8 *(Released Thu, 07 Nov 2013)*
10 - DRBD: ensure peers are UpToDate for dual-primary
11 - Improve error message for replace-disks
12 - More dependency checks at configure time
13 - Placate warnings on ganeti.outils_unittest.py
19 *(Released Thu, 17 Oct 2013)*
21 - Correctly start/stop luxid during gnt-cluster master-failover
22 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
23 - Fix path for the job queue serial file
24 - Improved harep man page
25 - Minor documentation improvements
31 *(Released Mon, 30 Sep 2013)*
33 Incompatible/important changes
34 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
36 - Instance policy can contain multiple instance specs, as described in
37 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
38 <design-partitioned>`. As a consequence, it's not possible to partially change
39 or override instance specs. Bounding specs (min and max) can be specified as a
40 whole using the new option ``--ipolicy-bounds-specs``, while standard
41 specs use the new option ``--ipolicy-std-specs``.
42 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
43 gnt-instance is a valid YAML object.
44 - hail now honors network restrictions when allocating nodes. This led to an
45 update of the IAllocator protocol. See the IAllocator documentation for
47 - confd now only answers static configuration request over the network. luxid
48 was extracted, listens on the local LUXI socket and responds to live queries.
49 This allows finer grained permissions if using separate users.
54 - Hotplug support. Introduce new option --hotplug to gnt-instance modify
55 so that disk and NIC modifications take effect without the need of actual
56 reboot. This feature is currently supported only for KVM hypervisor with
57 version greater than 1.0.
58 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
59 to always require authentication, ``--require-authentication``. It can
60 be specified in ``$sysconfdir/default/ganeti``.
61 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
62 be used to manage the disk templates to be used by instances in the cluster.
63 Initially, it will be set to a list that includes plain, drbd, if they were
64 enabled by specifying a volume group name, and file and sharedfile, if those
65 were enabled at configure time. Additionally, it will include all disk
66 templates that are currently used by instances. The order of disk templates
67 will be based on Ganeti's history of supporting them. In the future, the
68 first entry of the list will be used as a default disk template on instance
70 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
71 configuration back to the previous stable version.
72 - Disk templates in group ipolicy can be restored to the default value.
73 - Initial support for diskless instances and virtual clusters in QA.
74 - More QA and unit tests for instance policies.
75 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
76 describing why the opcode itself was executed.
77 - The monitoring daemon is now available. It allows users to query the cluster
78 for obtaining information about the status of the system. The daemon is only
79 responsible for providing the information over the network: the actual data
80 gathering is performed by data collectors (currently, only the DRBD status
81 collector is available).
82 - In order to help developers work on Ganeti, a new script
83 (``devel/build_chroot``) is provided, for building a chroot that contains all
84 the required development libraries and tools for compiling Ganeti on a Debian
86 - A new tool, ``harep``, for performing self-repair and recreation of instances
87 in Ganeti has been added.
88 - Split queries are enabled for tags, network, exports, cluster info, groups,
90 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
91 It prints the command line to set the current policies, to ease
93 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
94 devices for KVM virtio-net interfaces will get created with VNET_HDR
95 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
96 virtio-net interfaces, which prevents host kernel tainting and log
97 flooding, when dealing with broken or malicious virtio-net drivers.
98 It's set to true by default.
99 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
101 - Support 'viridian' parameter in Xen HVM
102 - Support DSA SSH keys in bootstrap
103 - To simplify the work of packaging frameworks that want to add the needed users
104 and groups in a split-user setup themselves, at build time three files in
105 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
106 the groups to be generated, the ``users`` file contains, one per line, the
107 users to be generated, optionally followed by their primary group, where
108 important. The ``groupmemberships`` file contains, one per line, additional
109 user-group membership relations that need to be established. The syntax of
110 these files will remain stable in all future versions.
115 The following new dependencies have been added:
118 - The ``curl`` library is not optional anymore for compiling the Haskell code.
119 - ``snap-server`` library (if monitoring is enabled).
122 - The minimum Python version needed to run Ganeti is now 2.6.
123 - ``yaml`` library (only for running the QA).
127 - Perform proper cleanup on termination of Haskell daemons
128 - Fix corner-case in handling of remaining retry time
134 *(Released Tue, 17 Sep 2013)*
136 - To simplify the work of packaging frameworks that want to add the needed users
137 and groups in a split-user setup themselves, at build time three files in
138 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
139 the groups to be generated, the ``users`` file contains, one per line, the
140 users to be generated, optionally followed by their primary group, where
141 important. The ``groupmemberships`` file contains, one per line, additional
142 user-group membership relations that need to be established. The syntax of
143 these files will remain stable in all future versions.
144 - Add a default to file-driver when unspecified over RAPI (Issue 571)
145 - Mark the DSA host pubkey as optional, and remove it during config downgrade
147 - Some documentation fixes
153 *(Released Tue, 27 Aug 2013)*
155 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
157 - Support 'viridian' parameter in Xen HVM (Issue 233)
158 - Include VCS version in ``gnt-cluster version``
159 - Support DSA SSH keys in bootstrap (Issue 338)
160 - Fix batch creation of instances
161 - Use FQDN to check master node status (Issue 551)
162 - Make the DRBD collector more failure-resilient
168 *(Released Fri, 2 Aug 2013)*
170 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
172 - Fix upgrading/downgrading from 2.7
173 - Increase maximum RAPI message size
174 - Documentation updates
175 - Split ``confd`` between ``luxid`` and ``confd``
176 - Merge 2.7 series up to the 2.7.1 release
177 - Allow the ``modify_etc_hosts`` option to be changed
178 - Add better debugging for ``luxid`` queries
179 - Expose bulk parameter for GetJobs in RAPI client
180 - Expose missing ``network`` fields in RAPI
181 - Add some ``cluster verify`` tests
182 - Some unittest fixes
183 - Fix a malfunction in ``hspace``'s tiered allocation
184 - Fix query compatibility between haskell and python implementations
185 - Add the ``vnet_hdr`` HV parameter for KVM
186 - Add ``--cleanup`` to instance failover
187 - Change the connected groups format in ``gnt-network info`` output; it
188 was previously displayed as a raw list by mistake. (Merged from 2.7)
194 *(Released Mon, 24 Jun 2013)*
196 This was the first beta release of the 2.8 series. All important changes
197 are listed in the latest 2.8 entry.
203 *(Released Thu, 26 Sep 2013)*
205 - Change the connected groups format in ``gnt-network info`` output; it
206 was previously displayed as a raw list by mistake
207 - Check disk template in right dict when copying
208 - Support multi-instance allocs without iallocator
209 - Fix some errors in the documentation
210 - Fix formatting of tuple in an error message
216 *(Released Thu, 25 Jul 2013)*
218 - Add logrotate functionality in daemon-util
219 - Add logrotate example file
220 - Add missing fields to network queries over rapi
221 - Fix network object timestamps
222 - Add support for querying network timestamps
223 - Fix a typo in the example crontab
224 - Fix a documentation typo
230 *(Released Thu, 04 Jul 2013)*
232 Incompatible/important changes
233 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
235 - Instance policies for disk size were documented to be on a per-disk
236 basis, but hail applied them to the sum of all disks. This has been
238 - ``hbal`` will now exit with status 0 if, during job execution over
239 LUXI, early exit has been requested and all jobs are successful;
240 before, exit status 1 was used, which cannot be differentiated from
242 - Compatibility with newer versions of rbd has been fixed
243 - ``gnt-instance batch-create`` has been changed to use the bulk create
244 opcode from Ganeti. This lead to incompatible changes in the format of
245 the JSON file. It's now not a custom dict anymore but a dict
246 compatible with the ``OpInstanceCreate`` opcode.
247 - Parent directories for file storage need to be listed in
248 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
249 write the file automatically based on old configuration values, but it
250 can not distribute it across all nodes and the file contents should be
251 verified. Use ``gnt-cluster copyfile
252 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
253 upgraded. The reason for requiring this list of paths now is that
254 before it would have been possible to inject new paths via RPC,
255 allowing files to be created in arbitrary locations. The RPC protocol
256 is protected using SSL/X.509 certificates, but as a design principle
257 Ganeti does not permit arbitrary paths to be passed.
258 - The parsing of the variants file for OSes (see
259 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
260 lines and comment lines (starting with ``#``) are ignored for better
262 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
263 longer available. ``gnt-node add`` now invokes a new tool on the
264 destination node, named ``prepare-node-join``, to configure the SSH
265 daemon. Paramiko is no longer necessary to configure nodes' SSH
266 daemons via ``gnt-node add``.
267 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
268 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
269 the flag is not reset after a master failover.
270 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
271 is recommended. 2.6 will be mandatory from the 2.8 series.
277 - New network management functionality to support automatic allocation
278 of IP addresses and managing of network parameters. See
279 :manpage:`gnt-network(8)` for more details.
280 - New external storage backend, to allow managing arbitrary storage
281 systems external to the cluster. See
282 :manpage:`ganeti-extstorage-interface(7)`.
283 - New ``exclusive-storage`` node parameter added, restricted to
284 nodegroup level. When it's set to true, physical disks are assigned in
285 an exclusive fashion to instances, as documented in :doc:`Partitioned
286 Ganeti <design-partitioned>`. Currently, only instances using the
287 ``plain`` disk template are supported.
288 - The KVM hypervisor has been updated with many new hypervisor
289 parameters, including a generic one for passing arbitrary command line
290 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
291 compatible up to qemu 1.4.
292 - A new tool, called ``mon-collector``, is the stand-alone executor of
293 the data collectors for a monitoring system. As of this version, it
294 just includes the DRBD data collector, that can be executed by calling
295 ``mon-collector`` using the ``drbd`` parameter. See
296 :manpage:`mon-collector(7)`.
297 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
298 for RAPI users. It allows granting permissions to query for
299 information to a specific user without giving
300 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
301 - A new tool named ``node-cleanup`` has been added. It cleans remains of
302 a cluster from a machine by stopping all daemons, removing
303 certificates and ssconf files. Unless the ``--no-backup`` option is
304 given, copies of the certificates are made.
305 - Instance creations now support the use of opportunistic locking,
306 potentially speeding up the (parallel) creation of multiple instances.
307 This feature is currently only available via the :doc:`RAPI
308 <rapi>` interface and when an instance allocator is used. If the
309 ``opportunistic_locking`` parameter is set the opcode will try to
310 acquire as many locks as possible, but will not wait for any locks
311 held by other opcodes. If not enough resources can be found to
312 allocate the instance, the temporary error code
313 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
314 retried thereafter, with or without opportunistic locking.
315 - New experimental linux-ha resource scripts.
316 - Restricted-commands support: ganeti can now be asked (via command line
317 or rapi) to perform commands on a node. These are passed via ganeti
318 RPC rather than ssh. This functionality is restricted to commands
319 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
320 reasons. The file is not copied automatically.
326 - Diskless instances are now externally mirrored (Issue 237). This for
327 now has only been tested in conjunction with explicit target nodes for
329 - Queries not needing locks or RPC access to the node can now be
330 performed by the confd daemon, making them independent from jobs, and
331 thus faster to execute. This is selectable at configure time.
332 - The functionality for allocating multiple instances at once has been
333 overhauled and is now also available through :doc:`RAPI <rapi>`.
