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.
28 - KVM hypervisors can now access RBD storage directly without having to
29 go through a block device.
30 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
31 procedure between two Ganeti versions that are both 2.10 or higher.
32 - The move-instance command can now change disk templates when moving
33 instances, and does not require any node placement options to be
34 specified if the destination cluster has a default iallocator.
35 - Users can now change the soundhw and cpuid settings for XEN hypervisors.
36 - Hail and hbal now have the (optional) capability of accessing average CPU
37 load information through the monitoring deamon, and to use it to dynamically
38 adapt the allocation of instances.
39 - Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
40 so that disk and NIC modifications take effect without the need of actual
41 reboot. There are a couple of constrains currently for this feature:
43 - only KVM hypervisor (versions >= 1.0) supports it,
44 - one can not (yet) hotplug a disk using userspace access mode for RBD
45 - in case of a downgrade instances should suffer a reboot in order to
46 be migratable (due to core change of runtime files)
51 - A new test framework for logical units was introduced and the test
52 coverage for logical units was improved significantly.
53 - Opcodes are entirely generated from Haskell using the tool 'hs2py' and
54 the module 'src/Ganeti/OpCodes.hs'.
55 - Constants are also generated from Haskell using the tool
56 'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
57 exception of socket related constants, which require changing the
58 cluster configuration file, and HVS related constants, because they
59 are part of a port of instance queries to Haskell. As a result, these
60 changes will be part of the next release of Ganeti.
65 The following new dependencies have been added/updated.
69 - The version requirements for ``python-mock`` have increased to at least
70 version 1.0.1. It is still used for testing only.
75 - All known issues in 2.10.0 beta1 have been resolved (see changes from
77 - Improve handling of KVM runtime files from earlier Ganeti versions
80 Inherited from the 2.9 branch:
82 - use custom KVM path if set for version checking
83 - SingleNotifyPipeCondition: don't share pollers
85 Inherited from the 2.8 branch:
87 - Fixed Luxi daemon socket permissions after master-failover
88 - Improve IP version detection code directly checking for colons rather than
89 passing the family from the cluster object
90 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
91 opportunistically anymore (Issue 622)
92 - Allow link local IPv6 gateways (Issue 624)
93 - Fix error printing (Issue 616)
94 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
95 disk modifications, keep the old one. If name=none then set disk name to
97 - Update build_chroot script to work with the latest hackage packages
98 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
99 - Fix evacuation out of drained node (Issue 615)
100 - Add default file_driver if missing (Issue 571)
101 - Fix job error message after unclean master shutdown (Issue 618)
102 - Lock group(s) when creating instances (Issue 621)
103 - SetDiskID() before accepting an instance (Issue 633)
104 - Allow the ext template disks to receive arbitrary parameters, both at creation
105 time and while being modified
106 - Xen handle domain shutdown (future proofing cherry-pick)
107 - Refactor reading live data in htools (future proofing cherry-pick)
113 *(Released Wed, 27 Nov 2013)*
115 This was the first beta release of the 2.10 series. All important changes
116 are listed in the latest 2.10 entry.
121 The following issues are known to be present in the beta and will be fixed
124 - Issue 477: Wrong permissions for confd LUXI socket
125 - Issue 621: Instance related opcodes do not aquire network/group locks
126 - Issue 622: Assertion Error: Node locks differ from node resource locks
127 - Issue 623: IPv6 Masterd <-> Luxid communication error
133 *(Released Fri, 13 Dec 2013)*
135 - use custom KVM path if set for version checking
136 - SingleNotifyPipeCondition: don't share pollers
138 Inherited from the 2.8 branch:
140 - Fixed Luxi daemon socket permissions after master-failover
141 - Improve IP version detection code directly checking for colons rather than
142 passing the family from the cluster object
143 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
144 opportunistically anymore (Issue 622)
145 - Allow link local IPv6 gateways (Issue 624)
146 - Fix error printing (Issue 616)
147 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
148 disk modifications, keep the old one. If name=none then set disk name to
150 - Update build_chroot script to work with the latest hackage packages
151 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
152 - Fix evacuation out of drained node (Issue 615)
153 - Add default file_driver if missing (Issue 571)
154 - Fix job error message after unclean master shutdown (Issue 618)
155 - Lock group(s) when creating instances (Issue 621)
156 - SetDiskID() before accepting an instance (Issue 633)
157 - Allow the ext template disks to receive arbitrary parameters, both at creation
158 time and while being modified
159 - Xen handle domain shutdown (future proofing cherry-pick)
160 - Refactor reading live data in htools (future proofing cherry-pick)
166 *(Released Wed, 13 Nov 2013)*
168 - fix bug, that kept nodes offline when readding
169 - when verifying DRBD versions, ignore unavailable nodes
170 - fix bug that made the console unavailable on kvm in split-user
172 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
178 *(Released Tue, 5 Nov 2013)*
180 Incompatible/important changes
181 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
183 - hroller now also plans for capacity to move non-redundant instances off
184 any node to be rebooted; the old behavior of completely ignoring any
185 non-redundant instances can be restored by adding the --ignore-non-redundant
187 - The cluster option '--no-lvm-storage' was removed in favor of the new option
188 '--enabled-disk-templates'.
189 - On instance creation, disk templates no longer need to be specified
190 with '-t'. The default disk template will be taken from the list of
191 enabled disk templates.
192 - The monitoring daemon is now running as root, in order to be able to collect
193 information only available to root (such as the state of Xen instances).
194 - The ConfD client is now IPv6 compatible.
195 - File and shared file storage is no longer dis/enabled at configure time,
196 but using the option '--enabled-disk-templates' at cluster initialization and
198 - The default directories for file and shared file storage are not anymore
199 specified at configure time, but taken from the cluster's configuration.
200 They can be set at cluster initialization and modification with
201 '--file-storage-dir' and '--shared-file-storage-dir'.
202 - Cluster verification now includes stricter checks regarding the
203 default file and shared file storage directories. It now checks that
204 the directories are explicitely allowed in the 'file-storage-paths' file and
205 that the directories exist on all nodes.
206 - The list of allowed disk templates in the instance policy and the list
207 of cluster-wide enabled disk templates is now checked for consistency
208 on cluster or group modification. On cluster initialization, the ipolicy
209 disk templates are ensured to be a subset of the cluster-wide enabled
215 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
216 the correct command syntax. It is possible to use different DRBD versions
217 on different nodes as long as they are compatible to each other. This
218 enables rolling upgrades of DRBD with no downtime. As permanent operation
219 of different DRBD versions within a node group is discouraged,
220 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
221 - New "inst-status-xen" data collector for the monitoring daemon, providing
222 information about the state of the xen instances on the nodes.
223 - New "lv" data collector for the monitoring daemon, collecting data about the
224 logical volumes on the nodes, and pairing them with the name of the instances
226 - New "diskstats" data collector, collecting the data from /proc/diskstats and
227 presenting them over the monitoring daemon interface.
228 - The ConfD client is now IPv6 compatible.
232 The following new dependencies have been added.
236 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
237 for the unit tests (and only used for testing).
241 - ``hslogger`` (http://software.complete.org/hslogger) is now always
242 required, even if confd is not enabled.
247 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
249 - Improved error messsages (inherited from stable-2.8)
255 *(Released Tue, 15 Oct 2013)*
257 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
259 - in implicit configuration upgrade, match ipolicy with enabled disk templates
260 - improved harep documentation (inherited from stable-2.8)
266 *(Released Wed, 9 Oct 2013)*
268 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
270 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
271 at least one DRBD instance.
272 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
279 *(Released Tue, 1 Oct 2013)*
281 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
284 - update of the documentation, in particular installation instructions
285 - merging of LD_* constants into DT_* constants
286 - python style changes to be compatible with newer versions of pylint
292 *(Released Thu, 29 Aug 2013)*
294 This was the first beta release of the 2.9 series. All important changes
295 are listed in the latest 2.9 entry.
301 *(Released Thu, 12 Dec 2013)*
303 - Fixed Luxi daemon socket permissions after master-failover
304 - Improve IP version detection code directly checking for colons rather than
305 passing the family from the cluster object
306 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
307 opportunistically anymore (Issue 622)
308 - Allow link local IPv6 gateways (Issue 624)
309 - Fix error printing (Issue 616)
310 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
311 disk modifications, keep the old one. If name=none then set disk name to
313 - Update build_chroot script to work with the latest hackage packages
314 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
315 - Fix evacuation out of drained node (Issue 615)
316 - Add default file_driver if missing (Issue 571)
317 - Fix job error message after unclean master shutdown (Issue 618)
318 - Lock group(s) when creating instances (Issue 621)
319 - SetDiskID() before accepting an instance (Issue 633)
320 - Allow the ext template disks to receive arbitrary parameters, both at creation
321 time and while being modified
322 - Xen handle domain shutdown (future proofing cherry-pick)
323 - Refactor reading live data in htools (future proofing cherry-pick)
329 *(Released Thu, 07 Nov 2013)*
331 - DRBD: ensure peers are UpToDate for dual-primary
332 - Improve error message for replace-disks
333 - More dependency checks at configure time
334 - Placate warnings on ganeti.outils_unittest.py
340 *(Released Thu, 17 Oct 2013)*
342 - Correctly start/stop luxid during gnt-cluster master-failover
343 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
344 - Fix path for the job queue serial file
345 - Improved harep man page
346 - Minor documentation improvements
352 *(Released Mon, 30 Sep 2013)*
354 Incompatible/important changes
355 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
357 - Instance policy can contain multiple instance specs, as described in
358 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
359 <design-partitioned>`. As a consequence, it's not possible to partially change
360 or override instance specs. Bounding specs (min and max) can be specified as a
361 whole using the new option ``--ipolicy-bounds-specs``, while standard
362 specs use the new option ``--ipolicy-std-specs``.
363 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
364 gnt-instance is a valid YAML object.
365 - hail now honors network restrictions when allocating nodes. This led to an
366 update of the IAllocator protocol. See the IAllocator documentation for
368 - confd now only answers static configuration request over the network. luxid
369 was extracted, listens on the local LUXI socket and responds to live queries.
370 This allows finer grained permissions if using separate users.
375 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
376 to always require authentication, ``--require-authentication``. It can
377 be specified in ``$sysconfdir/default/ganeti``.
378 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
379 be used to manage the disk templates to be used by instances in the cluster.
380 Initially, it will be set to a list that includes plain, drbd, if they were
381 enabled by specifying a volume group name, and file and sharedfile, if those
382 were enabled at configure time. Additionally, it will include all disk
383 templates that are currently used by instances. The order of disk templates
384 will be based on Ganeti's history of supporting them. In the future, the
385 first entry of the list will be used as a default disk template on instance
387 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
388 configuration back to the previous stable version.
389 - Disk templates in group ipolicy can be restored to the default value.
390 - Initial support for diskless instances and virtual clusters in QA.
391 - More QA and unit tests for instance policies.
392 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
393 describing why the opcode itself was executed.
394 - The monitoring daemon is now available. It allows users to query the cluster
395 for obtaining information about the status of the system. The daemon is only
396 responsible for providing the information over the network: the actual data
397 gathering is performed by data collectors (currently, only the DRBD status
398 collector is available).
399 - In order to help developers work on Ganeti, a new script
400 (``devel/build_chroot``) is provided, for building a chroot that contains all
401 the required development libraries and tools for compiling Ganeti on a Debian
403 - A new tool, ``harep``, for performing self-repair and recreation of instances
404 in Ganeti has been added.
405 - Split queries are enabled for tags, network, exports, cluster info, groups,
407 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
408 It prints the command line to set the current policies, to ease
410 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
411 devices for KVM virtio-net interfaces will get created with VNET_HDR
412 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
413 virtio-net interfaces, which prevents host kernel tainting and log
414 flooding, when dealing with broken or malicious virtio-net drivers.
415 It's set to true by default.
416 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
418 - Support 'viridian' parameter in Xen HVM
419 - Support DSA SSH keys in bootstrap
420 - To simplify the work of packaging frameworks that want to add the needed users
421 and groups in a split-user setup themselves, at build time three files in
422 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
423 the groups to be generated, the ``users`` file contains, one per line, the
424 users to be generated, optionally followed by their primary group, where
425 important. The ``groupmemberships`` file contains, one per line, additional
426 user-group membership relations that need to be established. The syntax of
427 these files will remain stable in all future versions.
432 The following new dependencies have been added:
435 - The ``curl`` library is not optional anymore for compiling the Haskell code.
436 - ``snap-server`` library (if monitoring is enabled).
