10 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
12 - Disks added through ``gnt-instance modify`` or created through
13 ``gnt-instance recreate-disks`` are wiped, if the
14 ``prealloc_wipe_disks`` flag is set.
15 - If wiping newly created disks fails, the disks are removed. Also,
16 partial failures in creating disks through ``gnt-instance modify``
17 triggers a cleanup of the partially-created disks.
18 - Removing the master IP address doesn't fail if the address has been
25 *(Released Fri, 3 May 2013)*
27 Incompatible/important changes
28 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
30 - Instance policies for disk size were documented to be on a per-disk
31 basis, but hail applied them to the sum of all disks. This has been
33 - ``hbal`` will now exit with status 0 if, during job execution over
34 LUXI, early exit has been requested and all jobs are successful;
35 before, exit status 1 was used, which cannot be differentiated from
37 - Compatibility with newer versions of rbd has been fixed
38 - ``gnt-instance batch-create`` has been changed to use the bulk create
39 opcode from Ganeti. This lead to incompatible changes in the format of
40 the JSON file. It's now not a custom dict anymore but a dict
41 compatible with the ``OpInstanceCreate`` opcode.
42 - Parent directories for file storage need to be listed in
43 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
44 write the file automatically based on old configuration values, but it
45 can not distribute it across all nodes and the file contents should be
46 verified. Use ``gnt-cluster copyfile
47 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
48 upgraded. The reason for requiring this list of paths now is that
49 before it would have been possible to inject new paths via RPC,
50 allowing files to be created in arbitrary locations. The RPC protocol
51 is protected using SSL/X.509 certificates, but as a design principle
52 Ganeti does not permit arbitrary paths to be passed.
53 - The parsing of the variants file for OSes (see
54 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
55 lines and comment lines (starting with ``#``) are ignored for better
57 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
58 longer available. ``gnt-node add`` now invokes a new tool on the
59 destination node, named ``prepare-node-join``, to configure the SSH
60 daemon. Paramiko is no longer necessary to configure nodes' SSH
61 daemons via ``gnt-node add``.
62 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
63 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
64 the flag is not reset after a master failover.
65 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
66 is recommended. 2.6 will be mandatory from the 2.8 series.
72 - New network management functionality to support automatic allocation
73 of IP addresses and managing of network parameters. See
74 :manpage:`gnt-network(8)` for more details.
75 - New external storage backend, to allow managing arbitrary storage
76 systems external to the cluster. See
77 :manpage:`ganeti-extstorage-interface(7)`.
78 - New ``exclusive-storage`` node parameter added, restricted to
79 nodegroup level. When it's set to true, physical disks are assigned in
80 an exclusive fashion to instances, as documented in :doc:`Partitioned
81 Ganeti <design-partitioned>`. Currently, only instances using the
82 ``plain`` disk template are supported.
83 - The KVM hypervisor has been updated with many new hypervisor
84 parameters, including a generic one for passing arbitrary command line
85 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
86 compatible up to qemu 1.4.
87 - A new tool, called ``mon-collector``, is the stand-alone executor of
88 the data collectors for a monitoring system. As of this version, it
89 just includes the DRBD data collector, that can be executed by calling
90 ``mon-collector`` using the ``drbd`` parameter. See
91 :manpage:`mon-collector(7)`.
92 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
93 for RAPI users. It allows granting permissions to query for
94 information to a specific user without giving
95 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
96 - A new tool named ``node-cleanup`` has been added. It cleans remains of
97 a cluster from a machine by stopping all daemons, removing
98 certificates and ssconf files. Unless the ``--no-backup`` option is
99 given, copies of the certificates are made.
100 - Instance creations now support the use of opportunistic locking,
101 potentially speeding up the (parallel) creation of multiple instances.
102 This feature is currently only available via the :doc:`RAPI
103 <rapi>` interface and when an instance allocator is used. If the
104 ``opportunistic_locking`` parameter is set the opcode will try to
105 acquire as many locks as possible, but will not wait for any locks
106 held by other opcodes. If not enough resources can be found to
107 allocate the instance, the temporary error code
108 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
109 retried thereafter, with or without opportunistic locking.
110 - New experimental linux-ha resource scripts.
111 - Restricted-commands support: ganeti can now be asked (via command line
112 or rapi) to perform commands on a node. These are passed via ganeti
113 RPC rather than ssh. This functionality is restricted to commands
114 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
115 reasons. The file is not copied automatically.
121 - Diskless instances are now externally mirrored (Issue 237). This for
122 now has only been tested in conjunction with explicit target nodes for
124 - Queries not needing locks or RPC access to the node can now be
125 performed by the confd daemon, making them independent from jobs, and
126 thus faster to execute. This is selectable at configure time.