335 There are no significant changes from version 2.7.0~rc3.
341 *(Released Tue, 25 Jun 2013)*
343 - Fix permissions on the confd query socket (Issue 477)
344 - Fix permissions on the job archive dir (Issue 498)
345 - Fix handling of an internal exception in replace-disks (Issue 472)
346 - Fix gnt-node info handling of shortened names (Issue 497)
347 - Fix gnt-instance grow-disk when wiping is enabled
348 - Documentation improvements, and support for newer pandoc
349 - Fix hspace honoring ipolicy for disks (Issue 484)
350 - Improve handling of the ``kvm_extra`` HV parameter
356 *(Released Fri, 24 May 2013)*
358 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
360 - Disks added through ``gnt-instance modify`` or created through
361 ``gnt-instance recreate-disks`` are wiped, if the
362 ``prealloc_wipe_disks`` flag is set.
363 - If wiping newly created disks fails, the disks are removed. Also,
364 partial failures in creating disks through ``gnt-instance modify``
365 triggers a cleanup of the partially-created disks.
366 - Removing the master IP address doesn't fail if the address has been
368 - Fix ownership of the OS log dir
369 - Workaround missing SO_PEERCRED constant (Issue 191)
375 *(Released Fri, 3 May 2013)*
377 This was the first release candidate of the 2.7 series. Since beta3:
379 - Fix kvm compatibility with qemu 1.4 (Issue 389)
380 - Documentation updates (admin guide, upgrade notes, install
381 instructions) (Issue 372)
382 - Fix gnt-group list nodes and instances count (Issue 436)
383 - Fix compilation without non-mandatory libraries (Issue 441)
384 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
385 - Make confd logging more verbose at INFO level (Issue 435)
386 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
387 - Fix failure path for instance storage type conversion (Issue 229)
388 - Update htools text backend documentation
389 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
390 - Disable inter-cluster instance move for file-based instances, because
391 it is dependant on instance export, which is not supported for
392 file-based instances. (Issue 414)
393 - Fix gnt-job crashes on non-ascii characters (Issue 427)
394 - Fix volume group checks on non-vm-capable nodes (Issue 432)
400 *(Released Mon, 22 Apr 2013)*
402 This was the third beta release of the 2.7 series. Since beta2:
404 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
405 - Fix data loss on wrong usage of ``gnt-instance move``
406 - Properly export errors in confd-based job queries
407 - Add ``users-setup`` tool
408 - Fix iallocator protocol to report 0 as a disk size for diskless
409 instances. This avoids hail breaking when a diskless instance is
411 - Fix job queue directory permission problem that made confd job queries
412 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
413 for access to archived jobs (Issue 406).
414 - Limit the sizes of networks supported by ``gnt-network`` to something
415 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
416 - Fix bugs in instance disk template conversion
417 - Fix GHC 7 compatibility
418 - Fix ``burnin`` install path (Issue 426).
419 - Allow very small disk grows (Issue 347).
420 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
421 that noded doesn't import masterd code (Issue 419).
422 - Make sure the default metavg at cluster init is the same as the vg, if
423 unspecified (Issue 358).
424 - Fix cleanup of partially created disks (part of Issue 416)
430 *(Released Tue, 2 Apr 2013)*
432 This was the second beta release of the 2.7 series. Since beta1:
434 - Networks no longer have a "type" slot, since this information was
435 unused in Ganeti: instead of it tags should be used.
436 - The rapi client now has a ``target_node`` option to MigrateInstance.
437 - Fix early exit return code for hbal (Issue 386).
438 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
439 - Fix ``rbd showmapped`` output parsing (Issue 312).
440 - Networks are now referenced indexed by UUID, rather than name. This
441 will require running cfgupgrade, from 2.7.0beta1, if networks are in
443 - The OS environment now includes network information.
444 - Deleting of a network is now disallowed if any instance nic is using
445 it, to prevent dangling references.
446 - External storage is now documented in man pages.
447 - The exclusive_storage flag can now only be set at nodegroup level.
448 - Hbal can now submit an explicit priority with its jobs.
449 - Many network related locking fixes.
450 - Bump up the required pylint version to 0.25.1.
451 - Fix the ``no_remember`` option in RAPI client.
452 - Many ipolicy related tests, qa, and fixes.
453 - Many documentation improvements and fixes.
454 - Fix building with ``--disable-file-storage``.
455 - Fix ``-q`` option in htools, which was broken if passed more than
457 - Some haskell/python interaction improvements and fixes.
458 - Fix iallocator in case of missing LVM storage.
459 - Fix confd config load in case of ``--no-lvm-storage``.
460 - The confd/query functionality is now mentioned in the security
467 *(Released Wed, 6 Feb 2013)*
469 This was the first beta release of the 2.7 series. All important changes
470 are listed in the latest 2.7 entry.
476 *(Released Fri, 21 Dec 2012)*
478 Important behaviour change: hbal won't rebalance anymore instances which
479 have the ``auto_balance`` attribute set to false. This was the intention
480 all along, but until now it only skipped those from the N+1 memory
481 reservation (DRBD-specific).
483 A significant number of bug fixes in this release:
485 - Fixed disk adoption interaction with ipolicy checks.
486 - Fixed networking issues when instances are started, stopped or
487 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
488 - Fixed the warning in cluster verify for shared storage instances not
490 - Fixed removal of storage directory on shared file storage (issue 262).
491 - Fixed validation of LVM volume group name in OpClusterSetParams
492 (``gnt-cluster modify``) (issue 285).
493 - Fixed runtime memory increases (``gnt-instance modify -m``).
494 - Fixed live migration under Xen's ``xl`` mode.
495 - Fixed ``gnt-instance console`` with ``xl``.
496 - Fixed building with newer Haskell compiler/libraries.
497 - Fixed PID file writing in Haskell daemons (confd); this prevents
498 restart issues if confd was launched manually (outside of
499 ``daemon-util``) while another copy of it was running
500 - Fixed a type error when doing live migrations with KVM (issue 297) and
501 the error messages for failing migrations have been improved.
502 - Fixed opcode validation for the out-of-band commands (``gnt-node
504 - Fixed a type error when unsetting OS hypervisor parameters (issue
505 311); now it's possible to unset all OS-specific hypervisor
507 - Fixed the ``dry-run`` mode for many operations: verification of
508 results was over-zealous but didn't take into account the ``dry-run``
509 operation, resulting in "wrong" failures.
510 - Fixed bash completion in ``gnt-job list`` when the job queue has
511 hundreds of entries; especially with older ``bash`` versions, this
512 results in significant CPU usage.
514 And lastly, a few other improvements have been made:
516 - Added option to force master-failover without voting (issue 282).
517 - Clarified error message on lock conflict (issue 287).
518 - Logging of newly submitted jobs has been improved (issue 290).
519 - Hostname checks have been made uniform between instance rename and
521 - The ``--submit`` option is now supported by ``gnt-debug delay``.
522 - Shutting down the master daemon by sending SIGTERM now stops it from
523 processing jobs waiting for locks; instead, those jobs will be started
524 once again after the master daemon is started the next time (issue
526 - Support for Xen's ``xl`` program has been improved (besides the fixes
528 - Reduced logging noise in the Haskell confd daemon (only show one log
529 entry for each config reload, instead of two).
530 - Several man page updates and typo fixes.
536 *(Released Fri, 12 Oct 2012)*
538 A small bugfix release. Among the bugs fixed:
540 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
541 made the command unusable.
542 - Commands that issue many jobs don't fail anymore just because some jobs
543 take so long that other jobs are archived.
544 - Failures during ``gnt-instance reinstall`` are reflected by the exit
546 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
548 - When ``always_failover`` is set, ``--allow-failover`` is not required
549 in migrate commands anymore.
550 - ``bash_completion`` works even if extglob is disabled.
551 - Fixed bug with locks that made failover for RDB-based instances fail.
552 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
553 a random node instead of one based on the allocator metric.
554 - Support for newer versions of pylint and pep8.
555 - Hail doesn't fail anymore when trying to add an instance of type
556 ``file``, ``sharedfile`` or ``rbd``.
557 - Added new Makefile target to rebuild the whole distribution, so that
558 all files are included.
564 *(Released Fri, 27 Jul 2012)*
567 .. attention:: The ``LUXI`` protocol has been made more consistent
568 regarding its handling of command arguments. This, however, leads to
569 incompatibility issues with previous versions. Please ensure that you
570 restart Ganeti daemons soon after the upgrade, otherwise most
571 ``LUXI`` calls (job submission, setting/resetting the drain flag,
572 pausing/resuming the watcher, cancelling and archiving jobs, querying
573 the cluster configuration) will fail.
582 The current ``admin_up`` field, which used to denote whether an instance
583 should be running or not, has been removed. Instead, ``admin_state`` is
584 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
586 The rational behind this is that an instance being “down” can have
589 - it could be down during a reboot
590 - it could be temporarily be down for a reinstall
591 - or it could be down because it is deprecated and kept just for its
594 The previous Boolean state was making it difficult to do capacity
595 calculations: should Ganeti reserve memory for a down instance? Now, the
596 tri-state field makes it clear:
598 - in ``up`` and ``down`` state, all resources are reserved for the
599 instance, and it can be at any time brought up if it is down
600 - in ``offline`` state, only disk space is reserved for it, but not
603 The field can have an extra use: since the transition between ``up`` and
604 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
605 transition between ``offline`` and ``down`` is done via ``gnt-instance
606 modify``, it is possible to given different rights to users. For
607 example, owners of an instance could be allowed to start/stop it, but
608 not transition it out of the offline state.
610 Instance policies and specs
611 +++++++++++++++++++++++++++
613 In previous Ganeti versions, an instance creation request was not
614 limited on the minimum size and on the maximum size just by the cluster
615 resources. As such, any policy could be implemented only in third-party
616 clients (RAPI clients, or shell wrappers over ``gnt-*``
617 tools). Furthermore, calculating cluster capacity via ``hspace`` again
618 required external input with regards to instance sizes.
620 In order to improve these workflows and to allow for example better
621 per-node group differentiation, we introduced instance specs, which
624 - minimum instance disk size, disk count, memory size, cpu count
625 - maximum values for the above metrics
626 - and “standard” values (used in ``hspace`` to calculate the standard
629 The minimum/maximum values can be also customised at node-group level,
630 for example allowing more powerful hardware to support bigger instance
633 Beside the instance specs, there are a few other settings belonging to
634 the instance policy framework. It is possible now to customise, per
635 cluster and node-group:
637 - the list of allowed disk templates
638 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
639 - the maximum ratio of instance to spindles (see below for more
640 information) for local storage
642 All these together should allow all tools that talk to Ganeti to know
643 what are the ranges of allowed values for instances and the
644 over-subscription that is allowed.
646 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
647 instance configuration, and the physical CPU configuration from the
648 node. For the spindle ratios however, we didn't track before these
649 values, so new parameters have been added:
651 - a new node parameter ``spindle_count``, defaults to 1, customisable at
652 node group or node level
653 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
655 Note that spindles in this context doesn't need to mean actual
656 mechanical hard-drives; it's just a relative number for both the node
657 I/O capacity and instance I/O consumption.
659 Instance migration behaviour
660 ++++++++++++++++++++++++++++
662 While live-migration is in general desirable over failover, it is
663 possible that for some workloads it is actually worse, due to the
664 variable time of the “suspend” phase during live migration.
666 To allow the tools to work consistently over such instances (without
667 having to hard-code instance names), a new backend parameter
668 ``always_failover`` has been added to control the migration/failover
669 behaviour. When set to True, all migration requests for an instance will
670 instead fall-back to failover.
