node group support has been added to ``gnt-cluster verify-disks``,
1507 which now operates per node group.
1508 - Watcher has been changed to work better with node groups.
1510 - One process and state file per node group.
1511 - Slow watcher in one group doesn't block other group's watcher.
1513 - Added new command, ``gnt-group evacuate``, to move all instances in a
1514 node group to other groups.
1515 - Added ``gnt-instance change-group`` to move an instance to another
1517 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1518 per-group operations.
1519 - Node groups can be tagged.
1520 - Some operations switch from an exclusive to a shared lock as soon as
1522 - Instance's primary and secondary nodes' groups are now available as
1523 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1524 and ``snodes.group.uuid``).
1529 - Numerous updates to documentation and manpages.
1531 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1533 - Some opcode/job results are now also documented, see :doc:`RAPI
1536 - A lockset's internal lock is now also visible in lock monitor.
1537 - Log messages from job queue workers now contain information about the
1538 opcode they're processing.
1539 - ``gnt-instance console`` no longer requires the instance lock.
1540 - A short delay when waiting for job changes reduces the number of LUXI
1541 requests significantly.
1542 - DRBD metadata volumes are overwritten with zeros during disk creation.
1543 - Out-of-band commands no longer acquire the cluster lock in exclusive
1545 - ``devel/upload`` now uses correct permissions for directories.
1551 *(Released Fri, 23 Mar 2012)*
1553 This was the sixth release candidate of the 2.5 series.
1559 *(Released Mon, 9 Jan 2012)*
1561 This was the fifth release candidate of the 2.5 series.
1567 *(Released Thu, 27 Oct 2011)*
1569 This was the fourth release candidate of the 2.5 series.
1575 *(Released Wed, 26 Oct 2011)*
1577 This was the third release candidate of the 2.5 series.
1583 *(Released Tue, 18 Oct 2011)*
1585 This was the second release candidate of the 2.5 series.
1591 *(Released Tue, 4 Oct 2011)*
1593 This was the first release candidate of the 2.5 series.
1599 *(Released Wed, 31 Aug 2011)*
1601 This was the third beta release of the 2.5 series.
1607 *(Released Mon, 22 Aug 2011)*
1609 This was the second beta release of the 2.5 series.
1615 *(Released Fri, 12 Aug 2011)*
1617 This was the first beta release of the 2.5 series.
1623 *(Released Thu, 27 Oct 2011)*
1625 - Fixed bug when parsing command line parameter values ending in
1627 - Fixed assertion error after unclean master shutdown
1628 - Disable HTTP client pool for RPC, significantly reducing memory usage
1630 - Fixed queue archive creation with wrong permissions
1636 *(Released Tue, 23 Aug 2011)*
1640 - Fixed documentation for importing with ``--src-dir`` option
1641 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1642 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1648 *(Released Fri, 5 Aug 2011)*
1650 Many bug-fixes and a few small features:
1652 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1653 issues when you tried to add an instance with two MAC addresses in one
1655 - KVM: fixed per-instance stored UID value
1656 - KVM: configure bridged NICs at migration start
1657 - KVM: Fix a bug where instance will not start with never KVM versions
1659 - Added OS search path to ``gnt-cluster info``
1660 - Fixed an issue with ``file_storage_dir`` where you were forced to
1661 provide an absolute path, but the documentation states it is a
1662 relative path, the documentation was right
1663 - Added a new parameter to instance stop/start called ``--no-remember``
1664 that will make the state change to not be remembered
1665 - Implemented ``no_remember`` at RAPI level
1666 - Improved the documentation
1667 - Node evacuation: don't call IAllocator if node is already empty
1668 - Fixed bug in DRBD8 replace disks on current nodes
1669 - Fixed bug in recreate-disks for DRBD instances
1670 - Moved assertion checking locks in ``gnt-instance replace-disks``
1671 causing it to abort with not owning the right locks for some situation
1672 - Job queue: Fixed potential race condition when cancelling queued jobs
1673 - Fixed off-by-one bug in job serial generation
1674 - ``gnt-node volumes``: Fix instance names
1675 - Fixed aliases in bash completion
1676 - Fixed a bug in reopening log files after being sent a SIGHUP
1677 - Added a flag to burnin to allow specifying VCPU count
1678 - Bugfixes to non-root Ganeti configuration
1684 *(Released Thu, 12 May 2011)*
1686 Many bug-fixes and a few new small features:
1688 - Fixed a bug related to log opening failures
1689 - Fixed a bug in instance listing with orphan instances
1690 - Fixed a bug which prevented resetting the cluster-level node parameter
1691 ``oob_program`` to the default
1692 - Many fixes related to the ``cluster-merge`` tool
1693 - Fixed a race condition in the lock monitor, which caused failures
1694 during (at least) creation of many instances in parallel
1695 - Improved output for gnt-job info
1696 - Removed the quiet flag on some ssh calls which prevented debugging
1698 - Improved the N+1 failure messages in cluster verify by actually
1699 showing the memory values (needed and available)
1700 - Increased lock attempt timeouts so that when executing long operations
1701 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1702 too early and thus prevent the use of the 'fair' mechanism
1703 - Changed instance query data (``gnt-instance info``) to not acquire
1704 locks unless needed, thus allowing its use on locked instance if only
1705 static information is asked for
1706 - Improved behaviour with filesystems that do not support rename on an
1708 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1709 kept locks on all nodes during the wipe, which is unneeded
1710 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1711 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1712 10GiB) which caused the wipe to fail right at the end in some cases
1713 - Fixed master IP activation when doing master failover with no-voting
1714 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1715 the master node itself
1716 - Fixed potential data-loss in under disk full conditions, where Ganeti
1717 wouldn't check correctly the return code and would consider
1718 partially-written files 'correct'
1719 - Fixed bug related to multiple VGs and DRBD disk replacing
1720 - Added new disk parameter ``metavg`` that allows placement of the meta
1721 device for DRBD in a different volume group
1722 - Fixed error handling in the node daemon when the system libc doesn't
1723 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1724 - Fixed lock release during replace-disks, which kept cluster-wide locks
1725 when doing disk replaces with an iallocator script
1726 - Added check for missing bridges in cluster verify
1727 - Handle EPIPE errors while writing to the terminal better, so that
1728 piping the output to e.g. ``less`` doesn't cause a backtrace
1729 - Fixed rare case where a ^C during Luxi calls could have been
1730 interpreted as server errors, instead of simply terminating
1731 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1733 - Added a few more parameters to the KVM hypervisor, allowing a second
1734 CDROM, custom disk type for CDROMs and a floppy image
1735 - Removed redundant message in instance rename when the name is given
1737 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1738 disks on new nodes, allowing recreation when the original instance
1739 nodes are completely gone
1740 - Added option when converting disk templates to DRBD to skip waiting
1741 for the resync, in order to make the instance available sooner
1742 - Added two new variables to the OS scripts environment (containing the
1744 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1745 to allow use of ``pvgrub`` as bootloader
1746 - Changed the instance memory modifications to only check out-of-memory
1747 conditions on memory increases, and turned the secondary node warnings
1748 into errors (they can still be overridden via ``--force``)
1749 - Fixed the handling of a corner case when the Python installation gets
1750 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1751 to execute a command that doesn't exist
1752 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1753 node of the instance returned failures during instance shutdown; this
1754 adds the option ``--ignore-consistency`` to gnt-instance move
1756 And as usual, various improvements to the error messages, documentation
1763 *(Released Wed, 09 Mar 2011)*
1765 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1766 the RAPI users file if run twice (even with ``--dry-run``).
