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. This feature is currently supported only for KVM hypervisor with
101 version greater than 1.0.
102 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
103 to always require authentication, ``--require-authentication``. It can
104 be specified in ``$sysconfdir/default/ganeti``.
105 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
106 be used to manage the disk templates to be used by instances in the cluster.
107 Initially, it will be set to a list that includes plain, drbd, if they were
108 enabled by specifying a volume group name, and file and sharedfile, if those
109 were enabled at configure time. Additionally, it will include all disk
110 templates that are currently used by instances. The order of disk templates
111 will be based on Ganeti's history of supporting them. In the future, the
112 first entry of the list will be used as a default disk template on instance
114 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
115 configuration back to the previous stable version.
116 - Disk templates in group ipolicy can be restored to the default value.
117 - Initial support for diskless instances and virtual clusters in QA.
118 - More QA and unit tests for instance policies.
119 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
120 describing why the opcode itself was executed.
121 - The monitoring daemon is now available. It allows users to query the cluster
122 for obtaining information about the status of the system. The daemon is only
123 responsible for providing the information over the network: the actual data
124 gathering is performed by data collectors (currently, only the DRBD status
125 collector is available).
126 - In order to help developers work on Ganeti, a new script
127 (``devel/build_chroot``) is provided, for building a chroot that contains all
128 the required development libraries and tools for compiling Ganeti on a Debian
130 - A new tool, ``harep``, for performing self-repair and recreation of instances
131 in Ganeti has been added.
132 - Split queries are enabled for tags, network, exports, cluster info, groups,
134 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
135 It prints the command line to set the current policies, to ease
137 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
138 devices for KVM virtio-net interfaces will get created with VNET_HDR
139 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
140 virtio-net interfaces, which prevents host kernel tainting and log
141 flooding, when dealing with broken or malicious virtio-net drivers.
142 It's set to true by default.
143 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
145 - Support 'viridian' parameter in Xen HVM
146 - Support DSA SSH keys in bootstrap
147 - To simplify the work of packaging frameworks that want to add the needed users
148 and groups in a split-user setup themselves, at build time three files in
149 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
150 the groups to be generated, the ``users`` file contains, one per line, the
151 users to be generated, optionally followed by their primary group, where
152 important. The ``groupmemberships`` file contains, one per line, additional
153 user-group membership relations that need to be established. The syntax of
154 these files will remain stable in all future versions.
159 The following new dependencies have been added:
162 - The ``curl`` library is not optional anymore for compiling the Haskell code.
163 - ``snap-server`` library (if monitoring is enabled).
166 - The minimum Python version needed to run Ganeti is now 2.6.
167 - ``yaml`` library (only for running the QA).
171 - Perform proper cleanup on termination of Haskell daemons
172 - Fix corner-case in handling of remaining retry time
178 *(Released Tue, 17 Sep 2013)*
180 - To simplify the work of packaging frameworks that want to add the needed users
181 and groups in a split-user setup themselves, at build time three files in
182 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
183 the groups to be generated, the ``users`` file contains, one per line, the
184 users to be generated, optionally followed by their primary group, where
185 important. The ``groupmemberships`` file contains, one per line, additional
186 user-group membership relations that need to be established. The syntax of
187 these files will remain stable in all future versions.
188 - Add a default to file-driver when unspecified over RAPI (Issue 571)
189 - Mark the DSA host pubkey as optional, and remove it during config downgrade
191 - Some documentation fixes
197 *(Released Tue, 27 Aug 2013)*
199 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
201 - Support 'viridian' parameter in Xen HVM (Issue 233)
202 - Include VCS version in ``gnt-cluster version``
203 - Support DSA SSH keys in bootstrap (Issue 338)
204 - Fix batch creation of instances
205 - Use FQDN to check master node status (Issue 551)
206 - Make the DRBD collector more failure-resilient
212 *(Released Fri, 2 Aug 2013)*
214 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
216 - Fix upgrading/downgrading from 2.7
217 - Increase maximum RAPI message size
218 - Documentation updates
219 - Split ``confd`` between ``luxid`` and ``confd``
220 - Merge 2.7 series up to the 2.7.1 release
221 - Allow the ``modify_etc_hosts`` option to be changed
222 - Add better debugging for ``luxid`` queries
223 - Expose bulk parameter for GetJobs in RAPI client
224 - Expose missing ``network`` fields in RAPI
225 - Add some ``cluster verify`` tests
226 - Some unittest fixes
227 - Fix a malfunction in ``hspace``'s tiered allocation
228 - Fix query compatibility between haskell and python implementations
229 - Add the ``vnet_hdr`` HV parameter for KVM
230 - Add ``--cleanup`` to instance failover
231 - Change the connected groups format in ``gnt-network info`` output; it
232 was previously displayed as a raw list by mistake. (Merged from 2.7)
238 *(Released Mon, 24 Jun 2013)*
240 This was the first beta release of the 2.8 series. All important changes
241 are listed in the latest 2.8 entry.
247 *(Released Thu, 26 Sep 2013)*
249 - Change the connected groups format in ``gnt-network info`` output; it
250 was previously displayed as a raw list by mistake
251 - Check disk template in right dict when copying
252 - Support multi-instance allocs without iallocator
253 - Fix some errors in the documentation
254 - Fix formatting of tuple in an error message
260 *(Released Thu, 25 Jul 2013)*
262 - Add logrotate functionality in daemon-util
263 - Add logrotate example file
264 - Add missing fields to network queries over rapi
265 - Fix network object timestamps
266 - Add support for querying network timestamps
267 - Fix a typo in the example crontab
268 - Fix a documentation typo
274 *(Released Thu, 04 Jul 2013)*
276 Incompatible/important changes
277 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
279 - Instance policies for disk size were documented to be on a per-disk
280 basis, but hail applied them to the sum of all disks. This has been
282 - ``hbal`` will now exit with status 0 if, during job execution over
283 LUXI, early exit has been requested and all jobs are successful;
284 before, exit status 1 was used, which cannot be differentiated from
286 - Compatibility with newer versions of rbd has been fixed
287 - ``gnt-instance batch-create`` has been changed to use the bulk create
288 opcode from Ganeti. This lead to incompatible changes in the format of
289 the JSON file. It's now not a custom dict anymore but a dict
290 compatible with the ``OpInstanceCreate`` opcode.
291 - Parent directories for file storage need to be listed in
292 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
293 write the file automatically based on old configuration values, but it
294 can not distribute it across all nodes and the file contents should be
295 verified. Use ``gnt-cluster copyfile
296 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
297 upgraded. The reason for requiring this list of paths now is that
298 before it would have been possible to inject new paths via RPC,
299 allowing files to be created in arbitrary locations. The RPC protocol
300 is protected using SSL/X.509 certificates, but as a design principle
301 Ganeti does not permit arbitrary paths to be passed.
302 - The parsing of the variants file for OSes (see
303 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
304 lines and comment lines (starting with ``#``) are ignored for better
306 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
307 longer available. ``gnt-node add`` now invokes a new tool on the
308 destination node, named ``prepare-node-join``, to configure the SSH
309 daemon. Paramiko is no longer necessary to configure nodes' SSH
310 daemons via ``gnt-node add``.
311 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
312 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
313 the flag is not reset after a master failover.
314 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
315 is recommended. 2.6 will be mandatory from the 2.8 series.
321 - New network management functionality to support automatic allocation
322 of IP addresses and managing of network parameters. See
323 :manpage:`gnt-network(8)` for more details.
324 - New external storage backend, to allow managing arbitrary storage
325 systems external to the cluster. See
326 :manpage:`ganeti-extstorage-interface(7)`.
327 - New ``exclusive-storage`` node parameter added, restricted to
328 nodegroup level. When it's set to true, physical disks are assigned in
329 an exclusive fashion to instances, as documented in :doc:`Partitioned
330 Ganeti <design-partitioned>`. Currently, only instances using the
331 ``plain`` disk template are supported.
332 - The KVM hypervisor has been updated with many new hypervisor
333 parameters, including a generic one for passing arbitrary command line
334 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
335 compatible up to qemu 1.4.
336 - A new tool, called ``mon-collector``, is the stand-alone executor of
337 the data collectors for a monitoring system. As of this version, it
338 just includes the DRBD data collector, that can be executed by calling
339 ``mon-collector`` using the ``drbd`` parameter. See
340 :manpage:`mon-collector(7)`.
341 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
342 for RAPI users. It allows granting permissions to query for
343 information to a specific user without giving
344 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
345 - A new tool named ``node-cleanup`` has been added. It cleans remains of
346 a cluster from a machine by stopping all daemons, removing
347 certificates and ssconf files. Unless the ``--no-backup`` option is
348 given, copies of the certificates are made.
349 - Instance creations now support the use of opportunistic locking,
350 potentially speeding up the (parallel) creation of multiple instances.
351 This feature is currently only available via the :doc:`RAPI
352 <rapi>` interface and when an instance allocator is used. If the
353 ``opportunistic_locking`` parameter is set the opcode will try to
354 acquire as many locks as possible, but will not wait for any locks
355 held by other opcodes. If not enough resources can be found to
356 allocate the instance, the temporary error code
357 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
358 retried thereafter, with or without opportunistic locking.
359 - New experimental linux-ha resource scripts.
360 - Restricted-commands support: ganeti can now be asked (via command line
361 or rapi) to perform commands on a node. These are passed via ganeti
362 RPC rather than ssh. This functionality is restricted to commands
363 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
364 reasons. The file is not copied automatically.
370 - Diskless instances are now externally mirrored (Issue 237). This for
371 now has only been tested in conjunction with explicit target nodes for
373 - Queries not needing locks or RPC access to the node can now be
374 performed by the confd daemon, making them independent from jobs, and
375 thus faster to execute. This is selectable at configure time.
376 - The functionality for allocating multiple instances at once has been
377 overhauled and is now also available through :doc:`RAPI <rapi>`.
379 There are no significant changes from version 2.7.0~rc3.
385 *(Released Tue, 25 Jun 2013)*
387 - Fix permissions on the confd query socket (Issue 477)
388 - Fix permissions on the job archive dir (Issue 498)
389 - Fix handling of an internal exception in replace-disks (Issue 472)
390 - Fix gnt-node info handling of shortened names (Issue 497)
391 - Fix gnt-instance grow-disk when wiping is enabled
392 - Documentation improvements, and support for newer pandoc
393 - Fix hspace honoring ipolicy for disks (Issue 484)
394 - Improve handling of the ``kvm_extra`` HV parameter
400 *(Released Fri, 24 May 2013)*
402 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
404 - Disks added through ``gnt-instance modify`` or created through
405 ``gnt-instance recreate-disks`` are wiped, if the
406 ``prealloc_wipe_disks`` flag is set.
407 - If wiping newly created disks fails, the disks are removed. Also,
408 partial failures in creating disks through ``gnt-instance modify``
409 triggers a cleanup of the partially-created disks.