672 Instance memory ballooning
673 ++++++++++++++++++++++++++
675 Initial support for memory ballooning has been added. The memory for an
676 instance is no longer fixed (backend parameter ``memory``), but instead
677 can vary between minimum and maximum values (backend parameters
678 ``minmem`` and ``maxmem``). Currently we only change an instance's
681 - live migrating or failing over and instance and the target node
682 doesn't have enough memory
683 - user requests changing the memory via ``gnt-instance modify
689 In order to control the use of specific CPUs by instance, support for
690 controlling CPU pinning has been added for the Xen, HVM and LXC
691 hypervisors. This is controlled by a new hypervisor parameter
692 ``cpu_mask``; details about possible values for this are in the
693 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
694 VCPU-to-CPU mapping) form will work well only when all nodes in your
695 cluster have the same amount of CPUs.
700 Another area in which Ganeti was not customisable were the parameters
701 used for storage configuration, e.g. how many stripes to use for LVM,
702 DRBD resync configuration, etc.
704 To improve this area, we've added disks parameters, which are
705 customisable at cluster and node group level, and which allow to
706 specify various parameters for disks (DRBD has the most parameters
707 currently), for example:
709 - DRBD resync algorithm and parameters (e.g. speed)
710 - the default VG for meta-data volumes for DRBD
711 - number of stripes for LVM (plain disk template)
714 These parameters can be modified via ``gnt-cluster modify -D …`` and
715 ``gnt-group modify -D …``, and are used at either instance creation (in
716 case of LVM stripes, for example) or at disk “activation” time
719 Rados block device support
720 ++++++++++++++++++++++++++
722 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
723 denoted by the ``rbd`` disk template type. This is considered
724 experimental, feedback is welcome. For details on configuring it, see
725 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
730 The existing master IP functionality works well only in simple setups (a
731 single network shared by all nodes); however, if nodes belong to
732 different networks, then the ``/32`` setup and lack of routing
733 information is not enough.
735 To allow the master IP to function well in more complex cases, the
736 system was reworked as follows:
738 - a master IP netmask setting has been added
739 - the master IP activation/turn-down code was moved from the node daemon
741 - whether to run the Ganeti-supplied master IP script or a user-supplied
742 on is a ``gnt-cluster init`` setting
744 Details about the location of the standard and custom setup scripts are
745 in the man page :manpage:`gnt-cluster(8)`; for information about the
746 setup script protocol, look at the Ganeti-supplied script.
751 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
754 It is now possible to use TLS-protected connections, and when renewing
755 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
756 it is now possible to specify spice or spice CA certificates. Also, it
757 is possible to configure a password for SPICE sessions via the
758 hypervisor parameter ``spice_password_file``.
760 There are also new parameters to control the compression and streaming
761 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
762 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
763 for the spice parameters.
765 Lastly, it is now possible to see the SPICE connection information via
766 ``gnt-instance console``.
771 A new tool (``tools/ovfconverter``) has been added that supports
772 conversion between Ganeti and the `Open Virtualization Format
773 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
776 This relies on the ``qemu-img`` tool to convert the disk formats, so the
777 actual compatibility with other virtualization solutions depends on it.
782 The configuration query daemon (``ganeti-confd``) is now optional, and
783 has been rewritten in Haskell; whether to use the daemon at all, use the
784 Python (default) or the Haskell version is selectable at configure time
785 via the ``--enable-confd`` parameter, which can take one of the
786 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
787 daemon will result in a smaller footprint; for larger systems, we
788 welcome feedback on the Haskell version which might become the default
791 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
792 daemon running. The Python version doesn't implement the new call.
795 User interface changes
796 ~~~~~~~~~~~~~~~~~~~~~~
798 We have replaced the ``--disks`` option of ``gnt-instance
799 replace-disks`` with a more flexible ``--disk`` option, which allows
800 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
801 disk size and mode can be changed upon recreation (via ``gnt-instance
802 recreate-disks``, which accepts the same ``--disk`` option).
804 As many people are used to a ``show`` command, we have added that as an
805 alias to ``info`` on all ``gnt-*`` commands.
807 The ``gnt-instance grow-disk`` command has a new mode in which it can
808 accept the target size of the disk, instead of the delta; this can be
809 more safe since two runs in absolute mode will be idempotent, and
810 sometimes it's also easier to specify the desired size directly.
812 Also the handling of instances with regard to offline secondaries has
813 been improved. Instance operations should not fail because one of it's
814 secondary nodes is offline, even though it's safe to proceed.
816 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
817 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
818 minors to instance name.
823 RAPI coverage has improved, with (for example) new resources for
824 recreate-disks, node power-cycle, etc.
829 There is partial support for ``xl`` in the Xen hypervisor; feedback is
832 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
833 whether to still support Python 2.4 or move to Python 2.6 as minimum
836 Support for Fedora has been slightly improved; the provided example
837 init.d script should work better on it and the INSTALL file should
838 document the needed dependencies.
843 The deprecated ``QueryLocks`` LUXI request has been removed. Use
844 ``Query(what=QR_LOCK, ...)`` instead.
846 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
847 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
848 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
849 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
850 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
852 RAPI client: ``CertificateError`` now derives from
853 ``GanetiApiError``. This should make it more easy to handle Ganeti
856 Deprecation warnings due to PyCrypto/paramiko import in
857 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
858 make sure to run an up-to-date paramiko version, if you use this tool.
860 The QA scripts now depend on Python 2.5 or above (the main code base
861 still works with Python 2.4).
863 The configuration file (``config.data``) is now written without
864 indentation for performance reasons; if you want to edit it, it can be
865 re-formatted via ``tools/fmtjson``.
867 A number of bugs has been fixed in the cluster merge tool.
869 ``x509`` certification verification (used in import-export) has been
870 changed to allow the same clock skew as permitted by the cluster
871 verification. This will remove some rare but hard to diagnose errors in
878 *(Released Thu, 19 Jul 2012)*
880 Very few changes from rc4 to the final release, only bugfixes:
882 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
883 instance, fix CDROM booting for KVM instances)
884 - fixed node group modification of node parameters
885 - fixed issue in LUClusterVerifyGroup with multi-group clusters
886 - fixed generation of bash completion to ensure a stable ordering
893 *(Released Fri, 13 Jul 2012)*
895 Third release candidate for 2.6. The following changes were done from
898 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
899 - Fixed an inconsistency in the LUXI protocol with the provided
900 arguments (NOT backwards compatible)
901 - Fixed a bug with node groups ipolicy where ``min`` was greater than
902 the cluster ``std`` value
903 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
904 easier instance debugging on nodes (requires ``hconfd`` to work)
910 *(Released Tue, 03 Jul 2012)*
912 Second release candidate for 2.6. The following changes were done from
915 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
917 - Fixed a RAPI regression on missing beparams/memory
918 - Fixed redistribution of files on offline nodes
919 - Added possibility to run activate-disks even though secondaries are
920 offline. With this change it relaxes also the strictness on some other
921 commands which use activate disks internally:
922 * ``gnt-instance start|reboot|rename|backup|export``
923 - Made it possible to remove safely an instance if its secondaries are
925 - Made it possible to reinstall even though secondaries are offline
931 *(Released Mon, 25 Jun 2012)*
933 First release candidate for 2.6. The following changes were done from
936 - Fixed bugs with disk parameters and ``rbd`` templates as well as
938 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
939 behaviour. It supports now the modify operation
940 - ``hcheck`` implemented to analyze cluster health and possibility of
941 improving health by rebalance
942 - ``hbal`` has been improved in dealing with split instances
948 *(Released Mon, 11 Jun 2012)*
950 Second beta release of 2.6. The following changes were done from beta2
953 - Fixed ``daemon-util`` with non-root user models
954 - Fixed creation of plain instances with ``--no-wait-for-sync``
955 - Fix wrong iv_names when running ``cfgupgrade``
956 - Export more information in RAPI group queries
957 - Fixed bug when changing instance network interfaces
958 - Extended burnin to do NIC changes
959 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
960 - Changed default for DRBD barriers
961 - Fixed DRBD error reporting for syncer rate
962 - Verify the options on disk parameters
964 And of course various fixes to documentation and improved unittests and
971 *(Released Wed, 23 May 2012)*
973 First beta release of 2.6. The following changes were done from beta1 to
976 - integrated patch for distributions without ``start-stop-daemon``
977 - adapted example init.d script to work on Fedora
978 - fixed log handling in Haskell daemons
979 - adapted checks in the watcher for pycurl linked against libnss
980 - add partial support for ``xl`` instead of ``xm`` for Xen
981 - fixed a type issue in cluster verification
982 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
985 Plus integrated fixes from the 2.5 branch:
987 - fixed ``kvm-ifup`` to use ``/bin/bash``
988 - fixed parallel build failures
989 - KVM live migration when using a custom keymap
995 *(Released Tue, 24 Jul 2012)*
997 A small bugfix release, with no new features:
999 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1000 different default shell (e.g. Debian, Ubuntu)
1001 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1002 243 and Debian bug #650664)
1003 - fixed compatibility with KVM versions that don't support multiple boot
1004 devices (fixes Issue 230 and Debian bug #624256)
1006 Additionally, a few fixes were done to the build system (fixed parallel
1007 build failures) and to the unittests (fixed race condition in test for
1008 FileID functions, and the default enable/disable mode for QA test is now
1015 *(Released Fri, 11 May 2012)*
1017 A small bugfix release.
1019 The main issues solved are on the topic of compatibility with newer LVM
1022 - fixed parsing of ``lv_attr`` field
1023 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1024 the ``--force`` flag is needed
1026 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1027 to accept kernel 3.x too (was hardcoded to 2.6.*).
1029 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1030 of hook results and that also made display of other errors suboptimal
1031 was fixed; the code behaves now like 2.4 and earlier.
1033 Another change in 2.5, the cleanup of the OS scripts environment, is too
1034 aggressive: it removed even the ``PATH`` variable, which requires the OS
1035 scripts to *always* need to export it. Since this is a bit too strict,
1036 we now export a minimal PATH, the same that we export for hooks.
1038 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1039 integrated into this release.
1041 Finally, a few other miscellaneous changes were done (no new features,
1042 just small improvements):
1044 - Fix ``gnt-group --help`` display
1045 - Fix hardcoded Xen kernel path
1046 - Fix grow-disk handling of invalid units
1047 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1048 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1054 *(Released Thu, 12 Apr 2012)*
1056 Incompatible/important changes and bugfixes
1057 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1059 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1060 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1061 to match the underlying LUXI interface.
1062 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1063 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1064 not maintain backwards-compatibility as the underlying operation
1065 changed in an incompatible way. The RAPI client can talk to old
1066 servers, but it needs to be told so as the return value changed.
1067 - When creating file-based instances via RAPI, the ``file_driver``
1068 parameter no longer defaults to ``loop`` and must be specified.
1069 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1071 - Support for the undocumented and deprecated RAPI instance creation
1072 request format version 0 has been dropped. Use version 1, supported
1073 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1074 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1075 documentation <install>`.
1076 - The "cluster-verify" hooks are now executed per group by the
1077 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1078 if you just run ``gnt-cluster verify``, which generates one opcode per
1080 - The environment as passed to the OS scripts is cleared, and thus no
1081 environment variables defined in the node daemon's environment will be
1082 inherited by the scripts.
1083 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1085 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1086 support operations involving multiple node groups.
