8 *(Released Thu, 23 Jan 2014)*
10 - Change the list separator for the usb_devices parameter from comma to space.
11 Commas could not work because they are already the hypervisor option
13 - Add support for blktap2 file-driver (Issue 638)
14 - Add network tag definitions to the haskell codebase (Issue 641)
15 - Fix RAPI network tag handling
16 - Add the network tags to the tags searched by gnt-cluster search-tags
17 - Fix caching bug preventing jobs from being cancelled
18 - Start-master/stop-master was always failing of ConfD was disabled. (Issue 685)
24 *(Released Thu, 12 Dec 2013)*
26 - Fixed Luxi daemon socket permissions after master-failover
27 - Improve IP version detection code directly checking for colons rather than
28 passing the family from the cluster object
29 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
30 opportunistically anymore (Issue 622)
31 - Allow link local IPv6 gateways (Issue 624)
32 - Fix error printing (Issue 616)
33 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
34 disk modifications, keep the old one. If name=none then set disk name to
36 - Update build_chroot script to work with the latest hackage packages
37 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
38 - Fix evacuation out of drained node (Issue 615)
39 - Add default file_driver if missing (Issue 571)
40 - Fix job error message after unclean master shutdown (Issue 618)
41 - Lock group(s) when creating instances (Issue 621)
42 - SetDiskID() before accepting an instance (Issue 633)
43 - Allow the ext template disks to receive arbitrary parameters, both at creation
44 time and while being modified
45 - Xen handle domain shutdown (future proofing cherry-pick)
46 - Refactor reading live data in htools (future proofing cherry-pick)
52 *(Released Thu, 07 Nov 2013)*
54 - DRBD: ensure peers are UpToDate for dual-primary
55 - Improve error message for replace-disks
56 - More dependency checks at configure time
57 - Placate warnings on ganeti.outils_unittest.py
63 *(Released Thu, 17 Oct 2013)*
65 - Correctly start/stop luxid during gnt-cluster master-failover
66 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
67 - Fix path for the job queue serial file
68 - Improved harep man page
69 - Minor documentation improvements
75 *(Released Mon, 30 Sep 2013)*
77 Incompatible/important changes
78 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
80 - Instance policy can contain multiple instance specs, as described in
81 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
82 <design-partitioned>`. As a consequence, it's not possible to partially change
83 or override instance specs. Bounding specs (min and max) can be specified as a
84 whole using the new option ``--ipolicy-bounds-specs``, while standard
85 specs use the new option ``--ipolicy-std-specs``.
86 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
87 gnt-instance is a valid YAML object.
88 - hail now honors network restrictions when allocating nodes. This led to an
89 update of the IAllocator protocol. See the IAllocator documentation for
91 - confd now only answers static configuration request over the network. luxid
92 was extracted, listens on the local LUXI socket and responds to live queries.
93 This allows finer grained permissions if using separate users.
98 - Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
99 so that disk and NIC modifications take effect without the need of actual
100 reboot. There are a couple of constrains currently for this feature:
102 - only KVM hypervisor (versions >= 1.0) supports it,
103 - one can not (yet) hotplug a disk using userspace access mode for RBD
104 - in case of a downgrade instances should suffer a reboot in order to
105 be migratable (due to core change of runtime files)
106 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
107 to always require authentication, ``--require-authentication``. It can
108 be specified in ``$sysconfdir/default/ganeti``.
109 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
110 be used to manage the disk templates to be used by instances in the cluster.
111 Initially, it will be set to a list that includes plain, drbd, if they were
112 enabled by specifying a volume group name, and file and sharedfile, if those
113 were enabled at configure time. Additionally, it will include all disk
114 templates that are currently used by instances. The order of disk templates
115 will be based on Ganeti's history of supporting them. In the future, the
116 first entry of the list will be used as a default disk template on instance
118 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
119 configuration back to the previous stable version.
120 - Disk templates in group ipolicy can be restored to the default value.
121 - Initial support for diskless instances and virtual clusters in QA.
122 - More QA and unit tests for instance policies.
123 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
124 describing why the opcode itself was executed.
125 - The monitoring daemon is now available. It allows users to query the cluster
126 for obtaining information about the status of the system. The daemon is only
127 responsible for providing the information over the network: the actual data
128 gathering is performed by data collectors (currently, only the DRBD status
129 collector is available).
130 - In order to help developers work on Ganeti, a new script
131 (``devel/build_chroot``) is provided, for building a chroot that contains all
132 the required development libraries and tools for compiling Ganeti on a Debian
134 - A new tool, ``harep``, for performing self-repair and recreation of instances
135 in Ganeti has been added.
136 - Split queries are enabled for tags, network, exports, cluster info, groups,
138 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
139 It prints the command line to set the current policies, to ease
141 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
142 devices for KVM virtio-net interfaces will get created with VNET_HDR
143 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
144 virtio-net interfaces, which prevents host kernel tainting and log
145 flooding, when dealing with broken or malicious virtio-net drivers.
146 It's set to true by default.
147 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
149 - Support 'viridian' parameter in Xen HVM
150 - Support DSA SSH keys in bootstrap
151 - To simplify the work of packaging frameworks that want to add the needed users
152 and groups in a split-user setup themselves, at build time three files in
153 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
154 the groups to be generated, the ``users`` file contains, one per line, the
155 users to be generated, optionally followed by their primary group, where
156 important. The ``groupmemberships`` file contains, one per line, additional
157 user-group membership relations that need to be established. The syntax of
158 these files will remain stable in all future versions.
163 The following new dependencies have been added:
166 - The ``curl`` library is not optional anymore for compiling the Haskell code.
167 - ``snap-server`` library (if monitoring is enabled).
170 - The minimum Python version needed to run Ganeti is now 2.6.
171 - ``yaml`` library (only for running the QA).
175 - Perform proper cleanup on termination of Haskell daemons
176 - Fix corner-case in handling of remaining retry time
182 *(Released Tue, 17 Sep 2013)*
184 - To simplify the work of packaging frameworks that want to add the needed users
185 and groups in a split-user setup themselves, at build time three files in
186 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
187 the groups to be generated, the ``users`` file contains, one per line, the
188 users to be generated, optionally followed by their primary group, where
189 important. The ``groupmemberships`` file contains, one per line, additional
190 user-group membership relations that need to be established. The syntax of
191 these files will remain stable in all future versions.
192 - Add a default to file-driver when unspecified over RAPI (Issue 571)
193 - Mark the DSA host pubkey as optional, and remove it during config downgrade
195 - Some documentation fixes
201 *(Released Tue, 27 Aug 2013)*
203 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
205 - Support 'viridian' parameter in Xen HVM (Issue 233)
206 - Include VCS version in ``gnt-cluster version``
207 - Support DSA SSH keys in bootstrap (Issue 338)
208 - Fix batch creation of instances
209 - Use FQDN to check master node status (Issue 551)
210 - Make the DRBD collector more failure-resilient
216 *(Released Fri, 2 Aug 2013)*
218 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
220 - Fix upgrading/downgrading from 2.7
221 - Increase maximum RAPI message size
222 - Documentation updates
223 - Split ``confd`` between ``luxid`` and ``confd``
224 - Merge 2.7 series up to the 2.7.1 release
225 - Allow the ``modify_etc_hosts`` option to be changed
226 - Add better debugging for ``luxid`` queries
227 - Expose bulk parameter for GetJobs in RAPI client
228 - Expose missing ``network`` fields in RAPI
229 - Add some ``cluster verify`` tests
230 - Some unittest fixes
231 - Fix a malfunction in ``hspace``'s tiered allocation
232 - Fix query compatibility between haskell and python implementations
233 - Add the ``vnet_hdr`` HV parameter for KVM
234 - Add ``--cleanup`` to instance failover
235 - Change the connected groups format in ``gnt-network info`` output; it
236 was previously displayed as a raw list by mistake. (Merged from 2.7)
242 *(Released Mon, 24 Jun 2013)*
244 This was the first beta release of the 2.8 series. All important changes
245 are listed in the latest 2.8 entry.
251 *(Released Thu, 26 Sep 2013)*
253 - Change the connected groups format in ``gnt-network info`` output; it
254 was previously displayed as a raw list by mistake
255 - Check disk template in right dict when copying
256 - Support multi-instance allocs without iallocator
257 - Fix some errors in the documentation
258 - Fix formatting of tuple in an error message
264 *(Released Thu, 25 Jul 2013)*
266 - Add logrotate functionality in daemon-util
267 - Add logrotate example file
268 - Add missing fields to network queries over rapi
269 - Fix network object timestamps
270 - Add support for querying network timestamps
271 - Fix a typo in the example crontab
272 - Fix a documentation typo
278 *(Released Thu, 04 Jul 2013)*
280 Incompatible/important changes
281 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
283 - Instance policies for disk size were documented to be on a per-disk
284 basis, but hail applied them to the sum of all disks. This has been
286 - ``hbal`` will now exit with status 0 if, during job execution over
287 LUXI, early exit has been requested and all jobs are successful;
288 before, exit status 1 was used, which cannot be differentiated from
290 - Compatibility with newer versions of rbd has been fixed
291 - ``gnt-instance batch-create`` has been changed to use the bulk create
292 opcode from Ganeti. This lead to incompatible changes in the format of
293 the JSON file. It's now not a custom dict anymore but a dict
294 compatible with the ``OpInstanceCreate`` opcode.
295 - Parent directories for file storage need to be listed in
296 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
297 write the file automatically based on old configuration values, but it
298 can not distribute it across all nodes and the file contents should be
299 verified. Use ``gnt-cluster copyfile
300 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
301 upgraded. The reason for requiring this list of paths now is that
302 before it would have been possible to inject new paths via RPC,
303 allowing files to be created in arbitrary locations. The RPC protocol
304 is protected using SSL/X.509 certificates, but as a design principle
305 Ganeti does not permit arbitrary paths to be passed.
306 - The parsing of the variants file for OSes (see
307 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
308 lines and comment lines (starting with ``#``) are ignored for better
310 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
311 longer available. ``gnt-node add`` now invokes a new tool on the
312 destination node, named ``prepare-node-join``, to configure the SSH
313 daemon. Paramiko is no longer necessary to configure nodes' SSH
314 daemons via ``gnt-node add``.
315 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
316 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
317 the flag is not reset after a master failover.
318 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
319 is recommended. 2.6 will be mandatory from the 2.8 series.
325 - New network management functionality to support automatic allocation
326 of IP addresses and managing of network parameters. See
327 :manpage:`gnt-network(8)` for more details.
328 - New external storage backend, to allow managing arbitrary storage
329 systems external to the cluster. See
330 :manpage:`ganeti-extstorage-interface(7)`.
331 - New ``exclusive-storage`` node parameter added, restricted to
332 nodegroup level. When it's set to true, physical disks are assigned in
333 an exclusive fashion to instances, as documented in :doc:`Partitioned
334 Ganeti <design-partitioned>`. Currently, only instances using the
335 ``plain`` disk template are supported.
336 - The KVM hypervisor has been updated with many new hypervisor
337 parameters, including a generic one for passing arbitrary command line
338 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
339 compatible up to qemu 1.4.
340 - A new tool, called ``mon-collector``, is the stand-alone executor of
341 the data collectors for a monitoring system. As of this version, it
342 just includes the DRBD data collector, that can be executed by calling
343 ``mon-collector`` using the ``drbd`` parameter. See
344 :manpage:`mon-collector(7)`.
345 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
346 for RAPI users. It allows granting permissions to query for
347 information to a specific user without giving
348 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
349 - A new tool named ``node-cleanup`` has been added. It cleans remains of
350 a cluster from a machine by stopping all daemons, removing
351 certificates and ssconf files. Unless the ``--no-backup`` option is
352 given, copies of the certificates are made.