439 - The minimum Python version needed to run Ganeti is now 2.6.
440 - ``yaml`` library (only for running the QA).
444 - Perform proper cleanup on termination of Haskell daemons
445 - Fix corner-case in handling of remaining retry time
451 *(Released Tue, 17 Sep 2013)*
453 - To simplify the work of packaging frameworks that want to add the needed users
454 and groups in a split-user setup themselves, at build time three files in
455 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
456 the groups to be generated, the ``users`` file contains, one per line, the
457 users to be generated, optionally followed by their primary group, where
458 important. The ``groupmemberships`` file contains, one per line, additional
459 user-group membership relations that need to be established. The syntax of
460 these files will remain stable in all future versions.
461 - Add a default to file-driver when unspecified over RAPI (Issue 571)
462 - Mark the DSA host pubkey as optional, and remove it during config downgrade
464 - Some documentation fixes
470 *(Released Tue, 27 Aug 2013)*
472 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
474 - Support 'viridian' parameter in Xen HVM (Issue 233)
475 - Include VCS version in ``gnt-cluster version``
476 - Support DSA SSH keys in bootstrap (Issue 338)
477 - Fix batch creation of instances
478 - Use FQDN to check master node status (Issue 551)
479 - Make the DRBD collector more failure-resilient
485 *(Released Fri, 2 Aug 2013)*
487 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
489 - Fix upgrading/downgrading from 2.7
490 - Increase maximum RAPI message size
491 - Documentation updates
492 - Split ``confd`` between ``luxid`` and ``confd``
493 - Merge 2.7 series up to the 2.7.1 release
494 - Allow the ``modify_etc_hosts`` option to be changed
495 - Add better debugging for ``luxid`` queries
496 - Expose bulk parameter for GetJobs in RAPI client
497 - Expose missing ``network`` fields in RAPI
498 - Add some ``cluster verify`` tests
499 - Some unittest fixes
500 - Fix a malfunction in ``hspace``'s tiered allocation
501 - Fix query compatibility between haskell and python implementations
502 - Add the ``vnet_hdr`` HV parameter for KVM
503 - Add ``--cleanup`` to instance failover
504 - Change the connected groups format in ``gnt-network info`` output; it
505 was previously displayed as a raw list by mistake. (Merged from 2.7)
511 *(Released Mon, 24 Jun 2013)*
513 This was the first beta release of the 2.8 series. All important changes
514 are listed in the latest 2.8 entry.
520 *(Released Thu, 26 Sep 2013)*
522 - Change the connected groups format in ``gnt-network info`` output; it
523 was previously displayed as a raw list by mistake
524 - Check disk template in right dict when copying
525 - Support multi-instance allocs without iallocator
526 - Fix some errors in the documentation
527 - Fix formatting of tuple in an error message
533 *(Released Thu, 25 Jul 2013)*
535 - Add logrotate functionality in daemon-util
536 - Add logrotate example file
537 - Add missing fields to network queries over rapi
538 - Fix network object timestamps
539 - Add support for querying network timestamps
540 - Fix a typo in the example crontab
541 - Fix a documentation typo
547 *(Released Thu, 04 Jul 2013)*
549 Incompatible/important changes
550 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
552 - Instance policies for disk size were documented to be on a per-disk
553 basis, but hail applied them to the sum of all disks. This has been
555 - ``hbal`` will now exit with status 0 if, during job execution over
556 LUXI, early exit has been requested and all jobs are successful;
557 before, exit status 1 was used, which cannot be differentiated from
559 - Compatibility with newer versions of rbd has been fixed
560 - ``gnt-instance batch-create`` has been changed to use the bulk create
561 opcode from Ganeti. This lead to incompatible changes in the format of
562 the JSON file. It's now not a custom dict anymore but a dict
563 compatible with the ``OpInstanceCreate`` opcode.
564 - Parent directories for file storage need to be listed in
565 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
566 write the file automatically based on old configuration values, but it
567 can not distribute it across all nodes and the file contents should be
568 verified. Use ``gnt-cluster copyfile
569 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
570 upgraded. The reason for requiring this list of paths now is that
571 before it would have been possible to inject new paths via RPC,
572 allowing files to be created in arbitrary locations. The RPC protocol
573 is protected using SSL/X.509 certificates, but as a design principle
574 Ganeti does not permit arbitrary paths to be passed.
575 - The parsing of the variants file for OSes (see
576 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
577 lines and comment lines (starting with ``#``) are ignored for better
579 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
580 longer available. ``gnt-node add`` now invokes a new tool on the
581 destination node, named ``prepare-node-join``, to configure the SSH
582 daemon. Paramiko is no longer necessary to configure nodes' SSH
583 daemons via ``gnt-node add``.
584 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
585 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
586 the flag is not reset after a master failover.
587 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
588 is recommended. 2.6 will be mandatory from the 2.8 series.
594 - New network management functionality to support automatic allocation
595 of IP addresses and managing of network parameters. See
596 :manpage:`gnt-network(8)` for more details.
597 - New external storage backend, to allow managing arbitrary storage
598 systems external to the cluster. See
599 :manpage:`ganeti-extstorage-interface(7)`.
600 - New ``exclusive-storage`` node parameter added, restricted to
601 nodegroup level. When it's set to true, physical disks are assigned in
602 an exclusive fashion to instances, as documented in :doc:`Partitioned
603 Ganeti <design-partitioned>`. Currently, only instances using the
604 ``plain`` disk template are supported.
605 - The KVM hypervisor has been updated with many new hypervisor
606 parameters, including a generic one for passing arbitrary command line
607 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
608 compatible up to qemu 1.4.
609 - A new tool, called ``mon-collector``, is the stand-alone executor of
610 the data collectors for a monitoring system. As of this version, it
611 just includes the DRBD data collector, that can be executed by calling
612 ``mon-collector`` using the ``drbd`` parameter. See
613 :manpage:`mon-collector(7)`.
614 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
615 for RAPI users. It allows granting permissions to query for
616 information to a specific user without giving
617 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
618 - A new tool named ``node-cleanup`` has been added. It cleans remains of
619 a cluster from a machine by stopping all daemons, removing
620 certificates and ssconf files. Unless the ``--no-backup`` option is
621 given, copies of the certificates are made.
622 - Instance creations now support the use of opportunistic locking,
623 potentially speeding up the (parallel) creation of multiple instances.
624 This feature is currently only available via the :doc:`RAPI
625 <rapi>` interface and when an instance allocator is used. If the
626 ``opportunistic_locking`` parameter is set the opcode will try to
627 acquire as many locks as possible, but will not wait for any locks
628 held by other opcodes. If not enough resources can be found to
629 allocate the instance, the temporary error code
630 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
631 retried thereafter, with or without opportunistic locking.
632 - New experimental linux-ha resource scripts.
633 - Restricted-commands support: ganeti can now be asked (via command line
634 or rapi) to perform commands on a node. These are passed via ganeti
635 RPC rather than ssh. This functionality is restricted to commands
636 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
637 reasons. The file is not copied automatically.
643 - Diskless instances are now externally mirrored (Issue 237). This for
644 now has only been tested in conjunction with explicit target nodes for
646 - Queries not needing locks or RPC access to the node can now be
647 performed by the confd daemon, making them independent from jobs, and
648 thus faster to execute. This is selectable at configure time.
649 - The functionality for allocating multiple instances at once has been
650 overhauled and is now also available through :doc:`RAPI <rapi>`.
652 There are no significant changes from version 2.7.0~rc3.
658 *(Released Tue, 25 Jun 2013)*
660 - Fix permissions on the confd query socket (Issue 477)
661 - Fix permissions on the job archive dir (Issue 498)
662 - Fix handling of an internal exception in replace-disks (Issue 472)
663 - Fix gnt-node info handling of shortened names (Issue 497)
664 - Fix gnt-instance grow-disk when wiping is enabled
665 - Documentation improvements, and support for newer pandoc
666 - Fix hspace honoring ipolicy for disks (Issue 484)
667 - Improve handling of the ``kvm_extra`` HV parameter
673 *(Released Fri, 24 May 2013)*
675 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
677 - Disks added through ``gnt-instance modify`` or created through
678 ``gnt-instance recreate-disks`` are wiped, if the
679 ``prealloc_wipe_disks`` flag is set.
680 - If wiping newly created disks fails, the disks are removed. Also,
681 partial failures in creating disks through ``gnt-instance modify``
682 triggers a cleanup of the partially-created disks.
683 - Removing the master IP address doesn't fail if the address has been
685 - Fix ownership of the OS log dir
686 - Workaround missing SO_PEERCRED constant (Issue 191)
692 *(Released Fri, 3 May 2013)*
694 This was the first release candidate of the 2.7 series. Since beta3:
696 - Fix kvm compatibility with qemu 1.4 (Issue 389)
697 - Documentation updates (admin guide, upgrade notes, install
698 instructions) (Issue 372)
699 - Fix gnt-group list nodes and instances count (Issue 436)
700 - Fix compilation without non-mandatory libraries (Issue 441)
701 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
702 - Make confd logging more verbose at INFO level (Issue 435)
703 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
704 - Fix failure path for instance storage type conversion (Issue 229)
705 - Update htools text backend documentation
706 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
707 - Disable inter-cluster instance move for file-based instances, because
708 it is dependant on instance export, which is not supported for
709 file-based instances. (Issue 414)
710 - Fix gnt-job crashes on non-ascii characters (Issue 427)
711 - Fix volume group checks on non-vm-capable nodes (Issue 432)
717 *(Released Mon, 22 Apr 2013)*
719 This was the third beta release of the 2.7 series. Since beta2:
721 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
722 - Fix data loss on wrong usage of ``gnt-instance move``
723 - Properly export errors in confd-based job queries
724 - Add ``users-setup`` tool
725 - Fix iallocator protocol to report 0 as a disk size for diskless
726 instances. This avoids hail breaking when a diskless instance is
728 - Fix job queue directory permission problem that made confd job queries
729 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
730 for access to archived jobs (Issue 406).
731 - Limit the sizes of networks supported by ``gnt-network`` to something
732 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
733 - Fix bugs in instance disk template conversion
734 - Fix GHC 7 compatibility
735 - Fix ``burnin`` install path (Issue 426).
736 - Allow very small disk grows (Issue 347).
737 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
738 that noded doesn't import masterd code (Issue 419).
739 - Make sure the default metavg at cluster init is the same as the vg, if
740 unspecified (Issue 358).
741 - Fix cleanup of partially created disks (part of Issue 416)
747 *(Released Tue, 2 Apr 2013)*
749 This was the second beta release of the 2.7 series. Since beta1:
751 - Networks no longer have a "type" slot, since this information was
752 unused in Ganeti: instead of it tags should be used.
753 - The rapi client now has a ``target_node`` option to MigrateInstance.
754 - Fix early exit return code for hbal (Issue 386).
755 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
756 - Fix ``rbd showmapped`` output parsing (Issue 312).
757 - Networks are now referenced indexed by UUID, rather than name. This
758 will require running cfgupgrade, from 2.7.0beta1, if networks are in
760 - The OS environment now includes network information.
761 - Deleting of a network is now disallowed if any instance nic is using
762 it, to prevent dangling references.
763 - External storage is now documented in man pages.
764 - The exclusive_storage flag can now only be set at nodegroup level.
765 - Hbal can now submit an explicit priority with its jobs.
766 - Many network related locking fixes.
767 - Bump up the required pylint version to 0.25.1.
768 - Fix the ``no_remember`` option in RAPI client.
769 - Many ipolicy related tests, qa, and fixes.
770 - Many documentation improvements and fixes.
771 - Fix building with ``--disable-file-storage``.
772 - Fix ``-q`` option in htools, which was broken if passed more than
774 - Some haskell/python interaction improvements and fixes.
775 - Fix iallocator in case of missing LVM storage.
776 - Fix confd config load in case of ``--no-lvm-storage``.
777 - The confd/query functionality is now mentioned in the security
784 *(Released Wed, 6 Feb 2013)*
786 This was the first beta release of the 2.7 series. All important changes
787 are listed in the latest 2.7 entry.
793 *(Released Fri, 21 Dec 2012)*
795 Important behaviour change: hbal won't rebalance anymore instances which
796 have the ``auto_balance`` attribute set to false. This was the intention
797 all along, but until now it only skipped those from the N+1 memory
798 reservation (DRBD-specific).
800 A significant number of bug fixes in this release:
802 - Fixed disk adoption interaction with ipolicy checks.
803 - Fixed networking issues when instances are started, stopped or
804 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
805 - Fixed the warning in cluster verify for shared storage instances not
807 - Fixed removal of storage directory on shared file storage (issue 262).