1087 - Offline nodes are ignored when failing over an instance.
1088 - Support for KVM version 1.0, which changed the version reporting format
1090 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1092 - ``Makefile`` is now compatible with Automake 1.11.2
1093 - Includes all bugfixes made in the 2.4 series
1098 - The ganeti-htools project has been merged into the ganeti-core source
1099 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1100 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1101 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1102 GPT-style partition tables (requires `parted
1103 <http://www.gnu.org/s/parted/>`_).
1104 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1105 - Allowed adding tags on instance creation.
1106 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1107 - Allow instances to be started in a paused state, enabling the user to
1108 see the complete console output on boot using the console.
1109 - Added new hypervisor flag to control default reboot behaviour
1110 (``reboot_behavior``).
1111 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1112 - Improved out-of-band management support:
1114 - Added ``gnt-node health`` command reporting the health status of
1116 - Added ``gnt-node power`` command to manage power status of nodes.
1117 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1119 - Instance migration can fall back to failover if instance is not
1121 - Filters can be used when listing nodes, instances, groups and locks;
1122 see :manpage:`ganeti(7)` manpage.
1123 - Added post-execution status as variables to :doc:`hooks <hooks>`
1125 - Instance tags are exported/imported together with the instance.
1126 - When given an explicit job ID, ``gnt-job info`` will work for archived
1128 - Jobs can define dependencies on other jobs (not yet supported via
1129 RAPI or command line, but used by internal commands and usable via
1132 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1135 - Instance failover is now available as a RAPI resource
1136 (``/2/instances/[instance_name]/failover``).
1137 - ``gnt-instance info`` defaults to static information if primary node
1139 - Opcodes have a new ``comment`` attribute.
1140 - Added basic SPICE support to KVM hypervisor.
1141 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1143 Node group improvements
1144 ~~~~~~~~~~~~~~~~~~~~~~~
1146 - ``gnt-cluster verify`` has been modified to check groups separately,
1147 thereby improving performance.
1148 - Node group support has been added to ``gnt-cluster verify-disks``,
1149 which now operates per node group.
1150 - Watcher has been changed to work better with node groups.
1152 - One process and state file per node group.
1153 - Slow watcher in one group doesn't block other group's watcher.
1155 - Added new command, ``gnt-group evacuate``, to move all instances in a
1156 node group to other groups.
1157 - Added ``gnt-instance change-group`` to move an instance to another
1159 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1160 per-group operations.
1161 - Node groups can be tagged.
1162 - Some operations switch from an exclusive to a shared lock as soon as
1164 - Instance's primary and secondary nodes' groups are now available as
1165 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1166 and ``snodes.group.uuid``).
1171 - Numerous updates to documentation and manpages.
1173 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1175 - Some opcode/job results are now also documented, see :doc:`RAPI
1178 - A lockset's internal lock is now also visible in lock monitor.
1179 - Log messages from job queue workers now contain information about the
1180 opcode they're processing.
1181 - ``gnt-instance console`` no longer requires the instance lock.
1182 - A short delay when waiting for job changes reduces the number of LUXI
1183 requests significantly.
1184 - DRBD metadata volumes are overwritten with zeros during disk creation.
1185 - Out-of-band commands no longer acquire the cluster lock in exclusive
1187 - ``devel/upload`` now uses correct permissions for directories.
1193 *(Released Fri, 23 Mar 2012)*
1195 This was the sixth release candidate of the 2.5 series.
1201 *(Released Mon, 9 Jan 2012)*
1203 This was the fifth release candidate of the 2.5 series.
1209 *(Released Thu, 27 Oct 2011)*
1211 This was the fourth release candidate of the 2.5 series.
1217 *(Released Wed, 26 Oct 2011)*
1219 This was the third release candidate of the 2.5 series.
1225 *(Released Tue, 18 Oct 2011)*
1227 This was the second release candidate of the 2.5 series.
1233 *(Released Tue, 4 Oct 2011)*
1235 This was the first release candidate of the 2.5 series.
1241 *(Released Wed, 31 Aug 2011)*
1243 This was the third beta release of the 2.5 series.
1249 *(Released Mon, 22 Aug 2011)*
1251 This was the second beta release of the 2.5 series.
1257 *(Released Fri, 12 Aug 2011)*
1259 This was the first beta release of the 2.5 series.
1265 *(Released Thu, 27 Oct 2011)*
1267 - Fixed bug when parsing command line parameter values ending in
1269 - Fixed assertion error after unclean master shutdown
1270 - Disable HTTP client pool for RPC, significantly reducing memory usage
1272 - Fixed queue archive creation with wrong permissions
1278 *(Released Tue, 23 Aug 2011)*
1282 - Fixed documentation for importing with ``--src-dir`` option
1283 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1284 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1290 *(Released Fri, 5 Aug 2011)*
1292 Many bug-fixes and a few small features:
1294 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1295 issues when you tried to add an instance with two MAC addresses in one
1297 - KVM: fixed per-instance stored UID value
1298 - KVM: configure bridged NICs at migration start
1299 - KVM: Fix a bug where instance will not start with never KVM versions
1301 - Added OS search path to ``gnt-cluster info``
1302 - Fixed an issue with ``file_storage_dir`` where you were forced to
1303 provide an absolute path, but the documentation states it is a
1304 relative path, the documentation was right
1305 - Added a new parameter to instance stop/start called ``--no-remember``
1306 that will make the state change to not be remembered
1307 - Implemented ``no_remember`` at RAPI level
1308 - Improved the documentation
1309 - Node evacuation: don't call IAllocator if node is already empty
1310 - Fixed bug in DRBD8 replace disks on current nodes
1311 - Fixed bug in recreate-disks for DRBD instances
1312 - Moved assertion checking locks in ``gnt-instance replace-disks``
1313 causing it to abort with not owning the right locks for some situation
1314 - Job queue: Fixed potential race condition when cancelling queued jobs
1315 - Fixed off-by-one bug in job serial generation
1316 - ``gnt-node volumes``: Fix instance names
1317 - Fixed aliases in bash completion
1318 - Fixed a bug in reopening log files after being sent a SIGHUP
1319 - Added a flag to burnin to allow specifying VCPU count
1320 - Bugfixes to non-root Ganeti configuration
1326 *(Released Thu, 12 May 2011)*
1328 Many bug-fixes and a few new small features:
1330 - Fixed a bug related to log opening failures
1331 - Fixed a bug in instance listing with orphan instances
1332 - Fixed a bug which prevented resetting the cluster-level node parameter
1333 ``oob_program`` to the default
1334 - Many fixes related to the ``cluster-merge`` tool
1335 - Fixed a race condition in the lock monitor, which caused failures
1336 during (at least) creation of many instances in parallel
1337 - Improved output for gnt-job info
1338 - Removed the quiet flag on some ssh calls which prevented debugging
1340 - Improved the N+1 failure messages in cluster verify by actually
1341 showing the memory values (needed and available)
1342 - Increased lock attempt timeouts so that when executing long operations
1343 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1344 too early and thus prevent the use of the 'fair' mechanism
1345 - Changed instance query data (``gnt-instance info``) to not acquire
1346 locks unless needed, thus allowing its use on locked instance if only
1347 static information is asked for
1348 - Improved behaviour with filesystems that do not support rename on an
1350 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1351 kept locks on all nodes during the wipe, which is unneeded
1352 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1353 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1354 10GiB) which caused the wipe to fail right at the end in some cases
1355 - Fixed master IP activation when doing master failover with no-voting
1356 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1357 the master node itself
1358 - Fixed potential data-loss in under disk full conditions, where Ganeti
1359 wouldn't check correctly the return code and would consider
1360 partially-written files 'correct'
1361 - Fixed bug related to multiple VGs and DRBD disk replacing
1362 - Added new disk parameter ``metavg`` that allows placement of the meta
1363 device for DRBD in a different volume group
1364 - Fixed error handling in the node daemon when the system libc doesn't
1365 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1366 - Fixed lock release during replace-disks, which kept cluster-wide locks
1367 when doing disk replaces with an iallocator script
1368 - Added check for missing bridges in cluster verify
1369 - Handle EPIPE errors while writing to the terminal better, so that
1370 piping the output to e.g. ``less`` doesn't cause a backtrace
1371 - Fixed rare case where a ^C during Luxi calls could have been
1372 interpreted as server errors, instead of simply terminating
1373 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1375 - Added a few more parameters to the KVM hypervisor, allowing a second
1376 CDROM, custom disk type for CDROMs and a floppy image
1377 - Removed redundant message in instance rename when the name is given
1379 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1380 disks on new nodes, allowing recreation when the original instance
1381 nodes are completely gone
1382 - Added option when converting disk templates to DRBD to skip waiting
1383 for the resync, in order to make the instance available sooner
1384 - Added two new variables to the OS scripts environment (containing the
1386 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1387 to allow use of ``pvgrub`` as bootloader
1388 - Changed the instance memory modifications to only check out-of-memory
1389 conditions on memory increases, and turned the secondary node warnings
1390 into errors (they can still be overridden via ``--force``)
1391 - Fixed the handling of a corner case when the Python installation gets
1392 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1393 to execute a command that doesn't exist
1394 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1395 node of the instance returned failures during instance shutdown; this
1396 adds the option ``--ignore-consistency`` to gnt-instance move
1398 And as usual, various improvements to the error messages, documentation
1405 *(Released Wed, 09 Mar 2011)*
1407 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1408 the RAPI users file if run twice (even with ``--dry-run``).
1410 The release fixes that bug (nothing else changed).
1416 *(Released Mon, 07 Mar 2011)*
1418 Final 2.4.0 release. Just a few small fixes:
1420 - Fixed RAPI node evacuate
1421 - Fixed the kvm-ifup script
1422 - Fixed internal error handling for special job cases
1423 - Updated man page to specify the escaping feature for options
1429 *(Released Mon, 28 Feb 2011)*
1431 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1432 that this feature wiped the disks of the wrong instance, leading to loss
1437 - Fixed title of query field containing instance name
1438 - Expanded the glossary in the documentation
1439 - Fixed one unittest (internal issue)
1445 *(Released Mon, 21 Feb 2011)*
1447 A number of bug fixes plus just a couple functionality changes.
1449 On the user-visible side, the ``gnt-* list`` command output has changed
1450 with respect to "special" field states. The current rc1 style of display
1451 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1452 the default output mode special fields states are displayed as follows:
1454 - Offline resource: ``*``
1455 - Unavailable/not applicable: ``-``
1456 - Data missing (RPC failure): ``?``
1457 - Unknown field: ``??``
1459 Another user-visible change is the addition of ``--force-join`` to
1464 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1465 - Fixed regression in RAPI/instance reinstall where all parameters were
1466 required (instead of optional)
1467 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1468 - Fixed iallocator usage (offline nodes were not considered offline)
1469 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1470 - Fixed hypervisor and OS parameter validation with respect to
1471 non-vm_capable nodes
1472 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1474 - Fixed ``tools/listrunner`` with respect to agent-based usage
1480 *(Released Fri, 4 Feb 2011)*
1482 Many changes and fixes since the beta1 release. While there were some
1483 internal changes, the code has been mostly stabilised for the RC
1486 Note: the dumb allocator was removed in this release, as it was not kept
1487 up-to-date with the IAllocator protocol changes. It is recommended to
1488 use the ``hail`` command from the ganeti-htools package.