410 - Removing the master IP address doesn't fail if the address has been
412 - Fix ownership of the OS log dir
413 - Workaround missing SO_PEERCRED constant (Issue 191)
419 *(Released Fri, 3 May 2013)*
421 This was the first release candidate of the 2.7 series. Since beta3:
423 - Fix kvm compatibility with qemu 1.4 (Issue 389)
424 - Documentation updates (admin guide, upgrade notes, install
425 instructions) (Issue 372)
426 - Fix gnt-group list nodes and instances count (Issue 436)
427 - Fix compilation without non-mandatory libraries (Issue 441)
428 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
429 - Make confd logging more verbose at INFO level (Issue 435)
430 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
431 - Fix failure path for instance storage type conversion (Issue 229)
432 - Update htools text backend documentation
433 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
434 - Disable inter-cluster instance move for file-based instances, because
435 it is dependant on instance export, which is not supported for
436 file-based instances. (Issue 414)
437 - Fix gnt-job crashes on non-ascii characters (Issue 427)
438 - Fix volume group checks on non-vm-capable nodes (Issue 432)
444 *(Released Mon, 22 Apr 2013)*
446 This was the third beta release of the 2.7 series. Since beta2:
448 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
449 - Fix data loss on wrong usage of ``gnt-instance move``
450 - Properly export errors in confd-based job queries
451 - Add ``users-setup`` tool
452 - Fix iallocator protocol to report 0 as a disk size for diskless
453 instances. This avoids hail breaking when a diskless instance is
455 - Fix job queue directory permission problem that made confd job queries
456 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
457 for access to archived jobs (Issue 406).
458 - Limit the sizes of networks supported by ``gnt-network`` to something
459 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
460 - Fix bugs in instance disk template conversion
461 - Fix GHC 7 compatibility
462 - Fix ``burnin`` install path (Issue 426).
463 - Allow very small disk grows (Issue 347).
464 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
465 that noded doesn't import masterd code (Issue 419).
466 - Make sure the default metavg at cluster init is the same as the vg, if
467 unspecified (Issue 358).
468 - Fix cleanup of partially created disks (part of Issue 416)
474 *(Released Tue, 2 Apr 2013)*
476 This was the second beta release of the 2.7 series. Since beta1:
478 - Networks no longer have a "type" slot, since this information was
479 unused in Ganeti: instead of it tags should be used.
480 - The rapi client now has a ``target_node`` option to MigrateInstance.
481 - Fix early exit return code for hbal (Issue 386).
482 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
483 - Fix ``rbd showmapped`` output parsing (Issue 312).
484 - Networks are now referenced indexed by UUID, rather than name. This
485 will require running cfgupgrade, from 2.7.0beta1, if networks are in
487 - The OS environment now includes network information.
488 - Deleting of a network is now disallowed if any instance nic is using
489 it, to prevent dangling references.
490 - External storage is now documented in man pages.
491 - The exclusive_storage flag can now only be set at nodegroup level.
492 - Hbal can now submit an explicit priority with its jobs.
493 - Many network related locking fixes.
494 - Bump up the required pylint version to 0.25.1.
495 - Fix the ``no_remember`` option in RAPI client.
496 - Many ipolicy related tests, qa, and fixes.
497 - Many documentation improvements and fixes.
498 - Fix building with ``--disable-file-storage``.
499 - Fix ``-q`` option in htools, which was broken if passed more than
501 - Some haskell/python interaction improvements and fixes.
502 - Fix iallocator in case of missing LVM storage.
503 - Fix confd config load in case of ``--no-lvm-storage``.
504 - The confd/query functionality is now mentioned in the security
511 *(Released Wed, 6 Feb 2013)*
513 This was the first beta release of the 2.7 series. All important changes
514 are listed in the latest 2.7 entry.
520 *(Released Fri, 21 Dec 2012)*
522 Important behaviour change: hbal won't rebalance anymore instances which
523 have the ``auto_balance`` attribute set to false. This was the intention
524 all along, but until now it only skipped those from the N+1 memory
525 reservation (DRBD-specific).
527 A significant number of bug fixes in this release:
529 - Fixed disk adoption interaction with ipolicy checks.
530 - Fixed networking issues when instances are started, stopped or
531 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
532 - Fixed the warning in cluster verify for shared storage instances not
534 - Fixed removal of storage directory on shared file storage (issue 262).
535 - Fixed validation of LVM volume group name in OpClusterSetParams
536 (``gnt-cluster modify``) (issue 285).
537 - Fixed runtime memory increases (``gnt-instance modify -m``).
538 - Fixed live migration under Xen's ``xl`` mode.
539 - Fixed ``gnt-instance console`` with ``xl``.
540 - Fixed building with newer Haskell compiler/libraries.
541 - Fixed PID file writing in Haskell daemons (confd); this prevents
542 restart issues if confd was launched manually (outside of
543 ``daemon-util``) while another copy of it was running
544 - Fixed a type error when doing live migrations with KVM (issue 297) and
545 the error messages for failing migrations have been improved.
546 - Fixed opcode validation for the out-of-band commands (``gnt-node
548 - Fixed a type error when unsetting OS hypervisor parameters (issue
549 311); now it's possible to unset all OS-specific hypervisor
551 - Fixed the ``dry-run`` mode for many operations: verification of
552 results was over-zealous but didn't take into account the ``dry-run``
553 operation, resulting in "wrong" failures.
554 - Fixed bash completion in ``gnt-job list`` when the job queue has
555 hundreds of entries; especially with older ``bash`` versions, this
556 results in significant CPU usage.
558 And lastly, a few other improvements have been made:
560 - Added option to force master-failover without voting (issue 282).
561 - Clarified error message on lock conflict (issue 287).
562 - Logging of newly submitted jobs has been improved (issue 290).
563 - Hostname checks have been made uniform between instance rename and
565 - The ``--submit`` option is now supported by ``gnt-debug delay``.
566 - Shutting down the master daemon by sending SIGTERM now stops it from
567 processing jobs waiting for locks; instead, those jobs will be started
568 once again after the master daemon is started the next time (issue
570 - Support for Xen's ``xl`` program has been improved (besides the fixes
572 - Reduced logging noise in the Haskell confd daemon (only show one log
573 entry for each config reload, instead of two).
574 - Several man page updates and typo fixes.
580 *(Released Fri, 12 Oct 2012)*
582 A small bugfix release. Among the bugs fixed:
584 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
585 made the command unusable.
586 - Commands that issue many jobs don't fail anymore just because some jobs
587 take so long that other jobs are archived.
588 - Failures during ``gnt-instance reinstall`` are reflected by the exit
590 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
592 - When ``always_failover`` is set, ``--allow-failover`` is not required
593 in migrate commands anymore.
594 - ``bash_completion`` works even if extglob is disabled.
595 - Fixed bug with locks that made failover for RDB-based instances fail.
596 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
597 a random node instead of one based on the allocator metric.
598 - Support for newer versions of pylint and pep8.
599 - Hail doesn't fail anymore when trying to add an instance of type
600 ``file``, ``sharedfile`` or ``rbd``.
601 - Added new Makefile target to rebuild the whole distribution, so that
602 all files are included.
608 *(Released Fri, 27 Jul 2012)*
611 .. attention:: The ``LUXI`` protocol has been made more consistent
612 regarding its handling of command arguments. This, however, leads to
613 incompatibility issues with previous versions. Please ensure that you
614 restart Ganeti daemons soon after the upgrade, otherwise most
615 ``LUXI`` calls (job submission, setting/resetting the drain flag,
616 pausing/resuming the watcher, cancelling and archiving jobs, querying
617 the cluster configuration) will fail.
626 The current ``admin_up`` field, which used to denote whether an instance
627 should be running or not, has been removed. Instead, ``admin_state`` is
628 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
630 The rational behind this is that an instance being “down” can have
633 - it could be down during a reboot
634 - it could be temporarily be down for a reinstall
635 - or it could be down because it is deprecated and kept just for its
638 The previous Boolean state was making it difficult to do capacity
639 calculations: should Ganeti reserve memory for a down instance? Now, the
640 tri-state field makes it clear:
642 - in ``up`` and ``down`` state, all resources are reserved for the
643 instance, and it can be at any time brought up if it is down
644 - in ``offline`` state, only disk space is reserved for it, but not
647 The field can have an extra use: since the transition between ``up`` and
648 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
649 transition between ``offline`` and ``down`` is done via ``gnt-instance
650 modify``, it is possible to given different rights to users. For
651 example, owners of an instance could be allowed to start/stop it, but
652 not transition it out of the offline state.
654 Instance policies and specs
655 +++++++++++++++++++++++++++
657 In previous Ganeti versions, an instance creation request was not
658 limited on the minimum size and on the maximum size just by the cluster
659 resources. As such, any policy could be implemented only in third-party
660 clients (RAPI clients, or shell wrappers over ``gnt-*``
661 tools). Furthermore, calculating cluster capacity via ``hspace`` again
662 required external input with regards to instance sizes.
664 In order to improve these workflows and to allow for example better
665 per-node group differentiation, we introduced instance specs, which
668 - minimum instance disk size, disk count, memory size, cpu count
669 - maximum values for the above metrics
670 - and “standard” values (used in ``hspace`` to calculate the standard
673 The minimum/maximum values can be also customised at node-group level,
674 for example allowing more powerful hardware to support bigger instance
677 Beside the instance specs, there are a few other settings belonging to
678 the instance policy framework. It is possible now to customise, per
679 cluster and node-group:
681 - the list of allowed disk templates
682 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
683 - the maximum ratio of instance to spindles (see below for more
684 information) for local storage
686 All these together should allow all tools that talk to Ganeti to know
687 what are the ranges of allowed values for instances and the
688 over-subscription that is allowed.
690 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
691 instance configuration, and the physical CPU configuration from the
692 node. For the spindle ratios however, we didn't track before these
693 values, so new parameters have been added:
695 - a new node parameter ``spindle_count``, defaults to 1, customisable at
696 node group or node level
697 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
699 Note that spindles in this context doesn't need to mean actual
700 mechanical hard-drives; it's just a relative number for both the node
701 I/O capacity and instance I/O consumption.
703 Instance migration behaviour
704 ++++++++++++++++++++++++++++
706 While live-migration is in general desirable over failover, it is
707 possible that for some workloads it is actually worse, due to the
708 variable time of the “suspend” phase during live migration.
710 To allow the tools to work consistently over such instances (without
711 having to hard-code instance names), a new backend parameter
712 ``always_failover`` has been added to control the migration/failover
713 behaviour. When set to True, all migration requests for an instance will
714 instead fall-back to failover.
716 Instance memory ballooning
717 ++++++++++++++++++++++++++
719 Initial support for memory ballooning has been added. The memory for an
720 instance is no longer fixed (backend parameter ``memory``), but instead
721 can vary between minimum and maximum values (backend parameters
722 ``minmem`` and ``maxmem``). Currently we only change an instance's
725 - live migrating or failing over and instance and the target node
726 doesn't have enough memory
727 - user requests changing the memory via ``gnt-instance modify
733 In order to control the use of specific CPUs by instance, support for
734 controlling CPU pinning has been added for the Xen, HVM and LXC
735 hypervisors. This is controlled by a new hypervisor parameter
736 ``cpu_mask``; details about possible values for this are in the
737 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
738 VCPU-to-CPU mapping) form will work well only when all nodes in your
739 cluster have the same amount of CPUs.