808 - Fixed validation of LVM volume group name in OpClusterSetParams
809 (``gnt-cluster modify``) (issue 285).
810 - Fixed runtime memory increases (``gnt-instance modify -m``).
811 - Fixed live migration under Xen's ``xl`` mode.
812 - Fixed ``gnt-instance console`` with ``xl``.
813 - Fixed building with newer Haskell compiler/libraries.
814 - Fixed PID file writing in Haskell daemons (confd); this prevents
815 restart issues if confd was launched manually (outside of
816 ``daemon-util``) while another copy of it was running
817 - Fixed a type error when doing live migrations with KVM (issue 297) and
818 the error messages for failing migrations have been improved.
819 - Fixed opcode validation for the out-of-band commands (``gnt-node
821 - Fixed a type error when unsetting OS hypervisor parameters (issue
822 311); now it's possible to unset all OS-specific hypervisor
824 - Fixed the ``dry-run`` mode for many operations: verification of
825 results was over-zealous but didn't take into account the ``dry-run``
826 operation, resulting in "wrong" failures.
827 - Fixed bash completion in ``gnt-job list`` when the job queue has
828 hundreds of entries; especially with older ``bash`` versions, this
829 results in significant CPU usage.
831 And lastly, a few other improvements have been made:
833 - Added option to force master-failover without voting (issue 282).
834 - Clarified error message on lock conflict (issue 287).
835 - Logging of newly submitted jobs has been improved (issue 290).
836 - Hostname checks have been made uniform between instance rename and
838 - The ``--submit`` option is now supported by ``gnt-debug delay``.
839 - Shutting down the master daemon by sending SIGTERM now stops it from
840 processing jobs waiting for locks; instead, those jobs will be started
841 once again after the master daemon is started the next time (issue
843 - Support for Xen's ``xl`` program has been improved (besides the fixes
845 - Reduced logging noise in the Haskell confd daemon (only show one log
846 entry for each config reload, instead of two).
847 - Several man page updates and typo fixes.
853 *(Released Fri, 12 Oct 2012)*
855 A small bugfix release. Among the bugs fixed:
857 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
858 made the command unusable.
859 - Commands that issue many jobs don't fail anymore just because some jobs
860 take so long that other jobs are archived.
861 - Failures during ``gnt-instance reinstall`` are reflected by the exit
863 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
865 - When ``always_failover`` is set, ``--allow-failover`` is not required
866 in migrate commands anymore.
867 - ``bash_completion`` works even if extglob is disabled.
868 - Fixed bug with locks that made failover for RDB-based instances fail.
869 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
870 a random node instead of one based on the allocator metric.
871 - Support for newer versions of pylint and pep8.
872 - Hail doesn't fail anymore when trying to add an instance of type
873 ``file``, ``sharedfile`` or ``rbd``.
874 - Added new Makefile target to rebuild the whole distribution, so that
875 all files are included.
881 *(Released Fri, 27 Jul 2012)*
884 .. attention:: The ``LUXI`` protocol has been made more consistent
885 regarding its handling of command arguments. This, however, leads to
886 incompatibility issues with previous versions. Please ensure that you
887 restart Ganeti daemons soon after the upgrade, otherwise most
888 ``LUXI`` calls (job submission, setting/resetting the drain flag,
889 pausing/resuming the watcher, cancelling and archiving jobs, querying
890 the cluster configuration) will fail.
899 The current ``admin_up`` field, which used to denote whether an instance
900 should be running or not, has been removed. Instead, ``admin_state`` is
901 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
903 The rational behind this is that an instance being “down” can have
906 - it could be down during a reboot
907 - it could be temporarily be down for a reinstall
908 - or it could be down because it is deprecated and kept just for its
911 The previous Boolean state was making it difficult to do capacity
912 calculations: should Ganeti reserve memory for a down instance? Now, the
913 tri-state field makes it clear:
915 - in ``up`` and ``down`` state, all resources are reserved for the
916 instance, and it can be at any time brought up if it is down
917 - in ``offline`` state, only disk space is reserved for it, but not
920 The field can have an extra use: since the transition between ``up`` and
921 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
922 transition between ``offline`` and ``down`` is done via ``gnt-instance
923 modify``, it is possible to given different rights to users. For
924 example, owners of an instance could be allowed to start/stop it, but
925 not transition it out of the offline state.
927 Instance policies and specs
928 +++++++++++++++++++++++++++
930 In previous Ganeti versions, an instance creation request was not
931 limited on the minimum size and on the maximum size just by the cluster
932 resources. As such, any policy could be implemented only in third-party
933 clients (RAPI clients, or shell wrappers over ``gnt-*``
934 tools). Furthermore, calculating cluster capacity via ``hspace`` again
935 required external input with regards to instance sizes.
937 In order to improve these workflows and to allow for example better
938 per-node group differentiation, we introduced instance specs, which
941 - minimum instance disk size, disk count, memory size, cpu count
942 - maximum values for the above metrics
943 - and “standard” values (used in ``hspace`` to calculate the standard
946 The minimum/maximum values can be also customised at node-group level,
947 for example allowing more powerful hardware to support bigger instance
950 Beside the instance specs, there are a few other settings belonging to
951 the instance policy framework. It is possible now to customise, per
952 cluster and node-group:
954 - the list of allowed disk templates
955 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
956 - the maximum ratio of instance to spindles (see below for more
957 information) for local storage
959 All these together should allow all tools that talk to Ganeti to know
960 what are the ranges of allowed values for instances and the
961 over-subscription that is allowed.
963 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
964 instance configuration, and the physical CPU configuration from the
965 node. For the spindle ratios however, we didn't track before these
966 values, so new parameters have been added:
968 - a new node parameter ``spindle_count``, defaults to 1, customisable at
969 node group or node level
970 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
972 Note that spindles in this context doesn't need to mean actual
973 mechanical hard-drives; it's just a relative number for both the node
974 I/O capacity and instance I/O consumption.
976 Instance migration behaviour
977 ++++++++++++++++++++++++++++
979 While live-migration is in general desirable over failover, it is
980 possible that for some workloads it is actually worse, due to the
981 variable time of the “suspend” phase during live migration.
983 To allow the tools to work consistently over such instances (without
984 having to hard-code instance names), a new backend parameter
985 ``always_failover`` has been added to control the migration/failover
986 behaviour. When set to True, all migration requests for an instance will
987 instead fall-back to failover.
989 Instance memory ballooning
990 ++++++++++++++++++++++++++
992 Initial support for memory ballooning has been added. The memory for an
993 instance is no longer fixed (backend parameter ``memory``), but instead
994 can vary between minimum and maximum values (backend parameters
995 ``minmem`` and ``maxmem``). Currently we only change an instance's
998 - live migrating or failing over and instance and the target node
999 doesn't have enough memory
1000 - user requests changing the memory via ``gnt-instance modify
1003 Instance CPU pinning
1004 ++++++++++++++++++++
1006 In order to control the use of specific CPUs by instance, support for
1007 controlling CPU pinning has been added for the Xen, HVM and LXC
1008 hypervisors. This is controlled by a new hypervisor parameter
1009 ``cpu_mask``; details about possible values for this are in the
1010 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
1011 VCPU-to-CPU mapping) form will work well only when all nodes in your
1012 cluster have the same amount of CPUs.
1017 Another area in which Ganeti was not customisable were the parameters
1018 used for storage configuration, e.g. how many stripes to use for LVM,
1019 DRBD resync configuration, etc.
1021 To improve this area, we've added disks parameters, which are
1022 customisable at cluster and node group level, and which allow to
1023 specify various parameters for disks (DRBD has the most parameters
1024 currently), for example:
1026 - DRBD resync algorithm and parameters (e.g. speed)
1027 - the default VG for meta-data volumes for DRBD
1028 - number of stripes for LVM (plain disk template)
1031 These parameters can be modified via ``gnt-cluster modify -D …`` and
1032 ``gnt-group modify -D …``, and are used at either instance creation (in
1033 case of LVM stripes, for example) or at disk “activation” time
1034 (e.g. resync speed).
1036 Rados block device support
1037 ++++++++++++++++++++++++++
1039 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
1040 denoted by the ``rbd`` disk template type. This is considered
1041 experimental, feedback is welcome. For details on configuring it, see
1042 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
1047 The existing master IP functionality works well only in simple setups (a
1048 single network shared by all nodes); however, if nodes belong to
1049 different networks, then the ``/32`` setup and lack of routing
1050 information is not enough.
1052 To allow the master IP to function well in more complex cases, the
1053 system was reworked as follows:
1055 - a master IP netmask setting has been added
1056 - the master IP activation/turn-down code was moved from the node daemon
1057 to a separate script
1058 - whether to run the Ganeti-supplied master IP script or a user-supplied
1059 on is a ``gnt-cluster init`` setting
1061 Details about the location of the standard and custom setup scripts are
1062 in the man page :manpage:`gnt-cluster(8)`; for information about the
1063 setup script protocol, look at the Ganeti-supplied script.
1068 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
1071 It is now possible to use TLS-protected connections, and when renewing
1072 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
1073 it is now possible to specify spice or spice CA certificates. Also, it
1074 is possible to configure a password for SPICE sessions via the
1075 hypervisor parameter ``spice_password_file``.
1077 There are also new parameters to control the compression and streaming
1078 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1079 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1080 for the spice parameters.
1082 Lastly, it is now possible to see the SPICE connection information via
1083 ``gnt-instance console``.
1088 A new tool (``tools/ovfconverter``) has been added that supports
1089 conversion between Ganeti and the `Open Virtualization Format
1090 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1093 This relies on the ``qemu-img`` tool to convert the disk formats, so the
1094 actual compatibility with other virtualization solutions depends on it.
1096 Confd daemon changes
1097 ++++++++++++++++++++
1099 The configuration query daemon (``ganeti-confd``) is now optional, and
1100 has been rewritten in Haskell; whether to use the daemon at all, use the
1101 Python (default) or the Haskell version is selectable at configure time
1102 via the ``--enable-confd`` parameter, which can take one of the
1103 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1104 daemon will result in a smaller footprint; for larger systems, we
1105 welcome feedback on the Haskell version which might become the default
1108 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1109 daemon running. The Python version doesn't implement the new call.
1112 User interface changes
1113 ~~~~~~~~~~~~~~~~~~~~~~
1115 We have replaced the ``--disks`` option of ``gnt-instance
1116 replace-disks`` with a more flexible ``--disk`` option, which allows
1117 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1118 disk size and mode can be changed upon recreation (via ``gnt-instance
1119 recreate-disks``, which accepts the same ``--disk`` option).
1121 As many people are used to a ``show`` command, we have added that as an
1122 alias to ``info`` on all ``gnt-*`` commands.
1124 The ``gnt-instance grow-disk`` command has a new mode in which it can
1125 accept the target size of the disk, instead of the delta; this can be
1126 more safe since two runs in absolute mode will be idempotent, and
1127 sometimes it's also easier to specify the desired size directly.
1129 Also the handling of instances with regard to offline secondaries has
1130 been improved. Instance operations should not fail because one of it's
1131 secondary nodes is offline, even though it's safe to proceed.
1133 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1134 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1135 minors to instance name.
1140 RAPI coverage has improved, with (for example) new resources for
1141 recreate-disks, node power-cycle, etc.
1146 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1149 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1150 whether to still support Python 2.4 or move to Python 2.6 as minimum
1153 Support for Fedora has been slightly improved; the provided example
1154 init.d script should work better on it and the INSTALL file should
1155 document the needed dependencies.
1160 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1161 ``Query(what=QR_LOCK, ...)`` instead.
1163 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1164 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1165 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1166 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1167 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1169 RAPI client: ``CertificateError`` now derives from
1170 ``GanetiApiError``. This should make it more easy to handle Ganeti
1173 Deprecation warnings due to PyCrypto/paramiko import in
1174 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1175 make sure to run an up-to-date paramiko version, if you use this tool.
1177 The QA scripts now depend on Python 2.5 or above (the main code base
1178 still works with Python 2.4).
1180 The configuration file (``config.data``) is now written without
1181 indentation for performance reasons; if you want to edit it, it can be
1182 re-formatted via ``tools/fmtjson``.