353 - Instance creations now support the use of opportunistic locking,
354 potentially speeding up the (parallel) creation of multiple instances.
355 This feature is currently only available via the :doc:`RAPI
356 <rapi>` interface and when an instance allocator is used. If the
357 ``opportunistic_locking`` parameter is set the opcode will try to
358 acquire as many locks as possible, but will not wait for any locks
359 held by other opcodes. If not enough resources can be found to
360 allocate the instance, the temporary error code
361 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
362 retried thereafter, with or without opportunistic locking.
363 - New experimental linux-ha resource scripts.
364 - Restricted-commands support: ganeti can now be asked (via command line
365 or rapi) to perform commands on a node. These are passed via ganeti
366 RPC rather than ssh. This functionality is restricted to commands
367 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
368 reasons. The file is not copied automatically.
374 - Diskless instances are now externally mirrored (Issue 237). This for
375 now has only been tested in conjunction with explicit target nodes for
377 - Queries not needing locks or RPC access to the node can now be
378 performed by the confd daemon, making them independent from jobs, and
379 thus faster to execute. This is selectable at configure time.
380 - The functionality for allocating multiple instances at once has been
381 overhauled and is now also available through :doc:`RAPI <rapi>`.
383 There are no significant changes from version 2.7.0~rc3.
389 *(Released Tue, 25 Jun 2013)*
391 - Fix permissions on the confd query socket (Issue 477)
392 - Fix permissions on the job archive dir (Issue 498)
393 - Fix handling of an internal exception in replace-disks (Issue 472)
394 - Fix gnt-node info handling of shortened names (Issue 497)
395 - Fix gnt-instance grow-disk when wiping is enabled
396 - Documentation improvements, and support for newer pandoc
397 - Fix hspace honoring ipolicy for disks (Issue 484)
398 - Improve handling of the ``kvm_extra`` HV parameter
404 *(Released Fri, 24 May 2013)*
406 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
408 - Disks added through ``gnt-instance modify`` or created through
409 ``gnt-instance recreate-disks`` are wiped, if the
410 ``prealloc_wipe_disks`` flag is set.
411 - If wiping newly created disks fails, the disks are removed. Also,
412 partial failures in creating disks through ``gnt-instance modify``
413 triggers a cleanup of the partially-created disks.
414 - Removing the master IP address doesn't fail if the address has been
416 - Fix ownership of the OS log dir
417 - Workaround missing SO_PEERCRED constant (Issue 191)
423 *(Released Fri, 3 May 2013)*
425 This was the first release candidate of the 2.7 series. Since beta3:
427 - Fix kvm compatibility with qemu 1.4 (Issue 389)
428 - Documentation updates (admin guide, upgrade notes, install
429 instructions) (Issue 372)
430 - Fix gnt-group list nodes and instances count (Issue 436)
431 - Fix compilation without non-mandatory libraries (Issue 441)
432 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
433 - Make confd logging more verbose at INFO level (Issue 435)
434 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
435 - Fix failure path for instance storage type conversion (Issue 229)
436 - Update htools text backend documentation
437 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
438 - Disable inter-cluster instance move for file-based instances, because
439 it is dependant on instance export, which is not supported for
440 file-based instances. (Issue 414)
441 - Fix gnt-job crashes on non-ascii characters (Issue 427)
442 - Fix volume group checks on non-vm-capable nodes (Issue 432)
448 *(Released Mon, 22 Apr 2013)*
450 This was the third beta release of the 2.7 series. Since beta2:
452 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
453 - Fix data loss on wrong usage of ``gnt-instance move``
454 - Properly export errors in confd-based job queries
455 - Add ``users-setup`` tool
456 - Fix iallocator protocol to report 0 as a disk size for diskless
457 instances. This avoids hail breaking when a diskless instance is
459 - Fix job queue directory permission problem that made confd job queries
460 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
461 for access to archived jobs (Issue 406).
462 - Limit the sizes of networks supported by ``gnt-network`` to something
463 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
464 - Fix bugs in instance disk template conversion
465 - Fix GHC 7 compatibility
466 - Fix ``burnin`` install path (Issue 426).
467 - Allow very small disk grows (Issue 347).
468 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
469 that noded doesn't import masterd code (Issue 419).
470 - Make sure the default metavg at cluster init is the same as the vg, if
471 unspecified (Issue 358).
472 - Fix cleanup of partially created disks (part of Issue 416)
478 *(Released Tue, 2 Apr 2013)*
480 This was the second beta release of the 2.7 series. Since beta1:
482 - Networks no longer have a "type" slot, since this information was
483 unused in Ganeti: instead of it tags should be used.
484 - The rapi client now has a ``target_node`` option to MigrateInstance.
485 - Fix early exit return code for hbal (Issue 386).
486 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
487 - Fix ``rbd showmapped`` output parsing (Issue 312).
488 - Networks are now referenced indexed by UUID, rather than name. This
489 will require running cfgupgrade, from 2.7.0beta1, if networks are in
491 - The OS environment now includes network information.
492 - Deleting of a network is now disallowed if any instance nic is using
493 it, to prevent dangling references.
494 - External storage is now documented in man pages.
495 - The exclusive_storage flag can now only be set at nodegroup level.
496 - Hbal can now submit an explicit priority with its jobs.
497 - Many network related locking fixes.
498 - Bump up the required pylint version to 0.25.1.
499 - Fix the ``no_remember`` option in RAPI client.
500 - Many ipolicy related tests, qa, and fixes.
501 - Many documentation improvements and fixes.
502 - Fix building with ``--disable-file-storage``.
503 - Fix ``-q`` option in htools, which was broken if passed more than
505 - Some haskell/python interaction improvements and fixes.
506 - Fix iallocator in case of missing LVM storage.
507 - Fix confd config load in case of ``--no-lvm-storage``.
508 - The confd/query functionality is now mentioned in the security
515 *(Released Wed, 6 Feb 2013)*
517 This was the first beta release of the 2.7 series. All important changes
518 are listed in the latest 2.7 entry.
524 *(Released Fri, 21 Dec 2012)*
526 Important behaviour change: hbal won't rebalance anymore instances which
527 have the ``auto_balance`` attribute set to false. This was the intention
528 all along, but until now it only skipped those from the N+1 memory
529 reservation (DRBD-specific).
531 A significant number of bug fixes in this release:
533 - Fixed disk adoption interaction with ipolicy checks.
534 - Fixed networking issues when instances are started, stopped or
535 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
536 - Fixed the warning in cluster verify for shared storage instances not
538 - Fixed removal of storage directory on shared file storage (issue 262).
539 - Fixed validation of LVM volume group name in OpClusterSetParams
540 (``gnt-cluster modify``) (issue 285).
541 - Fixed runtime memory increases (``gnt-instance modify -m``).
542 - Fixed live migration under Xen's ``xl`` mode.
543 - Fixed ``gnt-instance console`` with ``xl``.
544 - Fixed building with newer Haskell compiler/libraries.
545 - Fixed PID file writing in Haskell daemons (confd); this prevents
546 restart issues if confd was launched manually (outside of
547 ``daemon-util``) while another copy of it was running
548 - Fixed a type error when doing live migrations with KVM (issue 297) and
549 the error messages for failing migrations have been improved.
550 - Fixed opcode validation for the out-of-band commands (``gnt-node
552 - Fixed a type error when unsetting OS hypervisor parameters (issue
553 311); now it's possible to unset all OS-specific hypervisor
555 - Fixed the ``dry-run`` mode for many operations: verification of
556 results was over-zealous but didn't take into account the ``dry-run``
557 operation, resulting in "wrong" failures.
558 - Fixed bash completion in ``gnt-job list`` when the job queue has
559 hundreds of entries; especially with older ``bash`` versions, this
560 results in significant CPU usage.
562 And lastly, a few other improvements have been made:
564 - Added option to force master-failover without voting (issue 282).
565 - Clarified error message on lock conflict (issue 287).
566 - Logging of newly submitted jobs has been improved (issue 290).
567 - Hostname checks have been made uniform between instance rename and
569 - The ``--submit`` option is now supported by ``gnt-debug delay``.
570 - Shutting down the master daemon by sending SIGTERM now stops it from
571 processing jobs waiting for locks; instead, those jobs will be started
572 once again after the master daemon is started the next time (issue
574 - Support for Xen's ``xl`` program has been improved (besides the fixes
576 - Reduced logging noise in the Haskell confd daemon (only show one log
577 entry for each config reload, instead of two).
578 - Several man page updates and typo fixes.
584 *(Released Fri, 12 Oct 2012)*
586 A small bugfix release. Among the bugs fixed:
588 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
589 made the command unusable.
590 - Commands that issue many jobs don't fail anymore just because some jobs
591 take so long that other jobs are archived.
592 - Failures during ``gnt-instance reinstall`` are reflected by the exit
594 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
596 - When ``always_failover`` is set, ``--allow-failover`` is not required
597 in migrate commands anymore.
598 - ``bash_completion`` works even if extglob is disabled.
599 - Fixed bug with locks that made failover for RDB-based instances fail.
600 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
601 a random node instead of one based on the allocator metric.
602 - Support for newer versions of pylint and pep8.
603 - Hail doesn't fail anymore when trying to add an instance of type
604 ``file``, ``sharedfile`` or ``rbd``.
605 - Added new Makefile target to rebuild the whole distribution, so that
606 all files are included.
612 *(Released Fri, 27 Jul 2012)*
615 .. attention:: The ``LUXI`` protocol has been made more consistent
616 regarding its handling of command arguments. This, however, leads to
617 incompatibility issues with previous versions. Please ensure that you
618 restart Ganeti daemons soon after the upgrade, otherwise most
619 ``LUXI`` calls (job submission, setting/resetting the drain flag,
620 pausing/resuming the watcher, cancelling and archiving jobs, querying
621 the cluster configuration) will fail.
630 The current ``admin_up`` field, which used to denote whether an instance
631 should be running or not, has been removed. Instead, ``admin_state`` is
632 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
634 The rational behind this is that an instance being “down” can have
637 - it could be down during a reboot
638 - it could be temporarily be down for a reinstall
639 - or it could be down because it is deprecated and kept just for its
642 The previous Boolean state was making it difficult to do capacity
643 calculations: should Ganeti reserve memory for a down instance? Now, the
644 tri-state field makes it clear:
646 - in ``up`` and ``down`` state, all resources are reserved for the
647 instance, and it can be at any time brought up if it is down
648 - in ``offline`` state, only disk space is reserved for it, but not
651 The field can have an extra use: since the transition between ``up`` and
652 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
653 transition between ``offline`` and ``down`` is done via ``gnt-instance
654 modify``, it is possible to given different rights to users. For
655 example, owners of an instance could be allowed to start/stop it, but
656 not transition it out of the offline state.
658 Instance policies and specs
659 +++++++++++++++++++++++++++
661 In previous Ganeti versions, an instance creation request was not
662 limited on the minimum size and on the maximum size just by the cluster
663 resources. As such, any policy could be implemented only in third-party
664 clients (RAPI clients, or shell wrappers over ``gnt-*``
665 tools). Furthermore, calculating cluster capacity via ``hspace`` again
666 required external input with regards to instance sizes.