744 Another area in which Ganeti was not customisable were the parameters
745 used for storage configuration, e.g. how many stripes to use for LVM,
746 DRBD resync configuration, etc.
748 To improve this area, we've added disks parameters, which are
749 customisable at cluster and node group level, and which allow to
750 specify various parameters for disks (DRBD has the most parameters
751 currently), for example:
753 - DRBD resync algorithm and parameters (e.g. speed)
754 - the default VG for meta-data volumes for DRBD
755 - number of stripes for LVM (plain disk template)
758 These parameters can be modified via ``gnt-cluster modify -D …`` and
759 ``gnt-group modify -D …``, and are used at either instance creation (in
760 case of LVM stripes, for example) or at disk “activation” time
763 Rados block device support
764 ++++++++++++++++++++++++++
766 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
767 denoted by the ``rbd`` disk template type. This is considered
768 experimental, feedback is welcome. For details on configuring it, see
769 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
774 The existing master IP functionality works well only in simple setups (a
775 single network shared by all nodes); however, if nodes belong to
776 different networks, then the ``/32`` setup and lack of routing
777 information is not enough.
779 To allow the master IP to function well in more complex cases, the
780 system was reworked as follows:
782 - a master IP netmask setting has been added
783 - the master IP activation/turn-down code was moved from the node daemon
785 - whether to run the Ganeti-supplied master IP script or a user-supplied
786 on is a ``gnt-cluster init`` setting
788 Details about the location of the standard and custom setup scripts are
789 in the man page :manpage:`gnt-cluster(8)`; for information about the
790 setup script protocol, look at the Ganeti-supplied script.
795 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
798 It is now possible to use TLS-protected connections, and when renewing
799 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
800 it is now possible to specify spice or spice CA certificates. Also, it
801 is possible to configure a password for SPICE sessions via the
802 hypervisor parameter ``spice_password_file``.
804 There are also new parameters to control the compression and streaming
805 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
806 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
807 for the spice parameters.
809 Lastly, it is now possible to see the SPICE connection information via
810 ``gnt-instance console``.
815 A new tool (``tools/ovfconverter``) has been added that supports
816 conversion between Ganeti and the `Open Virtualization Format
817 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
820 This relies on the ``qemu-img`` tool to convert the disk formats, so the
821 actual compatibility with other virtualization solutions depends on it.
826 The configuration query daemon (``ganeti-confd``) is now optional, and
827 has been rewritten in Haskell; whether to use the daemon at all, use the
828 Python (default) or the Haskell version is selectable at configure time
829 via the ``--enable-confd`` parameter, which can take one of the
830 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
831 daemon will result in a smaller footprint; for larger systems, we
832 welcome feedback on the Haskell version which might become the default
835 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
836 daemon running. The Python version doesn't implement the new call.
839 User interface changes
840 ~~~~~~~~~~~~~~~~~~~~~~
842 We have replaced the ``--disks`` option of ``gnt-instance
843 replace-disks`` with a more flexible ``--disk`` option, which allows
844 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
845 disk size and mode can be changed upon recreation (via ``gnt-instance
846 recreate-disks``, which accepts the same ``--disk`` option).
848 As many people are used to a ``show`` command, we have added that as an
849 alias to ``info`` on all ``gnt-*`` commands.
851 The ``gnt-instance grow-disk`` command has a new mode in which it can
852 accept the target size of the disk, instead of the delta; this can be
853 more safe since two runs in absolute mode will be idempotent, and
854 sometimes it's also easier to specify the desired size directly.
856 Also the handling of instances with regard to offline secondaries has
857 been improved. Instance operations should not fail because one of it's
858 secondary nodes is offline, even though it's safe to proceed.
860 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
861 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
862 minors to instance name.
867 RAPI coverage has improved, with (for example) new resources for
868 recreate-disks, node power-cycle, etc.
873 There is partial support for ``xl`` in the Xen hypervisor; feedback is
876 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
877 whether to still support Python 2.4 or move to Python 2.6 as minimum
880 Support for Fedora has been slightly improved; the provided example
881 init.d script should work better on it and the INSTALL file should
882 document the needed dependencies.
887 The deprecated ``QueryLocks`` LUXI request has been removed. Use
888 ``Query(what=QR_LOCK, ...)`` instead.
890 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
891 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
892 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
893 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
894 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
896 RAPI client: ``CertificateError`` now derives from
897 ``GanetiApiError``. This should make it more easy to handle Ganeti
900 Deprecation warnings due to PyCrypto/paramiko import in
901 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
902 make sure to run an up-to-date paramiko version, if you use this tool.
904 The QA scripts now depend on Python 2.5 or above (the main code base
905 still works with Python 2.4).
907 The configuration file (``config.data``) is now written without
908 indentation for performance reasons; if you want to edit it, it can be
909 re-formatted via ``tools/fmtjson``.
911 A number of bugs has been fixed in the cluster merge tool.
913 ``x509`` certification verification (used in import-export) has been
914 changed to allow the same clock skew as permitted by the cluster
915 verification. This will remove some rare but hard to diagnose errors in
922 *(Released Thu, 19 Jul 2012)*
924 Very few changes from rc4 to the final release, only bugfixes:
926 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
927 instance, fix CDROM booting for KVM instances)
928 - fixed node group modification of node parameters
929 - fixed issue in LUClusterVerifyGroup with multi-group clusters
930 - fixed generation of bash completion to ensure a stable ordering
937 *(Released Fri, 13 Jul 2012)*
939 Third release candidate for 2.6. The following changes were done from
942 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
943 - Fixed an inconsistency in the LUXI protocol with the provided
944 arguments (NOT backwards compatible)
945 - Fixed a bug with node groups ipolicy where ``min`` was greater than
946 the cluster ``std`` value
947 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
948 easier instance debugging on nodes (requires ``hconfd`` to work)
954 *(Released Tue, 03 Jul 2012)*
956 Second release candidate for 2.6. The following changes were done from
959 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
961 - Fixed a RAPI regression on missing beparams/memory
962 - Fixed redistribution of files on offline nodes
963 - Added possibility to run activate-disks even though secondaries are
964 offline. With this change it relaxes also the strictness on some other
965 commands which use activate disks internally:
966 * ``gnt-instance start|reboot|rename|backup|export``
967 - Made it possible to remove safely an instance if its secondaries are
969 - Made it possible to reinstall even though secondaries are offline
975 *(Released Mon, 25 Jun 2012)*
977 First release candidate for 2.6. The following changes were done from
980 - Fixed bugs with disk parameters and ``rbd`` templates as well as
982 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
983 behaviour. It supports now the modify operation
984 - ``hcheck`` implemented to analyze cluster health and possibility of
985 improving health by rebalance
986 - ``hbal`` has been improved in dealing with split instances
992 *(Released Mon, 11 Jun 2012)*
994 Second beta release of 2.6. The following changes were done from beta2
997 - Fixed ``daemon-util`` with non-root user models
998 - Fixed creation of plain instances with ``--no-wait-for-sync``
999 - Fix wrong iv_names when running ``cfgupgrade``
1000 - Export more information in RAPI group queries
1001 - Fixed bug when changing instance network interfaces
1002 - Extended burnin to do NIC changes
1003 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1004 - Changed default for DRBD barriers
1005 - Fixed DRBD error reporting for syncer rate
1006 - Verify the options on disk parameters
1008 And of course various fixes to documentation and improved unittests and
1015 *(Released Wed, 23 May 2012)*
1017 First beta release of 2.6. The following changes were done from beta1 to
1020 - integrated patch for distributions without ``start-stop-daemon``
1021 - adapted example init.d script to work on Fedora
1022 - fixed log handling in Haskell daemons
1023 - adapted checks in the watcher for pycurl linked against libnss
1024 - add partial support for ``xl`` instead of ``xm`` for Xen
1025 - fixed a type issue in cluster verification
1026 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1029 Plus integrated fixes from the 2.5 branch:
1031 - fixed ``kvm-ifup`` to use ``/bin/bash``
1032 - fixed parallel build failures
1033 - KVM live migration when using a custom keymap
1039 *(Released Tue, 24 Jul 2012)*
1041 A small bugfix release, with no new features:
1043 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1044 different default shell (e.g. Debian, Ubuntu)
1045 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1046 243 and Debian bug #650664)
1047 - fixed compatibility with KVM versions that don't support multiple boot
1048 devices (fixes Issue 230 and Debian bug #624256)
1050 Additionally, a few fixes were done to the build system (fixed parallel
1051 build failures) and to the unittests (fixed race condition in test for
1052 FileID functions, and the default enable/disable mode for QA test is now
1059 *(Released Fri, 11 May 2012)*
1061 A small bugfix release.
1063 The main issues solved are on the topic of compatibility with newer LVM
1066 - fixed parsing of ``lv_attr`` field
1067 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1068 the ``--force`` flag is needed
1070 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1071 to accept kernel 3.x too (was hardcoded to 2.6.*).
1073 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1074 of hook results and that also made display of other errors suboptimal
1075 was fixed; the code behaves now like 2.4 and earlier.
1077 Another change in 2.5, the cleanup of the OS scripts environment, is too
1078 aggressive: it removed even the ``PATH`` variable, which requires the OS
1079 scripts to *always* need to export it. Since this is a bit too strict,
1080 we now export a minimal PATH, the same that we export for hooks.
1082 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1083 integrated into this release.
1085 Finally, a few other miscellaneous changes were done (no new features,
1086 just small improvements):
1088 - Fix ``gnt-group --help`` display
1089 - Fix hardcoded Xen kernel path
1090 - Fix grow-disk handling of invalid units
1091 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1092 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1098 *(Released Thu, 12 Apr 2012)*
1100 Incompatible/important changes and bugfixes
1101 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1103 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1104 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1105 to match the underlying LUXI interface.
1106 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1107 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1108 not maintain backwards-compatibility as the underlying operation
1109 changed in an incompatible way. The RAPI client can talk to old
1110 servers, but it needs to be told so as the return value changed.
1111 - When creating file-based instances via RAPI, the ``file_driver``
1112 parameter no longer defaults to ``loop`` and must be specified.
1113 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1115 - Support for the undocumented and deprecated RAPI instance creation
1116 request format version 0 has been dropped. Use version 1, supported
1117 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1118 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1119 documentation <install>`.
1120 - The "cluster-verify" hooks are now executed per group by the
1121 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1122 if you just run ``gnt-cluster verify``, which generates one opcode per
1124 - The environment as passed to the OS scripts is cleared, and thus no
1125 environment variables defined in the node daemon's environment will be
1126 inherited by the scripts.
1127 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1129 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1130 support operations involving multiple node groups.
1131 - Offline nodes are ignored when failing over an instance.
1132 - Support for KVM version 1.0, which changed the version reporting format
1134 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1136 - ``Makefile`` is now compatible with Automake 1.11.2
1137 - Includes all bugfixes made in the 2.4 series
1142 - The ganeti-htools project has been merged into the ganeti-core source
1143 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1144 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1145 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1146 GPT-style partition tables (requires `parted
1147 <http://www.gnu.org/s/parted/>`_).
1148 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1149 - Allowed adding tags on instance creation.
1150 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1151 - Allow instances to be started in a paused state, enabling the user to
1152 see the complete console output on boot using the console.
1153 - Added new hypervisor flag to control default reboot behaviour
1154 (``reboot_behavior``).
1155 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1156 - Improved out-of-band management support:
1158 - Added ``gnt-node health`` command reporting the health status of
1160 - Added ``gnt-node power`` command to manage power status of nodes.