1184 A number of bugs has been fixed in the cluster merge tool.
1186 ``x509`` certification verification (used in import-export) has been
1187 changed to allow the same clock skew as permitted by the cluster
1188 verification. This will remove some rare but hard to diagnose errors in
1195 *(Released Thu, 19 Jul 2012)*
1197 Very few changes from rc4 to the final release, only bugfixes:
1199 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1200 instance, fix CDROM booting for KVM instances)
1201 - fixed node group modification of node parameters
1202 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1203 - fixed generation of bash completion to ensure a stable ordering
1210 *(Released Fri, 13 Jul 2012)*
1212 Third release candidate for 2.6. The following changes were done from
1215 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1216 - Fixed an inconsistency in the LUXI protocol with the provided
1217 arguments (NOT backwards compatible)
1218 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1219 the cluster ``std`` value
1220 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1221 easier instance debugging on nodes (requires ``hconfd`` to work)
1227 *(Released Tue, 03 Jul 2012)*
1229 Second release candidate for 2.6. The following changes were done from
1232 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1234 - Fixed a RAPI regression on missing beparams/memory
1235 - Fixed redistribution of files on offline nodes
1236 - Added possibility to run activate-disks even though secondaries are
1237 offline. With this change it relaxes also the strictness on some other
1238 commands which use activate disks internally:
1239 * ``gnt-instance start|reboot|rename|backup|export``
1240 - Made it possible to remove safely an instance if its secondaries are
1242 - Made it possible to reinstall even though secondaries are offline
1248 *(Released Mon, 25 Jun 2012)*
1250 First release candidate for 2.6. The following changes were done from
1253 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1255 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1256 behaviour. It supports now the modify operation
1257 - ``hcheck`` implemented to analyze cluster health and possibility of
1258 improving health by rebalance
1259 - ``hbal`` has been improved in dealing with split instances
1265 *(Released Mon, 11 Jun 2012)*
1267 Second beta release of 2.6. The following changes were done from beta2
1270 - Fixed ``daemon-util`` with non-root user models
1271 - Fixed creation of plain instances with ``--no-wait-for-sync``
1272 - Fix wrong iv_names when running ``cfgupgrade``
1273 - Export more information in RAPI group queries
1274 - Fixed bug when changing instance network interfaces
1275 - Extended burnin to do NIC changes
1276 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1277 - Changed default for DRBD barriers
1278 - Fixed DRBD error reporting for syncer rate
1279 - Verify the options on disk parameters
1281 And of course various fixes to documentation and improved unittests and
1288 *(Released Wed, 23 May 2012)*
1290 First beta release of 2.6. The following changes were done from beta1 to
1293 - integrated patch for distributions without ``start-stop-daemon``
1294 - adapted example init.d script to work on Fedora
1295 - fixed log handling in Haskell daemons
1296 - adapted checks in the watcher for pycurl linked against libnss
1297 - add partial support for ``xl`` instead of ``xm`` for Xen
1298 - fixed a type issue in cluster verification
1299 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1302 Plus integrated fixes from the 2.5 branch:
1304 - fixed ``kvm-ifup`` to use ``/bin/bash``
1305 - fixed parallel build failures
1306 - KVM live migration when using a custom keymap
1312 *(Released Tue, 24 Jul 2012)*
1314 A small bugfix release, with no new features:
1316 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1317 different default shell (e.g. Debian, Ubuntu)
1318 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1319 243 and Debian bug #650664)
1320 - fixed compatibility with KVM versions that don't support multiple boot
1321 devices (fixes Issue 230 and Debian bug #624256)
1323 Additionally, a few fixes were done to the build system (fixed parallel
1324 build failures) and to the unittests (fixed race condition in test for
1325 FileID functions, and the default enable/disable mode for QA test is now
1332 *(Released Fri, 11 May 2012)*
1334 A small bugfix release.
1336 The main issues solved are on the topic of compatibility with newer LVM
1339 - fixed parsing of ``lv_attr`` field
1340 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1341 the ``--force`` flag is needed
1343 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1344 to accept kernel 3.x too (was hardcoded to 2.6.*).
1346 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1347 of hook results and that also made display of other errors suboptimal
1348 was fixed; the code behaves now like 2.4 and earlier.
1350 Another change in 2.5, the cleanup of the OS scripts environment, is too
1351 aggressive: it removed even the ``PATH`` variable, which requires the OS
1352 scripts to *always* need to export it. Since this is a bit too strict,
1353 we now export a minimal PATH, the same that we export for hooks.
1355 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1356 integrated into this release.
1358 Finally, a few other miscellaneous changes were done (no new features,
1359 just small improvements):
1361 - Fix ``gnt-group --help`` display
1362 - Fix hardcoded Xen kernel path
1363 - Fix grow-disk handling of invalid units
1364 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1365 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1371 *(Released Thu, 12 Apr 2012)*
1373 Incompatible/important changes and bugfixes
1374 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1376 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1377 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1378 to match the underlying LUXI interface.
1379 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1380 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1381 not maintain backwards-compatibility as the underlying operation
1382 changed in an incompatible way. The RAPI client can talk to old
1383 servers, but it needs to be told so as the return value changed.
1384 - When creating file-based instances via RAPI, the ``file_driver``
1385 parameter no longer defaults to ``loop`` and must be specified.
1386 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1388 - Support for the undocumented and deprecated RAPI instance creation
1389 request format version 0 has been dropped. Use version 1, supported
1390 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1391 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1392 documentation <install>`.
1393 - The "cluster-verify" hooks are now executed per group by the
1394 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1395 if you just run ``gnt-cluster verify``, which generates one opcode per
1397 - The environment as passed to the OS scripts is cleared, and thus no
1398 environment variables defined in the node daemon's environment will be
1399 inherited by the scripts.
1400 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1402 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1403 support operations involving multiple node groups.
1404 - Offline nodes are ignored when failing over an instance.
1405 - Support for KVM version 1.0, which changed the version reporting format
1407 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1409 - ``Makefile`` is now compatible with Automake 1.11.2
1410 - Includes all bugfixes made in the 2.4 series
1415 - The ganeti-htools project has been merged into the ganeti-core source
1416 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1417 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1418 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1419 GPT-style partition tables (requires `parted
1420 <http://www.gnu.org/s/parted/>`_).
1421 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1422 - Allowed adding tags on instance creation.
1423 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1424 - Allow instances to be started in a paused state, enabling the user to
1425 see the complete console output on boot using the console.
1426 - Added new hypervisor flag to control default reboot behaviour
1427 (``reboot_behavior``).
1428 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1429 - Improved out-of-band management support:
1431 - Added ``gnt-node health`` command reporting the health status of
1433 - Added ``gnt-node power`` command to manage power status of nodes.
1434 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1436 - Instance migration can fall back to failover if instance is not
1438 - Filters can be used when listing nodes, instances, groups and locks;
1439 see :manpage:`ganeti(7)` manpage.
1440 - Added post-execution status as variables to :doc:`hooks <hooks>`
1442 - Instance tags are exported/imported together with the instance.
1443 - When given an explicit job ID, ``gnt-job info`` will work for archived
1445 - Jobs can define dependencies on other jobs (not yet supported via
1446 RAPI or command line, but used by internal commands and usable via
1449 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1452 - Instance failover is now available as a RAPI resource
1453 (``/2/instances/[instance_name]/failover``).
1454 - ``gnt-instance info`` defaults to static information if primary node
1456 - Opcodes have a new ``comment`` attribute.
1457 - Added basic SPICE support to KVM hypervisor.
1458 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1460 Node group improvements
1461 ~~~~~~~~~~~~~~~~~~~~~~~
1463 - ``gnt-cluster verify`` has been modified to check groups separately,
1464 thereby improving performance.
1465 - Node group support has been added to ``gnt-cluster verify-disks``,
1466 which now operates per node group.
1467 - Watcher has been changed to work better with node groups.
1469 - One process and state file per node group.
1470 - Slow watcher in one group doesn't block other group's watcher.
1472 - Added new command, ``gnt-group evacuate``, to move all instances in a
1473 node group to other groups.
1474 - Added ``gnt-instance change-group`` to move an instance to another
1476 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1477 per-group operations.
1478 - Node groups can be tagged.
1479 - Some operations switch from an exclusive to a shared lock as soon as
1481 - Instance's primary and secondary nodes' groups are now available as
1482 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1483 and ``snodes.group.uuid``).
1488 - Numerous updates to documentation and manpages.
1490 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1492 - Some opcode/job results are now also documented, see :doc:`RAPI
1495 - A lockset's internal lock is now also visible in lock monitor.
1496 - Log messages from job queue workers now contain information about the
1497 opcode they're processing.
1498 - ``gnt-instance console`` no longer requires the instance lock.
1499 - A short delay when waiting for job changes reduces the number of LUXI
1500 requests significantly.
1501 - DRBD metadata volumes are overwritten with zeros during disk creation.
1502 - Out-of-band commands no longer acquire the cluster lock in exclusive
1504 - ``devel/upload`` now uses correct permissions for directories.
1510 *(Released Fri, 23 Mar 2012)*
1512 This was the sixth release candidate of the 2.5 series.
1518 *(Released Mon, 9 Jan 2012)*
1520 This was the fifth release candidate of the 2.5 series.
1526 *(Released Thu, 27 Oct 2011)*
1528 This was the fourth release candidate of the 2.5 series.
1534 *(Released Wed, 26 Oct 2011)*
1536 This was the third release candidate of the 2.5 series.
1542 *(Released Tue, 18 Oct 2011)*
1544 This was the second release candidate of the 2.5 series.
1550 *(Released Tue, 4 Oct 2011)*
1552 This was the first release candidate of the 2.5 series.
1558 *(Released Wed, 31 Aug 2011)*
1560 This was the third beta release of the 2.5 series.
1566 *(Released Mon, 22 Aug 2011)*
1568 This was the second beta release of the 2.5 series.
1574 *(Released Fri, 12 Aug 2011)*
1576 This was the first beta release of the 2.5 series.
1582 *(Released Thu, 27 Oct 2011)*
1584 - Fixed bug when parsing command line parameter values ending in
1586 - Fixed assertion error after unclean master shutdown
1587 - Disable HTTP client pool for RPC, significantly reducing memory usage
1589 - Fixed queue archive creation with wrong permissions
1595 *(Released Tue, 23 Aug 2011)*
1599 - Fixed documentation for importing with ``--src-dir`` option
1600 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1601 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1607 *(Released Fri, 5 Aug 2011)*
1609 Many bug-fixes and a few small features:
1611 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1612 issues when you tried to add an instance with two MAC addresses in one
1614 - KVM: fixed per-instance stored UID value
1615 - KVM: configure bridged NICs at migration start
1616 - KVM: Fix a bug where instance will not start with never KVM versions
1618 - Added OS search path to ``gnt-cluster info``
1619 - Fixed an issue with ``file_storage_dir`` where you were forced to
1620 provide an absolute path, but the documentation states it is a
1621 relative path, the documentation was right
1622 - Added a new parameter to instance stop/start called ``--no-remember``
1623 that will make the state change to not be remembered
1624 - Implemented ``no_remember`` at RAPI level
1625 - Improved the documentation
1626 - Node evacuation: don't call IAllocator if node is already empty
1627 - Fixed bug in DRBD8 replace disks on current nodes
1628 - Fixed bug in recreate-disks for DRBD instances
1629 - Moved assertion checking locks in ``gnt-instance replace-disks``
1630 causing it to abort with not owning the right locks for some situation
1631 - Job queue: Fixed potential race condition when cancelling queued jobs
1632 - Fixed off-by-one bug in job serial generation
1633 - ``gnt-node volumes``: Fix instance names
1634 - Fixed aliases in bash completion
1635 - Fixed a bug in reopening log files after being sent a SIGHUP
1636 - Added a flag to burnin to allow specifying VCPU count
1637 - Bugfixes to non-root Ganeti configuration
1643 *(Released Thu, 12 May 2011)*
1645 Many bug-fixes and a few new small features:
1647 - Fixed a bug related to log opening failures
1648 - Fixed a bug in instance listing with orphan instances
1649 - Fixed a bug which prevented resetting the cluster-level node parameter
1650 ``oob_program`` to the default
1651 - Many fixes related to the ``cluster-merge`` tool
1652 - Fixed a race condition in the lock monitor, which caused failures
1653 during (at least) creation of many instances in parallel
1654 - Improved output for gnt-job info
1655 - Removed the quiet flag on some ssh calls which prevented debugging
1657 - Improved the N+1 failure messages in cluster verify by actually
1658 showing the memory values (needed and available)
1659 - Increased lock attempt timeouts so that when executing long operations
1660 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1661 too early and thus prevent the use of the 'fair' mechanism
1662 - Changed instance query data (``gnt-instance info``) to not acquire
1663 locks unless needed, thus allowing its use on locked instance if only
1664 static information is asked for
1665 - Improved behaviour with filesystems that do not support rename on an
1667 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1668 kept locks on all nodes during the wipe, which is unneeded
1669 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1670 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1671 10GiB) which caused the wipe to fail right at the end in some cases
1672 - Fixed master IP activation when doing master failover with no-voting
1673 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1674 the master node itself
1675 - Fixed potential data-loss in under disk full conditions, where Ganeti
1676 wouldn't check correctly the return code and would consider
1677 partially-written files 'correct'
1678 - Fixed bug related to multiple VGs and DRBD disk replacing
1679 - Added new disk parameter ``metavg`` that allows placement of the meta
1680 device for DRBD in a different volume group
1681 - Fixed error handling in the node daemon when the system libc doesn't
1682 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1683 - Fixed lock release during replace-disks, which kept cluster-wide locks
1684 when doing disk replaces with an iallocator script
1685 - Added check for missing bridges in cluster verify
1686 - Handle EPIPE errors while writing to the terminal better, so that
1687 piping the output to e.g. ``less`` doesn't cause a backtrace
1688 - Fixed rare case where a ^C during Luxi calls could have been
1689 interpreted as server errors, instead of simply terminating
1690 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1692 - Added a few more parameters to the KVM hypervisor, allowing a second
1693 CDROM, custom disk type for CDROMs and a floppy image
1694 - Removed redundant message in instance rename when the name is given
1696 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1697 disks on new nodes, allowing recreation when the original instance
1698 nodes are completely gone
1699 - Added option when converting disk templates to DRBD to skip waiting
1700 for the resync, in order to make the instance available sooner
1701 - Added two new variables to the OS scripts environment (containing the
1703 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1704 to allow use of ``pvgrub`` as bootloader
1705 - Changed the instance memory modifications to only check out-of-memory
1706 conditions on memory increases, and turned the secondary node warnings
1707 into errors (they can still be overridden via ``--force``)
1708 - Fixed the handling of a corner case when the Python installation gets
1709 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1710 to execute a command that doesn't exist
1711 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1712 node of the instance returned failures during instance shutdown; this
1713 adds the option ``--ignore-consistency`` to gnt-instance move
1715 And as usual, various improvements to the error messages, documentation
1722 *(Released Wed, 09 Mar 2011)*
1724 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1725 the RAPI users file if run twice (even with ``--dry-run``).