668 In order to improve these workflows and to allow for example better
669 per-node group differentiation, we introduced instance specs, which
672 - minimum instance disk size, disk count, memory size, cpu count
673 - maximum values for the above metrics
674 - and “standard” values (used in ``hspace`` to calculate the standard
677 The minimum/maximum values can be also customised at node-group level,
678 for example allowing more powerful hardware to support bigger instance
681 Beside the instance specs, there are a few other settings belonging to
682 the instance policy framework. It is possible now to customise, per
683 cluster and node-group:
685 - the list of allowed disk templates
686 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
687 - the maximum ratio of instance to spindles (see below for more
688 information) for local storage
690 All these together should allow all tools that talk to Ganeti to know
691 what are the ranges of allowed values for instances and the
692 over-subscription that is allowed.
694 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
695 instance configuration, and the physical CPU configuration from the
696 node. For the spindle ratios however, we didn't track before these
697 values, so new parameters have been added:
699 - a new node parameter ``spindle_count``, defaults to 1, customisable at
700 node group or node level
701 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
703 Note that spindles in this context doesn't need to mean actual
704 mechanical hard-drives; it's just a relative number for both the node
705 I/O capacity and instance I/O consumption.
707 Instance migration behaviour
708 ++++++++++++++++++++++++++++
710 While live-migration is in general desirable over failover, it is
711 possible that for some workloads it is actually worse, due to the
712 variable time of the “suspend” phase during live migration.
714 To allow the tools to work consistently over such instances (without
715 having to hard-code instance names), a new backend parameter
716 ``always_failover`` has been added to control the migration/failover
717 behaviour. When set to True, all migration requests for an instance will
718 instead fall-back to failover.
720 Instance memory ballooning
721 ++++++++++++++++++++++++++
723 Initial support for memory ballooning has been added. The memory for an
724 instance is no longer fixed (backend parameter ``memory``), but instead
725 can vary between minimum and maximum values (backend parameters
726 ``minmem`` and ``maxmem``). Currently we only change an instance's
729 - live migrating or failing over and instance and the target node
730 doesn't have enough memory
731 - user requests changing the memory via ``gnt-instance modify
737 In order to control the use of specific CPUs by instance, support for
738 controlling CPU pinning has been added for the Xen, HVM and LXC
739 hypervisors. This is controlled by a new hypervisor parameter
740 ``cpu_mask``; details about possible values for this are in the
741 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
742 VCPU-to-CPU mapping) form will work well only when all nodes in your
743 cluster have the same amount of CPUs.
748 Another area in which Ganeti was not customisable were the parameters
749 used for storage configuration, e.g. how many stripes to use for LVM,
750 DRBD resync configuration, etc.
752 To improve this area, we've added disks parameters, which are
753 customisable at cluster and node group level, and which allow to
754 specify various parameters for disks (DRBD has the most parameters
755 currently), for example:
757 - DRBD resync algorithm and parameters (e.g. speed)
758 - the default VG for meta-data volumes for DRBD
759 - number of stripes for LVM (plain disk template)
762 These parameters can be modified via ``gnt-cluster modify -D …`` and
763 ``gnt-group modify -D …``, and are used at either instance creation (in
764 case of LVM stripes, for example) or at disk “activation” time
767 Rados block device support
768 ++++++++++++++++++++++++++
770 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
771 denoted by the ``rbd`` disk template type. This is considered
772 experimental, feedback is welcome. For details on configuring it, see
773 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
778 The existing master IP functionality works well only in simple setups (a
779 single network shared by all nodes); however, if nodes belong to
780 different networks, then the ``/32`` setup and lack of routing
781 information is not enough.
783 To allow the master IP to function well in more complex cases, the
784 system was reworked as follows:
786 - a master IP netmask setting has been added
787 - the master IP activation/turn-down code was moved from the node daemon
789 - whether to run the Ganeti-supplied master IP script or a user-supplied
790 on is a ``gnt-cluster init`` setting
792 Details about the location of the standard and custom setup scripts are
793 in the man page :manpage:`gnt-cluster(8)`; for information about the
794 setup script protocol, look at the Ganeti-supplied script.
799 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
802 It is now possible to use TLS-protected connections, and when renewing
803 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
804 it is now possible to specify spice or spice CA certificates. Also, it
805 is possible to configure a password for SPICE sessions via the
806 hypervisor parameter ``spice_password_file``.
808 There are also new parameters to control the compression and streaming
809 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
810 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
811 for the spice parameters.
813 Lastly, it is now possible to see the SPICE connection information via
814 ``gnt-instance console``.
819 A new tool (``tools/ovfconverter``) has been added that supports
820 conversion between Ganeti and the `Open Virtualization Format
821 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
824 This relies on the ``qemu-img`` tool to convert the disk formats, so the
825 actual compatibility with other virtualization solutions depends on it.
830 The configuration query daemon (``ganeti-confd``) is now optional, and
831 has been rewritten in Haskell; whether to use the daemon at all, use the
832 Python (default) or the Haskell version is selectable at configure time
833 via the ``--enable-confd`` parameter, which can take one of the
834 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
835 daemon will result in a smaller footprint; for larger systems, we
836 welcome feedback on the Haskell version which might become the default
839 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
840 daemon running. The Python version doesn't implement the new call.
843 User interface changes
844 ~~~~~~~~~~~~~~~~~~~~~~
846 We have replaced the ``--disks`` option of ``gnt-instance
847 replace-disks`` with a more flexible ``--disk`` option, which allows
848 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
849 disk size and mode can be changed upon recreation (via ``gnt-instance
850 recreate-disks``, which accepts the same ``--disk`` option).
852 As many people are used to a ``show`` command, we have added that as an
853 alias to ``info`` on all ``gnt-*`` commands.
855 The ``gnt-instance grow-disk`` command has a new mode in which it can
856 accept the target size of the disk, instead of the delta; this can be
857 more safe since two runs in absolute mode will be idempotent, and
858 sometimes it's also easier to specify the desired size directly.
860 Also the handling of instances with regard to offline secondaries has
861 been improved. Instance operations should not fail because one of it's
862 secondary nodes is offline, even though it's safe to proceed.
864 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
865 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
866 minors to instance name.
871 RAPI coverage has improved, with (for example) new resources for
872 recreate-disks, node power-cycle, etc.
877 There is partial support for ``xl`` in the Xen hypervisor; feedback is
880 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
881 whether to still support Python 2.4 or move to Python 2.6 as minimum
884 Support for Fedora has been slightly improved; the provided example
885 init.d script should work better on it and the INSTALL file should
886 document the needed dependencies.
891 The deprecated ``QueryLocks`` LUXI request has been removed. Use
892 ``Query(what=QR_LOCK, ...)`` instead.
894 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
895 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
896 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
897 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
898 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
900 RAPI client: ``CertificateError`` now derives from
901 ``GanetiApiError``. This should make it more easy to handle Ganeti
904 Deprecation warnings due to PyCrypto/paramiko import in
905 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
906 make sure to run an up-to-date paramiko version, if you use this tool.
908 The QA scripts now depend on Python 2.5 or above (the main code base
909 still works with Python 2.4).
911 The configuration file (``config.data``) is now written without
912 indentation for performance reasons; if you want to edit it, it can be
913 re-formatted via ``tools/fmtjson``.
915 A number of bugs has been fixed in the cluster merge tool.
917 ``x509`` certification verification (used in import-export) has been
918 changed to allow the same clock skew as permitted by the cluster
919 verification. This will remove some rare but hard to diagnose errors in
926 *(Released Thu, 19 Jul 2012)*
928 Very few changes from rc4 to the final release, only bugfixes:
930 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
931 instance, fix CDROM booting for KVM instances)
932 - fixed node group modification of node parameters
933 - fixed issue in LUClusterVerifyGroup with multi-group clusters
934 - fixed generation of bash completion to ensure a stable ordering
941 *(Released Fri, 13 Jul 2012)*
943 Third release candidate for 2.6. The following changes were done from
946 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
947 - Fixed an inconsistency in the LUXI protocol with the provided
948 arguments (NOT backwards compatible)
949 - Fixed a bug with node groups ipolicy where ``min`` was greater than
950 the cluster ``std`` value
951 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
952 easier instance debugging on nodes (requires ``hconfd`` to work)
958 *(Released Tue, 03 Jul 2012)*
960 Second release candidate for 2.6. The following changes were done from
963 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
965 - Fixed a RAPI regression on missing beparams/memory
966 - Fixed redistribution of files on offline nodes
967 - Added possibility to run activate-disks even though secondaries are
968 offline. With this change it relaxes also the strictness on some other
969 commands which use activate disks internally:
970 * ``gnt-instance start|reboot|rename|backup|export``
971 - Made it possible to remove safely an instance if its secondaries are
973 - Made it possible to reinstall even though secondaries are offline
979 *(Released Mon, 25 Jun 2012)*
981 First release candidate for 2.6. The following changes were done from
984 - Fixed bugs with disk parameters and ``rbd`` templates as well as
986 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
987 behaviour. It supports now the modify operation
988 - ``hcheck`` implemented to analyze cluster health and possibility of
989 improving health by rebalance
990 - ``hbal`` has been improved in dealing with split instances
996 *(Released Mon, 11 Jun 2012)*
998 Second beta release of 2.6. The following changes were done from beta2
1001 - Fixed ``daemon-util`` with non-root user models
1002 - Fixed creation of plain instances with ``--no-wait-for-sync``
1003 - Fix wrong iv_names when running ``cfgupgrade``
1004 - Export more information in RAPI group queries
1005 - Fixed bug when changing instance network interfaces
1006 - Extended burnin to do NIC changes
1007 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1008 - Changed default for DRBD barriers
1009 - Fixed DRBD error reporting for syncer rate
1010 - Verify the options on disk parameters
1012 And of course various fixes to documentation and improved unittests and
1019 *(Released Wed, 23 May 2012)*
1021 First beta release of 2.6. The following changes were done from beta1 to
1024 - integrated patch for distributions without ``start-stop-daemon``
1025 - adapted example init.d script to work on Fedora
1026 - fixed log handling in Haskell daemons
1027 - adapted checks in the watcher for pycurl linked against libnss
1028 - add partial support for ``xl`` instead of ``xm`` for Xen
1029 - fixed a type issue in cluster verification
1030 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1033 Plus integrated fixes from the 2.5 branch:
1035 - fixed ``kvm-ifup`` to use ``/bin/bash``
1036 - fixed parallel build failures
1037 - KVM live migration when using a custom keymap
1043 *(Released Tue, 24 Jul 2012)*
1045 A small bugfix release, with no new features:
1047 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1048 different default shell (e.g. Debian, Ubuntu)
1049 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1050 243 and Debian bug #650664)
1051 - fixed compatibility with KVM versions that don't support multiple boot
1052 devices (fixes Issue 230 and Debian bug #624256)
1054 Additionally, a few fixes were done to the build system (fixed parallel
1055 build failures) and to the unittests (fixed race condition in test for
1056 FileID functions, and the default enable/disable mode for QA test is now
1063 *(Released Fri, 11 May 2012)*
1065 A small bugfix release.
1067 The main issues solved are on the topic of compatibility with newer LVM
1070 - fixed parsing of ``lv_attr`` field
1071 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1072 the ``--force`` flag is needed
1074 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1075 to accept kernel 3.x too (was hardcoded to 2.6.*).
1077 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1078 of hook results and that also made display of other errors suboptimal
1079 was fixed; the code behaves now like 2.4 and earlier.
1081 Another change in 2.5, the cleanup of the OS scripts environment, is too
1082 aggressive: it removed even the ``PATH`` variable, which requires the OS
1083 scripts to *always* need to export it. Since this is a bit too strict,
1084 we now export a minimal PATH, the same that we export for hooks.
1086 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1087 integrated into this release.