1161 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1163 - Instance migration can fall back to failover if instance is not
1165 - Filters can be used when listing nodes, instances, groups and locks;
1166 see :manpage:`ganeti(7)` manpage.
1167 - Added post-execution status as variables to :doc:`hooks <hooks>`
1169 - Instance tags are exported/imported together with the instance.
1170 - When given an explicit job ID, ``gnt-job info`` will work for archived
1172 - Jobs can define dependencies on other jobs (not yet supported via
1173 RAPI or command line, but used by internal commands and usable via
1176 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1179 - Instance failover is now available as a RAPI resource
1180 (``/2/instances/[instance_name]/failover``).
1181 - ``gnt-instance info`` defaults to static information if primary node
1183 - Opcodes have a new ``comment`` attribute.
1184 - Added basic SPICE support to KVM hypervisor.
1185 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1187 Node group improvements
1188 ~~~~~~~~~~~~~~~~~~~~~~~
1190 - ``gnt-cluster verify`` has been modified to check groups separately,
1191 thereby improving performance.
1192 - Node group support has been added to ``gnt-cluster verify-disks``,
1193 which now operates per node group.
1194 - Watcher has been changed to work better with node groups.
1196 - One process and state file per node group.
1197 - Slow watcher in one group doesn't block other group's watcher.
1199 - Added new command, ``gnt-group evacuate``, to move all instances in a
1200 node group to other groups.
1201 - Added ``gnt-instance change-group`` to move an instance to another
1203 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1204 per-group operations.
1205 - Node groups can be tagged.
1206 - Some operations switch from an exclusive to a shared lock as soon as
1208 - Instance's primary and secondary nodes' groups are now available as
1209 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1210 and ``snodes.group.uuid``).
1215 - Numerous updates to documentation and manpages.
1217 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1219 - Some opcode/job results are now also documented, see :doc:`RAPI
1222 - A lockset's internal lock is now also visible in lock monitor.
1223 - Log messages from job queue workers now contain information about the
1224 opcode they're processing.
1225 - ``gnt-instance console`` no longer requires the instance lock.
1226 - A short delay when waiting for job changes reduces the number of LUXI
1227 requests significantly.
1228 - DRBD metadata volumes are overwritten with zeros during disk creation.
1229 - Out-of-band commands no longer acquire the cluster lock in exclusive
1231 - ``devel/upload`` now uses correct permissions for directories.
1237 *(Released Fri, 23 Mar 2012)*
1239 This was the sixth release candidate of the 2.5 series.
1245 *(Released Mon, 9 Jan 2012)*
1247 This was the fifth release candidate of the 2.5 series.
1253 *(Released Thu, 27 Oct 2011)*
1255 This was the fourth release candidate of the 2.5 series.
1261 *(Released Wed, 26 Oct 2011)*
1263 This was the third release candidate of the 2.5 series.
1269 *(Released Tue, 18 Oct 2011)*
1271 This was the second release candidate of the 2.5 series.
1277 *(Released Tue, 4 Oct 2011)*
1279 This was the first release candidate of the 2.5 series.
1285 *(Released Wed, 31 Aug 2011)*
1287 This was the third beta release of the 2.5 series.
1293 *(Released Mon, 22 Aug 2011)*
1295 This was the second beta release of the 2.5 series.
1301 *(Released Fri, 12 Aug 2011)*
1303 This was the first beta release of the 2.5 series.
1309 *(Released Thu, 27 Oct 2011)*
1311 - Fixed bug when parsing command line parameter values ending in
1313 - Fixed assertion error after unclean master shutdown
1314 - Disable HTTP client pool for RPC, significantly reducing memory usage
1316 - Fixed queue archive creation with wrong permissions
1322 *(Released Tue, 23 Aug 2011)*
1326 - Fixed documentation for importing with ``--src-dir`` option
1327 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1328 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1334 *(Released Fri, 5 Aug 2011)*
1336 Many bug-fixes and a few small features:
1338 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1339 issues when you tried to add an instance with two MAC addresses in one
1341 - KVM: fixed per-instance stored UID value
1342 - KVM: configure bridged NICs at migration start
1343 - KVM: Fix a bug where instance will not start with never KVM versions
1345 - Added OS search path to ``gnt-cluster info``
1346 - Fixed an issue with ``file_storage_dir`` where you were forced to
1347 provide an absolute path, but the documentation states it is a
1348 relative path, the documentation was right
1349 - Added a new parameter to instance stop/start called ``--no-remember``
1350 that will make the state change to not be remembered
1351 - Implemented ``no_remember`` at RAPI level
1352 - Improved the documentation
1353 - Node evacuation: don't call IAllocator if node is already empty
1354 - Fixed bug in DRBD8 replace disks on current nodes
1355 - Fixed bug in recreate-disks for DRBD instances
1356 - Moved assertion checking locks in ``gnt-instance replace-disks``
1357 causing it to abort with not owning the right locks for some situation
1358 - Job queue: Fixed potential race condition when cancelling queued jobs
1359 - Fixed off-by-one bug in job serial generation
1360 - ``gnt-node volumes``: Fix instance names
1361 - Fixed aliases in bash completion
1362 - Fixed a bug in reopening log files after being sent a SIGHUP
1363 - Added a flag to burnin to allow specifying VCPU count
1364 - Bugfixes to non-root Ganeti configuration
1370 *(Released Thu, 12 May 2011)*
1372 Many bug-fixes and a few new small features:
1374 - Fixed a bug related to log opening failures
1375 - Fixed a bug in instance listing with orphan instances
1376 - Fixed a bug which prevented resetting the cluster-level node parameter
1377 ``oob_program`` to the default
1378 - Many fixes related to the ``cluster-merge`` tool
1379 - Fixed a race condition in the lock monitor, which caused failures
1380 during (at least) creation of many instances in parallel
1381 - Improved output for gnt-job info
1382 - Removed the quiet flag on some ssh calls which prevented debugging
1384 - Improved the N+1 failure messages in cluster verify by actually
1385 showing the memory values (needed and available)
1386 - Increased lock attempt timeouts so that when executing long operations
1387 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1388 too early and thus prevent the use of the 'fair' mechanism
1389 - Changed instance query data (``gnt-instance info``) to not acquire
1390 locks unless needed, thus allowing its use on locked instance if only
1391 static information is asked for
1392 - Improved behaviour with filesystems that do not support rename on an
1394 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1395 kept locks on all nodes during the wipe, which is unneeded
1396 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1397 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1398 10GiB) which caused the wipe to fail right at the end in some cases
1399 - Fixed master IP activation when doing master failover with no-voting
1400 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1401 the master node itself
1402 - Fixed potential data-loss in under disk full conditions, where Ganeti
1403 wouldn't check correctly the return code and would consider
1404 partially-written files 'correct'
1405 - Fixed bug related to multiple VGs and DRBD disk replacing
1406 - Added new disk parameter ``metavg`` that allows placement of the meta
1407 device for DRBD in a different volume group
1408 - Fixed error handling in the node daemon when the system libc doesn't
1409 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1410 - Fixed lock release during replace-disks, which kept cluster-wide locks
1411 when doing disk replaces with an iallocator script
1412 - Added check for missing bridges in cluster verify
1413 - Handle EPIPE errors while writing to the terminal better, so that
1414 piping the output to e.g. ``less`` doesn't cause a backtrace
1415 - Fixed rare case where a ^C during Luxi calls could have been
1416 interpreted as server errors, instead of simply terminating
1417 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1419 - Added a few more parameters to the KVM hypervisor, allowing a second
1420 CDROM, custom disk type for CDROMs and a floppy image
1421 - Removed redundant message in instance rename when the name is given
1423 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1424 disks on new nodes, allowing recreation when the original instance
1425 nodes are completely gone
1426 - Added option when converting disk templates to DRBD to skip waiting
1427 for the resync, in order to make the instance available sooner
1428 - Added two new variables to the OS scripts environment (containing the
1430 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1431 to allow use of ``pvgrub`` as bootloader
1432 - Changed the instance memory modifications to only check out-of-memory
1433 conditions on memory increases, and turned the secondary node warnings
1434 into errors (they can still be overridden via ``--force``)
1435 - Fixed the handling of a corner case when the Python installation gets
1436 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1437 to execute a command that doesn't exist
1438 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1439 node of the instance returned failures during instance shutdown; this
1440 adds the option ``--ignore-consistency`` to gnt-instance move
1442 And as usual, various improvements to the error messages, documentation
1449 *(Released Wed, 09 Mar 2011)*
1451 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1452 the RAPI users file if run twice (even with ``--dry-run``).
1454 The release fixes that bug (nothing else changed).
1460 *(Released Mon, 07 Mar 2011)*
1462 Final 2.4.0 release. Just a few small fixes:
1464 - Fixed RAPI node evacuate
1465 - Fixed the kvm-ifup script
1466 - Fixed internal error handling for special job cases
1467 - Updated man page to specify the escaping feature for options
1473 *(Released Mon, 28 Feb 2011)*
1475 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1476 that this feature wiped the disks of the wrong instance, leading to loss
1481 - Fixed title of query field containing instance name
1482 - Expanded the glossary in the documentation
1483 - Fixed one unittest (internal issue)
1489 *(Released Mon, 21 Feb 2011)*
1491 A number of bug fixes plus just a couple functionality changes.
1493 On the user-visible side, the ``gnt-* list`` command output has changed
1494 with respect to "special" field states. The current rc1 style of display
1495 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1496 the default output mode special fields states are displayed as follows:
1498 - Offline resource: ``*``
1499 - Unavailable/not applicable: ``-``
1500 - Data missing (RPC failure): ``?``
1501 - Unknown field: ``??``
1503 Another user-visible change is the addition of ``--force-join`` to
1508 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1509 - Fixed regression in RAPI/instance reinstall where all parameters were
1510 required (instead of optional)
1511 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1512 - Fixed iallocator usage (offline nodes were not considered offline)
1513 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1514 - Fixed hypervisor and OS parameter validation with respect to
1515 non-vm_capable nodes
1516 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1518 - Fixed ``tools/listrunner`` with respect to agent-based usage
1524 *(Released Fri, 4 Feb 2011)*
1526 Many changes and fixes since the beta1 release. While there were some
1527 internal changes, the code has been mostly stabilised for the RC
1530 Note: the dumb allocator was removed in this release, as it was not kept
1531 up-to-date with the IAllocator protocol changes. It is recommended to
1532 use the ``hail`` command from the ganeti-htools package.
1534 Note: the 2.4 and up versions of Ganeti are not compatible with the
1535 0.2.x branch of ganeti-htools. You need to upgrade to
1536 ganeti-htools-0.3.0 (or later).