1727 The release fixes that bug (nothing else changed).
1733 *(Released Mon, 07 Mar 2011)*
1735 Final 2.4.0 release. Just a few small fixes:
1737 - Fixed RAPI node evacuate
1738 - Fixed the kvm-ifup script
1739 - Fixed internal error handling for special job cases
1740 - Updated man page to specify the escaping feature for options
1746 *(Released Mon, 28 Feb 2011)*
1748 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1749 that this feature wiped the disks of the wrong instance, leading to loss
1754 - Fixed title of query field containing instance name
1755 - Expanded the glossary in the documentation
1756 - Fixed one unittest (internal issue)
1762 *(Released Mon, 21 Feb 2011)*
1764 A number of bug fixes plus just a couple functionality changes.
1766 On the user-visible side, the ``gnt-* list`` command output has changed
1767 with respect to "special" field states. The current rc1 style of display
1768 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1769 the default output mode special fields states are displayed as follows:
1771 - Offline resource: ``*``
1772 - Unavailable/not applicable: ``-``
1773 - Data missing (RPC failure): ``?``
1774 - Unknown field: ``??``
1776 Another user-visible change is the addition of ``--force-join`` to
1781 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1782 - Fixed regression in RAPI/instance reinstall where all parameters were
1783 required (instead of optional)
1784 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1785 - Fixed iallocator usage (offline nodes were not considered offline)
1786 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1787 - Fixed hypervisor and OS parameter validation with respect to
1788 non-vm_capable nodes
1789 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1791 - Fixed ``tools/listrunner`` with respect to agent-based usage
1797 *(Released Fri, 4 Feb 2011)*
1799 Many changes and fixes since the beta1 release. While there were some
1800 internal changes, the code has been mostly stabilised for the RC
1803 Note: the dumb allocator was removed in this release, as it was not kept
1804 up-to-date with the IAllocator protocol changes. It is recommended to
1805 use the ``hail`` command from the ganeti-htools package.
1807 Note: the 2.4 and up versions of Ganeti are not compatible with the
1808 0.2.x branch of ganeti-htools. You need to upgrade to
1809 ganeti-htools-0.3.0 (or later).
1811 Regressions fixed from 2.3
1812 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1814 - Fixed the ``gnt-cluster verify-disks`` command
1815 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1817 - Fixed disk adoption breakage
1818 - Fixed wrong headers in instance listing for field aliases
1823 - Fixed corner case in KVM handling of NICs
1824 - Fixed many cases of wrong handling of non-vm_capable nodes
1825 - Fixed a bug where a missing instance symlink was not possible to
1826 recreate with any ``gnt-*`` command (now ``gnt-instance
1827 activate-disks`` does it)
1828 - Fixed the volume group name as reported by ``gnt-cluster
1830 - Increased timeouts for the import-export code, hopefully leading to
1831 fewer aborts due network or instance timeouts
1832 - Fixed bug in ``gnt-node list-storage``
1833 - Fixed bug where not all daemons were started on cluster
1834 initialisation, but only at the first watcher run
1835 - Fixed many bugs in the OOB implementation
1836 - Fixed watcher behaviour in presence of instances with offline
1838 - Fixed instance list output for instances running on the wrong node
1839 - a few fixes to the cluster-merge tool, but it still cannot merge
1840 multi-node groups (currently it is not recommended to use this tool)
1846 - Improved network configuration for the KVM hypervisor
1847 - Added e1000 as a supported NIC for Xen-HVM
1848 - Improved the lvmstrap tool to also be able to use partitions, as
1849 opposed to full disks
1850 - Improved speed of disk wiping (the cluster parameter
1851 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1852 of instance creations
1853 - Added documentation for the OS parameters
1854 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1855 hypervisor is not responding
1856 - Added display of blacklisted and hidden OS information in
1857 ``gnt-cluster info``
1858 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1859 NIC and node parameters, which might create problems with currently
1860 invalid (but undetected) configuration files, but prevents validation
1861 failures when unrelated parameters are modified
1862 - Changed cluster initialisation to wait for the master daemon to become
1864 - Expanded the RAPI interface:
1866 - Added config redistribution resource
1867 - Added activation/deactivation of instance disks
1868 - Added export of console information
1870 - Implemented log file reopening on SIGHUP, which allows using
1871 logrotate(8) for the Ganeti log files
1872 - Added a basic OOB helper script as an example
1878 *(Released Fri, 14 Jan 2011)*
1883 - Fixed timezone issues when formatting timestamps
1884 - Added support for node groups, available via ``gnt-group`` and other
1886 - Added out-of-band framework and management, see :doc:`design
1887 document <design-oob>`
1888 - Removed support for roman numbers from ``gnt-node list`` and
1889 ``gnt-instance list``.
1890 - Allowed modification of master network interface via ``gnt-cluster
1891 modify --master-netdev``
1892 - Accept offline secondaries while shutting down instance disks
1893 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1894 - Added support for multiple LVM volume groups
1895 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1897 - Added commands to list available fields:
1899 - ``gnt-node list-fields``
1900 - ``gnt-group list-fields``
1901 - ``gnt-instance list-fields``
1903 - Updated documentation and man pages
1908 - Moved ``rapi_users`` file into separate directory, now named
1909 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1911 - Added new tool for running commands on many machines,
1912 ``tools/ganeti-listrunner``
1913 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1914 improving the ``gnt-instance console`` output
1915 - Allowed customisation of disk index separator at ``configure`` time
1916 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1917 - Added support for non-partitioned md disks in ``lvmstrap``
1918 - Added script to gracefully power off KVM instances
1919 - Split ``utils`` module into smaller parts
1920 - Changed query operations to return more detailed information, e.g.
1921 whether an information is unavailable due to an offline node. To use
1922 this new functionality, the LUXI call ``Query`` must be used. Field
1923 information is now stored by the master daemon and can be retrieved
1924 using ``QueryFields``. Instances, nodes and groups can also be queried
1925 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1926 exposed via RAPI). The following commands make use of this
1927 infrastructure change:
1929 - ``gnt-group list``
1930 - ``gnt-group list-fields``
1932 - ``gnt-node list-fields``
1933 - ``gnt-instance list``
1934 - ``gnt-instance list-fields``
1935 - ``gnt-debug locks``
1940 - New RAPI resources (see :doc:`rapi`):
1944 - ``/2/groups/[group_name]``
1945 - ``/2/groups/[group_name]/assign-nodes``
1946 - ``/2/groups/[group_name]/modify``
1947 - ``/2/groups/[group_name]/rename``
1948 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1952 - Implemented ``no_install`` for instance creation
1953 - Implemented OS parameters for instance reinstallation, allowing
1954 use of special settings on reinstallation (e.g. for preserving data)
1959 - Added IPv6 support in import/export
1960 - Pause DRBD synchronization while wiping disks on instance creation
1961 - Updated unittests and QA scripts
1962 - Improved network parameters passed to KVM
1963 - Converted man pages from docbook to reStructuredText
1969 *(Released Mon, 20 Dec 2010)*
1971 Released version 2.3.1~rc1 without any changes.
1977 *(Released Wed, 1 Dec 2010)*
1979 - impexpd: Disable OpenSSL compression in socat if possible (backport
1980 from master, commit e90739d625b, see :doc:`installation guide
1981 <install-quick>` for details)