1768 The release fixes that bug (nothing else changed).
1774 *(Released Mon, 07 Mar 2011)*
1776 Final 2.4.0 release. Just a few small fixes:
1778 - Fixed RAPI node evacuate
1779 - Fixed the kvm-ifup script
1780 - Fixed internal error handling for special job cases
1781 - Updated man page to specify the escaping feature for options
1787 *(Released Mon, 28 Feb 2011)*
1789 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1790 that this feature wiped the disks of the wrong instance, leading to loss
1795 - Fixed title of query field containing instance name
1796 - Expanded the glossary in the documentation
1797 - Fixed one unittest (internal issue)
1803 *(Released Mon, 21 Feb 2011)*
1805 A number of bug fixes plus just a couple functionality changes.
1807 On the user-visible side, the ``gnt-* list`` command output has changed
1808 with respect to "special" field states. The current rc1 style of display
1809 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1810 the default output mode special fields states are displayed as follows:
1812 - Offline resource: ``*``
1813 - Unavailable/not applicable: ``-``
1814 - Data missing (RPC failure): ``?``
1815 - Unknown field: ``??``
1817 Another user-visible change is the addition of ``--force-join`` to
1822 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1823 - Fixed regression in RAPI/instance reinstall where all parameters were
1824 required (instead of optional)
1825 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1826 - Fixed iallocator usage (offline nodes were not considered offline)
1827 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1828 - Fixed hypervisor and OS parameter validation with respect to
1829 non-vm_capable nodes
1830 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1832 - Fixed ``tools/listrunner`` with respect to agent-based usage
1838 *(Released Fri, 4 Feb 2011)*
1840 Many changes and fixes since the beta1 release. While there were some
1841 internal changes, the code has been mostly stabilised for the RC
1844 Note: the dumb allocator was removed in this release, as it was not kept
1845 up-to-date with the IAllocator protocol changes. It is recommended to
1846 use the ``hail`` command from the ganeti-htools package.
1848 Note: the 2.4 and up versions of Ganeti are not compatible with the
1849 0.2.x branch of ganeti-htools. You need to upgrade to
1850 ganeti-htools-0.3.0 (or later).
1852 Regressions fixed from 2.3
1853 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1855 - Fixed the ``gnt-cluster verify-disks`` command
1856 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1858 - Fixed disk adoption breakage
1859 - Fixed wrong headers in instance listing for field aliases
1864 - Fixed corner case in KVM handling of NICs
1865 - Fixed many cases of wrong handling of non-vm_capable nodes
1866 - Fixed a bug where a missing instance symlink was not possible to
1867 recreate with any ``gnt-*`` command (now ``gnt-instance
1868 activate-disks`` does it)
1869 - Fixed the volume group name as reported by ``gnt-cluster
1871 - Increased timeouts for the import-export code, hopefully leading to
1872 fewer aborts due network or instance timeouts
1873 - Fixed bug in ``gnt-node list-storage``
1874 - Fixed bug where not all daemons were started on cluster
1875 initialisation, but only at the first watcher run
1876 - Fixed many bugs in the OOB implementation
1877 - Fixed watcher behaviour in presence of instances with offline
1879 - Fixed instance list output for instances running on the wrong node
1880 - a few fixes to the cluster-merge tool, but it still cannot merge
1881 multi-node groups (currently it is not recommended to use this tool)
1887 - Improved network configuration for the KVM hypervisor
1888 - Added e1000 as a supported NIC for Xen-HVM
1889 - Improved the lvmstrap tool to also be able to use partitions, as
1890 opposed to full disks
1891 - Improved speed of disk wiping (the cluster parameter
1892 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1893 of instance creations
1894 - Added documentation for the OS parameters
1895 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1896 hypervisor is not responding
1897 - Added display of blacklisted and hidden OS information in
1898 ``gnt-cluster info``
1899 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1900 NIC and node parameters, which might create problems with currently
1901 invalid (but undetected) configuration files, but prevents validation
1902 failures when unrelated parameters are modified
1903 - Changed cluster initialisation to wait for the master daemon to become
1905 - Expanded the RAPI interface:
1907 - Added config redistribution resource
1908 - Added activation/deactivation of instance disks
1909 - Added export of console information
1911 - Implemented log file reopening on SIGHUP, which allows using
1912 logrotate(8) for the Ganeti log files
1913 - Added a basic OOB helper script as an example
1919 *(Released Fri, 14 Jan 2011)*
1924 - Fixed timezone issues when formatting timestamps
1925 - Added support for node groups, available via ``gnt-group`` and other
1927 - Added out-of-band framework and management, see :doc:`design
1928 document <design-oob>`
1929 - Removed support for roman numbers from ``gnt-node list`` and
1930 ``gnt-instance list``.
1931 - Allowed modification of master network interface via ``gnt-cluster
1932 modify --master-netdev``
1933 - Accept offline secondaries while shutting down instance disks
1934 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1935 - Added support for multiple LVM volume groups
1936 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1938 - Added commands to list available fields:
1940 - ``gnt-node list-fields``
1941 - ``gnt-group list-fields``
1942 - ``gnt-instance list-fields``
1944 - Updated documentation and man pages
1949 - Moved ``rapi_users`` file into separate directory, now named
1950 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1952 - Added new tool for running commands on many machines,
1953 ``tools/ganeti-listrunner``
1954 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1955 improving the ``gnt-instance console`` output
1956 - Allowed customisation of disk index separator at ``configure`` time
1957 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1958 - Added support for non-partitioned md disks in ``lvmstrap``
1959 - Added script to gracefully power off KVM instances
1960 - Split ``utils`` module into smaller parts
1961 - Changed query operations to return more detailed information, e.g.