1538 Regressions fixed from 2.3
1539 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1541 - Fixed the ``gnt-cluster verify-disks`` command
1542 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1544 - Fixed disk adoption breakage
1545 - Fixed wrong headers in instance listing for field aliases
1550 - Fixed corner case in KVM handling of NICs
1551 - Fixed many cases of wrong handling of non-vm_capable nodes
1552 - Fixed a bug where a missing instance symlink was not possible to
1553 recreate with any ``gnt-*`` command (now ``gnt-instance
1554 activate-disks`` does it)
1555 - Fixed the volume group name as reported by ``gnt-cluster
1557 - Increased timeouts for the import-export code, hopefully leading to
1558 fewer aborts due network or instance timeouts
1559 - Fixed bug in ``gnt-node list-storage``
1560 - Fixed bug where not all daemons were started on cluster
1561 initialisation, but only at the first watcher run
1562 - Fixed many bugs in the OOB implementation
1563 - Fixed watcher behaviour in presence of instances with offline
1565 - Fixed instance list output for instances running on the wrong node
1566 - a few fixes to the cluster-merge tool, but it still cannot merge
1567 multi-node groups (currently it is not recommended to use this tool)
1573 - Improved network configuration for the KVM hypervisor
1574 - Added e1000 as a supported NIC for Xen-HVM
1575 - Improved the lvmstrap tool to also be able to use partitions, as
1576 opposed to full disks
1577 - Improved speed of disk wiping (the cluster parameter
1578 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1579 of instance creations
1580 - Added documentation for the OS parameters
1581 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1582 hypervisor is not responding
1583 - Added display of blacklisted and hidden OS information in
1584 ``gnt-cluster info``
1585 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1586 NIC and node parameters, which might create problems with currently
1587 invalid (but undetected) configuration files, but prevents validation
1588 failures when unrelated parameters are modified
1589 - Changed cluster initialisation to wait for the master daemon to become
1591 - Expanded the RAPI interface:
1593 - Added config redistribution resource
1594 - Added activation/deactivation of instance disks
1595 - Added export of console information
1597 - Implemented log file reopening on SIGHUP, which allows using
1598 logrotate(8) for the Ganeti log files
1599 - Added a basic OOB helper script as an example
1605 *(Released Fri, 14 Jan 2011)*
1610 - Fixed timezone issues when formatting timestamps
1611 - Added support for node groups, available via ``gnt-group`` and other
1613 - Added out-of-band framework and management, see :doc:`design
1614 document <design-oob>`
1615 - Removed support for roman numbers from ``gnt-node list`` and
1616 ``gnt-instance list``.
1617 - Allowed modification of master network interface via ``gnt-cluster
1618 modify --master-netdev``
1619 - Accept offline secondaries while shutting down instance disks
1620 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1621 - Added support for multiple LVM volume groups
1622 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1624 - Added commands to list available fields:
1626 - ``gnt-node list-fields``
1627 - ``gnt-group list-fields``
1628 - ``gnt-instance list-fields``
1630 - Updated documentation and man pages
1635 - Moved ``rapi_users`` file into separate directory, now named
1636 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1638 - Added new tool for running commands on many machines,
1639 ``tools/ganeti-listrunner``
1640 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1641 improving the ``gnt-instance console`` output
1642 - Allowed customisation of disk index separator at ``configure`` time
1643 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1644 - Added support for non-partitioned md disks in ``lvmstrap``
1645 - Added script to gracefully power off KVM instances
1646 - Split ``utils`` module into smaller parts
1647 - Changed query operations to return more detailed information, e.g.
1648 whether an information is unavailable due to an offline node. To use
1649 this new functionality, the LUXI call ``Query`` must be used. Field
1650 information is now stored by the master daemon and can be retrieved
1651 using ``QueryFields``. Instances, nodes and groups can also be queried
1652 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1653 exposed via RAPI). The following commands make use of this
1654 infrastructure change:
1656 - ``gnt-group list``
1657 - ``gnt-group list-fields``
1659 - ``gnt-node list-fields``
1660 - ``gnt-instance list``
1661 - ``gnt-instance list-fields``
1662 - ``gnt-debug locks``
1667 - New RAPI resources (see :doc:`rapi`):
1671 - ``/2/groups/[group_name]``
1672 - ``/2/groups/[group_name]/assign-nodes``
1673 - ``/2/groups/[group_name]/modify``
1674 - ``/2/groups/[group_name]/rename``
1675 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1679 - Implemented ``no_install`` for instance creation
1680 - Implemented OS parameters for instance reinstallation, allowing
1681 use of special settings on reinstallation (e.g. for preserving data)
1686 - Added IPv6 support in import/export
1687 - Pause DRBD synchronization while wiping disks on instance creation
1688 - Updated unittests and QA scripts
1689 - Improved network parameters passed to KVM
1690 - Converted man pages from docbook to reStructuredText
1696 *(Released Mon, 20 Dec 2010)*
1698 Released version 2.3.1~rc1 without any changes.
1704 *(Released Wed, 1 Dec 2010)*
1706 - impexpd: Disable OpenSSL compression in socat if possible (backport
1707 from master, commit e90739d625b, see :doc:`installation guide
1708 <install-quick>` for details)
1709 - Changed unittest coverage report to exclude test scripts
1710 - Added script to check version format
1716 *(Released Wed, 1 Dec 2010)*
1718 Released version 2.3.0~rc1 without any changes.
1724 *(Released Fri, 19 Nov 2010)*
1726 A number of bugfixes and documentation updates:
1728 - Update ganeti-os-interface documentation
1729 - Fixed a bug related to duplicate MACs or similar items which should be
1731 - Fix breakage in OS state modify
1732 - Reinstall instance: disallow offline secondaries (fixes bug related to
1733 OS changing but reinstall failing)
1734 - plus all the other fixes between 2.2.1 and 2.2.2
1740 *(Released Tue, 2 Nov 2010)*
1742 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1743 - Fixed master failover race with watcher
1744 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1746 - Accept previously stopped instance for export with instance removal
1747 - Simplify and extend the environment variables for instance OS scripts
1748 - Added new node flags, ``master_capable`` and ``vm_capable``
1749 - Added optional instance disk wiping prior during allocation. This is a
1750 cluster-wide option and can be set/modified using
1751 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1752 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1753 :doc:`install-quick`
1754 - Added a new watcher option (``--ignore-pause``)
1755 - Added option to ignore offline node on instance start/stop
1756 (``--ignore-offline``)
1757 - Allow overriding OS parameters with ``gnt-instance reinstall``
1758 - Added ability to change node's secondary IP address using ``gnt-node
1760 - Implemented privilege separation for all daemons except
1761 ``ganeti-noded``, see ``configure`` options
1762 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1764 - Implemented job priorities (see ``ganeti(7)`` manpage)
1765 - Ignore failures while shutting down instances during failover from
1767 - Exit daemon's bootstrap process only once daemon is ready
1768 - Export more information via ``LUInstanceQuery``/remote API
1769 - Improved documentation, QA and unittests
1770 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1771 restart if the file was created or changed
1772 - Added LUXI protocol version sent with each request and response,
1773 allowing detection of server/client mismatches
1774 - Moved the Python scripts among gnt-* and ganeti-* into modules
1775 - Moved all code related to setting up SSH to an external script,
1777 - Infrastructure changes for node group support in future versions
1783 *(Released Fri, 19 Nov 2010)*
1785 A few small bugs fixed, and some improvements to the build system:
1787 - Fix documentation regarding conversion to drbd
1788 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1790 - Fix error handling in node modify with multiple changes
1791 - Allow remote imports without checked names
1797 *(Released Tue, 19 Oct 2010)*
1799 - Disable SSL session ID cache in RPC client
1805 *(Released Thu, 14 Oct 2010)*
1807 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1808 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1810 - Fix problems with interaction between Curl and Python's HTTP server,
1811 resulting in increased speed in many RPC calls
1812 - Improve our release script to prevent breakage with older aclocal and
1819 *(Released Thu, 7 Oct 2010)*
1821 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1822 value retrieved from master
1823 - Added support for blacklisted or hidden OS definitions
1824 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1825 - Added support for -mem-path in KVM hypervisor abstraction layer
1826 - Allow overriding instance parameters in tool for inter-cluster
1827 instance moves (``tools/move-instance``)
1828 - Improved opcode summaries (e.g. in ``gnt-job list``)
1829 - Improve consistency of OS listing by sorting it
1830 - Documentation updates
1836 *(Released Fri, 8 Oct 2010)*
1838 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1844 *(Released Mon, 4 Oct 2010)*
1846 - Fixed regression in ``gnt-instance rename``
1852 *(Released Wed, 22 Sep 2010)*
1854 - Fixed OS_VARIANT variable for OS scripts
1855 - Fixed cluster tag operations via RAPI
1856 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1857 - Disabled RAPI CA checks in watcher
1863 *(Released Mon, 23 Aug 2010)*
1865 - Support DRBD versions of the format "a.b.c.d"
1867 - Re-introduce support for usage from multiple threads in RAPI client
1868 - Instance renames and modify via RAPI
1869 - Work around race condition between processing and archival in job
1871 - Mark opcodes following failed one as failed, too
1872 - Job field ``lock_status`` was removed due to difficulties making it
1873 work with the changed job queue in Ganeti 2.2; a better way to monitor
1874 locks is expected for a later 2.2.x release
1875 - Fixed dry-run behaviour with many commands
1876 - Support ``ssh-agent`` again when adding nodes
1877 - Many additional bugfixes
1883 *(Released Fri, 30 Jul 2010)*
1885 Important change: the internal RPC mechanism between Ganeti nodes has
1886 changed from using a home-grown http library (based on the Python base
1887 libraries) to use the PycURL library. This requires that PycURL is
1888 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1889 against GnuTLS by default. cURL's support for GnuTLS had known issues
1890 before cURL 7.21.0 and we recommend using the latest cURL release or
1891 linking against OpenSSL. Most other distributions already link PycURL
1892 and cURL against OpenSSL. The command::
1894 python -c 'import pycurl; print pycurl.version'
1896 can be used to determine the libraries PycURL and cURL are linked
1899 Other significant changes:
1901 - Rewrote much of the internals of the job queue, in order to achieve
1902 better parallelism; this decouples job query operations from the job
1903 processing, and it should allow much nicer behaviour of the master
1904 daemon under load, and it also has uncovered some long-standing bugs
1905 related to the job serialisation (now fixed)
1906 - Added a default iallocator setting to the cluster parameters,
1907 eliminating the need to always pass nodes or an iallocator for
1908 operations that require selection of new node(s)
1909 - Added experimental support for the LXC virtualization method
1910 - Added support for OS parameters, which allows the installation of
1911 instances to pass parameter to OS scripts in order to customise the
1913 - Added a hypervisor parameter controlling the migration type (live or
1914 non-live), since hypervisors have various levels of reliability; this
1915 has renamed the 'live' parameter to 'mode'
1916 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1917 logical volumes, meaning that cluster verify will ignore them and not
1918 flag their presence as errors
1919 - The watcher will now reset the error count for failed instances after
1920 8 hours, thus allowing self-healing if the problem that caused the
1921 instances to be down/fail to start has cleared in the meantime
1922 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1923 check for, and warn, if the drbd module parameter ``usermode_helper``
1924 is not consistent with the cluster-wide setting; this is needed to
1925 make diagnose easier of failed drbd creations
1926 - Started adding base IPv6 support, but this is not yet
1927 enabled/available for use
1928 - Rename operations (cluster, instance) will now return the new name,
1929 which is especially useful if a short name was passed in
1930 - Added support for instance migration in RAPI
1931 - Added a tool to pre-configure nodes for the SSH setup, before joining
1932 them to the cluster; this will allow in the future a simplified model
1933 for node joining (but not yet fully enabled in 2.2); this needs the
1934 paramiko python library
1935 - Fixed handling of name-resolving errors
1936 - Fixed consistency of job results on the error path
1937 - Fixed master-failover race condition when executed multiple times in
1939 - Fixed many bugs related to the job queue (mostly introduced during the
1940 2.2 development cycle, so not all are impacting 2.1)
1941 - Fixed instance migration with missing disk symlinks
1942 - Fixed handling of unknown jobs in ``gnt-job archive``
1943 - And many other small fixes/improvements
1947 - Enhanced both the unittest and the QA coverage
1948 - Switched the opcode validation to a generic model, and extended the
1949 validation to all opcode parameters
1950 - Changed more parts of the code that write shell scripts to use the
1952 - Switched the master daemon to use the asyncore library for the Luxi
1959 *(Released Thu, 17 Jun 2010)*
1961 - Added tool (``move-instance``) and infrastructure to move instances
1962 between separate clusters (see :doc:`separate documentation
1963 <move-instance>` and :doc:`design document <design-2.2>`)
1964 - Added per-request RPC timeout
1965 - RAPI now requires a Content-Type header for requests with a body (e.g.