1982 - Changed unittest coverage report to exclude test scripts
1983 - Added script to check version format
1989 *(Released Wed, 1 Dec 2010)*
1991 Released version 2.3.0~rc1 without any changes.
1997 *(Released Fri, 19 Nov 2010)*
1999 A number of bugfixes and documentation updates:
2001 - Update ganeti-os-interface documentation
2002 - Fixed a bug related to duplicate MACs or similar items which should be
2004 - Fix breakage in OS state modify
2005 - Reinstall instance: disallow offline secondaries (fixes bug related to
2006 OS changing but reinstall failing)
2007 - plus all the other fixes between 2.2.1 and 2.2.2
2013 *(Released Tue, 2 Nov 2010)*
2015 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
2016 - Fixed master failover race with watcher
2017 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
2019 - Accept previously stopped instance for export with instance removal
2020 - Simplify and extend the environment variables for instance OS scripts
2021 - Added new node flags, ``master_capable`` and ``vm_capable``
2022 - Added optional instance disk wiping prior during allocation. This is a
2023 cluster-wide option and can be set/modified using
2024 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
2025 - Added IPv6 support, see :doc:`design document <design-2.3>` and
2026 :doc:`install-quick`
2027 - Added a new watcher option (``--ignore-pause``)
2028 - Added option to ignore offline node on instance start/stop
2029 (``--ignore-offline``)
2030 - Allow overriding OS parameters with ``gnt-instance reinstall``
2031 - Added ability to change node's secondary IP address using ``gnt-node
2033 - Implemented privilege separation for all daemons except
2034 ``ganeti-noded``, see ``configure`` options
2035 - Complain if an instance's disk is marked faulty in ``gnt-cluster
2037 - Implemented job priorities (see ``ganeti(7)`` manpage)
2038 - Ignore failures while shutting down instances during failover from
2040 - Exit daemon's bootstrap process only once daemon is ready
2041 - Export more information via ``LUInstanceQuery``/remote API
2042 - Improved documentation, QA and unittests
2043 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
2044 restart if the file was created or changed
2045 - Added LUXI protocol version sent with each request and response,
2046 allowing detection of server/client mismatches
2047 - Moved the Python scripts among gnt-* and ganeti-* into modules
2048 - Moved all code related to setting up SSH to an external script,
2050 - Infrastructure changes for node group support in future versions
2056 *(Released Fri, 19 Nov 2010)*
2058 A few small bugs fixed, and some improvements to the build system:
2060 - Fix documentation regarding conversion to drbd
2061 - Fix validation of parameters in cluster modify (``gnt-cluster modify
2063 - Fix error handling in node modify with multiple changes
2064 - Allow remote imports without checked names
2070 *(Released Tue, 19 Oct 2010)*
2072 - Disable SSL session ID cache in RPC client
2078 *(Released Thu, 14 Oct 2010)*
2080 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
2081 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2083 - Fix problems with interaction between Curl and Python's HTTP server,
2084 resulting in increased speed in many RPC calls
2085 - Improve our release script to prevent breakage with older aclocal and
2092 *(Released Thu, 7 Oct 2010)*
2094 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2095 value retrieved from master
2096 - Added support for blacklisted or hidden OS definitions
2097 - Added simple lock monitor (accessible via (``gnt-debug locks``)
2098 - Added support for -mem-path in KVM hypervisor abstraction layer
2099 - Allow overriding instance parameters in tool for inter-cluster
2100 instance moves (``tools/move-instance``)
2101 - Improved opcode summaries (e.g. in ``gnt-job list``)
2102 - Improve consistency of OS listing by sorting it
2103 - Documentation updates
2109 *(Released Fri, 8 Oct 2010)*
2111 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2117 *(Released Mon, 4 Oct 2010)*
2119 - Fixed regression in ``gnt-instance rename``
2125 *(Released Wed, 22 Sep 2010)*
2127 - Fixed OS_VARIANT variable for OS scripts
2128 - Fixed cluster tag operations via RAPI
2129 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2130 - Disabled RAPI CA checks in watcher
2136 *(Released Mon, 23 Aug 2010)*
2138 - Support DRBD versions of the format "a.b.c.d"
2140 - Re-introduce support for usage from multiple threads in RAPI client
2141 - Instance renames and modify via RAPI
2142 - Work around race condition between processing and archival in job
2144 - Mark opcodes following failed one as failed, too
2145 - Job field ``lock_status`` was removed due to difficulties making it
2146 work with the changed job queue in Ganeti 2.2; a better way to monitor
2147 locks is expected for a later 2.2.x release
2148 - Fixed dry-run behaviour with many commands
2149 - Support ``ssh-agent`` again when adding nodes
2150 - Many additional bugfixes
2156 *(Released Fri, 30 Jul 2010)*
2158 Important change: the internal RPC mechanism between Ganeti nodes has
2159 changed from using a home-grown http library (based on the Python base
2160 libraries) to use the PycURL library. This requires that PycURL is
2161 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2162 against GnuTLS by default. cURL's support for GnuTLS had known issues
2163 before cURL 7.21.0 and we recommend using the latest cURL release or
2164 linking against OpenSSL. Most other distributions already link PycURL
2165 and cURL against OpenSSL. The command::
2167 python -c 'import pycurl; print pycurl.version'
2169 can be used to determine the libraries PycURL and cURL are linked
2172 Other significant changes:
2174 - Rewrote much of the internals of the job queue, in order to achieve
2175 better parallelism; this decouples job query operations from the job
2176 processing, and it should allow much nicer behaviour of the master
2177 daemon under load, and it also has uncovered some long-standing bugs
2178 related to the job serialisation (now fixed)
2179 - Added a default iallocator setting to the cluster parameters,
2180 eliminating the need to always pass nodes or an iallocator for
2181 operations that require selection of new node(s)
2182 - Added experimental support for the LXC virtualization method
2183 - Added support for OS parameters, which allows the installation of
2184 instances to pass parameter to OS scripts in order to customise the
2186 - Added a hypervisor parameter controlling the migration type (live or
2187 non-live), since hypervisors have various levels of reliability; this
2188 has renamed the 'live' parameter to 'mode'
2189 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2190 logical volumes, meaning that cluster verify will ignore them and not
2191 flag their presence as errors
2192 - The watcher will now reset the error count for failed instances after
2193 8 hours, thus allowing self-healing if the problem that caused the
2194 instances to be down/fail to start has cleared in the meantime
2195 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2196 check for, and warn, if the drbd module parameter ``usermode_helper``
2197 is not consistent with the cluster-wide setting; this is needed to
2198 make diagnose easier of failed drbd creations
2199 - Started adding base IPv6 support, but this is not yet
2200 enabled/available for use
2201 - Rename operations (cluster, instance) will now return the new name,
2202 which is especially useful if a short name was passed in
2203 - Added support for instance migration in RAPI
2204 - Added a tool to pre-configure nodes for the SSH setup, before joining
2205 them to the cluster; this will allow in the future a simplified model
2206 for node joining (but not yet fully enabled in 2.2); this needs the
2207 paramiko python library
2208 - Fixed handling of name-resolving errors
2209 - Fixed consistency of job results on the error path
2210 - Fixed master-failover race condition when executed multiple times in
2212 - Fixed many bugs related to the job queue (mostly introduced during the
2213 2.2 development cycle, so not all are impacting 2.1)
2214 - Fixed instance migration with missing disk symlinks
2215 - Fixed handling of unknown jobs in ``gnt-job archive``
2216 - And many other small fixes/improvements
2220 - Enhanced both the unittest and the QA coverage
2221 - Switched the opcode validation to a generic model, and extended the
2222 validation to all opcode parameters
2223 - Changed more parts of the code that write shell scripts to use the
2225 - Switched the master daemon to use the asyncore library for the Luxi
2232 *(Released Thu, 17 Jun 2010)*
2234 - Added tool (``move-instance``) and infrastructure to move instances
2235 between separate clusters (see :doc:`separate documentation
2236 <move-instance>` and :doc:`design document <design-2.2>`)
2237 - Added per-request RPC timeout
2238 - RAPI now requires a Content-Type header for requests with a body (e.g.
2239 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2240 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2241 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2243 - Implemented initial support for running Ganeti daemons as separate
2244 users, see configure-time flags ``--with-user-prefix`` and
2245 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2247 - Instances can be removed after export (``gnt-backup export
2248 --remove-instance``)
2249 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2250 key (instead of 1024 bit)
2251 - Added new cluster configuration file for cluster domain secret
2252 - Import/export now use SSL instead of SSH
2253 - Added support for showing estimated time when exporting an instance,
2254 see the ``ganeti-os-interface(7)`` manpage and look for
2261 *(Released Tue, 16 Nov 2010)*
2263 Some more bugfixes. Unless critical bugs occur, this will be the last
2266 - Fix case of MAC special-values
2267 - Fix mac checker regex
2268 - backend: Fix typo causing "out of range" error
2269 - Add missing --units in gnt-instance list man page
2275 *(Released Tue, 24 Aug 2010)*
2278 - Don't ignore secondary node silently on non-mirrored disk templates
2280 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2281 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2282 - Properly document the usb_mouse parameter
2283 - Fix path in ganeti-rapi(8) (issue 116)
2284 - Adjust error message when the ganeti user's .ssh directory is
2286 - Add same-node-check when changing the disk template to drbd
2292 *(Released Fri, 16 Jul 2010)*
2295 - Add an option to only select some reboot types during qa/burnin.
2296 (on some hypervisors consequent reboots are not supported)
2297 - Fix infrequent race condition in master failover. Sometimes the old
2298 master ip address would be still detected as up for a short time
2299 after it was removed, causing failover to fail.
2300 - Decrease mlockall warnings when the ctypes module is missing. On
2301 Python 2.4 we support running even if no ctypes module is installed,
2302 but we were too verbose about this issue.
2303 - Fix building on old distributions, on which man doesn't have a
2305 - Fix RAPI not to ignore the MAC address on instance creation
2306 - Implement the old instance creation format in the RAPI client.
2312 *(Released Thu, 01 Jul 2010)*
2314 A small bugfix release:
2315 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2316 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2317 a non-existing option
2318 - Fix instance create: the --force-variant option was ignored
2319 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2320 - Fix modify node storage with non-FQDN arguments
2321 - Fix RAPI client to authenticate under Python 2.6 when used
2322 for more than 5 requests needing authentication
2323 - Fix gnt-instance modify -t (storage) giving a wrong error message
2324 when converting a non-shutdown drbd instance to plain
2330 *(Released Fri, 18 Jun 2010)*
2332 A small bugfix release:
2334 - Fix live migration of KVM instances started with older Ganeti
2335 versions which had fewer hypervisor parameters
2336 - Fix gnt-instance grow-disk on down instances
2337 - Fix an error-reporting bug during instance migration
2338 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2339 silently ignoring broken ones
2340 - Fix an RPC error reporting bug affecting, for example, RAPI client
2342 - Fix bug triggered by different API version os-es on different nodes
2343 - Fix a bug in instance startup with custom hvparams: OS level
2344 parameters would fail to be applied.
2345 - Fix the RAPI client under Python 2.6 (but more work is needed to
2346 make it work completely well with OpenSSL)
2347 - Fix handling of errors when resolving names from DNS
2353 *(Released Thu, 3 Jun 2010)*
2355 A medium sized development cycle. Some new features, and some
2356 fixes/small improvements/cleanups.
2358 Significant features
2359 ~~~~~~~~~~~~~~~~~~~~
2361 The node deamon now tries to mlock itself into memory, unless the
2362 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2363 its logs, and falls back to console logging. This allows emergency
2364 features such as ``gnt-node powercycle`` to work even in the event of a
2365 broken node disk (tested offlining the disk hosting the node's
2366 filesystem and dropping its memory caches; don't try this at home)
2368 KVM: add vhost-net acceleration support. It can be tested with a new
2369 enough version of the kernel and of qemu-kvm.
2371 KVM: Add instance chrooting feature. If you use privilege dropping for
2372 your VMs you can also now force them to chroot to an empty directory,
2373 before starting the emulated guest.
2375 KVM: Add maximum migration bandwith and maximum downtime tweaking
2376 support (requires a new-enough version of qemu-kvm).
2378 Cluster verify will now warn if the master node doesn't have the master
2379 ip configured on it.
2381 Add a new (incompatible) instance creation request format to RAPI which
2382 supports all parameters (previously only a subset was supported, and it
2383 wasn't possible to extend the old format to accomodate all the new
2384 features. The old format is still supported, and a client can check for
2385 this feature, before using it, by checking for its presence in the
2386 ``features`` RAPI resource.
2388 Now with ancient latin support. Try it passing the ``--roman`` option to
2389 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2390 (requires the python-roman module to be installed, in order to work).
2395 As usual many internal code refactorings, documentation updates, and
2398 - Lots of improvements and cleanups to the experimental Remote API
2399 (RAPI) client library.
2400 - A new unit test suite for the core daemon libraries.
2401 - A fix to creating missing directories makes sure the umask is not
2402 applied anymore. This enforces the same directory permissions
2404 - Better handling terminating daemons with ctrl+c (used when running
2405 them in debugging mode).
2406 - Fix a race condition in live migrating a KVM instance, when stat()
2407 on the old proc status file returned EINVAL, which is an unexpected
2409 - Fixed manpage checking with newer man and utf-8 charachters. But now
2410 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2416 *(Released Fri, 7 May 2010)*
2418 Fix a bug which prevented untagged KVM instances from starting.
2424 *(Released Fri, 7 May 2010)*
2426 Another release with a long development cycle, during which many
2427 different features were added.
2429 Significant features
2430 ~~~~~~~~~~~~~~~~~~~~
2432 The KVM hypervisor now can run the individual instances as non-root, to
2433 reduce the impact of a VM being hijacked due to bugs in the
2434 hypervisor. It is possible to run all instances as a single (non-root)
2435 user, to manually specify a user for each instance, or to dynamically
2436 allocate a user out of a cluster-wide pool to each instance, with the
2437 guarantee that no two instances will run under the same user ID on any
2440 An experimental RAPI client library, that can be used standalone
2441 (without the other Ganeti libraries), is provided in the source tree as
2442 ``lib/rapi/client.py``. Note this client might change its interface in
2443 the future, as we iterate on its capabilities.
2445 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2446 replace the cluster's certificates and crypto keys. This might help in
2447 case they have been compromised, or have simply expired.
2449 A new disk option for instance creation has been added that allows one
2450 to "adopt" currently existing logical volumes, with data
2451 preservation. This should allow easier migration to Ganeti from
2452 unmanaged (or managed via other software) instances.
2454 Another disk improvement is the possibility to convert between redundant
2455 (DRBD) and plain (LVM) disk configuration for an instance. This should
2456 allow better scalability (starting with one node and growing the
2457 cluster, or shrinking a two-node cluster to one node).
2459 A new feature that could help with automated node failovers has been
2460 implemented: if a node sees itself as offline (by querying the master
2461 candidates), it will try to shutdown (hard) all instances and any active
2462 DRBD devices. This reduces the risk of duplicate instances if an
2463 external script automatically failovers the instances on such nodes. To
2464 enable this, the cluster parameter ``maintain_node_health`` should be
2465 enabled; in the future this option (per the name) will enable other
2466 automatic maintenance features.
2468 Instance export/import now will reuse the original instance
2469 specifications for all parameters; that means exporting an instance,
2470 deleting it and the importing it back should give an almost identical
2471 instance. Note that the default import behaviour has changed from
2472 before, where it created only one NIC; now it recreates the original
2475 Cluster verify has added a few new checks: SSL certificates validity,
2476 /etc/hosts consistency across the cluster, etc.