1089 Finally, a few other miscellaneous changes were done (no new features,
1090 just small improvements):
1092 - Fix ``gnt-group --help`` display
1093 - Fix hardcoded Xen kernel path
1094 - Fix grow-disk handling of invalid units
1095 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1096 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1102 *(Released Thu, 12 Apr 2012)*
1104 Incompatible/important changes and bugfixes
1105 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1107 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1108 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1109 to match the underlying LUXI interface.
1110 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1111 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1112 not maintain backwards-compatibility as the underlying operation
1113 changed in an incompatible way. The RAPI client can talk to old
1114 servers, but it needs to be told so as the return value changed.
1115 - When creating file-based instances via RAPI, the ``file_driver``
1116 parameter no longer defaults to ``loop`` and must be specified.
1117 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1119 - Support for the undocumented and deprecated RAPI instance creation
1120 request format version 0 has been dropped. Use version 1, supported
1121 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1122 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1123 documentation <install>`.
1124 - The "cluster-verify" hooks are now executed per group by the
1125 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1126 if you just run ``gnt-cluster verify``, which generates one opcode per
1128 - The environment as passed to the OS scripts is cleared, and thus no
1129 environment variables defined in the node daemon's environment will be
1130 inherited by the scripts.
1131 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1133 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1134 support operations involving multiple node groups.
1135 - Offline nodes are ignored when failing over an instance.
1136 - Support for KVM version 1.0, which changed the version reporting format
1138 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1140 - ``Makefile`` is now compatible with Automake 1.11.2
1141 - Includes all bugfixes made in the 2.4 series
1146 - The ganeti-htools project has been merged into the ganeti-core source
1147 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1148 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1149 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1150 GPT-style partition tables (requires `parted
1151 <http://www.gnu.org/s/parted/>`_).
1152 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1153 - Allowed adding tags on instance creation.
1154 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1155 - Allow instances to be started in a paused state, enabling the user to
1156 see the complete console output on boot using the console.
1157 - Added new hypervisor flag to control default reboot behaviour
1158 (``reboot_behavior``).
1159 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1160 - Improved out-of-band management support:
1162 - Added ``gnt-node health`` command reporting the health status of
1164 - Added ``gnt-node power`` command to manage power status of nodes.
1165 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1167 - Instance migration can fall back to failover if instance is not
1169 - Filters can be used when listing nodes, instances, groups and locks;
1170 see :manpage:`ganeti(7)` manpage.
1171 - Added post-execution status as variables to :doc:`hooks <hooks>`
1173 - Instance tags are exported/imported together with the instance.
1174 - When given an explicit job ID, ``gnt-job info`` will work for archived
1176 - Jobs can define dependencies on other jobs (not yet supported via
1177 RAPI or command line, but used by internal commands and usable via
1180 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1183 - Instance failover is now available as a RAPI resource
1184 (``/2/instances/[instance_name]/failover``).
1185 - ``gnt-instance info`` defaults to static information if primary node
1187 - Opcodes have a new ``comment`` attribute.
1188 - Added basic SPICE support to KVM hypervisor.
1189 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1191 Node group improvements
1192 ~~~~~~~~~~~~~~~~~~~~~~~
1194 - ``gnt-cluster verify`` has been modified to check groups separately,
1195 thereby improving performance.
1196 - Node group support has been added to ``gnt-cluster verify-disks``,
1197 which now operates per node group.
1198 - Watcher has been changed to work better with node groups.
1200 - One process and state file per node group.
1201 - Slow watcher in one group doesn't block other group's watcher.
1203 - Added new command, ``gnt-group evacuate``, to move all instances in a
1204 node group to other groups.
1205 - Added ``gnt-instance change-group`` to move an instance to another
1207 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1208 per-group operations.
1209 - Node groups can be tagged.
1210 - Some operations switch from an exclusive to a shared lock as soon as
1212 - Instance's primary and secondary nodes' groups are now available as
1213 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1214 and ``snodes.group.uuid``).
1219 - Numerous updates to documentation and manpages.
1221 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1223 - Some opcode/job results are now also documented, see :doc:`RAPI
1226 - A lockset's internal lock is now also visible in lock monitor.
1227 - Log messages from job queue workers now contain information about the
1228 opcode they're processing.
1229 - ``gnt-instance console`` no longer requires the instance lock.
1230 - A short delay when waiting for job changes reduces the number of LUXI
1231 requests significantly.
1232 - DRBD metadata volumes are overwritten with zeros during disk creation.
1233 - Out-of-band commands no longer acquire the cluster lock in exclusive
1235 - ``devel/upload`` now uses correct permissions for directories.
1241 *(Released Fri, 23 Mar 2012)*
1243 This was the sixth release candidate of the 2.5 series.
1249 *(Released Mon, 9 Jan 2012)*
1251 This was the fifth release candidate of the 2.5 series.
1257 *(Released Thu, 27 Oct 2011)*
1259 This was the fourth release candidate of the 2.5 series.
1265 *(Released Wed, 26 Oct 2011)*
1267 This was the third release candidate of the 2.5 series.
1273 *(Released Tue, 18 Oct 2011)*
1275 This was the second release candidate of the 2.5 series.
1281 *(Released Tue, 4 Oct 2011)*
1283 This was the first release candidate of the 2.5 series.
1289 *(Released Wed, 31 Aug 2011)*
1291 This was the third beta release of the 2.5 series.
1297 *(Released Mon, 22 Aug 2011)*
1299 This was the second beta release of the 2.5 series.
1305 *(Released Fri, 12 Aug 2011)*
1307 This was the first beta release of the 2.5 series.
1313 *(Released Thu, 27 Oct 2011)*
1315 - Fixed bug when parsing command line parameter values ending in
1317 - Fixed assertion error after unclean master shutdown
1318 - Disable HTTP client pool for RPC, significantly reducing memory usage
1320 - Fixed queue archive creation with wrong permissions
1326 *(Released Tue, 23 Aug 2011)*
1330 - Fixed documentation for importing with ``--src-dir`` option
1331 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1332 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1338 *(Released Fri, 5 Aug 2011)*
1340 Many bug-fixes and a few small features:
1342 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1343 issues when you tried to add an instance with two MAC addresses in one
1345 - KVM: fixed per-instance stored UID value
1346 - KVM: configure bridged NICs at migration start
1347 - KVM: Fix a bug where instance will not start with never KVM versions
1349 - Added OS search path to ``gnt-cluster info``
1350 - Fixed an issue with ``file_storage_dir`` where you were forced to
1351 provide an absolute path, but the documentation states it is a
1352 relative path, the documentation was right
1353 - Added a new parameter to instance stop/start called ``--no-remember``
1354 that will make the state change to not be remembered
1355 - Implemented ``no_remember`` at RAPI level
1356 - Improved the documentation
1357 - Node evacuation: don't call IAllocator if node is already empty
1358 - Fixed bug in DRBD8 replace disks on current nodes
1359 - Fixed bug in recreate-disks for DRBD instances
1360 - Moved assertion checking locks in ``gnt-instance replace-disks``
1361 causing it to abort with not owning the right locks for some situation
1362 - Job queue: Fixed potential race condition when cancelling queued jobs
1363 - Fixed off-by-one bug in job serial generation
1364 - ``gnt-node volumes``: Fix instance names
1365 - Fixed aliases in bash completion
1366 - Fixed a bug in reopening log files after being sent a SIGHUP
1367 - Added a flag to burnin to allow specifying VCPU count
1368 - Bugfixes to non-root Ganeti configuration
1374 *(Released Thu, 12 May 2011)*
1376 Many bug-fixes and a few new small features:
1378 - Fixed a bug related to log opening failures
1379 - Fixed a bug in instance listing with orphan instances
1380 - Fixed a bug which prevented resetting the cluster-level node parameter
1381 ``oob_program`` to the default
1382 - Many fixes related to the ``cluster-merge`` tool
1383 - Fixed a race condition in the lock monitor, which caused failures
1384 during (at least) creation of many instances in parallel
1385 - Improved output for gnt-job info
1386 - Removed the quiet flag on some ssh calls which prevented debugging
1388 - Improved the N+1 failure messages in cluster verify by actually
1389 showing the memory values (needed and available)
1390 - Increased lock attempt timeouts so that when executing long operations
1391 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1392 too early and thus prevent the use of the 'fair' mechanism
1393 - Changed instance query data (``gnt-instance info``) to not acquire
1394 locks unless needed, thus allowing its use on locked instance if only
1395 static information is asked for
1396 - Improved behaviour with filesystems that do not support rename on an
1398 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1399 kept locks on all nodes during the wipe, which is unneeded
1400 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1401 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1402 10GiB) which caused the wipe to fail right at the end in some cases
1403 - Fixed master IP activation when doing master failover with no-voting
1404 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1405 the master node itself
1406 - Fixed potential data-loss in under disk full conditions, where Ganeti
1407 wouldn't check correctly the return code and would consider
1408 partially-written files 'correct'
1409 - Fixed bug related to multiple VGs and DRBD disk replacing
1410 - Added new disk parameter ``metavg`` that allows placement of the meta
1411 device for DRBD in a different volume group
1412 - Fixed error handling in the node daemon when the system libc doesn't
1413 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1414 - Fixed lock release during replace-disks, which kept cluster-wide locks
1415 when doing disk replaces with an iallocator script
1416 - Added check for missing bridges in cluster verify
1417 - Handle EPIPE errors while writing to the terminal better, so that
1418 piping the output to e.g. ``less`` doesn't cause a backtrace
1419 - Fixed rare case where a ^C during Luxi calls could have been
1420 interpreted as server errors, instead of simply terminating
1421 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1423 - Added a few more parameters to the KVM hypervisor, allowing a second
1424 CDROM, custom disk type for CDROMs and a floppy image
1425 - Removed redundant message in instance rename when the name is given
1427 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1428 disks on new nodes, allowing recreation when the original instance
1429 nodes are completely gone
1430 - Added option when converting disk templates to DRBD to skip waiting
1431 for the resync, in order to make the instance available sooner
1432 - Added two new variables to the OS scripts environment (containing the
1434 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1435 to allow use of ``pvgrub`` as bootloader
1436 - Changed the instance memory modifications to only check out-of-memory
1437 conditions on memory increases, and turned the secondary node warnings
1438 into errors (they can still be overridden via ``--force``)
1439 - Fixed the handling of a corner case when the Python installation gets
1440 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1441 to execute a command that doesn't exist
1442 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1443 node of the instance returned failures during instance shutdown; this
1444 adds the option ``--ignore-consistency`` to gnt-instance move
1446 And as usual, various improvements to the error messages, documentation
1453 *(Released Wed, 09 Mar 2011)*
1455 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1456 the RAPI users file if run twice (even with ``--dry-run``).
1458 The release fixes that bug (nothing else changed).
1464 *(Released Mon, 07 Mar 2011)*
1466 Final 2.4.0 release. Just a few small fixes:
1468 - Fixed RAPI node evacuate
1469 - Fixed the kvm-ifup script
1470 - Fixed internal error handling for special job cases
1471 - Updated man page to specify the escaping feature for options
1477 *(Released Mon, 28 Feb 2011)*
1479 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1480 that this feature wiped the disks of the wrong instance, leading to loss
1485 - Fixed title of query field containing instance name
1486 - Expanded the glossary in the documentation
1487 - Fixed one unittest (internal issue)
1493 *(Released Mon, 21 Feb 2011)*
1495 A number of bug fixes plus just a couple functionality changes.
1497 On the user-visible side, the ``gnt-* list`` command output has changed
1498 with respect to "special" field states. The current rc1 style of display
1499 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1500 the default output mode special fields states are displayed as follows:
1502 - Offline resource: ``*``
1503 - Unavailable/not applicable: ``-``
1504 - Data missing (RPC failure): ``?``
1505 - Unknown field: ``??``
1507 Another user-visible change is the addition of ``--force-join`` to
1512 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1513 - Fixed regression in RAPI/instance reinstall where all parameters were
1514 required (instead of optional)
1515 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1516 - Fixed iallocator usage (offline nodes were not considered offline)
1517 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1518 - Fixed hypervisor and OS parameter validation with respect to
1519 non-vm_capable nodes
1520 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1522 - Fixed ``tools/listrunner`` with respect to agent-based usage
1528 *(Released Fri, 4 Feb 2011)*
1530 Many changes and fixes since the beta1 release. While there were some
1531 internal changes, the code has been mostly stabilised for the RC
1534 Note: the dumb allocator was removed in this release, as it was not kept
1535 up-to-date with the IAllocator protocol changes. It is recommended to
1536 use the ``hail`` command from the ganeti-htools package.