1962 whether an information is unavailable due to an offline node. To use
1963 this new functionality, the LUXI call ``Query`` must be used. Field
1964 information is now stored by the master daemon and can be retrieved
1965 using ``QueryFields``. Instances, nodes and groups can also be queried
1966 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1967 exposed via RAPI). The following commands make use of this
1968 infrastructure change:
1970 - ``gnt-group list``
1971 - ``gnt-group list-fields``
1973 - ``gnt-node list-fields``
1974 - ``gnt-instance list``
1975 - ``gnt-instance list-fields``
1976 - ``gnt-debug locks``
1981 - New RAPI resources (see :doc:`rapi`):
1985 - ``/2/groups/[group_name]``
1986 - ``/2/groups/[group_name]/assign-nodes``
1987 - ``/2/groups/[group_name]/modify``
1988 - ``/2/groups/[group_name]/rename``
1989 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1993 - Implemented ``no_install`` for instance creation
1994 - Implemented OS parameters for instance reinstallation, allowing
1995 use of special settings on reinstallation (e.g. for preserving data)
2000 - Added IPv6 support in import/export
2001 - Pause DRBD synchronization while wiping disks on instance creation
2002 - Updated unittests and QA scripts
2003 - Improved network parameters passed to KVM
2004 - Converted man pages from docbook to reStructuredText
2010 *(Released Mon, 20 Dec 2010)*
2012 Released version 2.3.1~rc1 without any changes.
2018 *(Released Wed, 1 Dec 2010)*
2020 - impexpd: Disable OpenSSL compression in socat if possible (backport
2021 from master, commit e90739d625b, see :doc:`installation guide
2022 <install-quick>` for details)
2023 - Changed unittest coverage report to exclude test scripts
2024 - Added script to check version format
2030 *(Released Wed, 1 Dec 2010)*
2032 Released version 2.3.0~rc1 without any changes.
2038 *(Released Fri, 19 Nov 2010)*
2040 A number of bugfixes and documentation updates:
2042 - Update ganeti-os-interface documentation
2043 - Fixed a bug related to duplicate MACs or similar items which should be
2045 - Fix breakage in OS state modify
2046 - Reinstall instance: disallow offline secondaries (fixes bug related to
2047 OS changing but reinstall failing)
2048 - plus all the other fixes between 2.2.1 and 2.2.2
2054 *(Released Tue, 2 Nov 2010)*
2056 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
2057 - Fixed master failover race with watcher
2058 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
2060 - Accept previously stopped instance for export with instance removal
2061 - Simplify and extend the environment variables for instance OS scripts
2062 - Added new node flags, ``master_capable`` and ``vm_capable``
2063 - Added optional instance disk wiping prior during allocation. This is a
2064 cluster-wide option and can be set/modified using
2065 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
2066 - Added IPv6 support, see :doc:`design document <design-2.3>` and
2067 :doc:`install-quick`
2068 - Added a new watcher option (``--ignore-pause``)
2069 - Added option to ignore offline node on instance start/stop
2070 (``--ignore-offline``)
2071 - Allow overriding OS parameters with ``gnt-instance reinstall``
2072 - Added ability to change node's secondary IP address using ``gnt-node
2074 - Implemented privilege separation for all daemons except
2075 ``ganeti-noded``, see ``configure`` options
2076 - Complain if an instance's disk is marked faulty in ``gnt-cluster
2078 - Implemented job priorities (see ``ganeti(7)`` manpage)
2079 - Ignore failures while shutting down instances during failover from
2081 - Exit daemon's bootstrap process only once daemon is ready
2082 - Export more information via ``LUInstanceQuery``/remote API
2083 - Improved documentation, QA and unittests
2084 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
2085 restart if the file was created or changed
2086 - Added LUXI protocol version sent with each request and response,
2087 allowing detection of server/client mismatches
2088 - Moved the Python scripts among gnt-* and ganeti-* into modules
2089 - Moved all code related to setting up SSH to an external script,
2091 - Infrastructure changes for node group support in future versions
2097 *(Released Fri, 19 Nov 2010)*
2099 A few small bugs fixed, and some improvements to the build system:
2101 - Fix documentation regarding conversion to drbd
2102 - Fix validation of parameters in cluster modify (``gnt-cluster modify
2104 - Fix error handling in node modify with multiple changes
2105 - Allow remote imports without checked names
2111 *(Released Tue, 19 Oct 2010)*
2113 - Disable SSL session ID cache in RPC client
2119 *(Released Thu, 14 Oct 2010)*
2121 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
2122 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2124 - Fix problems with interaction between Curl and Python's HTTP server,
2125 resulting in increased speed in many RPC calls
2126 - Improve our release script to prevent breakage with older aclocal and
2133 *(Released Thu, 7 Oct 2010)*
2135 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2136 value retrieved from master
2137 - Added support for blacklisted or hidden OS definitions
2138 - Added simple lock monitor (accessible via (``gnt-debug locks``)
2139 - Added support for -mem-path in KVM hypervisor abstraction layer
2140 - Allow overriding instance parameters in tool for inter-cluster
2141 instance moves (``tools/move-instance``)
2142 - Improved opcode summaries (e.g. in ``gnt-job list``)
2143 - Improve consistency of OS listing by sorting it
2144 - Documentation updates
2150 *(Released Fri, 8 Oct 2010)*
2152 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2158 *(Released Mon, 4 Oct 2010)*
2160 - Fixed regression in ``gnt-instance rename``
2166 *(Released Wed, 22 Sep 2010)*
2168 - Fixed OS_VARIANT variable for OS scripts
2169 - Fixed cluster tag operations via RAPI
2170 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2171 - Disabled RAPI CA checks in watcher
2177 *(Released Mon, 23 Aug 2010)*
2179 - Support DRBD versions of the format "a.b.c.d"
2181 - Re-introduce support for usage from multiple threads in RAPI client
2182 - Instance renames and modify via RAPI
2183 - Work around race condition between processing and archival in job
2185 - Mark opcodes following failed one as failed, too
2186 - Job field ``lock_status`` was removed due to difficulties making it
2187 work with the changed job queue in Ganeti 2.2; a better way to monitor
2188 locks is expected for a later 2.2.x release
2189 - Fixed dry-run behaviour with many commands
2190 - Support ``ssh-agent`` again when adding nodes
2191 - Many additional bugfixes
2197 *(Released Fri, 30 Jul 2010)*
2199 Important change: the internal RPC mechanism between Ganeti nodes has
2200 changed from using a home-grown http library (based on the Python base
2201 libraries) to use the PycURL library. This requires that PycURL is
2202 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2203 against GnuTLS by default. cURL's support for GnuTLS had known issues
2204 before cURL 7.21.0 and we recommend using the latest cURL release or
2205 linking against OpenSSL. Most other distributions already link PycURL
2206 and cURL against OpenSSL. The command::
2208 python -c 'import pycurl; print pycurl.version'
2210 can be used to determine the libraries PycURL and cURL are linked
2213 Other significant changes:
2215 - Rewrote much of the internals of the job queue, in order to achieve
2216 better parallelism; this decouples job query operations from the job
2217 processing, and it should allow much nicer behaviour of the master
2218 daemon under load, and it also has uncovered some long-standing bugs
2219 related to the job serialisation (now fixed)
2220 - Added a default iallocator setting to the cluster parameters,
2221 eliminating the need to always pass nodes or an iallocator for
2222 operations that require selection of new node(s)
2223 - Added experimental support for the LXC virtualization method
2224 - Added support for OS parameters, which allows the installation of
2225 instances to pass parameter to OS scripts in order to customise the
2227 - Added a hypervisor parameter controlling the migration type (live or
2228 non-live), since hypervisors have various levels of reliability; this
2229 has renamed the 'live' parameter to 'mode'
2230 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2231 logical volumes, meaning that cluster verify will ignore them and not
2232 flag their presence as errors
2233 - The watcher will now reset the error count for failed instances after
2234 8 hours, thus allowing self-healing if the problem that caused the
2235 instances to be down/fail to start has cleared in the meantime
2236 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2237 check for, and warn, if the drbd module parameter ``usermode_helper``
2238 is not consistent with the cluster-wide setting; this is needed to
2239 make diagnose easier of failed drbd creations
2240 - Started adding base IPv6 support, but this is not yet
2241 enabled/available for use
2242 - Rename operations (cluster, instance) will now return the new name,
2243 which is especially useful if a short name was passed in
2244 - Added support for instance migration in RAPI
2245 - Added a tool to pre-configure nodes for the SSH setup, before joining
2246 them to the cluster; this will allow in the future a simplified model
2247 for node joining (but not yet fully enabled in 2.2); this needs the
2248 paramiko python library
2249 - Fixed handling of name-resolving errors
2250 - Fixed consistency of job results on the error path
2251 - Fixed master-failover race condition when executed multiple times in
2253 - Fixed many bugs related to the job queue (mostly introduced during the
2254 2.2 development cycle, so not all are impacting 2.1)
2255 - Fixed instance migration with missing disk symlinks
2256 - Fixed handling of unknown jobs in ``gnt-job archive``
2257 - And many other small fixes/improvements
2261 - Enhanced both the unittest and the QA coverage
2262 - Switched the opcode validation to a generic model, and extended the
2263 validation to all opcode parameters
2264 - Changed more parts of the code that write shell scripts to use the
2266 - Switched the master daemon to use the asyncore library for the Luxi
2273 *(Released Thu, 17 Jun 2010)*
2275 - Added tool (``move-instance``) and infrastructure to move instances
2276 between separate clusters (see :doc:`separate documentation
2277 <move-instance>` and :doc:`design document <design-2.2>`)
2278 - Added per-request RPC timeout
2279 - RAPI now requires a Content-Type header for requests with a body (e.g.