2481 As usual, many internal changes were done, documentation fixes,
2484 - Fixed cluster initialization with disabled cluster storage (regression
2485 introduced in 2.1.1)
2486 - File-based storage supports growing the disks
2487 - Fixed behaviour of node role changes
2488 - Fixed cluster verify for some corner cases, plus a general rewrite of
2489 cluster verify to allow future extension with more checks
2490 - Fixed log spamming by watcher and node daemon (regression introduced
2492 - Fixed possible validation issues when changing the list of enabled
2494 - Fixed cleanup of /etc/hosts during node removal
2495 - Fixed RAPI response for invalid methods
2496 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2497 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2498 from ide disks, etc.)
2499 - Allow OS changes without re-installation (to record a changed OS
2500 outside of Ganeti, or to allow OS renames)
2501 - Allow instance creation without OS installation (useful for example if
2502 the OS will be installed manually, or restored from a backup not in
2504 - Implemented option to make cluster ``copyfile`` use the replication
2506 - Added list of enabled hypervisors to ssconf (possibly useful for
2508 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2510 - A partial form of node re-IP is possible via node readd, which now
2511 allows changed node primary IP
2512 - Command line utilities now show an informational message if the job is
2514 - The logs of the master daemon now show the PID/UID/GID of the
2521 *(Released Fri, 12 Mar 2010)*
2523 During the 2.1.0 long release candidate cycle, a lot of improvements and
2524 changes have accumulated with were released later as 2.1.1.
2529 The node evacuate command (``gnt-node evacuate``) was significantly
2530 rewritten, and as such the IAllocator protocol was changed - a new
2531 request type has been added. This unfortunate change during a stable
2532 series is designed to improve performance of node evacuations; on
2533 clusters with more than about five nodes and which are well-balanced,
2534 evacuation should proceed in parallel for all instances of the node
2535 being evacuated. As such, any existing IAllocator scripts need to be
2536 updated, otherwise the above command will fail due to the unknown
2537 request. The provided "dumb" allocator has not been updated; but the
2538 ganeti-htools package supports the new protocol since version 0.2.4.
2540 Another important change is increased validation of node and instance
2541 names. This might create problems in special cases, if invalid host
2542 names are being used.
2544 Also, a new layer of hypervisor parameters has been added, that sits at
2545 OS level between the cluster defaults and the instance ones. This allows
2546 customisation of virtualization parameters depending on the installed
2547 OS. For example instances with OS 'X' may have a different KVM kernel
2548 (or any other parameter) than the cluster defaults. This is intended to
2549 help managing a multiple OSes on the same cluster, without manual
2550 modification of each instance's parameters.
2552 A tool for merging clusters, ``cluster-merge``, has been added in the
2553 tools sub-directory.
2558 - Improved the int/float conversions that should make the code more
2559 robust in face of errors from the node daemons
2560 - Fixed the remove node code in case of internal configuration errors
2561 - Fixed the node daemon behaviour in face of inconsistent queue
2562 directory (e.g. read-only file-system where we can't open the files
2564 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2565 now it either aborts cleanly or, if given the new "auto_promote"
2566 parameter, will automatically promote other nodes as needed
2567 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2568 completely prevent Ganeti from working
2569 - Fixed bug for instance export when not all disks were successfully
2571 - Fixed behaviour of node add when the new node is slow in starting up
2573 - Fixed handling of signals in the LUXI client, which should improve
2574 behaviour of command-line scripts
2575 - Added checks for invalid node/instance names in the configuration (now
2576 flagged during cluster verify)
2577 - Fixed watcher behaviour for disk activation errors
2578 - Fixed two potentially endless loops in http library, which led to the
2579 RAPI daemon hanging and consuming 100% CPU in some cases
2580 - Fixed bug in RAPI daemon related to hashed passwords
2581 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2583 - Enhanced compatibility with non-Debian OSes, but not using absolute
2584 path in some commands and allowing customisation of the ssh
2585 configuration directory
2586 - Fixed possible future issue with new Python versions by abiding to the
2587 proper use of ``__slots__`` attribute on classes
2588 - Added checks that should prevent directory traversal attacks
2589 - Many documentation fixes based on feedback from users
2594 - Added an "early_release" more for instance replace disks and node
2595 evacuate, where we release locks earlier and thus allow higher
2596 parallelism within the cluster
2597 - Added watcher hooks, intended to allow the watcher to restart other
2598 daemons (e.g. from the ganeti-nbma project), but they can be used of
2599 course for any other purpose
2600 - Added a compile-time disable for DRBD barriers, to increase
2601 performance if the administrator trusts the power supply or the
2602 storage system to not lose writes
2603 - Added the option of using syslog for logging instead of, or in
2604 addition to, Ganeti's own log files
2605 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2606 can indeed boot from a paravirtual NIC
2607 - Added a generic debug level for many operations; while this is not
2608 used widely yet, it allows one to pass the debug value all the way to
2610 - Enhanced the hooks environment for instance moves (failovers,
2611 migrations) where the primary/secondary nodes changed during the
2612 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2613 - Enhanced data validations for many user-supplied values; one important
2614 item is the restrictions imposed on instance and node names, which
2615 might reject some (invalid) host names
2616 - Add a configure-time option to disable file-based storage, if it's not
2617 needed; this allows greater security separation between the master
2618 node and the other nodes from the point of view of the inter-node RPC
2620 - Added user notification in interactive tools if job is waiting in the
2621 job queue or trying to acquire locks
2622 - Added log messages when a job is waiting for locks
2623 - Added filtering by node tags in instance operations which admit
2624 multiple instances (start, stop, reboot, reinstall)
2625 - Added a new tool for cluster mergers, ``cluster-merge``
2626 - Parameters from command line which are of the form ``a=b,c=d`` can now
2627 use backslash escapes to pass in values which contain commas,
2628 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2630 - For KVM, the instance name is the first parameter passed to KVM, so
2631 that it's more visible in the process list
2637 *(Released Tue, 2 Mar 2010)*
2639 Ganeti 2.1 brings many improvements with it. Major changes:
2641 - Added infrastructure to ease automated disk repairs
2642 - Added new daemon to export configuration data in a cheaper way than
2643 using the remote API
2644 - Instance NICs can now be routed instead of being associated with a
2646 - Improved job locking logic to reduce impact of jobs acquiring multiple
2647 locks waiting for other long-running jobs
2649 In-depth implementation details can be found in the Ganeti 2.1 design
2655 - Added chroot hypervisor
2656 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2658 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2659 ``bootloader_path`` and ``bootloader_args``)
2660 - Added the ``use_localtime`` option for the xen-hvm and kvm
2661 hypervisors, and the default value for this has changed to false (in
2662 2.0 xen-hvm always enabled it)
2663 - Added luxi call to submit multiple jobs in one go
2664 - Added cluster initialization option to not modify ``/etc/hosts``
2666 - Added network interface parameters
2667 - Added dry run mode to some LUs
2668 - Added RAPI resources:
2670 - ``/2/instances/[instance_name]/info``
2671 - ``/2/instances/[instance_name]/replace-disks``
2672 - ``/2/nodes/[node_name]/evacuate``
2673 - ``/2/nodes/[node_name]/migrate``
2674 - ``/2/nodes/[node_name]/role``
2675 - ``/2/nodes/[node_name]/storage``
2676 - ``/2/nodes/[node_name]/storage/modify``
2677 - ``/2/nodes/[node_name]/storage/repair``
2679 - Added OpCodes to evacuate or migrate all instances on a node
2680 - Added new command to list storage elements on nodes (``gnt-node
2681 list-storage``) and modify them (``gnt-node modify-storage``)
2682 - Added new ssconf files with master candidate IP address
2683 (``ssconf_master_candidates_ips``), node primary IP address
2684 (``ssconf_node_primary_ips``) and node secondary IP address
2685 (``ssconf_node_secondary_ips``)
2686 - Added ``ganeti-confd`` and a client library to query the Ganeti
2687 configuration via UDP
2688 - Added ability to run hooks after cluster initialization and before
2690 - Added automatic mode for disk replace (``gnt-instance replace-disks
2692 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2693 after catastrophic data-loss
2694 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2696 - Added ``gnt-instance move`` command to move instances
2697 - Added ``gnt-cluster watcher`` command to control watcher
2698 - Added ``gnt-node powercycle`` command to powercycle nodes
2699 - Added new job status field ``lock_status``
2700 - Added parseable error codes to cluster verification (``gnt-cluster
2701 verify --error-codes``) and made output less verbose (use
2702 ``--verbose`` to restore previous behaviour)
2703 - Added UUIDs to the main config entities (cluster, nodes, instances)
2704 - Added support for OS variants
2705 - Added support for hashed passwords in the Ganeti remote API users file
2707 - Added option to specify maximum timeout on instance shutdown
2708 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2709 - Added new helper script to start and stop Ganeti daemons
2710 (``daemon-util``), with the intent to reduce the work necessary to
2711 adjust Ganeti for non-Debian distributions and to start/stop daemons
2713 - Added more unittests
2714 - Fixed critical bug in ganeti-masterd startup
2715 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2716 now customisable at the cluster level for both the KVM and Xen
2717 hypervisors using the new ``migration_port`` parameter
2718 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2719 reinstalling an instance
2720 - Allowed ``@`` in tag names
2721 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2722 - Many documentation updates
2723 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2724 - ``gnt-instance reinstall`` can now reinstall multiple instances
2725 - Updated many command line parameters
2726 - Introduced new OS API version 15
2727 - No longer support a default hypervisor
2728 - Treat virtual LVs as inexistent
2729 - Improved job locking logic to reduce lock contention
2730 - Match instance and node names case insensitively
2731 - Reimplemented bash completion script to be more complete
2738 *(Released Thu, 4 Feb 2010)*
2740 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2741 - Fix a string formatting bug
2742 - Improve safety of the code in some error paths
2743 - Improve data validation in the master of values returned from nodes
2749 *(Released Thu, 17 Dec 2009)*
2751 - Fix security issue due to missing validation of iallocator names; this
2752 allows local and remote execution of arbitrary executables
2753 - Fix failure of gnt-node list during instance removal
2754 - Ship the RAPI documentation in the archive
2760 *(Released Wed, 30 Sep 2009)*
2762 - Fixed many wrong messages
2763 - Fixed a few bugs related to the locking library
2764 - Fixed MAC checking at instance creation time
2765 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2766 - Fixed a few issues related to signal handling in both daemons and
2768 - Fixed the example startup script provided
2769 - Fixed insserv dependencies in the example startup script (patch from
2771 - Fixed handling of drained nodes in the iallocator framework
2772 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2774 - Fixed error related to invalid job IDs in job polling
2775 - Fixed job/opcode persistence on unclean master shutdown
2776 - Fixed handling of partial job processing after unclean master
2778 - Fixed error reporting from LUs, previously all errors were converted
2779 into execution errors
2780 - Fixed error reporting from burnin
2781 - Decreased significantly the memory usage of the job queue
2782 - Optimised slightly multi-job submission
2783 - Optimised slightly opcode loading
2784 - Backported the multi-job submit framework from the development
2785 branch; multi-instance start and stop should be faster
2786 - Added script to clean archived jobs after 21 days; this will reduce
2787 the size of the queue directory
2788 - Added some extra checks in disk size tracking
2789 - Added an example ethers hook script
2790 - Added a cluster parameter that prevents Ganeti from modifying of
2792 - Added more node information to RAPI responses
2793 - Added a ``gnt-job watch`` command that allows following the ouput of a
2795 - Added a bind-address option to ganeti-rapi
2796 - Added more checks to the configuration verify
2797 - Enhanced the burnin script such that some operations can be retried
2799 - Converted instance reinstall to multi-instance model
2805 *(Released Fri, 7 Aug 2009)*
2807 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2808 to allow using the pre-2.0.2 behaviour in activation, if any existing
2809 instances have mismatched disk sizes in the configuration
2810 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2811 any configuration mismatches for disk sizes
2812 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2813 failover to work on two-node clusters
2814 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2816 - Fixed detection of OS script errors in ``gnt-backup export``
2817 - Fixed exit code of ``gnt-backup export``
2823 *(Released Fri, 17 Jul 2009)*
2825 - Added experimental support for stripped logical volumes; this should
2826 enhance performance but comes with a higher complexity in the block
2827 device handling; stripping is only enabled when passing
2828 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2829 affected even in the non-stripped mode
2830 - Improved resiliency against transient failures at the end of DRBD
2831 resyncs, and in general of DRBD resync checks
2832 - Fixed a couple of issues with exports and snapshot errors
2833 - Fixed a couple of issues in instance listing
2834 - Added display of the disk size in ``gnt-instance info``