1490 Note: the 2.4 and up versions of Ganeti are not compatible with the
1491 0.2.x branch of ganeti-htools. You need to upgrade to
1492 ganeti-htools-0.3.0 (or later).
1494 Regressions fixed from 2.3
1495 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1497 - Fixed the ``gnt-cluster verify-disks`` command
1498 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1500 - Fixed disk adoption breakage
1501 - Fixed wrong headers in instance listing for field aliases
1506 - Fixed corner case in KVM handling of NICs
1507 - Fixed many cases of wrong handling of non-vm_capable nodes
1508 - Fixed a bug where a missing instance symlink was not possible to
1509 recreate with any ``gnt-*`` command (now ``gnt-instance
1510 activate-disks`` does it)
1511 - Fixed the volume group name as reported by ``gnt-cluster
1513 - Increased timeouts for the import-export code, hopefully leading to
1514 fewer aborts due network or instance timeouts
1515 - Fixed bug in ``gnt-node list-storage``
1516 - Fixed bug where not all daemons were started on cluster
1517 initialisation, but only at the first watcher run
1518 - Fixed many bugs in the OOB implementation
1519 - Fixed watcher behaviour in presence of instances with offline
1521 - Fixed instance list output for instances running on the wrong node
1522 - a few fixes to the cluster-merge tool, but it still cannot merge
1523 multi-node groups (currently it is not recommended to use this tool)
1529 - Improved network configuration for the KVM hypervisor
1530 - Added e1000 as a supported NIC for Xen-HVM
1531 - Improved the lvmstrap tool to also be able to use partitions, as
1532 opposed to full disks
1533 - Improved speed of disk wiping (the cluster parameter
1534 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1535 of instance creations
1536 - Added documentation for the OS parameters
1537 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1538 hypervisor is not responding
1539 - Added display of blacklisted and hidden OS information in
1540 ``gnt-cluster info``
1541 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1542 NIC and node parameters, which might create problems with currently
1543 invalid (but undetected) configuration files, but prevents validation
1544 failures when unrelated parameters are modified
1545 - Changed cluster initialisation to wait for the master daemon to become
1547 - Expanded the RAPI interface:
1549 - Added config redistribution resource
1550 - Added activation/deactivation of instance disks
1551 - Added export of console information
1553 - Implemented log file reopening on SIGHUP, which allows using
1554 logrotate(8) for the Ganeti log files
1555 - Added a basic OOB helper script as an example
1561 *(Released Fri, 14 Jan 2011)*
1566 - Fixed timezone issues when formatting timestamps
1567 - Added support for node groups, available via ``gnt-group`` and other
1569 - Added out-of-band framework and management, see :doc:`design
1570 document <design-oob>`
1571 - Removed support for roman numbers from ``gnt-node list`` and
1572 ``gnt-instance list``.
1573 - Allowed modification of master network interface via ``gnt-cluster
1574 modify --master-netdev``
1575 - Accept offline secondaries while shutting down instance disks
1576 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1577 - Added support for multiple LVM volume groups
1578 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1580 - Added commands to list available fields:
1582 - ``gnt-node list-fields``
1583 - ``gnt-group list-fields``
1584 - ``gnt-instance list-fields``
1586 - Updated documentation and man pages
1591 - Moved ``rapi_users`` file into separate directory, now named
1592 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1594 - Added new tool for running commands on many machines,
1595 ``tools/ganeti-listrunner``
1596 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1597 improving the ``gnt-instance console`` output
1598 - Allowed customisation of disk index separator at ``configure`` time
1599 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1600 - Added support for non-partitioned md disks in ``lvmstrap``
1601 - Added script to gracefully power off KVM instances
1602 - Split ``utils`` module into smaller parts
1603 - Changed query operations to return more detailed information, e.g.
1604 whether an information is unavailable due to an offline node. To use
1605 this new functionality, the LUXI call ``Query`` must be used. Field
1606 information is now stored by the master daemon and can be retrieved
1607 using ``QueryFields``. Instances, nodes and groups can also be queried
1608 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1609 exposed via RAPI). The following commands make use of this
1610 infrastructure change:
1612 - ``gnt-group list``
1613 - ``gnt-group list-fields``
1615 - ``gnt-node list-fields``
1616 - ``gnt-instance list``
1617 - ``gnt-instance list-fields``
1618 - ``gnt-debug locks``
1623 - New RAPI resources (see :doc:`rapi`):
1627 - ``/2/groups/[group_name]``
1628 - ``/2/groups/[group_name]/assign-nodes``
1629 - ``/2/groups/[group_name]/modify``
1630 - ``/2/groups/[group_name]/rename``
1631 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1635 - Implemented ``no_install`` for instance creation
1636 - Implemented OS parameters for instance reinstallation, allowing
1637 use of special settings on reinstallation (e.g. for preserving data)
1642 - Added IPv6 support in import/export
1643 - Pause DRBD synchronization while wiping disks on instance creation
1644 - Updated unittests and QA scripts
1645 - Improved network parameters passed to KVM
1646 - Converted man pages from docbook to reStructuredText
1652 *(Released Mon, 20 Dec 2010)*
1654 Released version 2.3.1~rc1 without any changes.
1660 *(Released Wed, 1 Dec 2010)*
1662 - impexpd: Disable OpenSSL compression in socat if possible (backport
1663 from master, commit e90739d625b, see :doc:`installation guide
1664 <install-quick>` for details)
1665 - Changed unittest coverage report to exclude test scripts
1666 - Added script to check version format
1672 *(Released Wed, 1 Dec 2010)*
1674 Released version 2.3.0~rc1 without any changes.
1680 *(Released Fri, 19 Nov 2010)*
1682 A number of bugfixes and documentation updates:
1684 - Update ganeti-os-interface documentation
1685 - Fixed a bug related to duplicate MACs or similar items which should be
1687 - Fix breakage in OS state modify
1688 - Reinstall instance: disallow offline secondaries (fixes bug related to
1689 OS changing but reinstall failing)
1690 - plus all the other fixes between 2.2.1 and 2.2.2
1696 *(Released Tue, 2 Nov 2010)*
1698 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1699 - Fixed master failover race with watcher
1700 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1702 - Accept previously stopped instance for export with instance removal
1703 - Simplify and extend the environment variables for instance OS scripts
1704 - Added new node flags, ``master_capable`` and ``vm_capable``
1705 - Added optional instance disk wiping prior during allocation. This is a
1706 cluster-wide option and can be set/modified using
1707 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1708 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1709 :doc:`install-quick`
1710 - Added a new watcher option (``--ignore-pause``)
1711 - Added option to ignore offline node on instance start/stop
1712 (``--ignore-offline``)
1713 - Allow overriding OS parameters with ``gnt-instance reinstall``
1714 - Added ability to change node's secondary IP address using ``gnt-node
1716 - Implemented privilege separation for all daemons except
1717 ``ganeti-noded``, see ``configure`` options
1718 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1720 - Implemented job priorities (see ``ganeti(7)`` manpage)
1721 - Ignore failures while shutting down instances during failover from
1723 - Exit daemon's bootstrap process only once daemon is ready
1724 - Export more information via ``LUInstanceQuery``/remote API
1725 - Improved documentation, QA and unittests
1726 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1727 restart if the file was created or changed
1728 - Added LUXI protocol version sent with each request and response,
1729 allowing detection of server/client mismatches
1730 - Moved the Python scripts among gnt-* and ganeti-* into modules
1731 - Moved all code related to setting up SSH to an external script,
1733 - Infrastructure changes for node group support in future versions
1739 *(Released Fri, 19 Nov 2010)*
1741 A few small bugs fixed, and some improvements to the build system:
1743 - Fix documentation regarding conversion to drbd
1744 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1746 - Fix error handling in node modify with multiple changes
1747 - Allow remote imports without checked names
1753 *(Released Tue, 19 Oct 2010)*
1755 - Disable SSL session ID cache in RPC client
1761 *(Released Thu, 14 Oct 2010)*
1763 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1764 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1766 - Fix problems with interaction between Curl and Python's HTTP server,
1767 resulting in increased speed in many RPC calls
1768 - Improve our release script to prevent breakage with older aclocal and
1775 *(Released Thu, 7 Oct 2010)*
1777 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1778 value retrieved from master
1779 - Added support for blacklisted or hidden OS definitions
1780 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1781 - Added support for -mem-path in KVM hypervisor abstraction layer
1782 - Allow overriding instance parameters in tool for inter-cluster
1783 instance moves (``tools/move-instance``)
1784 - Improved opcode summaries (e.g. in ``gnt-job list``)
1785 - Improve consistency of OS listing by sorting it
1786 - Documentation updates
1792 *(Released Fri, 8 Oct 2010)*
1794 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1800 *(Released Mon, 4 Oct 2010)*
1802 - Fixed regression in ``gnt-instance rename``
1808 *(Released Wed, 22 Sep 2010)*
1810 - Fixed OS_VARIANT variable for OS scripts
1811 - Fixed cluster tag operations via RAPI
1812 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1813 - Disabled RAPI CA checks in watcher
1819 *(Released Mon, 23 Aug 2010)*
1821 - Support DRBD versions of the format "a.b.c.d"
1823 - Re-introduce support for usage from multiple threads in RAPI client
1824 - Instance renames and modify via RAPI
1825 - Work around race condition between processing and archival in job
1827 - Mark opcodes following failed one as failed, too
1828 - Job field ``lock_status`` was removed due to difficulties making it
1829 work with the changed job queue in Ganeti 2.2; a better way to monitor
1830 locks is expected for a later 2.2.x release
1831 - Fixed dry-run behaviour with many commands
1832 - Support ``ssh-agent`` again when adding nodes
1833 - Many additional bugfixes
1839 *(Released Fri, 30 Jul 2010)*
1841 Important change: the internal RPC mechanism between Ganeti nodes has
1842 changed from using a home-grown http library (based on the Python base
1843 libraries) to use the PycURL library. This requires that PycURL is
1844 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1845 against GnuTLS by default. cURL's support for GnuTLS had known issues
1846 before cURL 7.21.0 and we recommend using the latest cURL release or
1847 linking against OpenSSL. Most other distributions already link PycURL
1848 and cURL against OpenSSL. The command::
1850 python -c 'import pycurl; print pycurl.version'
1852 can be used to determine the libraries PycURL and cURL are linked
1855 Other significant changes:
1857 - Rewrote much of the internals of the job queue, in order to achieve
1858 better parallelism; this decouples job query operations from the job
1859 processing, and it should allow much nicer behaviour of the master
1860 daemon under load, and it also has uncovered some long-standing bugs
1861 related to the job serialisation (now fixed)
1862 - Added a default iallocator setting to the cluster parameters,
1863 eliminating the need to always pass nodes or an iallocator for
1864 operations that require selection of new node(s)
1865 - Added experimental support for the LXC virtualization method
1866 - Added support for OS parameters, which allows the installation of
1867 instances to pass parameter to OS scripts in order to customise the
1869 - Added a hypervisor parameter controlling the migration type (live or
1870 non-live), since hypervisors have various levels of reliability; this
1871 has renamed the 'live' parameter to 'mode'
1872 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1873 logical volumes, meaning that cluster verify will ignore them and not
1874 flag their presence as errors
1875 - The watcher will now reset the error count for failed instances after
1876 8 hours, thus allowing self-healing if the problem that caused the
1877 instances to be down/fail to start has cleared in the meantime
1878 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1879 check for, and warn, if the drbd module parameter ``usermode_helper``
1880 is not consistent with the cluster-wide setting; this is needed to
1881 make diagnose easier of failed drbd creations
1882 - Started adding base IPv6 support, but this is not yet
1883 enabled/available for use
1884 - Rename operations (cluster, instance) will now return the new name,
1885 which is especially useful if a short name was passed in
1886 - Added support for instance migration in RAPI
1887 - Added a tool to pre-configure nodes for the SSH setup, before joining
1888 them to the cluster; this will allow in the future a simplified model
1889 for node joining (but not yet fully enabled in 2.2); this needs the
1890 paramiko python library
1891 - Fixed handling of name-resolving errors
1892 - Fixed consistency of job results on the error path
1893 - Fixed master-failover race condition when executed multiple times in
1895 - Fixed many bugs related to the job queue (mostly introduced during the
1896 2.2 development cycle, so not all are impacting 2.1)
1897 - Fixed instance migration with missing disk symlinks
1898 - Fixed handling of unknown jobs in ``gnt-job archive``
1899 - And many other small fixes/improvements
1903 - Enhanced both the unittest and the QA coverage
1904 - Switched the opcode validation to a generic model, and extended the
1905 validation to all opcode parameters
1906 - Changed more parts of the code that write shell scripts to use the
1908 - Switched the master daemon to use the asyncore library for the Luxi
1915 *(Released Thu, 17 Jun 2010)*
1917 - Added tool (``move-instance``) and infrastructure to move instances
1918 between separate clusters (see :doc:`separate documentation
1919 <move-instance>` and :doc:`design document <design-2.2>`)
1920 - Added per-request RPC timeout
1921 - RAPI now requires a Content-Type header for requests with a body (e.g.