2280 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2281 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2282 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2284 - Implemented initial support for running Ganeti daemons as separate
2285 users, see configure-time flags ``--with-user-prefix`` and
2286 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2288 - Instances can be removed after export (``gnt-backup export
2289 --remove-instance``)
2290 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2291 key (instead of 1024 bit)
2292 - Added new cluster configuration file for cluster domain secret
2293 - Import/export now use SSL instead of SSH
2294 - Added support for showing estimated time when exporting an instance,
2295 see the ``ganeti-os-interface(7)`` manpage and look for
2302 *(Released Tue, 16 Nov 2010)*
2304 Some more bugfixes. Unless critical bugs occur, this will be the last
2307 - Fix case of MAC special-values
2308 - Fix mac checker regex
2309 - backend: Fix typo causing "out of range" error
2310 - Add missing --units in gnt-instance list man page
2316 *(Released Tue, 24 Aug 2010)*
2319 - Don't ignore secondary node silently on non-mirrored disk templates
2321 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2322 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2323 - Properly document the usb_mouse parameter
2324 - Fix path in ganeti-rapi(8) (issue 116)
2325 - Adjust error message when the ganeti user's .ssh directory is
2327 - Add same-node-check when changing the disk template to drbd
2333 *(Released Fri, 16 Jul 2010)*
2336 - Add an option to only select some reboot types during qa/burnin.
2337 (on some hypervisors consequent reboots are not supported)
2338 - Fix infrequent race condition in master failover. Sometimes the old
2339 master ip address would be still detected as up for a short time
2340 after it was removed, causing failover to fail.
2341 - Decrease mlockall warnings when the ctypes module is missing. On
2342 Python 2.4 we support running even if no ctypes module is installed,
2343 but we were too verbose about this issue.
2344 - Fix building on old distributions, on which man doesn't have a
2346 - Fix RAPI not to ignore the MAC address on instance creation
2347 - Implement the old instance creation format in the RAPI client.
2353 *(Released Thu, 01 Jul 2010)*
2355 A small bugfix release:
2356 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2357 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2358 a non-existing option
2359 - Fix instance create: the --force-variant option was ignored
2360 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2361 - Fix modify node storage with non-FQDN arguments
2362 - Fix RAPI client to authenticate under Python 2.6 when used
2363 for more than 5 requests needing authentication
2364 - Fix gnt-instance modify -t (storage) giving a wrong error message
2365 when converting a non-shutdown drbd instance to plain
2371 *(Released Fri, 18 Jun 2010)*
2373 A small bugfix release:
2375 - Fix live migration of KVM instances started with older Ganeti
2376 versions which had fewer hypervisor parameters
2377 - Fix gnt-instance grow-disk on down instances
2378 - Fix an error-reporting bug during instance migration
2379 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2380 silently ignoring broken ones
2381 - Fix an RPC error reporting bug affecting, for example, RAPI client
2383 - Fix bug triggered by different API version os-es on different nodes
2384 - Fix a bug in instance startup with custom hvparams: OS level
2385 parameters would fail to be applied.
2386 - Fix the RAPI client under Python 2.6 (but more work is needed to
2387 make it work completely well with OpenSSL)
2388 - Fix handling of errors when resolving names from DNS
2394 *(Released Thu, 3 Jun 2010)*
2396 A medium sized development cycle. Some new features, and some
2397 fixes/small improvements/cleanups.
2399 Significant features
2400 ~~~~~~~~~~~~~~~~~~~~
2402 The node deamon now tries to mlock itself into memory, unless the
2403 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2404 its logs, and falls back to console logging. This allows emergency
2405 features such as ``gnt-node powercycle`` to work even in the event of a
2406 broken node disk (tested offlining the disk hosting the node's
2407 filesystem and dropping its memory caches; don't try this at home)
2409 KVM: add vhost-net acceleration support. It can be tested with a new
2410 enough version of the kernel and of qemu-kvm.
2412 KVM: Add instance chrooting feature. If you use privilege dropping for
2413 your VMs you can also now force them to chroot to an empty directory,
2414 before starting the emulated guest.
2416 KVM: Add maximum migration bandwith and maximum downtime tweaking
2417 support (requires a new-enough version of qemu-kvm).
2419 Cluster verify will now warn if the master node doesn't have the master
2420 ip configured on it.
2422 Add a new (incompatible) instance creation request format to RAPI which
2423 supports all parameters (previously only a subset was supported, and it
2424 wasn't possible to extend the old format to accomodate all the new
2425 features. The old format is still supported, and a client can check for
2426 this feature, before using it, by checking for its presence in the
2427 ``features`` RAPI resource.
2429 Now with ancient latin support. Try it passing the ``--roman`` option to
2430 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2431 (requires the python-roman module to be installed, in order to work).
2436 As usual many internal code refactorings, documentation updates, and
2439 - Lots of improvements and cleanups to the experimental Remote API
2440 (RAPI) client library.
2441 - A new unit test suite for the core daemon libraries.
2442 - A fix to creating missing directories makes sure the umask is not
2443 applied anymore. This enforces the same directory permissions
2445 - Better handling terminating daemons with ctrl+c (used when running
2446 them in debugging mode).