2835 - Fixed checking for valid OSes in instance creation
2836 - Fixed handling of the "vcpus" parameter in instance listing and in
2837 general of invalid parameters
2838 - Fixed http server library, and thus RAPI, to handle invalid
2839 username/password combinations correctly; this means that now they
2840 report unauthorized for queries too, not only for modifications,
2841 allowing earlier detect of configuration problems
2842 - Added a new "role" node list field, equivalent to the master/master
2843 candidate/drained/offline flags combinations
2844 - Fixed cluster modify and changes of candidate pool size
2845 - Fixed cluster verify error messages for wrong files on regular nodes
2846 - Fixed a couple of issues with node demotion from master candidate role
2847 - Fixed node readd issues
2848 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2849 - Added a new ``--no-voting`` option for masterfailover to fix failover
2850 on two-nodes clusters when the former master node is unreachable
2851 - Added instance reinstall over RAPI
2857 *(Released Tue, 16 Jun 2009)*
2859 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2860 allow re-adding the start in single-user option (regression from 1.2)
2861 - the watcher writes the instance status to a file, to allow monitoring
2862 to report the instance status (from the master) based on cached
2863 results of the watcher's queries; while this can get stale if the
2864 watcher is being locked due to other work on the cluster, this is
2865 still an improvement
2866 - the watcher now also restarts the node daemon and the rapi daemon if
2868 - fixed the watcher to handle full and drained queue cases
2869 - hooks export more instance data in the environment, which helps if
2870 hook scripts need to take action based on the instance's properties
2871 (no longer need to query back into ganeti)
2872 - instance failovers when the instance is stopped do not check for free
2873 RAM, so that failing over a stopped instance is possible in low memory
2875 - rapi uses queries for tags instead of jobs (for less job traffic), and
2876 for cluster tags it won't talk to masterd at all but read them from
2878 - a couple of error handling fixes in RAPI
2879 - drbd handling: improved the error handling of inconsistent disks after
2880 resync to reduce the frequency of "there are some degraded disks for
2881 this instance" messages
2882 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2883 error handling path called a wrong function name)
2889 *(Released Wed, 27 May 2009)*
2891 - no changes from rc5
2897 *(Released Wed, 20 May 2009)*
2899 - fix a couple of bugs (validation, argument checks)
2900 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2901 - some small improvements to RAPI and IAllocator
2902 - make watcher automatically start the master daemon if down
2908 *(Released Mon, 27 Apr 2009)*
2910 - change the OS list to not require locks; this helps with big clusters
2911 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2912 volume group is broken
2913 - ``gnt-instance info``, without any arguments, doesn't run for all
2914 instances anymore; either pass ``--all`` or pass the desired
2915 instances; this helps against mistakes on big clusters where listing
2916 the information for all instances takes a long time
2917 - miscellaneous doc and man pages fixes
2923 *(Released Wed, 8 Apr 2009)*
2925 - Change the internal locking model of some ``gnt-node`` commands, in
2926 order to reduce contention (and blocking of master daemon) when
2927 batching many creation/reinstall jobs
2928 - Fixes to Xen soft reboot
2929 - No longer build documentation at build time, instead distribute it in
2930 the archive, in order to reduce the need for the whole docbook/rst
2937 *(Released Fri, 27 Mar 2009)*
2939 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2940 - Fix watcher startup sequence, improves the behaviour of busy clusters
2941 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2942 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2943 bind address and other places
2944 - Some documentation fixes and updates
2950 *(Released Mon, 2 Mar 2009)*
2952 - More documentation updates, now all docs should be more-or-less
2954 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2956 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2958 - Fix an issue related to $libdir/run/ganeti and cluster creation
2964 *(Released Thu, 19 Feb 2009)*
2966 - Xen PVM and KVM have switched the default value for the instance root
2967 disk to the first partition on the first drive, instead of the whole
2968 drive; this means that the OS installation scripts must be changed
2970 - Man pages have been updated
2971 - RAPI has been switched by default to HTTPS, and the exported functions
2972 should all work correctly
2973 - RAPI v1 has been removed
2974 - Many improvements to the KVM hypervisor
2975 - Block device errors are now better reported
2976 - Many other bugfixes and small improvements
2982 *(Released Mon, 26 Jan 2009)*
2984 - Version 2 is a general rewrite of the code and therefore the
2985 differences are too many to list, see the design document for 2.0 in
2986 the ``doc/`` subdirectory for more details
2987 - In this beta version there is not yet a migration path from 1.2 (there
2988 will be one in the final 2.0 release)
2989 - A few significant changes are:
2991 - all commands are executed by a daemon (``ganeti-masterd``) and the
2992 various ``gnt-*`` commands are just front-ends to it
2993 - all the commands are entered into, and executed from a job queue,
2994 see the ``gnt-job(8)`` manpage
2995 - the RAPI daemon supports read-write operations, secured by basic
2996 HTTP authentication on top of HTTPS
2997 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2998 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2999 to migrate to DRBD 8 first while still running Ganeti 1.2)
3000 - DRBD devices are using statically allocated minor numbers, which
3001 will be assigned to existing instances during the migration process
3002 - there is support for both Xen PVM and Xen HVM instances running on
3004 - KVM virtualization is supported too
3005 - file-based storage has been implemented, which means that it is
3006 possible to run the cluster without LVM and DRBD storage, for
3007 example using a shared filesystem exported from shared storage (and
3008 still have live migration)
3014 *(Released Tue, 13 Jan 2009)*
3016 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
3017 - Reuse the old instance mac address by default on instance import, if
3018 the instance name is the same.
3019 - Handle situations in which the node info rpc returns incomplete
3021 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
3022 - Improved version of batcher:
3024 - state file support
3025 - instance mac address support
3026 - support for HVM clusters/instances
3028 - Add an option to show the number of cpu sockets and nodes in
3030 - Support OSes that handle more than one version of the OS api (but do
3031 not change the current API in any other way)
3032 - Fix ``gnt-node migrate``
3033 - ``gnt-debug`` man page
3034 - Fixes various more typos and small issues
3035 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
3041 *(Released Wed, 24 Sep 2008)*
3043 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
3044 type of disk exported to fully virtualized instances.
3045 - provide access to the serial console of HVM instances
3046 - instance auto_balance flag, set by default. If turned off it will
3047 avoid warnings on cluster verify if there is not enough memory to fail
3048 over an instance. in the future it will prevent automatically failing
3049 it over when we will support that.
3050 - batcher tool for instance creation, see ``tools/README.batcher``
3051 - ``gnt-instance reinstall --select-os`` to interactively select a new
3052 operating system when reinstalling an instance.
3053 - when changing the memory amount on instance modify a check has been
3054 added that the instance will be able to start. also warnings are
3055 emitted if the instance will not be able to fail over, if auto_balance
3057 - documentation fixes
3058 - sync fields between ``gnt-instance list/modify/add/import``
3059 - fix a race condition in drbd when the sync speed was set after giving
3060 the device a remote peer.
3066 *(Released Tue, 22 Jul 2008)*
3068 - note: the allowed size and number of tags per object were reduced
3069 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
3070 - fixed twisted 8.x compatibility
3071 - fixed ``gnt-instance replace-disks`` with iallocator
3072 - add TCP keepalives on twisted connections to detect restarted nodes
3073 - disk increase support, see ``gnt-instance grow-disk``
3074 - implement bulk node/instance query for RAPI
3075 - add tags in node/instance listing (optional)
3076 - experimental migration (and live migration) support, read the man page
3077 for ``gnt-instance migrate``
3078 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3079 has some small improvements in handling its state file
3085 *(Released Fri, 13 Jun 2008)*
3087 - Experimental readonly, REST-based remote API implementation;
3088 automatically started on master node, TCP port 5080, if enabled by
3089 ``--enable-rapi`` parameter to configure script.
3090 - Instance allocator support. Add and import instance accept a
3091 ``--iallocator`` parameter, and call that instance allocator to decide
3092 which node to use for the instance. The iallocator document describes
3093 what's expected from an allocator script.
3094 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3095 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3096 node is lost there is still enough memory to fail over the instances
3098 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3099 ``gnt-cluster verify``, to check for site-specific compliance. All the
3100 hooks will run, and their output, if any, will be displayed. Any
3101 failing hook will make the verification return an error value.
3102 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3103 primary and secondary interfaces
3104 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3105 that is still declared as part of the cluster and has failed.
3106 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3107 specifying output fields, that just adds the chosen fields to the
3109 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3110 export from the filesystem.
3111 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3112 have been added. Using them you can enable/disable acpi and pae
3113 support, and specify a path for a cd image to be exported to the
3114 instance. These parameters as the name suggest only work on HVM
3116 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3117 PAE support will be set to the previously hardcoded values, but the
3118 (previously hardcoded) path to the CDROM ISO image will be unset and
3119 if required, needs to be set manually with ``gnt-instance modify``
3121 - The address to which an instance's VNC console is bound is now
3122 selectable per-instance, rather than being cluster wide. Of course
3123 this only applies to instances controlled via VNC, so currently just
3124 applies to HVM clusters.
3130 *(Released Mon, 18 Feb 2008)*
3132 - more tweaks to the disk activation code (especially helpful for DRBD)
3133 - change the default ``gnt-instance list`` output format, now there is
3134 one combined status field (see the manpage for the exact values this
3136 - some more fixes for the mac export to hooks change
3137 - make Ganeti not break with DRBD 8.2.x (which changed the version
3138 format in ``/proc/drbd``) (issue 24)
3139 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3140 template, allowing migration from DRBD0.7+MD to DRBD8
3146 *(Released Wed, 30 Jan 2008)*
3148 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3150 - add command aliases infrastructure and a few aliases
3151 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3152 man pages and the ``--help`` option of ``gnt-node
3153 list``/``gnt-instance list``
3154 - fix ``gnt-backup list`` with down nodes (issue 21)
3155 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3156 - fix the dist archive and add a check for including svn/git files in
3158 - some developer-related changes: improve the burnin and the QA suite,
3159 add an upload script for testing during development
3165 *(Released Wed, 16 Jan 2008)*
3167 - experimental HVM support, read the install document, section
3168 "Initializing the cluster"
3169 - allow for the PVM hypervisor per-instance kernel and initrd paths
3170 - add a new command ``gnt-cluster verify-disks`` which uses a new
3171 algorithm to improve the reconnection of the DRBD pairs if the device
3172 on the secondary node has gone away
3173 - make logical volume code auto-activate LVs at disk activation time
3174 - slightly improve the speed of activating disks
3175 - allow specification of the MAC address at instance creation time, and
3176 changing it later via ``gnt-instance modify``
3177 - fix handling of external commands that generate lots of output on
3179 - update documentation with regard to minimum version of DRBD8 supported
3185 *(Released Tue, 4 Dec 2007)*
3187 - Log the ``xm create`` output to the node daemon log on failure (to
3188 help diagnosing the error)
3189 - In debug mode, log all external commands output if failed to the logs
3190 - Change parsing of lvm commands to ignore stderr
3196 *(Released Wed, 28 Nov 2007)*
3198 - Another round of updates to the DRBD 8 code to deal with more failures
3199 in the replace secondary node operation
3200 - Some more logging of failures in disk operations (lvm, drbd)
3201 - A few documentation updates
3208 *(Released Tue, 13 Nov 2007)*
3210 - Change configuration file format from Python's Pickle to JSON.
3211 Upgrading is possible using the cfgupgrade utility.
3212 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3213 faster replace disks and is more stable (DRBD 8 has many improvements
3214 compared to DRBD 0.7)
3215 - Added command line tags support (see man pages for ``gnt-instance``,
3216 ``gnt-node``, ``gnt-cluster``)
3217 - Added instance rename support
3218 - Added multi-instance startup/shutdown
3219 - Added cluster rename support
3220 - Added ``gnt-node evacuate`` to simplify some node operations
3221 - Added instance reboot operation that can speedup reboot as compared to
3223 - Soften the requirement that hostnames are in FQDN format
3224 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3226 - Removed dependency on debian's patched fping that uses the
3227 non-standard ``-S`` option
3228 - Now the OS definitions are searched for in multiple, configurable
3229 paths (easier for distros to package)
3230 - Some changes to the hooks infrastructure (especially the new
3231 post-configuration update hook)
3232 - Other small bugfixes
3234 .. vim: set textwidth=72 syntax=rst :