1966 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1967 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1968 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1970 - Implemented initial support for running Ganeti daemons as separate
1971 users, see configure-time flags ``--with-user-prefix`` and
1972 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1974 - Instances can be removed after export (``gnt-backup export
1975 --remove-instance``)
1976 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1977 key (instead of 1024 bit)
1978 - Added new cluster configuration file for cluster domain secret
1979 - Import/export now use SSL instead of SSH
1980 - Added support for showing estimated time when exporting an instance,
1981 see the ``ganeti-os-interface(7)`` manpage and look for
1988 *(Released Tue, 16 Nov 2010)*
1990 Some more bugfixes. Unless critical bugs occur, this will be the last
1993 - Fix case of MAC special-values
1994 - Fix mac checker regex
1995 - backend: Fix typo causing "out of range" error
1996 - Add missing --units in gnt-instance list man page
2002 *(Released Tue, 24 Aug 2010)*
2005 - Don't ignore secondary node silently on non-mirrored disk templates
2007 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2008 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2009 - Properly document the usb_mouse parameter
2010 - Fix path in ganeti-rapi(8) (issue 116)
2011 - Adjust error message when the ganeti user's .ssh directory is
2013 - Add same-node-check when changing the disk template to drbd
2019 *(Released Fri, 16 Jul 2010)*
2022 - Add an option to only select some reboot types during qa/burnin.
2023 (on some hypervisors consequent reboots are not supported)
2024 - Fix infrequent race condition in master failover. Sometimes the old
2025 master ip address would be still detected as up for a short time
2026 after it was removed, causing failover to fail.
2027 - Decrease mlockall warnings when the ctypes module is missing. On
2028 Python 2.4 we support running even if no ctypes module is installed,
2029 but we were too verbose about this issue.
2030 - Fix building on old distributions, on which man doesn't have a
2032 - Fix RAPI not to ignore the MAC address on instance creation
2033 - Implement the old instance creation format in the RAPI client.
2039 *(Released Thu, 01 Jul 2010)*
2041 A small bugfix release:
2042 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2043 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2044 a non-existing option
2045 - Fix instance create: the --force-variant option was ignored
2046 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2047 - Fix modify node storage with non-FQDN arguments
2048 - Fix RAPI client to authenticate under Python 2.6 when used
2049 for more than 5 requests needing authentication
2050 - Fix gnt-instance modify -t (storage) giving a wrong error message
2051 when converting a non-shutdown drbd instance to plain
2057 *(Released Fri, 18 Jun 2010)*
2059 A small bugfix release:
2061 - Fix live migration of KVM instances started with older Ganeti
2062 versions which had fewer hypervisor parameters
2063 - Fix gnt-instance grow-disk on down instances
2064 - Fix an error-reporting bug during instance migration
2065 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2066 silently ignoring broken ones
2067 - Fix an RPC error reporting bug affecting, for example, RAPI client
2069 - Fix bug triggered by different API version os-es on different nodes
2070 - Fix a bug in instance startup with custom hvparams: OS level
2071 parameters would fail to be applied.
2072 - Fix the RAPI client under Python 2.6 (but more work is needed to
2073 make it work completely well with OpenSSL)
2074 - Fix handling of errors when resolving names from DNS
2080 *(Released Thu, 3 Jun 2010)*
2082 A medium sized development cycle. Some new features, and some
2083 fixes/small improvements/cleanups.
2085 Significant features
2086 ~~~~~~~~~~~~~~~~~~~~
2088 The node deamon now tries to mlock itself into memory, unless the
2089 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2090 its logs, and falls back to console logging. This allows emergency
2091 features such as ``gnt-node powercycle`` to work even in the event of a
2092 broken node disk (tested offlining the disk hosting the node's
2093 filesystem and dropping its memory caches; don't try this at home)
2095 KVM: add vhost-net acceleration support. It can be tested with a new
2096 enough version of the kernel and of qemu-kvm.
2098 KVM: Add instance chrooting feature. If you use privilege dropping for
2099 your VMs you can also now force them to chroot to an empty directory,
2100 before starting the emulated guest.
2102 KVM: Add maximum migration bandwith and maximum downtime tweaking
2103 support (requires a new-enough version of qemu-kvm).
2105 Cluster verify will now warn if the master node doesn't have the master
2106 ip configured on it.
2108 Add a new (incompatible) instance creation request format to RAPI which
2109 supports all parameters (previously only a subset was supported, and it
2110 wasn't possible to extend the old format to accomodate all the new
2111 features. The old format is still supported, and a client can check for
2112 this feature, before using it, by checking for its presence in the
2113 ``features`` RAPI resource.
2115 Now with ancient latin support. Try it passing the ``--roman`` option to
2116 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2117 (requires the python-roman module to be installed, in order to work).
2122 As usual many internal code refactorings, documentation updates, and
2125 - Lots of improvements and cleanups to the experimental Remote API
2126 (RAPI) client library.
2127 - A new unit test suite for the core daemon libraries.
2128 - A fix to creating missing directories makes sure the umask is not
2129 applied anymore. This enforces the same directory permissions
2131 - Better handling terminating daemons with ctrl+c (used when running
2132 them in debugging mode).
2133 - Fix a race condition in live migrating a KVM instance, when stat()
2134 on the old proc status file returned EINVAL, which is an unexpected
2136 - Fixed manpage checking with newer man and utf-8 charachters. But now
2137 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2143 *(Released Fri, 7 May 2010)*
2145 Fix a bug which prevented untagged KVM instances from starting.
2151 *(Released Fri, 7 May 2010)*
2153 Another release with a long development cycle, during which many
2154 different features were added.
2156 Significant features
2157 ~~~~~~~~~~~~~~~~~~~~
2159 The KVM hypervisor now can run the individual instances as non-root, to
2160 reduce the impact of a VM being hijacked due to bugs in the
2161 hypervisor. It is possible to run all instances as a single (non-root)
2162 user, to manually specify a user for each instance, or to dynamically
2163 allocate a user out of a cluster-wide pool to each instance, with the
2164 guarantee that no two instances will run under the same user ID on any
2167 An experimental RAPI client library, that can be used standalone
2168 (without the other Ganeti libraries), is provided in the source tree as
2169 ``lib/rapi/client.py``. Note this client might change its interface in
2170 the future, as we iterate on its capabilities.
2172 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2173 replace the cluster's certificates and crypto keys. This might help in
2174 case they have been compromised, or have simply expired.
2176 A new disk option for instance creation has been added that allows one
2177 to "adopt" currently existing logical volumes, with data
2178 preservation. This should allow easier migration to Ganeti from
2179 unmanaged (or managed via other software) instances.
2181 Another disk improvement is the possibility to convert between redundant
2182 (DRBD) and plain (LVM) disk configuration for an instance. This should
2183 allow better scalability (starting with one node and growing the
2184 cluster, or shrinking a two-node cluster to one node).
2186 A new feature that could help with automated node failovers has been
2187 implemented: if a node sees itself as offline (by querying the master
2188 candidates), it will try to shutdown (hard) all instances and any active
2189 DRBD devices. This reduces the risk of duplicate instances if an
2190 external script automatically failovers the instances on such nodes. To
2191 enable this, the cluster parameter ``maintain_node_health`` should be
2192 enabled; in the future this option (per the name) will enable other
2193 automatic maintenance features.
2195 Instance export/import now will reuse the original instance
2196 specifications for all parameters; that means exporting an instance,
2197 deleting it and the importing it back should give an almost identical
2198 instance. Note that the default import behaviour has changed from
2199 before, where it created only one NIC; now it recreates the original
2202 Cluster verify has added a few new checks: SSL certificates validity,
2203 /etc/hosts consistency across the cluster, etc.
2208 As usual, many internal changes were done, documentation fixes,
2211 - Fixed cluster initialization with disabled cluster storage (regression
2212 introduced in 2.1.1)
2213 - File-based storage supports growing the disks
2214 - Fixed behaviour of node role changes
2215 - Fixed cluster verify for some corner cases, plus a general rewrite of
2216 cluster verify to allow future extension with more checks
2217 - Fixed log spamming by watcher and node daemon (regression introduced
2219 - Fixed possible validation issues when changing the list of enabled
2221 - Fixed cleanup of /etc/hosts during node removal
2222 - Fixed RAPI response for invalid methods
2223 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2224 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2225 from ide disks, etc.)
2226 - Allow OS changes without re-installation (to record a changed OS
2227 outside of Ganeti, or to allow OS renames)
2228 - Allow instance creation without OS installation (useful for example if
2229 the OS will be installed manually, or restored from a backup not in
2231 - Implemented option to make cluster ``copyfile`` use the replication
2233 - Added list of enabled hypervisors to ssconf (possibly useful for
2235 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2237 - A partial form of node re-IP is possible via node readd, which now
2238 allows changed node primary IP
2239 - Command line utilities now show an informational message if the job is
2241 - The logs of the master daemon now show the PID/UID/GID of the
2248 *(Released Fri, 12 Mar 2010)*
2250 During the 2.1.0 long release candidate cycle, a lot of improvements and
2251 changes have accumulated with were released later as 2.1.1.
2256 The node evacuate command (``gnt-node evacuate``) was significantly
2257 rewritten, and as such the IAllocator protocol was changed - a new
2258 request type has been added. This unfortunate change during a stable
2259 series is designed to improve performance of node evacuations; on
2260 clusters with more than about five nodes and which are well-balanced,
2261 evacuation should proceed in parallel for all instances of the node
2262 being evacuated. As such, any existing IAllocator scripts need to be
2263 updated, otherwise the above command will fail due to the unknown
2264 request. The provided "dumb" allocator has not been updated; but the
2265 ganeti-htools package supports the new protocol since version 0.2.4.
2267 Another important change is increased validation of node and instance
2268 names. This might create problems in special cases, if invalid host
2269 names are being used.
2271 Also, a new layer of hypervisor parameters has been added, that sits at
2272 OS level between the cluster defaults and the instance ones. This allows
2273 customisation of virtualization parameters depending on the installed
2274 OS. For example instances with OS 'X' may have a different KVM kernel
2275 (or any other parameter) than the cluster defaults. This is intended to
2276 help managing a multiple OSes on the same cluster, without manual
2277 modification of each instance's parameters.