127 - The functionality for allocating multiple instances at once has been
128 overhauled and is now also available through :doc:`RAPI <rapi>`.
133 - Fix kvm compatibility with qemu 1.4 (Issue 389)
134 - Documentation updates (admin guide, upgrade notes, install
135 instructions) (Issue 372)
136 - Fix gnt-group list nodes and instances count (Issue 436)
137 - Fix compilation without non-mandatory libraries (Issue 441)
138 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
139 - Make confd logging more verbose at INFO level (Issue 435)
140 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
141 - Fix failure path for instance storage type conversion (Issue 229)
142 - Update htools text backend documentation
143 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
144 - Disable inter-cluster instance move for file-based instances, because
145 it is dependant on instance export, which is not supported for
146 file-based instances. (Issue 414)
147 - Fix gnt-job crashes on non-ascii characters (Issue 427)
148 - Fix volume group checks on non-vm-capable nodes (Issue 432)
154 *(Released Mon, 22 Apr 2013)*
156 This was the third beta release of the 2.7 series. Since beta2:
158 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
159 - Fix data loss on wrong usage of ``gnt-instance move``
160 - Properly export errors in confd-based job queries
161 - Add ``users-setup`` tool
162 - Fix iallocator protocol to report 0 as a disk size for diskless
163 instances. This avoids hail breaking when a diskless instance is
165 - Fix job queue directory permission problem that made confd job queries
166 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
167 for access to archived jobs (Issue 406).
168 - Limit the sizes of networks supported by ``gnt-network`` to something
169 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
170 - Fix bugs in instance disk template conversion
171 - Fix GHC 7 compatibility
172 - Fix ``burnin`` install path (Issue 426).
173 - Allow very small disk grows (Issue 347).
174 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
175 that noded doesn't import masterd code (Issue 419).
176 - Make sure the default metavg at cluster init is the same as the vg, if
177 unspecified (Issue 358).
178 - Fix cleanup of partially created disks (part of Issue 416)
184 *(Released Tue, 2 Apr 2013)*
186 This was the second beta release of the 2.7 series. Since beta1:
188 - Networks no longer have a "type" slot, since this information was
189 unused in Ganeti: instead of it tags should be used.
190 - The rapi client now has a ``target_node`` option to MigrateInstance.
191 - Fix early exit return code for hbal (Issue 386).
192 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
193 - Fix ``rbd showmapped`` output parsing (Issue 312).
194 - Networks are now referenced indexed by UUID, rather than name. This
195 will require running cfgupgrade, from 2.7.0beta1, if networks are in
197 - The OS environment now includes network information.
198 - Deleting of a network is now disallowed if any instance nic is using
199 it, to prevent dangling references.
200 - External storage is now documented in man pages.
201 - The exclusive_storage flag can now only be set at nodegroup level.
202 - Hbal can now submit an explicit priority with its jobs.
203 - Many network related locking fixes.
204 - Bump up the required pylint version to 0.25.1.
205 - Fix the ``no_remember`` option in RAPI client.
206 - Many ipolicy related tests, qa, and fixes.
207 - Many documentation improvements and fixes.
208 - Fix building with ``--disable-file-storage``.
209 - Fix ``-q`` option in htools, which was broken if passed more than
211 - Some haskell/python interaction improvements and fixes.
212 - Fix iallocator in case of missing LVM storage.
213 - Fix confd config load in case of ``--no-lvm-storage``.
214 - The confd/query functionality is now mentioned in the security
221 *(Released Wed, 6 Feb 2013)*
223 This was the first beta release of the 2.7 series. All important changes
224 are listed in the latest 2.7 entry.
230 *(Released Fri, 21 Dec 2012)*
232 Important behaviour change: hbal won't rebalance anymore instances which
233 have the ``auto_balance`` attribute set to false. This was the intention
234 all along, but until now it only skipped those from the N+1 memory
235 reservation (DRBD-specific).
237 A significant number of bug fixes in this release:
239 - Fixed disk adoption interaction with ipolicy checks.
240 - Fixed networking issues when instances are started, stopped or
241 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
242 - Fixed the warning in cluster verify for shared storage instances not
244 - Fixed removal of storage directory on shared file storage (issue 262).
245 - Fixed validation of LVM volume group name in OpClusterSetParams
246 (``gnt-cluster modify``) (issue 285).
247 - Fixed runtime memory increases (``gnt-instance modify -m``).
248 - Fixed live migration under Xen's ``xl`` mode.
249 - Fixed ``gnt-instance console`` with ``xl``.
250 - Fixed building with newer Haskell compiler/libraries.