2447 - Fix a race condition in live migrating a KVM instance, when stat()
2448 on the old proc status file returned EINVAL, which is an unexpected
2450 - Fixed manpage checking with newer man and utf-8 charachters. But now
2451 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2457 *(Released Fri, 7 May 2010)*
2459 Fix a bug which prevented untagged KVM instances from starting.
2465 *(Released Fri, 7 May 2010)*
2467 Another release with a long development cycle, during which many
2468 different features were added.
2470 Significant features
2471 ~~~~~~~~~~~~~~~~~~~~
2473 The KVM hypervisor now can run the individual instances as non-root, to
2474 reduce the impact of a VM being hijacked due to bugs in the
2475 hypervisor. It is possible to run all instances as a single (non-root)
2476 user, to manually specify a user for each instance, or to dynamically
2477 allocate a user out of a cluster-wide pool to each instance, with the
2478 guarantee that no two instances will run under the same user ID on any
2481 An experimental RAPI client library, that can be used standalone
2482 (without the other Ganeti libraries), is provided in the source tree as
2483 ``lib/rapi/client.py``. Note this client might change its interface in
2484 the future, as we iterate on its capabilities.
2486 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2487 replace the cluster's certificates and crypto keys. This might help in
2488 case they have been compromised, or have simply expired.
2490 A new disk option for instance creation has been added that allows one
2491 to "adopt" currently existing logical volumes, with data
2492 preservation. This should allow easier migration to Ganeti from
2493 unmanaged (or managed via other software) instances.
2495 Another disk improvement is the possibility to convert between redundant
2496 (DRBD) and plain (LVM) disk configuration for an instance. This should
2497 allow better scalability (starting with one node and growing the
2498 cluster, or shrinking a two-node cluster to one node).
2500 A new feature that could help with automated node failovers has been
2501 implemented: if a node sees itself as offline (by querying the master
2502 candidates), it will try to shutdown (hard) all instances and any active
2503 DRBD devices. This reduces the risk of duplicate instances if an
2504 external script automatically failovers the instances on such nodes. To
2505 enable this, the cluster parameter ``maintain_node_health`` should be
2506 enabled; in the future this option (per the name) will enable other
2507 automatic maintenance features.
2509 Instance export/import now will reuse the original instance
2510 specifications for all parameters; that means exporting an instance,
2511 deleting it and the importing it back should give an almost identical
2512 instance. Note that the default import behaviour has changed from
2513 before, where it created only one NIC; now it recreates the original
2516 Cluster verify has added a few new checks: SSL certificates validity,
2517 /etc/hosts consistency across the cluster, etc.
2522 As usual, many internal changes were done, documentation fixes,
2525 - Fixed cluster initialization with disabled cluster storage (regression
2526 introduced in 2.1.1)
2527 - File-based storage supports growing the disks
2528 - Fixed behaviour of node role changes
2529 - Fixed cluster verify for some corner cases, plus a general rewrite of
2530 cluster verify to allow future extension with more checks
2531 - Fixed log spamming by watcher and node daemon (regression introduced
2533 - Fixed possible validation issues when changing the list of enabled
2535 - Fixed cleanup of /etc/hosts during node removal
2536 - Fixed RAPI response for invalid methods
2537 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2538 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2539 from ide disks, etc.)
2540 - Allow OS changes without re-installation (to record a changed OS
2541 outside of Ganeti, or to allow OS renames)
2542 - Allow instance creation without OS installation (useful for example if
2543 the OS will be installed manually, or restored from a backup not in
2545 - Implemented option to make cluster ``copyfile`` use the replication
2547 - Added list of enabled hypervisors to ssconf (possibly useful for
2549 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2551 - A partial form of node re-IP is possible via node readd, which now
2552 allows changed node primary IP
2553 - Command line utilities now show an informational message if the job is
2555 - The logs of the master daemon now show the PID/UID/GID of the
2562 *(Released Fri, 12 Mar 2010)*
2564 During the 2.1.0 long release candidate cycle, a lot of improvements and
2565 changes have accumulated with were released later as 2.1.1.
2570 The node evacuate command (``gnt-node evacuate``) was significantly
2571 rewritten, and as such the IAllocator protocol was changed - a new
2572 request type has been added. This unfortunate change during a stable
2573 series is designed to improve performance of node evacuations; on
2574 clusters with more than about five nodes and which are well-balanced,
2575 evacuation should proceed in parallel for all instances of the node
2576 being evacuated. As such, any existing IAllocator scripts need to be
2577 updated, otherwise the above command will fail due to the unknown
2578 request. The provided "dumb" allocator has not been updated; but the
2579 ganeti-htools package supports the new protocol since version 0.2.4.
2581 Another important change is increased validation of node and instance
2582 names. This might create problems in special cases, if invalid host
2583 names are being used.
2585 Also, a new layer of hypervisor parameters has been added, that sits at
2586 OS level between the cluster defaults and the instance ones. This allows
2587 customisation of virtualization parameters depending on the installed
2588 OS. For example instances with OS 'X' may have a different KVM kernel
2589 (or any other parameter) than the cluster defaults. This is intended to
2590 help managing a multiple OSes on the same cluster, without manual
2591 modification of each instance's parameters.
2593 A tool for merging clusters, ``cluster-merge``, has been added in the
2594 tools sub-directory.