2279 A tool for merging clusters, ``cluster-merge``, has been added in the
2280 tools sub-directory.
2285 - Improved the int/float conversions that should make the code more
2286 robust in face of errors from the node daemons
2287 - Fixed the remove node code in case of internal configuration errors
2288 - Fixed the node daemon behaviour in face of inconsistent queue
2289 directory (e.g. read-only file-system where we can't open the files
2291 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2292 now it either aborts cleanly or, if given the new "auto_promote"
2293 parameter, will automatically promote other nodes as needed
2294 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2295 completely prevent Ganeti from working
2296 - Fixed bug for instance export when not all disks were successfully
2298 - Fixed behaviour of node add when the new node is slow in starting up
2300 - Fixed handling of signals in the LUXI client, which should improve
2301 behaviour of command-line scripts
2302 - Added checks for invalid node/instance names in the configuration (now
2303 flagged during cluster verify)
2304 - Fixed watcher behaviour for disk activation errors
2305 - Fixed two potentially endless loops in http library, which led to the
2306 RAPI daemon hanging and consuming 100% CPU in some cases
2307 - Fixed bug in RAPI daemon related to hashed passwords
2308 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2310 - Enhanced compatibility with non-Debian OSes, but not using absolute
2311 path in some commands and allowing customisation of the ssh
2312 configuration directory
2313 - Fixed possible future issue with new Python versions by abiding to the
2314 proper use of ``__slots__`` attribute on classes
2315 - Added checks that should prevent directory traversal attacks
2316 - Many documentation fixes based on feedback from users
2321 - Added an "early_release" more for instance replace disks and node
2322 evacuate, where we release locks earlier and thus allow higher
2323 parallelism within the cluster
2324 - Added watcher hooks, intended to allow the watcher to restart other
2325 daemons (e.g. from the ganeti-nbma project), but they can be used of
2326 course for any other purpose
2327 - Added a compile-time disable for DRBD barriers, to increase
2328 performance if the administrator trusts the power supply or the
2329 storage system to not lose writes
2330 - Added the option of using syslog for logging instead of, or in
2331 addition to, Ganeti's own log files
2332 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2333 can indeed boot from a paravirtual NIC
2334 - Added a generic debug level for many operations; while this is not
2335 used widely yet, it allows one to pass the debug value all the way to
2337 - Enhanced the hooks environment for instance moves (failovers,
2338 migrations) where the primary/secondary nodes changed during the
2339 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2340 - Enhanced data validations for many user-supplied values; one important
2341 item is the restrictions imposed on instance and node names, which
2342 might reject some (invalid) host names
2343 - Add a configure-time option to disable file-based storage, if it's not
2344 needed; this allows greater security separation between the master
2345 node and the other nodes from the point of view of the inter-node RPC
2347 - Added user notification in interactive tools if job is waiting in the
2348 job queue or trying to acquire locks
2349 - Added log messages when a job is waiting for locks
2350 - Added filtering by node tags in instance operations which admit
2351 multiple instances (start, stop, reboot, reinstall)
2352 - Added a new tool for cluster mergers, ``cluster-merge``
2353 - Parameters from command line which are of the form ``a=b,c=d`` can now
2354 use backslash escapes to pass in values which contain commas,
2355 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2357 - For KVM, the instance name is the first parameter passed to KVM, so
2358 that it's more visible in the process list
2364 *(Released Tue, 2 Mar 2010)*
2366 Ganeti 2.1 brings many improvements with it. Major changes:
2368 - Added infrastructure to ease automated disk repairs
2369 - Added new daemon to export configuration data in a cheaper way than
2370 using the remote API
2371 - Instance NICs can now be routed instead of being associated with a
2373 - Improved job locking logic to reduce impact of jobs acquiring multiple
2374 locks waiting for other long-running jobs
2376 In-depth implementation details can be found in the Ganeti 2.1 design
2382 - Added chroot hypervisor
2383 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2385 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2386 ``bootloader_path`` and ``bootloader_args``)
2387 - Added the ``use_localtime`` option for the xen-hvm and kvm
2388 hypervisors, and the default value for this has changed to false (in
2389 2.0 xen-hvm always enabled it)
2390 - Added luxi call to submit multiple jobs in one go
2391 - Added cluster initialization option to not modify ``/etc/hosts``
2393 - Added network interface parameters
2394 - Added dry run mode to some LUs
2395 - Added RAPI resources:
2397 - ``/2/instances/[instance_name]/info``
2398 - ``/2/instances/[instance_name]/replace-disks``
2399 - ``/2/nodes/[node_name]/evacuate``
2400 - ``/2/nodes/[node_name]/migrate``
2401 - ``/2/nodes/[node_name]/role``
2402 - ``/2/nodes/[node_name]/storage``
2403 - ``/2/nodes/[node_name]/storage/modify``
2404 - ``/2/nodes/[node_name]/storage/repair``
2406 - Added OpCodes to evacuate or migrate all instances on a node
2407 - Added new command to list storage elements on nodes (``gnt-node
2408 list-storage``) and modify them (``gnt-node modify-storage``)
2409 - Added new ssconf files with master candidate IP address
2410 (``ssconf_master_candidates_ips``), node primary IP address
2411 (``ssconf_node_primary_ips``) and node secondary IP address
2412 (``ssconf_node_secondary_ips``)
2413 - Added ``ganeti-confd`` and a client library to query the Ganeti
2414 configuration via UDP
2415 - Added ability to run hooks after cluster initialization and before
2417 - Added automatic mode for disk replace (``gnt-instance replace-disks
2419 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2420 after catastrophic data-loss
2421 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2423 - Added ``gnt-instance move`` command to move instances
2424 - Added ``gnt-cluster watcher`` command to control watcher
2425 - Added ``gnt-node powercycle`` command to powercycle nodes
2426 - Added new job status field ``lock_status``
2427 - Added parseable error codes to cluster verification (``gnt-cluster
2428 verify --error-codes``) and made output less verbose (use
2429 ``--verbose`` to restore previous behaviour)
2430 - Added UUIDs to the main config entities (cluster, nodes, instances)
2431 - Added support for OS variants
2432 - Added support for hashed passwords in the Ganeti remote API users file
2434 - Added option to specify maximum timeout on instance shutdown
2435 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2436 - Added new helper script to start and stop Ganeti daemons
2437 (``daemon-util``), with the intent to reduce the work necessary to
2438 adjust Ganeti for non-Debian distributions and to start/stop daemons
2440 - Added more unittests
2441 - Fixed critical bug in ganeti-masterd startup
2442 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2443 now customisable at the cluster level for both the KVM and Xen
2444 hypervisors using the new ``migration_port`` parameter
2445 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2446 reinstalling an instance
2447 - Allowed ``@`` in tag names
2448 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2449 - Many documentation updates
2450 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2451 - ``gnt-instance reinstall`` can now reinstall multiple instances
2452 - Updated many command line parameters
2453 - Introduced new OS API version 15
2454 - No longer support a default hypervisor
2455 - Treat virtual LVs as inexistent
2456 - Improved job locking logic to reduce lock contention
2457 - Match instance and node names case insensitively
2458 - Reimplemented bash completion script to be more complete
2465 *(Released Thu, 4 Feb 2010)*
2467 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2468 - Fix a string formatting bug
2469 - Improve safety of the code in some error paths
2470 - Improve data validation in the master of values returned from nodes
2476 *(Released Thu, 17 Dec 2009)*
2478 - Fix security issue due to missing validation of iallocator names; this
2479 allows local and remote execution of arbitrary executables
2480 - Fix failure of gnt-node list during instance removal
2481 - Ship the RAPI documentation in the archive
2487 *(Released Wed, 30 Sep 2009)*
2489 - Fixed many wrong messages
2490 - Fixed a few bugs related to the locking library
2491 - Fixed MAC checking at instance creation time
2492 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2493 - Fixed a few issues related to signal handling in both daemons and
2495 - Fixed the example startup script provided
2496 - Fixed insserv dependencies in the example startup script (patch from
2498 - Fixed handling of drained nodes in the iallocator framework
2499 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2501 - Fixed error related to invalid job IDs in job polling
2502 - Fixed job/opcode persistence on unclean master shutdown
2503 - Fixed handling of partial job processing after unclean master
2505 - Fixed error reporting from LUs, previously all errors were converted
2506 into execution errors
2507 - Fixed error reporting from burnin
2508 - Decreased significantly the memory usage of the job queue
2509 - Optimised slightly multi-job submission
2510 - Optimised slightly opcode loading
2511 - Backported the multi-job submit framework from the development
2512 branch; multi-instance start and stop should be faster
2513 - Added script to clean archived jobs after 21 days; this will reduce
2514 the size of the queue directory
2515 - Added some extra checks in disk size tracking
2516 - Added an example ethers hook script
2517 - Added a cluster parameter that prevents Ganeti from modifying of
2519 - Added more node information to RAPI responses
2520 - Added a ``gnt-job watch`` command that allows following the ouput of a
2522 - Added a bind-address option to ganeti-rapi
2523 - Added more checks to the configuration verify
2524 - Enhanced the burnin script such that some operations can be retried
2526 - Converted instance reinstall to multi-instance model
2532 *(Released Fri, 7 Aug 2009)*
2534 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2535 to allow using the pre-2.0.2 behaviour in activation, if any existing
2536 instances have mismatched disk sizes in the configuration
2537 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2538 any configuration mismatches for disk sizes
2539 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2540 failover to work on two-node clusters
2541 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2543 - Fixed detection of OS script errors in ``gnt-backup export``
2544 - Fixed exit code of ``gnt-backup export``
2550 *(Released Fri, 17 Jul 2009)*
2552 - Added experimental support for stripped logical volumes; this should
2553 enhance performance but comes with a higher complexity in the block
2554 device handling; stripping is only enabled when passing
2555 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2556 affected even in the non-stripped mode
2557 - Improved resiliency against transient failures at the end of DRBD
2558 resyncs, and in general of DRBD resync checks
2559 - Fixed a couple of issues with exports and snapshot errors
2560 - Fixed a couple of issues in instance listing
2561 - Added display of the disk size in ``gnt-instance info``
2562 - Fixed checking for valid OSes in instance creation
2563 - Fixed handling of the "vcpus" parameter in instance listing and in
2564 general of invalid parameters
2565 - Fixed http server library, and thus RAPI, to handle invalid
2566 username/password combinations correctly; this means that now they
2567 report unauthorized for queries too, not only for modifications,
2568 allowing earlier detect of configuration problems
2569 - Added a new "role" node list field, equivalent to the master/master
2570 candidate/drained/offline flags combinations
2571 - Fixed cluster modify and changes of candidate pool size
2572 - Fixed cluster verify error messages for wrong files on regular nodes
2573 - Fixed a couple of issues with node demotion from master candidate role
2574 - Fixed node readd issues
2575 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2576 - Added a new ``--no-voting`` option for masterfailover to fix failover
2577 on two-nodes clusters when the former master node is unreachable
2578 - Added instance reinstall over RAPI
2584 *(Released Tue, 16 Jun 2009)*
2586 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2587 allow re-adding the start in single-user option (regression from 1.2)
2588 - the watcher writes the instance status to a file, to allow monitoring
2589 to report the instance status (from the master) based on cached
2590 results of the watcher's queries; while this can get stale if the
2591 watcher is being locked due to other work on the cluster, this is
2592 still an improvement