1922 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1923 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1924 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1926 - Implemented initial support for running Ganeti daemons as separate
1927 users, see configure-time flags ``--with-user-prefix`` and
1928 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1930 - Instances can be removed after export (``gnt-backup export
1931 --remove-instance``)
1932 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1933 key (instead of 1024 bit)
1934 - Added new cluster configuration file for cluster domain secret
1935 - Import/export now use SSL instead of SSH
1936 - Added support for showing estimated time when exporting an instance,
1937 see the ``ganeti-os-interface(7)`` manpage and look for
1944 *(Released Tue, 16 Nov 2010)*
1946 Some more bugfixes. Unless critical bugs occur, this will be the last
1949 - Fix case of MAC special-values
1950 - Fix mac checker regex
1951 - backend: Fix typo causing "out of range" error
1952 - Add missing --units in gnt-instance list man page
1958 *(Released Tue, 24 Aug 2010)*
1961 - Don't ignore secondary node silently on non-mirrored disk templates
1963 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1964 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1965 - Properly document the usb_mouse parameter
1966 - Fix path in ganeti-rapi(8) (issue 116)
1967 - Adjust error message when the ganeti user's .ssh directory is
1969 - Add same-node-check when changing the disk template to drbd
1975 *(Released Fri, 16 Jul 2010)*
1978 - Add an option to only select some reboot types during qa/burnin.
1979 (on some hypervisors consequent reboots are not supported)
1980 - Fix infrequent race condition in master failover. Sometimes the old
1981 master ip address would be still detected as up for a short time
1982 after it was removed, causing failover to fail.
1983 - Decrease mlockall warnings when the ctypes module is missing. On
1984 Python 2.4 we support running even if no ctypes module is installed,
1985 but we were too verbose about this issue.
1986 - Fix building on old distributions, on which man doesn't have a
1988 - Fix RAPI not to ignore the MAC address on instance creation
1989 - Implement the old instance creation format in the RAPI client.
1995 *(Released Thu, 01 Jul 2010)*
1997 A small bugfix release:
1998 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1999 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2000 a non-existing option
2001 - Fix instance create: the --force-variant option was ignored
2002 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2003 - Fix modify node storage with non-FQDN arguments
2004 - Fix RAPI client to authenticate under Python 2.6 when used
2005 for more than 5 requests needing authentication
2006 - Fix gnt-instance modify -t (storage) giving a wrong error message
2007 when converting a non-shutdown drbd instance to plain
2013 *(Released Fri, 18 Jun 2010)*
2015 A small bugfix release:
2017 - Fix live migration of KVM instances started with older Ganeti
2018 versions which had fewer hypervisor parameters
2019 - Fix gnt-instance grow-disk on down instances
2020 - Fix an error-reporting bug during instance migration
2021 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2022 silently ignoring broken ones
2023 - Fix an RPC error reporting bug affecting, for example, RAPI client
2025 - Fix bug triggered by different API version os-es on different nodes
2026 - Fix a bug in instance startup with custom hvparams: OS level
2027 parameters would fail to be applied.
2028 - Fix the RAPI client under Python 2.6 (but more work is needed to
2029 make it work completely well with OpenSSL)
2030 - Fix handling of errors when resolving names from DNS
2036 *(Released Thu, 3 Jun 2010)*
2038 A medium sized development cycle. Some new features, and some
2039 fixes/small improvements/cleanups.
2041 Significant features
2042 ~~~~~~~~~~~~~~~~~~~~
2044 The node deamon now tries to mlock itself into memory, unless the
2045 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2046 its logs, and falls back to console logging. This allows emergency
2047 features such as ``gnt-node powercycle`` to work even in the event of a
2048 broken node disk (tested offlining the disk hosting the node's
2049 filesystem and dropping its memory caches; don't try this at home)
2051 KVM: add vhost-net acceleration support. It can be tested with a new
2052 enough version of the kernel and of qemu-kvm.
2054 KVM: Add instance chrooting feature. If you use privilege dropping for
2055 your VMs you can also now force them to chroot to an empty directory,
2056 before starting the emulated guest.
2058 KVM: Add maximum migration bandwith and maximum downtime tweaking
2059 support (requires a new-enough version of qemu-kvm).
2061 Cluster verify will now warn if the master node doesn't have the master
2062 ip configured on it.
2064 Add a new (incompatible) instance creation request format to RAPI which
2065 supports all parameters (previously only a subset was supported, and it
2066 wasn't possible to extend the old format to accomodate all the new
2067 features. The old format is still supported, and a client can check for
2068 this feature, before using it, by checking for its presence in the
2069 ``features`` RAPI resource.
2071 Now with ancient latin support. Try it passing the ``--roman`` option to
2072 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2073 (requires the python-roman module to be installed, in order to work).
2078 As usual many internal code refactorings, documentation updates, and
2081 - Lots of improvements and cleanups to the experimental Remote API
2082 (RAPI) client library.
2083 - A new unit test suite for the core daemon libraries.
2084 - A fix to creating missing directories makes sure the umask is not
2085 applied anymore. This enforces the same directory permissions
2087 - Better handling terminating daemons with ctrl+c (used when running
2088 them in debugging mode).
2089 - Fix a race condition in live migrating a KVM instance, when stat()
2090 on the old proc status file returned EINVAL, which is an unexpected
2092 - Fixed manpage checking with newer man and utf-8 charachters. But now
2093 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2099 *(Released Fri, 7 May 2010)*
2101 Fix a bug which prevented untagged KVM instances from starting.
2107 *(Released Fri, 7 May 2010)*
2109 Another release with a long development cycle, during which many
2110 different features were added.
2112 Significant features
2113 ~~~~~~~~~~~~~~~~~~~~
2115 The KVM hypervisor now can run the individual instances as non-root, to
2116 reduce the impact of a VM being hijacked due to bugs in the
2117 hypervisor. It is possible to run all instances as a single (non-root)
2118 user, to manually specify a user for each instance, or to dynamically
2119 allocate a user out of a cluster-wide pool to each instance, with the
2120 guarantee that no two instances will run under the same user ID on any
2123 An experimental RAPI client library, that can be used standalone
2124 (without the other Ganeti libraries), is provided in the source tree as
2125 ``lib/rapi/client.py``. Note this client might change its interface in
2126 the future, as we iterate on its capabilities.
2128 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2129 replace the cluster's certificates and crypto keys. This might help in
2130 case they have been compromised, or have simply expired.
2132 A new disk option for instance creation has been added that allows one
2133 to "adopt" currently existing logical volumes, with data
2134 preservation. This should allow easier migration to Ganeti from
2135 unmanaged (or managed via other software) instances.
2137 Another disk improvement is the possibility to convert between redundant
2138 (DRBD) and plain (LVM) disk configuration for an instance. This should
2139 allow better scalability (starting with one node and growing the
2140 cluster, or shrinking a two-node cluster to one node).
2142 A new feature that could help with automated node failovers has been
2143 implemented: if a node sees itself as offline (by querying the master
2144 candidates), it will try to shutdown (hard) all instances and any active
2145 DRBD devices. This reduces the risk of duplicate instances if an
2146 external script automatically failovers the instances on such nodes. To
2147 enable this, the cluster parameter ``maintain_node_health`` should be
2148 enabled; in the future this option (per the name) will enable other
2149 automatic maintenance features.
2151 Instance export/import now will reuse the original instance
2152 specifications for all parameters; that means exporting an instance,
2153 deleting it and the importing it back should give an almost identical
2154 instance. Note that the default import behaviour has changed from
2155 before, where it created only one NIC; now it recreates the original
2158 Cluster verify has added a few new checks: SSL certificates validity,
2159 /etc/hosts consistency across the cluster, etc.
2164 As usual, many internal changes were done, documentation fixes,
2167 - Fixed cluster initialization with disabled cluster storage (regression
2168 introduced in 2.1.1)
2169 - File-based storage supports growing the disks
2170 - Fixed behaviour of node role changes
2171 - Fixed cluster verify for some corner cases, plus a general rewrite of
2172 cluster verify to allow future extension with more checks
2173 - Fixed log spamming by watcher and node daemon (regression introduced
2175 - Fixed possible validation issues when changing the list of enabled
2177 - Fixed cleanup of /etc/hosts during node removal
2178 - Fixed RAPI response for invalid methods
2179 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2180 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2181 from ide disks, etc.)
2182 - Allow OS changes without re-installation (to record a changed OS
2183 outside of Ganeti, or to allow OS renames)
2184 - Allow instance creation without OS installation (useful for example if
2185 the OS will be installed manually, or restored from a backup not in
2187 - Implemented option to make cluster ``copyfile`` use the replication
2189 - Added list of enabled hypervisors to ssconf (possibly useful for
2191 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2193 - A partial form of node re-IP is possible via node readd, which now
2194 allows changed node primary IP
2195 - Command line utilities now show an informational message if the job is
2197 - The logs of the master daemon now show the PID/UID/GID of the
2204 *(Released Fri, 12 Mar 2010)*
2206 During the 2.1.0 long release candidate cycle, a lot of improvements and
2207 changes have accumulated with were released later as 2.1.1.
2212 The node evacuate command (``gnt-node evacuate``) was significantly
2213 rewritten, and as such the IAllocator protocol was changed - a new
2214 request type has been added. This unfortunate change during a stable
2215 series is designed to improve performance of node evacuations; on
2216 clusters with more than about five nodes and which are well-balanced,
2217 evacuation should proceed in parallel for all instances of the node
2218 being evacuated. As such, any existing IAllocator scripts need to be
2219 updated, otherwise the above command will fail due to the unknown
2220 request. The provided "dumb" allocator has not been updated; but the
2221 ganeti-htools package supports the new protocol since version 0.2.4.
2223 Another important change is increased validation of node and instance
2224 names. This might create problems in special cases, if invalid host
2225 names are being used.