2599 - Improved the int/float conversions that should make the code more
2600 robust in face of errors from the node daemons
2601 - Fixed the remove node code in case of internal configuration errors
2602 - Fixed the node daemon behaviour in face of inconsistent queue
2603 directory (e.g. read-only file-system where we can't open the files
2605 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2606 now it either aborts cleanly or, if given the new "auto_promote"
2607 parameter, will automatically promote other nodes as needed
2608 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2609 completely prevent Ganeti from working
2610 - Fixed bug for instance export when not all disks were successfully
2612 - Fixed behaviour of node add when the new node is slow in starting up
2614 - Fixed handling of signals in the LUXI client, which should improve
2615 behaviour of command-line scripts
2616 - Added checks for invalid node/instance names in the configuration (now
2617 flagged during cluster verify)
2618 - Fixed watcher behaviour for disk activation errors
2619 - Fixed two potentially endless loops in http library, which led to the
2620 RAPI daemon hanging and consuming 100% CPU in some cases
2621 - Fixed bug in RAPI daemon related to hashed passwords
2622 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2624 - Enhanced compatibility with non-Debian OSes, but not using absolute
2625 path in some commands and allowing customisation of the ssh
2626 configuration directory
2627 - Fixed possible future issue with new Python versions by abiding to the
2628 proper use of ``__slots__`` attribute on classes
2629 - Added checks that should prevent directory traversal attacks
2630 - Many documentation fixes based on feedback from users
2635 - Added an "early_release" more for instance replace disks and node
2636 evacuate, where we release locks earlier and thus allow higher
2637 parallelism within the cluster
2638 - Added watcher hooks, intended to allow the watcher to restart other
2639 daemons (e.g. from the ganeti-nbma project), but they can be used of
2640 course for any other purpose
2641 - Added a compile-time disable for DRBD barriers, to increase
2642 performance if the administrator trusts the power supply or the
2643 storage system to not lose writes
2644 - Added the option of using syslog for logging instead of, or in
2645 addition to, Ganeti's own log files
2646 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2647 can indeed boot from a paravirtual NIC
2648 - Added a generic debug level for many operations; while this is not
2649 used widely yet, it allows one to pass the debug value all the way to
2651 - Enhanced the hooks environment for instance moves (failovers,
2652 migrations) where the primary/secondary nodes changed during the
2653 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2654 - Enhanced data validations for many user-supplied values; one important
2655 item is the restrictions imposed on instance and node names, which
2656 might reject some (invalid) host names
2657 - Add a configure-time option to disable file-based storage, if it's not
2658 needed; this allows greater security separation between the master
2659 node and the other nodes from the point of view of the inter-node RPC
2661 - Added user notification in interactive tools if job is waiting in the
2662 job queue or trying to acquire locks
2663 - Added log messages when a job is waiting for locks
2664 - Added filtering by node tags in instance operations which admit
2665 multiple instances (start, stop, reboot, reinstall)
2666 - Added a new tool for cluster mergers, ``cluster-merge``
2667 - Parameters from command line which are of the form ``a=b,c=d`` can now
2668 use backslash escapes to pass in values which contain commas,
2669 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2671 - For KVM, the instance name is the first parameter passed to KVM, so
2672 that it's more visible in the process list
2678 *(Released Tue, 2 Mar 2010)*
2680 Ganeti 2.1 brings many improvements with it. Major changes:
2682 - Added infrastructure to ease automated disk repairs
2683 - Added new daemon to export configuration data in a cheaper way than
2684 using the remote API
2685 - Instance NICs can now be routed instead of being associated with a
2687 - Improved job locking logic to reduce impact of jobs acquiring multiple
2688 locks waiting for other long-running jobs
2690 In-depth implementation details can be found in the Ganeti 2.1 design
2696 - Added chroot hypervisor
2697 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2699 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2700 ``bootloader_path`` and ``bootloader_args``)
2701 - Added the ``use_localtime`` option for the xen-hvm and kvm
2702 hypervisors, and the default value for this has changed to false (in
2703 2.0 xen-hvm always enabled it)
2704 - Added luxi call to submit multiple jobs in one go
2705 - Added cluster initialization option to not modify ``/etc/hosts``
2707 - Added network interface parameters
2708 - Added dry run mode to some LUs
2709 - Added RAPI resources:
2711 - ``/2/instances/[instance_name]/info``
2712 - ``/2/instances/[instance_name]/replace-disks``
2713 - ``/2/nodes/[node_name]/evacuate``
2714 - ``/2/nodes/[node_name]/migrate``
2715 - ``/2/nodes/[node_name]/role``
2716 - ``/2/nodes/[node_name]/storage``
2717 - ``/2/nodes/[node_name]/storage/modify``
2718 - ``/2/nodes/[node_name]/storage/repair``
2720 - Added OpCodes to evacuate or migrate all instances on a node
2721 - Added new command to list storage elements on nodes (``gnt-node
2722 list-storage``) and modify them (``gnt-node modify-storage``)
2723 - Added new ssconf files with master candidate IP address
2724 (``ssconf_master_candidates_ips``), node primary IP address
2725 (``ssconf_node_primary_ips``) and node secondary IP address
2726 (``ssconf_node_secondary_ips``)
2727 - Added ``ganeti-confd`` and a client library to query the Ganeti
2728 configuration via UDP
2729 - Added ability to run hooks after cluster initialization and before
2731 - Added automatic mode for disk replace (``gnt-instance replace-disks
2733 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2734 after catastrophic data-loss
2735 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2737 - Added ``gnt-instance move`` command to move instances
2738 - Added ``gnt-cluster watcher`` command to control watcher
2739 - Added ``gnt-node powercycle`` command to powercycle nodes
2740 - Added new job status field ``lock_status``
2741 - Added parseable error codes to cluster verification (``gnt-cluster
2742 verify --error-codes``) and made output less verbose (use
2743 ``--verbose`` to restore previous behaviour)
2744 - Added UUIDs to the main config entities (cluster, nodes, instances)
2745 - Added support for OS variants
2746 - Added support for hashed passwords in the Ganeti remote API users file
2748 - Added option to specify maximum timeout on instance shutdown
2749 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2750 - Added new helper script to start and stop Ganeti daemons
2751 (``daemon-util``), with the intent to reduce the work necessary to
2752 adjust Ganeti for non-Debian distributions and to start/stop daemons
2754 - Added more unittests
2755 - Fixed critical bug in ganeti-masterd startup
2756 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2757 now customisable at the cluster level for both the KVM and Xen
2758 hypervisors using the new ``migration_port`` parameter
2759 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2760 reinstalling an instance
2761 - Allowed ``@`` in tag names
2762 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2763 - Many documentation updates
2764 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2765 - ``gnt-instance reinstall`` can now reinstall multiple instances
2766 - Updated many command line parameters
2767 - Introduced new OS API version 15
2768 - No longer support a default hypervisor
2769 - Treat virtual LVs as inexistent
2770 - Improved job locking logic to reduce lock contention
2771 - Match instance and node names case insensitively
2772 - Reimplemented bash completion script to be more complete
2779 *(Released Thu, 4 Feb 2010)*
2781 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2782 - Fix a string formatting bug
2783 - Improve safety of the code in some error paths
2784 - Improve data validation in the master of values returned from nodes
2790 *(Released Thu, 17 Dec 2009)*
2792 - Fix security issue due to missing validation of iallocator names; this
2793 allows local and remote execution of arbitrary executables
2794 - Fix failure of gnt-node list during instance removal
2795 - Ship the RAPI documentation in the archive
2801 *(Released Wed, 30 Sep 2009)*
2803 - Fixed many wrong messages
2804 - Fixed a few bugs related to the locking library
2805 - Fixed MAC checking at instance creation time
2806 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2807 - Fixed a few issues related to signal handling in both daemons and
2809 - Fixed the example startup script provided
2810 - Fixed insserv dependencies in the example startup script (patch from
2812 - Fixed handling of drained nodes in the iallocator framework