2593 - the watcher now also restarts the node daemon and the rapi daemon if
2595 - fixed the watcher to handle full and drained queue cases
2596 - hooks export more instance data in the environment, which helps if
2597 hook scripts need to take action based on the instance's properties
2598 (no longer need to query back into ganeti)
2599 - instance failovers when the instance is stopped do not check for free
2600 RAM, so that failing over a stopped instance is possible in low memory
2602 - rapi uses queries for tags instead of jobs (for less job traffic), and
2603 for cluster tags it won't talk to masterd at all but read them from
2605 - a couple of error handling fixes in RAPI
2606 - drbd handling: improved the error handling of inconsistent disks after
2607 resync to reduce the frequency of "there are some degraded disks for
2608 this instance" messages
2609 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2610 error handling path called a wrong function name)
2616 *(Released Wed, 27 May 2009)*
2618 - no changes from rc5
2624 *(Released Wed, 20 May 2009)*
2626 - fix a couple of bugs (validation, argument checks)
2627 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2628 - some small improvements to RAPI and IAllocator
2629 - make watcher automatically start the master daemon if down
2635 *(Released Mon, 27 Apr 2009)*
2637 - change the OS list to not require locks; this helps with big clusters
2638 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2639 volume group is broken
2640 - ``gnt-instance info``, without any arguments, doesn't run for all
2641 instances anymore; either pass ``--all`` or pass the desired
2642 instances; this helps against mistakes on big clusters where listing
2643 the information for all instances takes a long time
2644 - miscellaneous doc and man pages fixes
2650 *(Released Wed, 8 Apr 2009)*
2652 - Change the internal locking model of some ``gnt-node`` commands, in
2653 order to reduce contention (and blocking of master daemon) when
2654 batching many creation/reinstall jobs
2655 - Fixes to Xen soft reboot
2656 - No longer build documentation at build time, instead distribute it in
2657 the archive, in order to reduce the need for the whole docbook/rst
2664 *(Released Fri, 27 Mar 2009)*
2666 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2667 - Fix watcher startup sequence, improves the behaviour of busy clusters
2668 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2669 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2670 bind address and other places
2671 - Some documentation fixes and updates
2677 *(Released Mon, 2 Mar 2009)*
2679 - More documentation updates, now all docs should be more-or-less
2681 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2683 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2685 - Fix an issue related to $libdir/run/ganeti and cluster creation
2691 *(Released Thu, 19 Feb 2009)*
2693 - Xen PVM and KVM have switched the default value for the instance root
2694 disk to the first partition on the first drive, instead of the whole
2695 drive; this means that the OS installation scripts must be changed
2697 - Man pages have been updated
2698 - RAPI has been switched by default to HTTPS, and the exported functions
2699 should all work correctly
2700 - RAPI v1 has been removed
2701 - Many improvements to the KVM hypervisor
2702 - Block device errors are now better reported
2703 - Many other bugfixes and small improvements
2709 *(Released Mon, 26 Jan 2009)*
2711 - Version 2 is a general rewrite of the code and therefore the
2712 differences are too many to list, see the design document for 2.0 in
2713 the ``doc/`` subdirectory for more details
2714 - In this beta version there is not yet a migration path from 1.2 (there
2715 will be one in the final 2.0 release)
2716 - A few significant changes are:
2718 - all commands are executed by a daemon (``ganeti-masterd``) and the
2719 various ``gnt-*`` commands are just front-ends to it
2720 - all the commands are entered into, and executed from a job queue,
2721 see the ``gnt-job(8)`` manpage
2722 - the RAPI daemon supports read-write operations, secured by basic
2723 HTTP authentication on top of HTTPS
2724 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2725 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2726 to migrate to DRBD 8 first while still running Ganeti 1.2)
2727 - DRBD devices are using statically allocated minor numbers, which
2728 will be assigned to existing instances during the migration process
2729 - there is support for both Xen PVM and Xen HVM instances running on
2731 - KVM virtualization is supported too
2732 - file-based storage has been implemented, which means that it is
2733 possible to run the cluster without LVM and DRBD storage, for
2734 example using a shared filesystem exported from shared storage (and
2735 still have live migration)
2741 *(Released Tue, 13 Jan 2009)*
2743 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2744 - Reuse the old instance mac address by default on instance import, if
2745 the instance name is the same.
2746 - Handle situations in which the node info rpc returns incomplete
2748 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2749 - Improved version of batcher:
2751 - state file support
2752 - instance mac address support
2753 - support for HVM clusters/instances
2755 - Add an option to show the number of cpu sockets and nodes in
2757 - Support OSes that handle more than one version of the OS api (but do
2758 not change the current API in any other way)
2759 - Fix ``gnt-node migrate``
2760 - ``gnt-debug`` man page
2761 - Fixes various more typos and small issues
2762 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2768 *(Released Wed, 24 Sep 2008)*
2770 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2771 type of disk exported to fully virtualized instances.
2772 - provide access to the serial console of HVM instances
2773 - instance auto_balance flag, set by default. If turned off it will
2774 avoid warnings on cluster verify if there is not enough memory to fail
2775 over an instance. in the future it will prevent automatically failing
2776 it over when we will support that.
2777 - batcher tool for instance creation, see ``tools/README.batcher``
2778 - ``gnt-instance reinstall --select-os`` to interactively select a new
2779 operating system when reinstalling an instance.
2780 - when changing the memory amount on instance modify a check has been
2781 added that the instance will be able to start. also warnings are
2782 emitted if the instance will not be able to fail over, if auto_balance
2784 - documentation fixes
2785 - sync fields between ``gnt-instance list/modify/add/import``
2786 - fix a race condition in drbd when the sync speed was set after giving
2787 the device a remote peer.
2793 *(Released Tue, 22 Jul 2008)*
2795 - note: the allowed size and number of tags per object were reduced
2796 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2797 - fixed twisted 8.x compatibility
2798 - fixed ``gnt-instance replace-disks`` with iallocator
2799 - add TCP keepalives on twisted connections to detect restarted nodes
2800 - disk increase support, see ``gnt-instance grow-disk``
2801 - implement bulk node/instance query for RAPI
2802 - add tags in node/instance listing (optional)
2803 - experimental migration (and live migration) support, read the man page
2804 for ``gnt-instance migrate``
2805 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2806 has some small improvements in handling its state file
2812 *(Released Fri, 13 Jun 2008)*
2814 - Experimental readonly, REST-based remote API implementation;
2815 automatically started on master node, TCP port 5080, if enabled by
2816 ``--enable-rapi`` parameter to configure script.
2817 - Instance allocator support. Add and import instance accept a
2818 ``--iallocator`` parameter, and call that instance allocator to decide
2819 which node to use for the instance. The iallocator document describes
2820 what's expected from an allocator script.
2821 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2822 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2823 node is lost there is still enough memory to fail over the instances
2825 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2826 ``gnt-cluster verify``, to check for site-specific compliance. All the
2827 hooks will run, and their output, if any, will be displayed. Any
2828 failing hook will make the verification return an error value.
2829 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2830 primary and secondary interfaces
2831 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2832 that is still declared as part of the cluster and has failed.
2833 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2834 specifying output fields, that just adds the chosen fields to the
2836 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2837 export from the filesystem.
2838 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2839 have been added. Using them you can enable/disable acpi and pae
2840 support, and specify a path for a cd image to be exported to the
2841 instance. These parameters as the name suggest only work on HVM
2843 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2844 PAE support will be set to the previously hardcoded values, but the
2845 (previously hardcoded) path to the CDROM ISO image will be unset and
2846 if required, needs to be set manually with ``gnt-instance modify``
2848 - The address to which an instance's VNC console is bound is now
2849 selectable per-instance, rather than being cluster wide. Of course
2850 this only applies to instances controlled via VNC, so currently just
2851 applies to HVM clusters.
2857 *(Released Mon, 18 Feb 2008)*
2859 - more tweaks to the disk activation code (especially helpful for DRBD)
2860 - change the default ``gnt-instance list`` output format, now there is
2861 one combined status field (see the manpage for the exact values this
2863 - some more fixes for the mac export to hooks change
2864 - make Ganeti not break with DRBD 8.2.x (which changed the version
2865 format in ``/proc/drbd``) (issue 24)
2866 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2867 template, allowing migration from DRBD0.7+MD to DRBD8
2873 *(Released Wed, 30 Jan 2008)*
2875 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2877 - add command aliases infrastructure and a few aliases
2878 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2879 man pages and the ``--help`` option of ``gnt-node
2880 list``/``gnt-instance list``
2881 - fix ``gnt-backup list`` with down nodes (issue 21)
2882 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2883 - fix the dist archive and add a check for including svn/git files in
2885 - some developer-related changes: improve the burnin and the QA suite,
2886 add an upload script for testing during development
2892 *(Released Wed, 16 Jan 2008)*
2894 - experimental HVM support, read the install document, section
2895 "Initializing the cluster"
2896 - allow for the PVM hypervisor per-instance kernel and initrd paths
2897 - add a new command ``gnt-cluster verify-disks`` which uses a new
2898 algorithm to improve the reconnection of the DRBD pairs if the device
2899 on the secondary node has gone away
2900 - make logical volume code auto-activate LVs at disk activation time
2901 - slightly improve the speed of activating disks
2902 - allow specification of the MAC address at instance creation time, and
2903 changing it later via ``gnt-instance modify``
2904 - fix handling of external commands that generate lots of output on
2906 - update documentation with regard to minimum version of DRBD8 supported
2912 *(Released Tue, 4 Dec 2007)*
2914 - Log the ``xm create`` output to the node daemon log on failure (to
2915 help diagnosing the error)
2916 - In debug mode, log all external commands output if failed to the logs
2917 - Change parsing of lvm commands to ignore stderr
2923 *(Released Wed, 28 Nov 2007)*
2925 - Another round of updates to the DRBD 8 code to deal with more failures
2926 in the replace secondary node operation
2927 - Some more logging of failures in disk operations (lvm, drbd)
2928 - A few documentation updates
2935 *(Released Tue, 13 Nov 2007)*
2937 - Change configuration file format from Python's Pickle to JSON.
2938 Upgrading is possible using the cfgupgrade utility.
2939 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2940 faster replace disks and is more stable (DRBD 8 has many improvements
2941 compared to DRBD 0.7)
2942 - Added command line tags support (see man pages for ``gnt-instance``,
2943 ``gnt-node``, ``gnt-cluster``)
2944 - Added instance rename support
2945 - Added multi-instance startup/shutdown
2946 - Added cluster rename support
2947 - Added ``gnt-node evacuate`` to simplify some node operations
2948 - Added instance reboot operation that can speedup reboot as compared to
2950 - Soften the requirement that hostnames are in FQDN format
2951 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2953 - Removed dependency on debian's patched fping that uses the
2954 non-standard ``-S`` option
2955 - Now the OS definitions are searched for in multiple, configurable
2956 paths (easier for distros to package)
2957 - Some changes to the hooks infrastructure (especially the new
2958 post-configuration update hook)
2959 - Other small bugfixes
2961 .. vim: set textwidth=72 syntax=rst :