1538 Note: the 2.4 and up versions of Ganeti are not compatible with the
1539 0.2.x branch of ganeti-htools. You need to upgrade to
1540 ganeti-htools-0.3.0 (or later).
1542 Regressions fixed from 2.3
1543 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1545 - Fixed the ``gnt-cluster verify-disks`` command
1546 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1548 - Fixed disk adoption breakage
1549 - Fixed wrong headers in instance listing for field aliases
1554 - Fixed corner case in KVM handling of NICs
1555 - Fixed many cases of wrong handling of non-vm_capable nodes
1556 - Fixed a bug where a missing instance symlink was not possible to
1557 recreate with any ``gnt-*`` command (now ``gnt-instance
1558 activate-disks`` does it)
1559 - Fixed the volume group name as reported by ``gnt-cluster
1561 - Increased timeouts for the import-export code, hopefully leading to
1562 fewer aborts due network or instance timeouts
1563 - Fixed bug in ``gnt-node list-storage``
1564 - Fixed bug where not all daemons were started on cluster
1565 initialisation, but only at the first watcher run
1566 - Fixed many bugs in the OOB implementation
1567 - Fixed watcher behaviour in presence of instances with offline
1569 - Fixed instance list output for instances running on the wrong node
1570 - a few fixes to the cluster-merge tool, but it still cannot merge
1571 multi-node groups (currently it is not recommended to use this tool)
1577 - Improved network configuration for the KVM hypervisor
1578 - Added e1000 as a supported NIC for Xen-HVM
1579 - Improved the lvmstrap tool to also be able to use partitions, as
1580 opposed to full disks
1581 - Improved speed of disk wiping (the cluster parameter
1582 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1583 of instance creations
1584 - Added documentation for the OS parameters
1585 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1586 hypervisor is not responding
1587 - Added display of blacklisted and hidden OS information in
1588 ``gnt-cluster info``
1589 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1590 NIC and node parameters, which might create problems with currently
1591 invalid (but undetected) configuration files, but prevents validation
1592 failures when unrelated parameters are modified
1593 - Changed cluster initialisation to wait for the master daemon to become
1595 - Expanded the RAPI interface:
1597 - Added config redistribution resource
1598 - Added activation/deactivation of instance disks
1599 - Added export of console information
1601 - Implemented log file reopening on SIGHUP, which allows using
1602 logrotate(8) for the Ganeti log files
1603 - Added a basic OOB helper script as an example
1609 *(Released Fri, 14 Jan 2011)*
1614 - Fixed timezone issues when formatting timestamps
1615 - Added support for node groups, available via ``gnt-group`` and other
1617 - Added out-of-band framework and management, see :doc:`design
1618 document <design-oob>`
1619 - Removed support for roman numbers from ``gnt-node list`` and
1620 ``gnt-instance list``.
1621 - Allowed modification of master network interface via ``gnt-cluster
1622 modify --master-netdev``
1623 - Accept offline secondaries while shutting down instance disks
1624 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1625 - Added support for multiple LVM volume groups
1626 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1628 - Added commands to list available fields:
1630 - ``gnt-node list-fields``
1631 - ``gnt-group list-fields``
1632 - ``gnt-instance list-fields``
1634 - Updated documentation and man pages
1639 - Moved ``rapi_users`` file into separate directory, now named
1640 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1642 - Added new tool for running commands on many machines,
1643 ``tools/ganeti-listrunner``
1644 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1645 improving the ``gnt-instance console`` output
1646 - Allowed customisation of disk index separator at ``configure`` time
1647 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1648 - Added support for non-partitioned md disks in ``lvmstrap``
1649 - Added script to gracefully power off KVM instances
1650 - Split ``utils`` module into smaller parts
1651 - Changed query operations to return more detailed information, e.g.
1652 whether an information is unavailable due to an offline node. To use
1653 this new functionality, the LUXI call ``Query`` must be used. Field
1654 information is now stored by the master daemon and can be retrieved
1655 using ``QueryFields``. Instances, nodes and groups can also be queried
1656 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1657 exposed via RAPI). The following commands make use of this
1658 infrastructure change:
1660 - ``gnt-group list``
1661 - ``gnt-group list-fields``
1663 - ``gnt-node list-fields``
1664 - ``gnt-instance list``
1665 - ``gnt-instance list-fields``
1666 - ``gnt-debug locks``
1671 - New RAPI resources (see :doc:`rapi`):
1675 - ``/2/groups/[group_name]``
1676 - ``/2/groups/[group_name]/assign-nodes``
1677 - ``/2/groups/[group_name]/modify``
1678 - ``/2/groups/[group_name]/rename``
1679 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1683 - Implemented ``no_install`` for instance creation
1684 - Implemented OS parameters for instance reinstallation, allowing
1685 use of special settings on reinstallation (e.g. for preserving data)
1690 - Added IPv6 support in import/export
1691 - Pause DRBD synchronization while wiping disks on instance creation
1692 - Updated unittests and QA scripts
1693 - Improved network parameters passed to KVM
1694 - Converted man pages from docbook to reStructuredText
1700 *(Released Mon, 20 Dec 2010)*
1702 Released version 2.3.1~rc1 without any changes.
1708 *(Released Wed, 1 Dec 2010)*
1710 - impexpd: Disable OpenSSL compression in socat if possible (backport
1711 from master, commit e90739d625b, see :doc:`installation guide
1712 <install-quick>` for details)
1713 - Changed unittest coverage report to exclude test scripts
1714 - Added script to check version format
1720 *(Released Wed, 1 Dec 2010)*
1722 Released version 2.3.0~rc1 without any changes.
1728 *(Released Fri, 19 Nov 2010)*
1730 A number of bugfixes and documentation updates:
1732 - Update ganeti-os-interface documentation
1733 - Fixed a bug related to duplicate MACs or similar items which should be
1735 - Fix breakage in OS state modify
1736 - Reinstall instance: disallow offline secondaries (fixes bug related to
1737 OS changing but reinstall failing)
1738 - plus all the other fixes between 2.2.1 and 2.2.2
1744 *(Released Tue, 2 Nov 2010)*
1746 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1747 - Fixed master failover race with watcher
1748 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1750 - Accept previously stopped instance for export with instance removal
1751 - Simplify and extend the environment variables for instance OS scripts
1752 - Added new node flags, ``master_capable`` and ``vm_capable``
1753 - Added optional instance disk wiping prior during allocation. This is a
1754 cluster-wide option and can be set/modified using
1755 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1756 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1757 :doc:`install-quick`
1758 - Added a new watcher option (``--ignore-pause``)
1759 - Added option to ignore offline node on instance start/stop
1760 (``--ignore-offline``)
1761 - Allow overriding OS parameters with ``gnt-instance reinstall``
1762 - Added ability to change node's secondary IP address using ``gnt-node
1764 - Implemented privilege separation for all daemons except
1765 ``ganeti-noded``, see ``configure`` options
1766 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1768 - Implemented job priorities (see ``ganeti(7)`` manpage)
1769 - Ignore failures while shutting down instances during failover from
1771 - Exit daemon's bootstrap process only once daemon is ready
1772 - Export more information via ``LUInstanceQuery``/remote API
1773 - Improved documentation, QA and unittests
1774 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1775 restart if the file was created or changed
1776 - Added LUXI protocol version sent with each request and response,
1777 allowing detection of server/client mismatches
1778 - Moved the Python scripts among gnt-* and ganeti-* into modules
1779 - Moved all code related to setting up SSH to an external script,
1781 - Infrastructure changes for node group support in future versions
1787 *(Released Fri, 19 Nov 2010)*
1789 A few small bugs fixed, and some improvements to the build system:
1791 - Fix documentation regarding conversion to drbd
1792 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1794 - Fix error handling in node modify with multiple changes
1795 - Allow remote imports without checked names
1801 *(Released Tue, 19 Oct 2010)*
1803 - Disable SSL session ID cache in RPC client
1809 *(Released Thu, 14 Oct 2010)*
1811 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1812 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1814 - Fix problems with interaction between Curl and Python's HTTP server,
1815 resulting in increased speed in many RPC calls
1816 - Improve our release script to prevent breakage with older aclocal and
1823 *(Released Thu, 7 Oct 2010)*
1825 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1826 value retrieved from master
1827 - Added support for blacklisted or hidden OS definitions
1828 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1829 - Added support for -mem-path in KVM hypervisor abstraction layer
1830 - Allow overriding instance parameters in tool for inter-cluster
1831 instance moves (``tools/move-instance``)
1832 - Improved opcode summaries (e.g. in ``gnt-job list``)
1833 - Improve consistency of OS listing by sorting it
1834 - Documentation updates
1840 *(Released Fri, 8 Oct 2010)*
1842 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1848 *(Released Mon, 4 Oct 2010)*
1850 - Fixed regression in ``gnt-instance rename``
1856 *(Released Wed, 22 Sep 2010)*
1858 - Fixed OS_VARIANT variable for OS scripts
1859 - Fixed cluster tag operations via RAPI
1860 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1861 - Disabled RAPI CA checks in watcher
1867 *(Released Mon, 23 Aug 2010)*
1869 - Support DRBD versions of the format "a.b.c.d"
1871 - Re-introduce support for usage from multiple threads in RAPI client
1872 - Instance renames and modify via RAPI
1873 - Work around race condition between processing and archival in job
1875 - Mark opcodes following failed one as failed, too
1876 - Job field ``lock_status`` was removed due to difficulties making it
1877 work with the changed job queue in Ganeti 2.2; a better way to monitor
1878 locks is expected for a later 2.2.x release
1879 - Fixed dry-run behaviour with many commands
1880 - Support ``ssh-agent`` again when adding nodes
1881 - Many additional bugfixes
1887 *(Released Fri, 30 Jul 2010)*
1889 Important change: the internal RPC mechanism between Ganeti nodes has
1890 changed from using a home-grown http library (based on the Python base
1891 libraries) to use the PycURL library. This requires that PycURL is
1892 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1893 against GnuTLS by default. cURL's support for GnuTLS had known issues
1894 before cURL 7.21.0 and we recommend using the latest cURL release or
1895 linking against OpenSSL. Most other distributions already link PycURL
1896 and cURL against OpenSSL. The command::
1898 python -c 'import pycurl; print pycurl.version'
1900 can be used to determine the libraries PycURL and cURL are linked
1903 Other significant changes:
1905 - Rewrote much of the internals of the job queue, in order to achieve
1906 better parallelism; this decouples job query operations from the job
1907 processing, and it should allow much nicer behaviour of the master
1908 daemon under load, and it also has uncovered some long-standing bugs
1909 related to the job serialisation (now fixed)
1910 - Added a default iallocator setting to the cluster parameters,
1911 eliminating the need to always pass nodes or an iallocator for
1912 operations that require selection of new node(s)
1913 - Added experimental support for the LXC virtualization method
1914 - Added support for OS parameters, which allows the installation of
1915 instances to pass parameter to OS scripts in order to customise the
1917 - Added a hypervisor parameter controlling the migration type (live or
1918 non-live), since hypervisors have various levels of reliability; this
1919 has renamed the 'live' parameter to 'mode'
1920 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1921 logical volumes, meaning that cluster verify will ignore them and not
1922 flag their presence as errors
1923 - The watcher will now reset the error count for failed instances after
1924 8 hours, thus allowing self-healing if the problem that caused the
1925 instances to be down/fail to start has cleared in the meantime
1926 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1927 check for, and warn, if the drbd module parameter ``usermode_helper``
1928 is not consistent with the cluster-wide setting; this is needed to
1929 make diagnose easier of failed drbd creations
1930 - Started adding base IPv6 support, but this is not yet
1931 enabled/available for use
1932 - Rename operations (cluster, instance) will now return the new name,
1933 which is especially useful if a short name was passed in
1934 - Added support for instance migration in RAPI
1935 - Added a tool to pre-configure nodes for the SSH setup, before joining
1936 them to the cluster; this will allow in the future a simplified model
1937 for node joining (but not yet fully enabled in 2.2); this needs the
1938 paramiko python library
1939 - Fixed handling of name-resolving errors
1940 - Fixed consistency of job results on the error path
1941 - Fixed master-failover race condition when executed multiple times in
1943 - Fixed many bugs related to the job queue (mostly introduced during the
1944 2.2 development cycle, so not all are impacting 2.1)
1945 - Fixed instance migration with missing disk symlinks
1946 - Fixed handling of unknown jobs in ``gnt-job archive``
1947 - And many other small fixes/improvements
1951 - Enhanced both the unittest and the QA coverage
1952 - Switched the opcode validation to a generic model, and extended the
1953 validation to all opcode parameters
1954 - Changed more parts of the code that write shell scripts to use the
1956 - Switched the master daemon to use the asyncore library for the Luxi
1963 *(Released Thu, 17 Jun 2010)*
1965 - Added tool (``move-instance``) and infrastructure to move instances
1966 between separate clusters (see :doc:`separate documentation
1967 <move-instance>` and :doc:`design document <design-2.2>`)
1968 - Added per-request RPC timeout
1969 - RAPI now requires a Content-Type header for requests with a body (e.g.