2813 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2815 - Fixed error related to invalid job IDs in job polling
2816 - Fixed job/opcode persistence on unclean master shutdown
2817 - Fixed handling of partial job processing after unclean master
2819 - Fixed error reporting from LUs, previously all errors were converted
2820 into execution errors
2821 - Fixed error reporting from burnin
2822 - Decreased significantly the memory usage of the job queue
2823 - Optimised slightly multi-job submission
2824 - Optimised slightly opcode loading
2825 - Backported the multi-job submit framework from the development
2826 branch; multi-instance start and stop should be faster
2827 - Added script to clean archived jobs after 21 days; this will reduce
2828 the size of the queue directory
2829 - Added some extra checks in disk size tracking
2830 - Added an example ethers hook script
2831 - Added a cluster parameter that prevents Ganeti from modifying of
2833 - Added more node information to RAPI responses
2834 - Added a ``gnt-job watch`` command that allows following the ouput of a
2836 - Added a bind-address option to ganeti-rapi
2837 - Added more checks to the configuration verify
2838 - Enhanced the burnin script such that some operations can be retried
2840 - Converted instance reinstall to multi-instance model
2846 *(Released Fri, 7 Aug 2009)*
2848 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2849 to allow using the pre-2.0.2 behaviour in activation, if any existing
2850 instances have mismatched disk sizes in the configuration
2851 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2852 any configuration mismatches for disk sizes
2853 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2854 failover to work on two-node clusters
2855 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2857 - Fixed detection of OS script errors in ``gnt-backup export``
2858 - Fixed exit code of ``gnt-backup export``
2864 *(Released Fri, 17 Jul 2009)*
2866 - Added experimental support for stripped logical volumes; this should
2867 enhance performance but comes with a higher complexity in the block
2868 device handling; stripping is only enabled when passing
2869 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2870 affected even in the non-stripped mode
2871 - Improved resiliency against transient failures at the end of DRBD
2872 resyncs, and in general of DRBD resync checks
2873 - Fixed a couple of issues with exports and snapshot errors
2874 - Fixed a couple of issues in instance listing
2875 - Added display of the disk size in ``gnt-instance info``
2876 - Fixed checking for valid OSes in instance creation
2877 - Fixed handling of the "vcpus" parameter in instance listing and in
2878 general of invalid parameters
2879 - Fixed http server library, and thus RAPI, to handle invalid
2880 username/password combinations correctly; this means that now they
2881 report unauthorized for queries too, not only for modifications,
2882 allowing earlier detect of configuration problems
2883 - Added a new "role" node list field, equivalent to the master/master
2884 candidate/drained/offline flags combinations
2885 - Fixed cluster modify and changes of candidate pool size
2886 - Fixed cluster verify error messages for wrong files on regular nodes
2887 - Fixed a couple of issues with node demotion from master candidate role
2888 - Fixed node readd issues
2889 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2890 - Added a new ``--no-voting`` option for masterfailover to fix failover
2891 on two-nodes clusters when the former master node is unreachable
2892 - Added instance reinstall over RAPI
2898 *(Released Tue, 16 Jun 2009)*
2900 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2901 allow re-adding the start in single-user option (regression from 1.2)
2902 - the watcher writes the instance status to a file, to allow monitoring
2903 to report the instance status (from the master) based on cached
2904 results of the watcher's queries; while this can get stale if the
2905 watcher is being locked due to other work on the cluster, this is
2906 still an improvement
2907 - the watcher now also restarts the node daemon and the rapi daemon if
2909 - fixed the watcher to handle full and drained queue cases
2910 - hooks export more instance data in the environment, which helps if
2911 hook scripts need to take action based on the instance's properties
2912 (no longer need to query back into ganeti)
2913 - instance failovers when the instance is stopped do not check for free
2914 RAM, so that failing over a stopped instance is possible in low memory
2916 - rapi uses queries for tags instead of jobs (for less job traffic), and
2917 for cluster tags it won't talk to masterd at all but read them from
2919 - a couple of error handling fixes in RAPI
2920 - drbd handling: improved the error handling of inconsistent disks after
2921 resync to reduce the frequency of "there are some degraded disks for
2922 this instance" messages
2923 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2924 error handling path called a wrong function name)
2930 *(Released Wed, 27 May 2009)*
2932 - no changes from rc5
2938 *(Released Wed, 20 May 2009)*
2940 - fix a couple of bugs (validation, argument checks)
2941 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2942 - some small improvements to RAPI and IAllocator
2943 - make watcher automatically start the master daemon if down
2949 *(Released Mon, 27 Apr 2009)*
2951 - change the OS list to not require locks; this helps with big clusters
2952 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2953 volume group is broken
2954 - ``gnt-instance info``, without any arguments, doesn't run for all
2955 instances anymore; either pass ``--all`` or pass the desired
2956 instances; this helps against mistakes on big clusters where listing
2957 the information for all instances takes a long time
2958 - miscellaneous doc and man pages fixes
2964 *(Released Wed, 8 Apr 2009)*
2966 - Change the internal locking model of some ``gnt-node`` commands, in
2967 order to reduce contention (and blocking of master daemon) when
2968 batching many creation/reinstall jobs
2969 - Fixes to Xen soft reboot
2970 - No longer build documentation at build time, instead distribute it in
2971 the archive, in order to reduce the need for the whole docbook/rst
2978 *(Released Fri, 27 Mar 2009)*
2980 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2981 - Fix watcher startup sequence, improves the behaviour of busy clusters
2982 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2983 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2984 bind address and other places
2985 - Some documentation fixes and updates
2991 *(Released Mon, 2 Mar 2009)*
2993 - More documentation updates, now all docs should be more-or-less
2995 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2997 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2999 - Fix an issue related to $libdir/run/ganeti and cluster creation
3005 *(Released Thu, 19 Feb 2009)*
3007 - Xen PVM and KVM have switched the default value for the instance root
3008 disk to the first partition on the first drive, instead of the whole
3009 drive; this means that the OS installation scripts must be changed
3011 - Man pages have been updated
3012 - RAPI has been switched by default to HTTPS, and the exported functions
3013 should all work correctly
3014 - RAPI v1 has been removed
3015 - Many improvements to the KVM hypervisor
3016 - Block device errors are now better reported
3017 - Many other bugfixes and small improvements
3023 *(Released Mon, 26 Jan 2009)*
3025 - Version 2 is a general rewrite of the code and therefore the
3026 differences are too many to list, see the design document for 2.0 in
3027 the ``doc/`` subdirectory for more details
3028 - In this beta version there is not yet a migration path from 1.2 (there
3029 will be one in the final 2.0 release)
3030 - A few significant changes are:
3032 - all commands are executed by a daemon (``ganeti-masterd``) and the
3033 various ``gnt-*`` commands are just front-ends to it
3034 - all the commands are entered into, and executed from a job queue,
3035 see the ``gnt-job(8)`` manpage
3036 - the RAPI daemon supports read-write operations, secured by basic
3037 HTTP authentication on top of HTTPS
3038 - DRBD version 0.7 support has been removed, DRBD 8 is the only
3039 supported version (when migrating from Ganeti 1.2 to 2.0, you need
3040 to migrate to DRBD 8 first while still running Ganeti 1.2)
3041 - DRBD devices are using statically allocated minor numbers, which
3042 will be assigned to existing instances during the migration process
3043 - there is support for both Xen PVM and Xen HVM instances running on
3045 - KVM virtualization is supported too
3046 - file-based storage has been implemented, which means that it is
3047 possible to run the cluster without LVM and DRBD storage, for
3048 example using a shared filesystem exported from shared storage (and
3049 still have live migration)
3055 *(Released Tue, 13 Jan 2009)*
3057 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
3058 - Reuse the old instance mac address by default on instance import, if
3059 the instance name is the same.