2227 Also, a new layer of hypervisor parameters has been added, that sits at
2228 OS level between the cluster defaults and the instance ones. This allows
2229 customisation of virtualization parameters depending on the installed
2230 OS. For example instances with OS 'X' may have a different KVM kernel
2231 (or any other parameter) than the cluster defaults. This is intended to
2232 help managing a multiple OSes on the same cluster, without manual
2233 modification of each instance's parameters.
2235 A tool for merging clusters, ``cluster-merge``, has been added in the
2236 tools sub-directory.
2241 - Improved the int/float conversions that should make the code more
2242 robust in face of errors from the node daemons
2243 - Fixed the remove node code in case of internal configuration errors
2244 - Fixed the node daemon behaviour in face of inconsistent queue
2245 directory (e.g. read-only file-system where we can't open the files
2247 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2248 now it either aborts cleanly or, if given the new "auto_promote"
2249 parameter, will automatically promote other nodes as needed
2250 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2251 completely prevent Ganeti from working
2252 - Fixed bug for instance export when not all disks were successfully
2254 - Fixed behaviour of node add when the new node is slow in starting up
2256 - Fixed handling of signals in the LUXI client, which should improve
2257 behaviour of command-line scripts
2258 - Added checks for invalid node/instance names in the configuration (now
2259 flagged during cluster verify)
2260 - Fixed watcher behaviour for disk activation errors
2261 - Fixed two potentially endless loops in http library, which led to the
2262 RAPI daemon hanging and consuming 100% CPU in some cases
2263 - Fixed bug in RAPI daemon related to hashed passwords
2264 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2266 - Enhanced compatibility with non-Debian OSes, but not using absolute
2267 path in some commands and allowing customisation of the ssh
2268 configuration directory
2269 - Fixed possible future issue with new Python versions by abiding to the
2270 proper use of ``__slots__`` attribute on classes
2271 - Added checks that should prevent directory traversal attacks
2272 - Many documentation fixes based on feedback from users
2277 - Added an "early_release" more for instance replace disks and node
2278 evacuate, where we release locks earlier and thus allow higher
2279 parallelism within the cluster
2280 - Added watcher hooks, intended to allow the watcher to restart other
2281 daemons (e.g. from the ganeti-nbma project), but they can be used of
2282 course for any other purpose
2283 - Added a compile-time disable for DRBD barriers, to increase
2284 performance if the administrator trusts the power supply or the
2285 storage system to not lose writes
2286 - Added the option of using syslog for logging instead of, or in
2287 addition to, Ganeti's own log files
2288 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2289 can indeed boot from a paravirtual NIC
2290 - Added a generic debug level for many operations; while this is not
2291 used widely yet, it allows one to pass the debug value all the way to
2293 - Enhanced the hooks environment for instance moves (failovers,
2294 migrations) where the primary/secondary nodes changed during the
2295 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2296 - Enhanced data validations for many user-supplied values; one important
2297 item is the restrictions imposed on instance and node names, which
2298 might reject some (invalid) host names
2299 - Add a configure-time option to disable file-based storage, if it's not
2300 needed; this allows greater security separation between the master
2301 node and the other nodes from the point of view of the inter-node RPC
2303 - Added user notification in interactive tools if job is waiting in the
2304 job queue or trying to acquire locks
2305 - Added log messages when a job is waiting for locks
2306 - Added filtering by node tags in instance operations which admit
2307 multiple instances (start, stop, reboot, reinstall)
2308 - Added a new tool for cluster mergers, ``cluster-merge``
2309 - Parameters from command line which are of the form ``a=b,c=d`` can now
2310 use backslash escapes to pass in values which contain commas,
2311 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2313 - For KVM, the instance name is the first parameter passed to KVM, so
2314 that it's more visible in the process list
2320 *(Released Tue, 2 Mar 2010)*
2322 Ganeti 2.1 brings many improvements with it. Major changes:
2324 - Added infrastructure to ease automated disk repairs
2325 - Added new daemon to export configuration data in a cheaper way than
2326 using the remote API
2327 - Instance NICs can now be routed instead of being associated with a
2329 - Improved job locking logic to reduce impact of jobs acquiring multiple
2330 locks waiting for other long-running jobs
2332 In-depth implementation details can be found in the Ganeti 2.1 design
2338 - Added chroot hypervisor
2339 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2341 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2342 ``bootloader_path`` and ``bootloader_args``)
2343 - Added the ``use_localtime`` option for the xen-hvm and kvm
2344 hypervisors, and the default value for this has changed to false (in
2345 2.0 xen-hvm always enabled it)
2346 - Added luxi call to submit multiple jobs in one go
2347 - Added cluster initialization option to not modify ``/etc/hosts``
2349 - Added network interface parameters
2350 - Added dry run mode to some LUs
2351 - Added RAPI resources:
2353 - ``/2/instances/[instance_name]/info``
2354 - ``/2/instances/[instance_name]/replace-disks``
2355 - ``/2/nodes/[node_name]/evacuate``
2356 - ``/2/nodes/[node_name]/migrate``
2357 - ``/2/nodes/[node_name]/role``
2358 - ``/2/nodes/[node_name]/storage``
2359 - ``/2/nodes/[node_name]/storage/modify``
2360 - ``/2/nodes/[node_name]/storage/repair``
2362 - Added OpCodes to evacuate or migrate all instances on a node
2363 - Added new command to list storage elements on nodes (``gnt-node
2364 list-storage``) and modify them (``gnt-node modify-storage``)
2365 - Added new ssconf files with master candidate IP address
2366 (``ssconf_master_candidates_ips``), node primary IP address
2367 (``ssconf_node_primary_ips``) and node secondary IP address
2368 (``ssconf_node_secondary_ips``)
2369 - Added ``ganeti-confd`` and a client library to query the Ganeti
2370 configuration via UDP
2371 - Added ability to run hooks after cluster initialization and before
2373 - Added automatic mode for disk replace (``gnt-instance replace-disks
2375 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2376 after catastrophic data-loss
2377 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2379 - Added ``gnt-instance move`` command to move instances
2380 - Added ``gnt-cluster watcher`` command to control watcher
2381 - Added ``gnt-node powercycle`` command to powercycle nodes
2382 - Added new job status field ``lock_status``
2383 - Added parseable error codes to cluster verification (``gnt-cluster
2384 verify --error-codes``) and made output less verbose (use
2385 ``--verbose`` to restore previous behaviour)
2386 - Added UUIDs to the main config entities (cluster, nodes, instances)
2387 - Added support for OS variants
2388 - Added support for hashed passwords in the Ganeti remote API users file
2390 - Added option to specify maximum timeout on instance shutdown
2391 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2392 - Added new helper script to start and stop Ganeti daemons
2393 (``daemon-util``), with the intent to reduce the work necessary to
2394 adjust Ganeti for non-Debian distributions and to start/stop daemons
2396 - Added more unittests
2397 - Fixed critical bug in ganeti-masterd startup
2398 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2399 now customisable at the cluster level for both the KVM and Xen
2400 hypervisors using the new ``migration_port`` parameter
2401 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2402 reinstalling an instance
2403 - Allowed ``@`` in tag names
2404 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2405 - Many documentation updates
2406 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2407 - ``gnt-instance reinstall`` can now reinstall multiple instances
2408 - Updated many command line parameters
2409 - Introduced new OS API version 15
2410 - No longer support a default hypervisor
2411 - Treat virtual LVs as inexistent
2412 - Improved job locking logic to reduce lock contention
2413 - Match instance and node names case insensitively
2414 - Reimplemented bash completion script to be more complete
2421 *(Released Thu, 4 Feb 2010)*
2423 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2424 - Fix a string formatting bug
2425 - Improve safety of the code in some error paths
2426 - Improve data validation in the master of values returned from nodes
2432 *(Released Thu, 17 Dec 2009)*
2434 - Fix security issue due to missing validation of iallocator names; this
2435 allows local and remote execution of arbitrary executables
2436 - Fix failure of gnt-node list during instance removal
2437 - Ship the RAPI documentation in the archive
2443 *(Released Wed, 30 Sep 2009)*
2445 - Fixed many wrong messages
2446 - Fixed a few bugs related to the locking library
2447 - Fixed MAC checking at instance creation time
2448 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2449 - Fixed a few issues related to signal handling in both daemons and
2451 - Fixed the example startup script provided
2452 - Fixed insserv dependencies in the example startup script (patch from
2454 - Fixed handling of drained nodes in the iallocator framework
2455 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2457 - Fixed error related to invalid job IDs in job polling
2458 - Fixed job/opcode persistence on unclean master shutdown
2459 - Fixed handling of partial job processing after unclean master
2461 - Fixed error reporting from LUs, previously all errors were converted
2462 into execution errors
2463 - Fixed error reporting from burnin
2464 - Decreased significantly the memory usage of the job queue
2465 - Optimised slightly multi-job submission
2466 - Optimised slightly opcode loading
2467 - Backported the multi-job submit framework from the development
2468 branch; multi-instance start and stop should be faster
2469 - Added script to clean archived jobs after 21 days; this will reduce
2470 the size of the queue directory
2471 - Added some extra checks in disk size tracking
2472 - Added an example ethers hook script
2473 - Added a cluster parameter that prevents Ganeti from modifying of
2475 - Added more node information to RAPI responses
2476 - Added a ``gnt-job watch`` command that allows following the ouput of a
2478 - Added a bind-address option to ganeti-rapi
2479 - Added more checks to the configuration verify
2480 - Enhanced the burnin script such that some operations can be retried
2482 - Converted instance reinstall to multi-instance model
2488 *(Released Fri, 7 Aug 2009)*
2490 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2491 to allow using the pre-2.0.2 behaviour in activation, if any existing
2492 instances have mismatched disk sizes in the configuration
2493 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2494 any configuration mismatches for disk sizes
2495 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2496 failover to work on two-node clusters
2497 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2499 - Fixed detection of OS script errors in ``gnt-backup export``
2500 - Fixed exit code of ``gnt-backup export``
2506 *(Released Fri, 17 Jul 2009)*
2508 - Added experimental support for stripped logical volumes; this should
2509 enhance performance but comes with a higher complexity in the block
2510 device handling; stripping is only enabled when passing
2511 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2512 affected even in the non-stripped mode
2513 - Improved resiliency against transient failures at the end of DRBD
2514 resyncs, and in general of DRBD resync checks
2515 - Fixed a couple of issues with exports and snapshot errors
2516 - Fixed a couple of issues in instance listing
2517 - Added display of the disk size in ``gnt-instance info``
2518 - Fixed checking for valid OSes in instance creation
2519 - Fixed handling of the "vcpus" parameter in instance listing and in
2520 general of invalid parameters
2521 - Fixed http server library, and thus RAPI, to handle invalid
2522 username/password combinations correctly; this means that now they
2523 report unauthorized for queries too, not only for modifications,
2524 allowing earlier detect of configuration problems
2525 - Added a new "role" node list field, equivalent to the master/master
2526 candidate/drained/offline flags combinations
2527 - Fixed cluster modify and changes of candidate pool size
2528 - Fixed cluster verify error messages for wrong files on regular nodes
2529 - Fixed a couple of issues with node demotion from master candidate role
2530 - Fixed node readd issues
2531 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2532 - Added a new ``--no-voting`` option for masterfailover to fix failover
2533 on two-nodes clusters when the former master node is unreachable
2534 - Added instance reinstall over RAPI
2540 *(Released Tue, 16 Jun 2009)*
2542 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2543 allow re-adding the start in single-user option (regression from 1.2)
2544 - the watcher writes the instance status to a file, to allow monitoring