1970 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1971 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1972 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1974 - Implemented initial support for running Ganeti daemons as separate
1975 users, see configure-time flags ``--with-user-prefix`` and
1976 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1978 - Instances can be removed after export (``gnt-backup export
1979 --remove-instance``)
1980 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1981 key (instead of 1024 bit)
1982 - Added new cluster configuration file for cluster domain secret
1983 - Import/export now use SSL instead of SSH
1984 - Added support for showing estimated time when exporting an instance,
1985 see the ``ganeti-os-interface(7)`` manpage and look for
1992 *(Released Tue, 16 Nov 2010)*
1994 Some more bugfixes. Unless critical bugs occur, this will be the last
1997 - Fix case of MAC special-values
1998 - Fix mac checker regex
1999 - backend: Fix typo causing "out of range" error
2000 - Add missing --units in gnt-instance list man page
2006 *(Released Tue, 24 Aug 2010)*
2009 - Don't ignore secondary node silently on non-mirrored disk templates
2011 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2012 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2013 - Properly document the usb_mouse parameter
2014 - Fix path in ganeti-rapi(8) (issue 116)
2015 - Adjust error message when the ganeti user's .ssh directory is
2017 - Add same-node-check when changing the disk template to drbd
2023 *(Released Fri, 16 Jul 2010)*
2026 - Add an option to only select some reboot types during qa/burnin.
2027 (on some hypervisors consequent reboots are not supported)
2028 - Fix infrequent race condition in master failover. Sometimes the old
2029 master ip address would be still detected as up for a short time
2030 after it was removed, causing failover to fail.
2031 - Decrease mlockall warnings when the ctypes module is missing. On
2032 Python 2.4 we support running even if no ctypes module is installed,
2033 but we were too verbose about this issue.
2034 - Fix building on old distributions, on which man doesn't have a
2036 - Fix RAPI not to ignore the MAC address on instance creation
2037 - Implement the old instance creation format in the RAPI client.
2043 *(Released Thu, 01 Jul 2010)*
2045 A small bugfix release:
2046 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2047 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2048 a non-existing option
2049 - Fix instance create: the --force-variant option was ignored
2050 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2051 - Fix modify node storage with non-FQDN arguments
2052 - Fix RAPI client to authenticate under Python 2.6 when used
2053 for more than 5 requests needing authentication
2054 - Fix gnt-instance modify -t (storage) giving a wrong error message
2055 when converting a non-shutdown drbd instance to plain
2061 *(Released Fri, 18 Jun 2010)*
2063 A small bugfix release:
2065 - Fix live migration of KVM instances started with older Ganeti
2066 versions which had fewer hypervisor parameters
2067 - Fix gnt-instance grow-disk on down instances
2068 - Fix an error-reporting bug during instance migration
2069 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2070 silently ignoring broken ones
2071 - Fix an RPC error reporting bug affecting, for example, RAPI client
2073 - Fix bug triggered by different API version os-es on different nodes
2074 - Fix a bug in instance startup with custom hvparams: OS level
2075 parameters would fail to be applied.
2076 - Fix the RAPI client under Python 2.6 (but more work is needed to
2077 make it work completely well with OpenSSL)
2078 - Fix handling of errors when resolving names from DNS
2084 *(Released Thu, 3 Jun 2010)*
2086 A medium sized development cycle. Some new features, and some
2087 fixes/small improvements/cleanups.
2089 Significant features
2090 ~~~~~~~~~~~~~~~~~~~~
2092 The node deamon now tries to mlock itself into memory, unless the
2093 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2094 its logs, and falls back to console logging. This allows emergency
2095 features such as ``gnt-node powercycle`` to work even in the event of a
2096 broken node disk (tested offlining the disk hosting the node's
2097 filesystem and dropping its memory caches; don't try this at home)
2099 KVM: add vhost-net acceleration support. It can be tested with a new
2100 enough version of the kernel and of qemu-kvm.
2102 KVM: Add instance chrooting feature. If you use privilege dropping for
2103 your VMs you can also now force them to chroot to an empty directory,
2104 before starting the emulated guest.
2106 KVM: Add maximum migration bandwith and maximum downtime tweaking
2107 support (requires a new-enough version of qemu-kvm).
2109 Cluster verify will now warn if the master node doesn't have the master
2110 ip configured on it.
2112 Add a new (incompatible) instance creation request format to RAPI which
2113 supports all parameters (previously only a subset was supported, and it
2114 wasn't possible to extend the old format to accomodate all the new
2115 features. The old format is still supported, and a client can check for
2116 this feature, before using it, by checking for its presence in the
2117 ``features`` RAPI resource.
2119 Now with ancient latin support. Try it passing the ``--roman`` option to
2120 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2121 (requires the python-roman module to be installed, in order to work).
2126 As usual many internal code refactorings, documentation updates, and
2129 - Lots of improvements and cleanups to the experimental Remote API
2130 (RAPI) client library.
2131 - A new unit test suite for the core daemon libraries.
2132 - A fix to creating missing directories makes sure the umask is not
2133 applied anymore. This enforces the same directory permissions
2135 - Better handling terminating daemons with ctrl+c (used when running
2136 them in debugging mode).
2137 - Fix a race condition in live migrating a KVM instance, when stat()
2138 on the old proc status file returned EINVAL, which is an unexpected
2140 - Fixed manpage checking with newer man and utf-8 charachters. But now
2141 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2147 *(Released Fri, 7 May 2010)*
2149 Fix a bug which prevented untagged KVM instances from starting.
2155 *(Released Fri, 7 May 2010)*
2157 Another release with a long development cycle, during which many
2158 different features were added.
2160 Significant features
2161 ~~~~~~~~~~~~~~~~~~~~
2163 The KVM hypervisor now can run the individual instances as non-root, to
2164 reduce the impact of a VM being hijacked due to bugs in the
2165 hypervisor. It is possible to run all instances as a single (non-root)
2166 user, to manually specify a user for each instance, or to dynamically
2167 allocate a user out of a cluster-wide pool to each instance, with the
2168 guarantee that no two instances will run under the same user ID on any
2171 An experimental RAPI client library, that can be used standalone
2172 (without the other Ganeti libraries), is provided in the source tree as
2173 ``lib/rapi/client.py``. Note this client might change its interface in
2174 the future, as we iterate on its capabilities.
2176 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2177 replace the cluster's certificates and crypto keys. This might help in
2178 case they have been compromised, or have simply expired.
2180 A new disk option for instance creation has been added that allows one
2181 to "adopt" currently existing logical volumes, with data
2182 preservation. This should allow easier migration to Ganeti from
2183 unmanaged (or managed via other software) instances.
2185 Another disk improvement is the possibility to convert between redundant
2186 (DRBD) and plain (LVM) disk configuration for an instance. This should
2187 allow better scalability (starting with one node and growing the
2188 cluster, or shrinking a two-node cluster to one node).
2190 A new feature that could help with automated node failovers has been
2191 implemented: if a node sees itself as offline (by querying the master
2192 candidates), it will try to shutdown (hard) all instances and any active
2193 DRBD devices. This reduces the risk of duplicate instances if an
2194 external script automatically failovers the instances on such nodes. To
2195 enable this, the cluster parameter ``maintain_node_health`` should be
2196 enabled; in the future this option (per the name) will enable other
2197 automatic maintenance features.
2199 Instance export/import now will reuse the original instance
2200 specifications for all parameters; that means exporting an instance,
2201 deleting it and the importing it back should give an almost identical
2202 instance. Note that the default import behaviour has changed from
2203 before, where it created only one NIC; now it recreates the original
2206 Cluster verify has added a few new checks: SSL certificates validity,
2207 /etc/hosts consistency across the cluster, etc.
2212 As usual, many internal changes were done, documentation fixes,
2215 - Fixed cluster initialization with disabled cluster storage (regression
2216 introduced in 2.1.1)
2217 - File-based storage supports growing the disks
2218 - Fixed behaviour of node role changes
2219 - Fixed cluster verify for some corner cases, plus a general rewrite of
2220 cluster verify to allow future extension with more checks
2221 - Fixed log spamming by watcher and node daemon (regression introduced
2223 - Fixed possible validation issues when changing the list of enabled
2225 - Fixed cleanup of /etc/hosts during node removal
2226 - Fixed RAPI response for invalid methods
2227 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2228 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2229 from ide disks, etc.)
2230 - Allow OS changes without re-installation (to record a changed OS
2231 outside of Ganeti, or to allow OS renames)
2232 - Allow instance creation without OS installation (useful for example if
2233 the OS will be installed manually, or restored from a backup not in
2235 - Implemented option to make cluster ``copyfile`` use the replication
2237 - Added list of enabled hypervisors to ssconf (possibly useful for
2239 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2241 - A partial form of node re-IP is possible via node readd, which now
2242 allows changed node primary IP
2243 - Command line utilities now show an informational message if the job is
2245 - The logs of the master daemon now show the PID/UID/GID of the
2252 *(Released Fri, 12 Mar 2010)*
2254 During the 2.1.0 long release candidate cycle, a lot of improvements and
2255 changes have accumulated with were released later as 2.1.1.