3060 - Handle situations in which the node info rpc returns incomplete
3062 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
3063 - Improved version of batcher:
3065 - state file support
3066 - instance mac address support
3067 - support for HVM clusters/instances
3069 - Add an option to show the number of cpu sockets and nodes in
3071 - Support OSes that handle more than one version of the OS api (but do
3072 not change the current API in any other way)
3073 - Fix ``gnt-node migrate``
3074 - ``gnt-debug`` man page
3075 - Fixes various more typos and small issues
3076 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
3082 *(Released Wed, 24 Sep 2008)*
3084 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
3085 type of disk exported to fully virtualized instances.
3086 - provide access to the serial console of HVM instances
3087 - instance auto_balance flag, set by default. If turned off it will
3088 avoid warnings on cluster verify if there is not enough memory to fail
3089 over an instance. in the future it will prevent automatically failing
3090 it over when we will support that.
3091 - batcher tool for instance creation, see ``tools/README.batcher``
3092 - ``gnt-instance reinstall --select-os`` to interactively select a new
3093 operating system when reinstalling an instance.
3094 - when changing the memory amount on instance modify a check has been
3095 added that the instance will be able to start. also warnings are
3096 emitted if the instance will not be able to fail over, if auto_balance
3098 - documentation fixes
3099 - sync fields between ``gnt-instance list/modify/add/import``
3100 - fix a race condition in drbd when the sync speed was set after giving
3101 the device a remote peer.
3107 *(Released Tue, 22 Jul 2008)*
3109 - note: the allowed size and number of tags per object were reduced
3110 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
3111 - fixed twisted 8.x compatibility
3112 - fixed ``gnt-instance replace-disks`` with iallocator
3113 - add TCP keepalives on twisted connections to detect restarted nodes
3114 - disk increase support, see ``gnt-instance grow-disk``
3115 - implement bulk node/instance query for RAPI
3116 - add tags in node/instance listing (optional)
3117 - experimental migration (and live migration) support, read the man page
3118 for ``gnt-instance migrate``
3119 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3120 has some small improvements in handling its state file
3126 *(Released Fri, 13 Jun 2008)*
3128 - Experimental readonly, REST-based remote API implementation;
3129 automatically started on master node, TCP port 5080, if enabled by
3130 ``--enable-rapi`` parameter to configure script.
3131 - Instance allocator support. Add and import instance accept a
3132 ``--iallocator`` parameter, and call that instance allocator to decide
3133 which node to use for the instance. The iallocator document describes
3134 what's expected from an allocator script.
3135 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3136 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3137 node is lost there is still enough memory to fail over the instances
3139 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3140 ``gnt-cluster verify``, to check for site-specific compliance. All the
3141 hooks will run, and their output, if any, will be displayed. Any
3142 failing hook will make the verification return an error value.
3143 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3144 primary and secondary interfaces
3145 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3146 that is still declared as part of the cluster and has failed.
3147 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3148 specifying output fields, that just adds the chosen fields to the
3150 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3151 export from the filesystem.
3152 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3153 have been added. Using them you can enable/disable acpi and pae
3154 support, and specify a path for a cd image to be exported to the
3155 instance. These parameters as the name suggest only work on HVM
3157 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3158 PAE support will be set to the previously hardcoded values, but the
3159 (previously hardcoded) path to the CDROM ISO image will be unset and
3160 if required, needs to be set manually with ``gnt-instance modify``
3162 - The address to which an instance's VNC console is bound is now
3163 selectable per-instance, rather than being cluster wide. Of course
3164 this only applies to instances controlled via VNC, so currently just
3165 applies to HVM clusters.
3171 *(Released Mon, 18 Feb 2008)*
3173 - more tweaks to the disk activation code (especially helpful for DRBD)
3174 - change the default ``gnt-instance list`` output format, now there is
3175 one combined status field (see the manpage for the exact values this
3177 - some more fixes for the mac export to hooks change
3178 - make Ganeti not break with DRBD 8.2.x (which changed the version
3179 format in ``/proc/drbd``) (issue 24)
3180 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3181 template, allowing migration from DRBD0.7+MD to DRBD8
3187 *(Released Wed, 30 Jan 2008)*
3189 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3191 - add command aliases infrastructure and a few aliases
3192 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3193 man pages and the ``--help`` option of ``gnt-node
3194 list``/``gnt-instance list``
3195 - fix ``gnt-backup list`` with down nodes (issue 21)
3196 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3197 - fix the dist archive and add a check for including svn/git files in
3199 - some developer-related changes: improve the burnin and the QA suite,
3200 add an upload script for testing during development
3206 *(Released Wed, 16 Jan 2008)*
3208 - experimental HVM support, read the install document, section
3209 "Initializing the cluster"
3210 - allow for the PVM hypervisor per-instance kernel and initrd paths
3211 - add a new command ``gnt-cluster verify-disks`` which uses a new
3212 algorithm to improve the reconnection of the DRBD pairs if the device
3213 on the secondary node has gone away
3214 - make logical volume code auto-activate LVs at disk activation time
3215 - slightly improve the speed of activating disks
3216 - allow specification of the MAC address at instance creation time, and
3217 changing it later via ``gnt-instance modify``
3218 - fix handling of external commands that generate lots of output on
3220 - update documentation with regard to minimum version of DRBD8 supported
3226 *(Released Tue, 4 Dec 2007)*
3228 - Log the ``xm create`` output to the node daemon log on failure (to
3229 help diagnosing the error)
3230 - In debug mode, log all external commands output if failed to the logs
3231 - Change parsing of lvm commands to ignore stderr
3237 *(Released Wed, 28 Nov 2007)*
3239 - Another round of updates to the DRBD 8 code to deal with more failures
3240 in the replace secondary node operation
3241 - Some more logging of failures in disk operations (lvm, drbd)
3242 - A few documentation updates
3249 *(Released Tue, 13 Nov 2007)*
3251 - Change configuration file format from Python's Pickle to JSON.
3252 Upgrading is possible using the cfgupgrade utility.
3253 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3254 faster replace disks and is more stable (DRBD 8 has many improvements
3255 compared to DRBD 0.7)
3256 - Added command line tags support (see man pages for ``gnt-instance``,
3257 ``gnt-node``, ``gnt-cluster``)
3258 - Added instance rename support
3259 - Added multi-instance startup/shutdown
3260 - Added cluster rename support
3261 - Added ``gnt-node evacuate`` to simplify some node operations
3262 - Added instance reboot operation that can speedup reboot as compared to
3264 - Soften the requirement that hostnames are in FQDN format
3265 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3267 - Removed dependency on debian's patched fping that uses the
3268 non-standard ``-S`` option
3269 - Now the OS definitions are searched for in multiple, configurable
3270 paths (easier for distros to package)
3271 - Some changes to the hooks infrastructure (especially the new
3272 post-configuration update hook)
3273 - Other small bugfixes
3275 .. vim: set textwidth=72 syntax=rst :