2545 to report the instance status (from the master) based on cached
2546 results of the watcher's queries; while this can get stale if the
2547 watcher is being locked due to other work on the cluster, this is
2548 still an improvement
2549 - the watcher now also restarts the node daemon and the rapi daemon if
2551 - fixed the watcher to handle full and drained queue cases
2552 - hooks export more instance data in the environment, which helps if
2553 hook scripts need to take action based on the instance's properties
2554 (no longer need to query back into ganeti)
2555 - instance failovers when the instance is stopped do not check for free
2556 RAM, so that failing over a stopped instance is possible in low memory
2558 - rapi uses queries for tags instead of jobs (for less job traffic), and
2559 for cluster tags it won't talk to masterd at all but read them from
2561 - a couple of error handling fixes in RAPI
2562 - drbd handling: improved the error handling of inconsistent disks after
2563 resync to reduce the frequency of "there are some degraded disks for
2564 this instance" messages
2565 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2566 error handling path called a wrong function name)
2572 *(Released Wed, 27 May 2009)*
2574 - no changes from rc5
2580 *(Released Wed, 20 May 2009)*
2582 - fix a couple of bugs (validation, argument checks)
2583 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2584 - some small improvements to RAPI and IAllocator
2585 - make watcher automatically start the master daemon if down
2591 *(Released Mon, 27 Apr 2009)*
2593 - change the OS list to not require locks; this helps with big clusters
2594 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2595 volume group is broken
2596 - ``gnt-instance info``, without any arguments, doesn't run for all
2597 instances anymore; either pass ``--all`` or pass the desired
2598 instances; this helps against mistakes on big clusters where listing
2599 the information for all instances takes a long time
2600 - miscellaneous doc and man pages fixes
2606 *(Released Wed, 8 Apr 2009)*
2608 - Change the internal locking model of some ``gnt-node`` commands, in
2609 order to reduce contention (and blocking of master daemon) when
2610 batching many creation/reinstall jobs
2611 - Fixes to Xen soft reboot
2612 - No longer build documentation at build time, instead distribute it in
2613 the archive, in order to reduce the need for the whole docbook/rst
2620 *(Released Fri, 27 Mar 2009)*
2622 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2623 - Fix watcher startup sequence, improves the behaviour of busy clusters
2624 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2625 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2626 bind address and other places
2627 - Some documentation fixes and updates
2633 *(Released Mon, 2 Mar 2009)*
2635 - More documentation updates, now all docs should be more-or-less
2637 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2639 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2641 - Fix an issue related to $libdir/run/ganeti and cluster creation
2647 *(Released Thu, 19 Feb 2009)*
2649 - Xen PVM and KVM have switched the default value for the instance root
2650 disk to the first partition on the first drive, instead of the whole
2651 drive; this means that the OS installation scripts must be changed
2653 - Man pages have been updated
2654 - RAPI has been switched by default to HTTPS, and the exported functions
2655 should all work correctly
2656 - RAPI v1 has been removed
2657 - Many improvements to the KVM hypervisor
2658 - Block device errors are now better reported
2659 - Many other bugfixes and small improvements
2665 *(Released Mon, 26 Jan 2009)*
2667 - Version 2 is a general rewrite of the code and therefore the
2668 differences are too many to list, see the design document for 2.0 in
2669 the ``doc/`` subdirectory for more details
2670 - In this beta version there is not yet a migration path from 1.2 (there
2671 will be one in the final 2.0 release)
2672 - A few significant changes are:
2674 - all commands are executed by a daemon (``ganeti-masterd``) and the
2675 various ``gnt-*`` commands are just front-ends to it
2676 - all the commands are entered into, and executed from a job queue,
2677 see the ``gnt-job(8)`` manpage
2678 - the RAPI daemon supports read-write operations, secured by basic
2679 HTTP authentication on top of HTTPS
2680 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2681 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2682 to migrate to DRBD 8 first while still running Ganeti 1.2)
2683 - DRBD devices are using statically allocated minor numbers, which
2684 will be assigned to existing instances during the migration process
2685 - there is support for both Xen PVM and Xen HVM instances running on
2687 - KVM virtualization is supported too
2688 - file-based storage has been implemented, which means that it is
2689 possible to run the cluster without LVM and DRBD storage, for
2690 example using a shared filesystem exported from shared storage (and
2691 still have live migration)
2697 *(Released Tue, 13 Jan 2009)*
2699 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2700 - Reuse the old instance mac address by default on instance import, if
2701 the instance name is the same.
2702 - Handle situations in which the node info rpc returns incomplete
2704 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2705 - Improved version of batcher:
2707 - state file support
2708 - instance mac address support
2709 - support for HVM clusters/instances
2711 - Add an option to show the number of cpu sockets and nodes in
2713 - Support OSes that handle more than one version of the OS api (but do
2714 not change the current API in any other way)
2715 - Fix ``gnt-node migrate``
2716 - ``gnt-debug`` man page
2717 - Fixes various more typos and small issues
2718 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2724 *(Released Wed, 24 Sep 2008)*
2726 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2727 type of disk exported to fully virtualized instances.
2728 - provide access to the serial console of HVM instances
2729 - instance auto_balance flag, set by default. If turned off it will
2730 avoid warnings on cluster verify if there is not enough memory to fail
2731 over an instance. in the future it will prevent automatically failing
2732 it over when we will support that.
2733 - batcher tool for instance creation, see ``tools/README.batcher``
2734 - ``gnt-instance reinstall --select-os`` to interactively select a new
2735 operating system when reinstalling an instance.
2736 - when changing the memory amount on instance modify a check has been
2737 added that the instance will be able to start. also warnings are
2738 emitted if the instance will not be able to fail over, if auto_balance
2740 - documentation fixes
2741 - sync fields between ``gnt-instance list/modify/add/import``
2742 - fix a race condition in drbd when the sync speed was set after giving
2743 the device a remote peer.
2749 *(Released Tue, 22 Jul 2008)*
2751 - note: the allowed size and number of tags per object were reduced
2752 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2753 - fixed twisted 8.x compatibility
2754 - fixed ``gnt-instance replace-disks`` with iallocator
2755 - add TCP keepalives on twisted connections to detect restarted nodes
2756 - disk increase support, see ``gnt-instance grow-disk``
2757 - implement bulk node/instance query for RAPI
2758 - add tags in node/instance listing (optional)
2759 - experimental migration (and live migration) support, read the man page
2760 for ``gnt-instance migrate``
2761 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2762 has some small improvements in handling its state file
2768 *(Released Fri, 13 Jun 2008)*
2770 - Experimental readonly, REST-based remote API implementation;
2771 automatically started on master node, TCP port 5080, if enabled by
2772 ``--enable-rapi`` parameter to configure script.
2773 - Instance allocator support. Add and import instance accept a
2774 ``--iallocator`` parameter, and call that instance allocator to decide
2775 which node to use for the instance. The iallocator document describes
2776 what's expected from an allocator script.
2777 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2778 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2779 node is lost there is still enough memory to fail over the instances
2781 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2782 ``gnt-cluster verify``, to check for site-specific compliance. All the
2783 hooks will run, and their output, if any, will be displayed. Any
2784 failing hook will make the verification return an error value.
2785 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2786 primary and secondary interfaces
2787 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2788 that is still declared as part of the cluster and has failed.
2789 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2790 specifying output fields, that just adds the chosen fields to the
2792 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2793 export from the filesystem.
2794 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2795 have been added. Using them you can enable/disable acpi and pae
2796 support, and specify a path for a cd image to be exported to the
2797 instance. These parameters as the name suggest only work on HVM
2799 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2800 PAE support will be set to the previously hardcoded values, but the
2801 (previously hardcoded) path to the CDROM ISO image will be unset and
2802 if required, needs to be set manually with ``gnt-instance modify``
2804 - The address to which an instance's VNC console is bound is now
2805 selectable per-instance, rather than being cluster wide. Of course
2806 this only applies to instances controlled via VNC, so currently just
2807 applies to HVM clusters.
2813 *(Released Mon, 18 Feb 2008)*
2815 - more tweaks to the disk activation code (especially helpful for DRBD)
2816 - change the default ``gnt-instance list`` output format, now there is
2817 one combined status field (see the manpage for the exact values this
2819 - some more fixes for the mac export to hooks change
2820 - make Ganeti not break with DRBD 8.2.x (which changed the version
2821 format in ``/proc/drbd``) (issue 24)
2822 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2823 template, allowing migration from DRBD0.7+MD to DRBD8
2829 *(Released Wed, 30 Jan 2008)*
2831 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2833 - add command aliases infrastructure and a few aliases
2834 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2835 man pages and the ``--help`` option of ``gnt-node
2836 list``/``gnt-instance list``
2837 - fix ``gnt-backup list`` with down nodes (issue 21)
2838 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2839 - fix the dist archive and add a check for including svn/git files in
2841 - some developer-related changes: improve the burnin and the QA suite,
2842 add an upload script for testing during development
2848 *(Released Wed, 16 Jan 2008)*
2850 - experimental HVM support, read the install document, section
2851 "Initializing the cluster"
2852 - allow for the PVM hypervisor per-instance kernel and initrd paths
2853 - add a new command ``gnt-cluster verify-disks`` which uses a new
2854 algorithm to improve the reconnection of the DRBD pairs if the device
2855 on the secondary node has gone away
2856 - make logical volume code auto-activate LVs at disk activation time
2857 - slightly improve the speed of activating disks
2858 - allow specification of the MAC address at instance creation time, and
2859 changing it later via ``gnt-instance modify``
2860 - fix handling of external commands that generate lots of output on
2862 - update documentation with regard to minimum version of DRBD8 supported
2868 *(Released Tue, 4 Dec 2007)*
2870 - Log the ``xm create`` output to the node daemon log on failure (to
2871 help diagnosing the error)
2872 - In debug mode, log all external commands output if failed to the logs
2873 - Change parsing of lvm commands to ignore stderr
2879 *(Released Wed, 28 Nov 2007)*
2881 - Another round of updates to the DRBD 8 code to deal with more failures
2882 in the replace secondary node operation
2883 - Some more logging of failures in disk operations (lvm, drbd)
2884 - A few documentation updates
2891 *(Released Tue, 13 Nov 2007)*
2893 - Change configuration file format from Python's Pickle to JSON.
2894 Upgrading is possible using the cfgupgrade utility.
2895 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2896 faster replace disks and is more stable (DRBD 8 has many improvements
2897 compared to DRBD 0.7)
2898 - Added command line tags support (see man pages for ``gnt-instance``,
2899 ``gnt-node``, ``gnt-cluster``)
2900 - Added instance rename support
2901 - Added multi-instance startup/shutdown
2902 - Added cluster rename support
2903 - Added ``gnt-node evacuate`` to simplify some node operations
2904 - Added instance reboot operation that can speedup reboot as compared to
2906 - Soften the requirement that hostnames are in FQDN format
2907 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2909 - Removed dependency on debian's patched fping that uses the
2910 non-standard ``-S`` option
2911 - Now the OS definitions are searched for in multiple, configurable
2912 paths (easier for distros to package)
2913 - Some changes to the hooks infrastructure (especially the new
2914 post-configuration update hook)
2915 - Other small bugfixes
2917 .. vim: set textwidth=72 syntax=rst :