2260 The node evacuate command (``gnt-node evacuate``) was significantly
2261 rewritten, and as such the IAllocator protocol was changed - a new
2262 request type has been added. This unfortunate change during a stable
2263 series is designed to improve performance of node evacuations; on
2264 clusters with more than about five nodes and which are well-balanced,
2265 evacuation should proceed in parallel for all instances of the node
2266 being evacuated. As such, any existing IAllocator scripts need to be
2267 updated, otherwise the above command will fail due to the unknown
2268 request. The provided "dumb" allocator has not been updated; but the
2269 ganeti-htools package supports the new protocol since version 0.2.4.
2271 Another important change is increased validation of node and instance
2272 names. This might create problems in special cases, if invalid host
2273 names are being used.
2275 Also, a new layer of hypervisor parameters has been added, that sits at
2276 OS level between the cluster defaults and the instance ones. This allows
2277 customisation of virtualization parameters depending on the installed
2278 OS. For example instances with OS 'X' may have a different KVM kernel
2279 (or any other parameter) than the cluster defaults. This is intended to
2280 help managing a multiple OSes on the same cluster, without manual
2281 modification of each instance's parameters.
2283 A tool for merging clusters, ``cluster-merge``, has been added in the
2284 tools sub-directory.
2289 - Improved the int/float conversions that should make the code more
2290 robust in face of errors from the node daemons
2291 - Fixed the remove node code in case of internal configuration errors
2292 - Fixed the node daemon behaviour in face of inconsistent queue
2293 directory (e.g. read-only file-system where we can't open the files
2295 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2296 now it either aborts cleanly or, if given the new "auto_promote"
2297 parameter, will automatically promote other nodes as needed
2298 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2299 completely prevent Ganeti from working
2300 - Fixed bug for instance export when not all disks were successfully
2302 - Fixed behaviour of node add when the new node is slow in starting up
2304 - Fixed handling of signals in the LUXI client, which should improve
2305 behaviour of command-line scripts
2306 - Added checks for invalid node/instance names in the configuration (now
2307 flagged during cluster verify)
2308 - Fixed watcher behaviour for disk activation errors
2309 - Fixed two potentially endless loops in http library, which led to the
2310 RAPI daemon hanging and consuming 100% CPU in some cases
2311 - Fixed bug in RAPI daemon related to hashed passwords
2312 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2314 - Enhanced compatibility with non-Debian OSes, but not using absolute
2315 path in some commands and allowing customisation of the ssh
2316 configuration directory
2317 - Fixed possible future issue with new Python versions by abiding to the
2318 proper use of ``__slots__`` attribute on classes
2319 - Added checks that should prevent directory traversal attacks
2320 - Many documentation fixes based on feedback from users
2325 - Added an "early_release" more for instance replace disks and node
2326 evacuate, where we release locks earlier and thus allow higher
2327 parallelism within the cluster
2328 - Added watcher hooks, intended to allow the watcher to restart other
2329 daemons (e.g. from the ganeti-nbma project), but they can be used of
2330 course for any other purpose
2331 - Added a compile-time disable for DRBD barriers, to increase
2332 performance if the administrator trusts the power supply or the
2333 storage system to not lose writes
2334 - Added the option of using syslog for logging instead of, or in
2335 addition to, Ganeti's own log files
2336 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2337 can indeed boot from a paravirtual NIC
2338 - Added a generic debug level for many operations; while this is not
2339 used widely yet, it allows one to pass the debug value all the way to
2341 - Enhanced the hooks environment for instance moves (failovers,
2342 migrations) where the primary/secondary nodes changed during the
2343 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2344 - Enhanced data validations for many user-supplied values; one important
2345 item is the restrictions imposed on instance and node names, which
2346 might reject some (invalid) host names
2347 - Add a configure-time option to disable file-based storage, if it's not
2348 needed; this allows greater security separation between the master
2349 node and the other nodes from the point of view of the inter-node RPC
2351 - Added user notification in interactive tools if job is waiting in the
2352 job queue or trying to acquire locks
2353 - Added log messages when a job is waiting for locks
2354 - Added filtering by node tags in instance operations which admit
2355 multiple instances (start, stop, reboot, reinstall)
2356 - Added a new tool for cluster mergers, ``cluster-merge``
2357 - Parameters from command line which are of the form ``a=b,c=d`` can now
2358 use backslash escapes to pass in values which contain commas,
2359 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2361 - For KVM, the instance name is the first parameter passed to KVM, so
2362 that it's more visible in the process list
2368 *(Released Tue, 2 Mar 2010)*
2370 Ganeti 2.1 brings many improvements with it. Major changes:
2372 - Added infrastructure to ease automated disk repairs
2373 - Added new daemon to export configuration data in a cheaper way than
2374 using the remote API
2375 - Instance NICs can now be routed instead of being associated with a
2377 - Improved job locking logic to reduce impact of jobs acquiring multiple
2378 locks waiting for other long-running jobs
2380 In-depth implementation details can be found in the Ganeti 2.1 design
2386 - Added chroot hypervisor
2387 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2389 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2390 ``bootloader_path`` and ``bootloader_args``)
2391 - Added the ``use_localtime`` option for the xen-hvm and kvm
2392 hypervisors, and the default value for this has changed to false (in
2393 2.0 xen-hvm always enabled it)
2394 - Added luxi call to submit multiple jobs in one go
2395 - Added cluster initialization option to not modify ``/etc/hosts``
2397 - Added network interface parameters
2398 - Added dry run mode to some LUs
2399 - Added RAPI resources:
2401 - ``/2/instances/[instance_name]/info``
2402 - ``/2/instances/[instance_name]/replace-disks``
2403 - ``/2/nodes/[node_name]/evacuate``
2404 - ``/2/nodes/[node_name]/migrate``
2405 - ``/2/nodes/[node_name]/role``
2406 - ``/2/nodes/[node_name]/storage``
2407 - ``/2/nodes/[node_name]/storage/modify``
2408 - ``/2/nodes/[node_name]/storage/repair``
2410 - Added OpCodes to evacuate or migrate all instances on a node
2411 - Added new command to list storage elements on nodes (``gnt-node
2412 list-storage``) and modify them (``gnt-node modify-storage``)
2413 - Added new ssconf files with master candidate IP address
2414 (``ssconf_master_candidates_ips``), node primary IP address
2415 (``ssconf_node_primary_ips``) and node secondary IP address
2416 (``ssconf_node_secondary_ips``)
2417 - Added ``ganeti-confd`` and a client library to query the Ganeti
2418 configuration via UDP
2419 - Added ability to run hooks after cluster initialization and before
2421 - Added automatic mode for disk replace (``gnt-instance replace-disks
2423 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2424 after catastrophic data-loss
2425 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2427 - Added ``gnt-instance move`` command to move instances
2428 - Added ``gnt-cluster watcher`` command to control watcher
2429 - Added ``gnt-node powercycle`` command to powercycle nodes
2430 - Added new job status field ``lock_status``
2431 - Added parseable error codes to cluster verification (``gnt-cluster
2432 verify --error-codes``) and made output less verbose (use
2433 ``--verbose`` to restore previous behaviour)
2434 - Added UUIDs to the main config entities (cluster, nodes, instances)
2435 - Added support for OS variants
2436 - Added support for hashed passwords in the Ganeti remote API users file
2438 - Added option to specify maximum timeout on instance shutdown
2439 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2440 - Added new helper script to start and stop Ganeti daemons
2441 (``daemon-util``), with the intent to reduce the work necessary to
2442 adjust Ganeti for non-Debian distributions and to start/stop daemons
2444 - Added more unittests
2445 - Fixed critical bug in ganeti-masterd startup
2446 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2447 now customisable at the cluster level for both the KVM and Xen
2448 hypervisors using the new ``migration_port`` parameter
2449 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2450 reinstalling an instance
2451 - Allowed ``@`` in tag names
2452 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2453 - Many documentation updates
2454 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2455 - ``gnt-instance reinstall`` can now reinstall multiple instances
2456 - Updated many command line parameters
2457 - Introduced new OS API version 15
2458 - No longer support a default hypervisor
2459 - Treat virtual LVs as inexistent
2460 - Improved job locking logic to reduce lock contention
2461 - Match instance and node names case insensitively
2462 - Reimplemented bash completion script to be more complete
2469 *(Released Thu, 4 Feb 2010)*
2471 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2472 - Fix a string formatting bug
2473 - Improve safety of the code in some error paths
2474 - Improve data validation in the master of values returned from nodes
2480 *(Released Thu, 17 Dec 2009)*
2482 - Fix security issue due to missing validation of iallocator names; this
2483 allows local and remote execution of arbitrary executables
2484 - Fix failure of gnt-node list during instance removal
2485 - Ship the RAPI documentation in the archive
2491 *(Released Wed, 30 Sep 2009)*
2493 - Fixed many wrong messages
2494 - Fixed a few bugs related to the locking library
2495 - Fixed MAC checking at instance creation time
2496 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2497 - Fixed a few issues related to signal handling in both daemons and
2499 - Fixed the example startup script provided
2500 - Fixed insserv dependencies in the example startup script (patch from
2502 - Fixed handling of drained nodes in the iallocator framework
2503 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2505 - Fixed error related to invalid job IDs in job polling
2506 - Fixed job/opcode persistence on unclean master shutdown
2507 - Fixed handling of partial job processing after unclean master
2509 - Fixed error reporting from LUs, previously all errors were converted
2510 into execution errors
2511 - Fixed error reporting from burnin
2512 - Decreased significantly the memory usage of the job queue
2513 - Optimised slightly multi-job submission
2514 - Optimised slightly opcode loading
2515 - Backported the multi-job submit framework from the development
2516 branch; multi-instance start and stop should be faster
2517 - Added script to clean archived jobs after 21 days; this will reduce
2518 the size of the queue directory
2519 - Added some extra checks in disk size tracking
2520 - Added an example ethers hook script
2521 - Added a cluster parameter that prevents Ganeti from modifying of
2523 - Added more node information to RAPI responses
2524 - Added a ``gnt-job watch`` command that allows following the ouput of a
2526 - Added a bind-address option to ganeti-rapi
2527 - Added more checks to the configuration verify
2528 - Enhanced the burnin script such that some operations can be retried
2530 - Converted instance reinstall to multi-instance model
2536 *(Released Fri, 7 Aug 2009)*
2538 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2539 to allow using the pre-2.0.2 behaviour in activation, if any existing
2540 instances have mismatched disk sizes in the configuration
2541 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2542 any configuration mismatches for disk sizes
2543 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2544 failover to work on two-node clusters
2545 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2547 - Fixed detection of OS script errors in ``gnt-backup export``
2548 - Fixed exit code of ``gnt-backup export``
2554 *(Released Fri, 17 Jul 2009)*
2556 - Added experimental support for stripped logical volumes; this should
2557 enhance performance but comes with a higher complexity in the block
2558 device handling; stripping is only enabled when passing
2559 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2560 affected even in the non-stripped mode
2561 - Improved resiliency against transient failures at the end of DRBD
2562 resyncs, and in general of DRBD resync checks
2563 - Fixed a couple of issues with exports and snapshot errors
2564 - Fixed a couple of issues in instance listing
2565 - Added display of the disk size in ``gnt-instance info``
2566 - Fixed checking for valid OSes in instance creation
2567 - Fixed handling of the "vcpus" parameter in instance listing and in
2568 general of invalid parameters
2569 - Fixed http server library, and thus RAPI, to handle invalid
2570 username/password combinations correctly; this means that now they
2571 report unauthorized for queries too, not only for modifications,
2572 allowing earlier detect of configuration problems
2573 - Added a new "role" node list field, equivalent to the master/master
2574 candidate/drained/offline flags combinations
2575 - Fixed cluster modify and changes of candidate pool size
2576 - Fixed cluster verify error messages for wrong files on regular nodes
2577 - Fixed a couple of issues with node demotion from master candidate role
2578 - Fixed node readd issues
2579 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2580 - Added a new ``--no-voting`` option for masterfailover to fix failover
2581 on two-nodes clusters when the former master node is unreachable
2582 - Added instance reinstall over RAPI
2588 *(Released Tue, 16 Jun 2009)*
2590 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2591 allow re-adding the start in single-user option (regression from 1.2)
2592 - the watcher writes the instance status to a file, to allow monitoring
2593 to report the instance status (from the master) based on cached
2594 results of the watcher's queries; while this can get stale if the
2595 watcher is being locked due to other work on the cluster, this is
2596 still an improvement
2597 - the watcher now also restarts the node daemon and the rapi daemon if
2599 - fixed the watcher to handle full and drained queue cases
2600 - hooks export more instance data in the environment, which helps if
2601 hook scripts need to take action based on the instance's properties
2602 (no longer need to query back into ganeti)
2603 - instance failovers when the instance is stopped do not check for free
2604 RAM, so that failing over a stopped instance is possible in low memory
2606 - rapi uses queries for tags instead of jobs (for less job traffic), and
2607 for cluster tags it won't talk to masterd at all but read them from
2609 - a couple of error handling fixes in RAPI
2610 - drbd handling: improved the error handling of inconsistent disks after
2611 resync to reduce the frequency of "there are some degraded disks for
2612 this instance" messages
2613 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2614 error handling path called a wrong function name)
2620 *(Released Wed, 27 May 2009)*
2622 - no changes from rc5
2628 *(Released Wed, 20 May 2009)*
2630 - fix a couple of bugs (validation, argument checks)
2631 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2632 - some small improvements to RAPI and IAllocator
2633 - make watcher automatically start the master daemon if down
2639 *(Released Mon, 27 Apr 2009)*
2641 - change the OS list to not require locks; this helps with big clusters
2642 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2643 volume group is broken
2644 - ``gnt-instance info``, without any arguments, doesn't run for all
2645 instances anymore; either pass ``--all`` or pass the desired
2646 instances; this helps against mistakes on big clusters where listing
2647 the information for all instances takes a long time
2648 - miscellaneous doc and man pages fixes
2654 *(Released Wed, 8 Apr 2009)*
2656 - Change the internal locking model of some ``gnt-node`` commands, in
2657 order to reduce contention (and blocking of master daemon) when
2658 batching many creation/reinstall jobs
2659 - Fixes to Xen soft reboot
2660 - No longer build documentation at build time, instead distribute it in
2661 the archive, in order to reduce the need for the whole docbook/rst
2668 *(Released Fri, 27 Mar 2009)*
2670 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2671 - Fix watcher startup sequence, improves the behaviour of busy clusters
2672 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2673 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2674 bind address and other places
2675 - Some documentation fixes and updates
2681 *(Released Mon, 2 Mar 2009)*
2683 - More documentation updates, now all docs should be more-or-less
2685 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2687 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2689 - Fix an issue related to $libdir/run/ganeti and cluster creation
2695 *(Released Thu, 19 Feb 2009)*
2697 - Xen PVM and KVM have switched the default value for the instance root
2698 disk to the first partition on the first drive, instead of the whole
2699 drive; this means that the OS installation scripts must be changed
2701 - Man pages have been updated
2702 - RAPI has been switched by default to HTTPS, and the exported functions
2703 should all work correctly
2704 - RAPI v1 has been removed
2705 - Many improvements to the KVM hypervisor
2706 - Block device errors are now better reported
2707 - Many other bugfixes and small improvements
2713 *(Released Mon, 26 Jan 2009)*
2715 - Version 2 is a general rewrite of the code and therefore the
2716 differences are too many to list, see the design document for 2.0 in
2717 the ``doc/`` subdirectory for more details
2718 - In this beta version there is not yet a migration path from 1.2 (there
2719 will be one in the final 2.0 release)
2720 - A few significant changes are:
2722 - all commands are executed by a daemon (``ganeti-masterd``) and the
2723 various ``gnt-*`` commands are just front-ends to it
2724 - all the commands are entered into, and executed from a job queue,
2725 see the ``gnt-job(8)`` manpage
2726 - the RAPI daemon supports read-write operations, secured by basic
2727 HTTP authentication on top of HTTPS
2728 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2729 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2730 to migrate to DRBD 8 first while still running Ganeti 1.2)
2731 - DRBD devices are using statically allocated minor numbers, which
2732 will be assigned to existing instances during the migration process
2733 - there is support for both Xen PVM and Xen HVM instances running on
2735 - KVM virtualization is supported too
2736 - file-based storage has been implemented, which means that it is
2737 possible to run the cluster without LVM and DRBD storage, for
2738 example using a shared filesystem exported from shared storage (and
2739 still have live migration)
2745 *(Released Tue, 13 Jan 2009)*
2747 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2748 - Reuse the old instance mac address by default on instance import, if
2749 the instance name is the same.
2750 - Handle situations in which the node info rpc returns incomplete
2752 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2753 - Improved version of batcher:
2755 - state file support
2756 - instance mac address support
2757 - support for HVM clusters/instances
2759 - Add an option to show the number of cpu sockets and nodes in
2761 - Support OSes that handle more than one version of the OS api (but do
2762 not change the current API in any other way)
2763 - Fix ``gnt-node migrate``
2764 - ``gnt-debug`` man page
2765 - Fixes various more typos and small issues
2766 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2772 *(Released Wed, 24 Sep 2008)*
2774 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2775 type of disk exported to fully virtualized instances.
2776 - provide access to the serial console of HVM instances
2777 - instance auto_balance flag, set by default. If turned off it will
2778 avoid warnings on cluster verify if there is not enough memory to fail
2779 over an instance. in the future it will prevent automatically failing
2780 it over when we will support that.
2781 - batcher tool for instance creation, see ``tools/README.batcher``
2782 - ``gnt-instance reinstall --select-os`` to interactively select a new
2783 operating system when reinstalling an instance.
2784 - when changing the memory amount on instance modify a check has been
2785 added that the instance will be able to start. also warnings are
2786 emitted if the instance will not be able to fail over, if auto_balance
2788 - documentation fixes
2789 - sync fields between ``gnt-instance list/modify/add/import``
2790 - fix a race condition in drbd when the sync speed was set after giving
2791 the device a remote peer.
2797 *(Released Tue, 22 Jul 2008)*
2799 - note: the allowed size and number of tags per object were reduced
2800 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2801 - fixed twisted 8.x compatibility
2802 - fixed ``gnt-instance replace-disks`` with iallocator
2803 - add TCP keepalives on twisted connections to detect restarted nodes
2804 - disk increase support, see ``gnt-instance grow-disk``
2805 - implement bulk node/instance query for RAPI
2806 - add tags in node/instance listing (optional)
2807 - experimental migration (and live migration) support, read the man page
2808 for ``gnt-instance migrate``
2809 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2810 has some small improvements in handling its state file
2816 *(Released Fri, 13 Jun 2008)*
2818 - Experimental readonly, REST-based remote API implementation;
2819 automatically started on master node, TCP port 5080, if enabled by
2820 ``--enable-rapi`` parameter to configure script.
2821 - Instance allocator support. Add and import instance accept a
2822 ``--iallocator`` parameter, and call that instance allocator to decide
2823 which node to use for the instance. The iallocator document describes
2824 what's expected from an allocator script.
2825 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2826 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2827 node is lost there is still enough memory to fail over the instances
2829 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2830 ``gnt-cluster verify``, to check for site-specific compliance. All the
2831 hooks will run, and their output, if any, will be displayed. Any
2832 failing hook will make the verification return an error value.
2833 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2834 primary and secondary interfaces
2835 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2836 that is still declared as part of the cluster and has failed.
2837 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2838 specifying output fields, that just adds the chosen fields to the
2840 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2841 export from the filesystem.
2842 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2843 have been added. Using them you can enable/disable acpi and pae
2844 support, and specify a path for a cd image to be exported to the
2845 instance. These parameters as the name suggest only work on HVM
2847 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2848 PAE support will be set to the previously hardcoded values, but the
2849 (previously hardcoded) path to the CDROM ISO image will be unset and
2850 if required, needs to be set manually with ``gnt-instance modify``
2852 - The address to which an instance's VNC console is bound is now
2853 selectable per-instance, rather than being cluster wide. Of course
2854 this only applies to instances controlled via VNC, so currently just
2855 applies to HVM clusters.
2861 *(Released Mon, 18 Feb 2008)*
2863 - more tweaks to the disk activation code (especially helpful for DRBD)
2864 - change the default ``gnt-instance list`` output format, now there is
2865 one combined status field (see the manpage for the exact values this
2867 - some more fixes for the mac export to hooks change
2868 - make Ganeti not break with DRBD 8.2.x (which changed the version
2869 format in ``/proc/drbd``) (issue 24)
2870 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2871 template, allowing migration from DRBD0.7+MD to DRBD8
2877 *(Released Wed, 30 Jan 2008)*
2879 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2881 - add command aliases infrastructure and a few aliases
2882 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2883 man pages and the ``--help`` option of ``gnt-node
2884 list``/``gnt-instance list``
2885 - fix ``gnt-backup list`` with down nodes (issue 21)
2886 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2887 - fix the dist archive and add a check for including svn/git files in
2889 - some developer-related changes: improve the burnin and the QA suite,
2890 add an upload script for testing during development
2896 *(Released Wed, 16 Jan 2008)*
2898 - experimental HVM support, read the install document, section
2899 "Initializing the cluster"
2900 - allow for the PVM hypervisor per-instance kernel and initrd paths
2901 - add a new command ``gnt-cluster verify-disks`` which uses a new
2902 algorithm to improve the reconnection of the DRBD pairs if the device
2903 on the secondary node has gone away
2904 - make logical volume code auto-activate LVs at disk activation time
2905 - slightly improve the speed of activating disks
2906 - allow specification of the MAC address at instance creation time, and
2907 changing it later via ``gnt-instance modify``
2908 - fix handling of external commands that generate lots of output on
2910 - update documentation with regard to minimum version of DRBD8 supported
2916 *(Released Tue, 4 Dec 2007)*
2918 - Log the ``xm create`` output to the node daemon log on failure (to
2919 help diagnosing the error)
2920 - In debug mode, log all external commands output if failed to the logs
2921 - Change parsing of lvm commands to ignore stderr
2927 *(Released Wed, 28 Nov 2007)*
2929 - Another round of updates to the DRBD 8 code to deal with more failures
2930 in the replace secondary node operation
2931 - Some more logging of failures in disk operations (lvm, drbd)
2932 - A few documentation updates
2939 *(Released Tue, 13 Nov 2007)*
2941 - Change configuration file format from Python's Pickle to JSON.
2942 Upgrading is possible using the cfgupgrade utility.
2943 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2944 faster replace disks and is more stable (DRBD 8 has many improvements
2945 compared to DRBD 0.7)
2946 - Added command line tags support (see man pages for ``gnt-instance``,
2947 ``gnt-node``, ``gnt-cluster``)
2948 - Added instance rename support
2949 - Added multi-instance startup/shutdown
2950 - Added cluster rename support
2951 - Added ``gnt-node evacuate`` to simplify some node operations
2952 - Added instance reboot operation that can speedup reboot as compared to
2954 - Soften the requirement that hostnames are in FQDN format
2955 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2957 - Removed dependency on debian's patched fping that uses the
2958 non-standard ``-S`` option
2959 - Now the OS definitions are searched for in multiple, configurable
2960 paths (easier for distros to package)
2961 - Some changes to the hooks infrastructure (especially the new
2962 post-configuration update hook)
2963 - Other small bugfixes
2965 .. vim: set textwidth=72 syntax=rst :