251 - Fixed PID file writing in Haskell daemons (confd); this prevents
252 restart issues if confd was launched manually (outside of
253 ``daemon-util``) while another copy of it was running
254 - Fixed a type error when doing live migrations with KVM (issue 297) and
255 the error messages for failing migrations have been improved.
256 - Fixed opcode validation for the out-of-band commands (``gnt-node
258 - Fixed a type error when unsetting OS hypervisor parameters (issue
259 311); now it's possible to unset all OS-specific hypervisor
261 - Fixed the ``dry-run`` mode for many operations: verification of
262 results was over-zealous but didn't take into account the ``dry-run``
263 operation, resulting in "wrong" failures.
264 - Fixed bash completion in ``gnt-job list`` when the job queue has
265 hundreds of entries; especially with older ``bash`` versions, this
266 results in significant CPU usage.
268 And lastly, a few other improvements have been made:
270 - Added option to force master-failover without voting (issue 282).
271 - Clarified error message on lock conflict (issue 287).
272 - Logging of newly submitted jobs has been improved (issue 290).
273 - Hostname checks have been made uniform between instance rename and
275 - The ``--submit`` option is now supported by ``gnt-debug delay``.
276 - Shutting down the master daemon by sending SIGTERM now stops it from
277 processing jobs waiting for locks; instead, those jobs will be started
278 once again after the master daemon is started the next time (issue
280 - Support for Xen's ``xl`` program has been improved (besides the fixes
282 - Reduced logging noise in the Haskell confd daemon (only show one log
283 entry for each config reload, instead of two).
284 - Several man page updates and typo fixes.
290 *(Released Fri, 12 Oct 2012)*
292 A small bugfix release. Among the bugs fixed:
294 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
295 made the command unusable.
296 - Commands that issue many jobs don't fail anymore just because some jobs
297 take so long that other jobs are archived.
298 - Failures during ``gnt-instance reinstall`` are reflected by the exit
300 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
302 - When ``always_failover`` is set, ``--allow-failover`` is not required
303 in migrate commands anymore.
304 - ``bash_completion`` works even if extglob is disabled.
305 - Fixed bug with locks that made failover for RDB-based instances fail.
306 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
307 a random node instead of one based on the allocator metric.
308 - Support for newer versions of pylint and pep8.
309 - Hail doesn't fail anymore when trying to add an instance of type
310 ``file``, ``sharedfile`` or ``rbd``.
311 - Added new Makefile target to rebuild the whole distribution, so that
312 all files are included.
318 *(Released Fri, 27 Jul 2012)*
321 .. attention:: The ``LUXI`` protocol has been made more consistent
322 regarding its handling of command arguments. This, however, leads to
323 incompatibility issues with previous versions. Please ensure that you
324 restart Ganeti daemons soon after the upgrade, otherwise most
325 ``LUXI`` calls (job submission, setting/resetting the drain flag,
326 pausing/resuming the watcher, cancelling and archiving jobs, querying
327 the cluster configuration) will fail.
336 The current ``admin_up`` field, which used to denote whether an instance
337 should be running or not, has been removed. Instead, ``admin_state`` is
338 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
340 The rational behind this is that an instance being “down” can have
343 - it could be down during a reboot
344 - it could be temporarily be down for a reinstall
345 - or it could be down because it is deprecated and kept just for its
348 The previous Boolean state was making it difficult to do capacity
349 calculations: should Ganeti reserve memory for a down instance? Now, the
350 tri-state field makes it clear:
352 - in ``up`` and ``down`` state, all resources are reserved for the
353 instance, and it can be at any time brought up if it is down
354 - in ``offline`` state, only disk space is reserved for it, but not
357 The field can have an extra use: since the transition between ``up`` and
358 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
359 transition between ``offline`` and ``down`` is done via ``gnt-instance
360 modify``, it is possible to given different rights to users. For
361 example, owners of an instance could be allowed to start/stop it, but
362 not transition it out of the offline state.
364 Instance policies and specs
365 +++++++++++++++++++++++++++
367 In previous Ganeti versions, an instance creation request was not
368 limited on the minimum size and on the maximum size just by the cluster
369 resources. As such, any policy could be implemented only in third-party
370 clients (RAPI clients, or shell wrappers over ``gnt-*``
371 tools). Furthermore, calculating cluster capacity via ``hspace`` again
372 required external input with regards to instance sizes.
374 In order to improve these workflows and to allow for example better
375 per-node group differentiation, we introduced instance specs, which
378 - minimum instance disk size, disk count, memory size, cpu count
379 - maximum values for the above metrics
380 - and “standard” values (used in ``hspace`` to calculate the standard
383 The minimum/maximum values can be also customised at node-group level,
384 for example allowing more powerful hardware to support bigger instance
387 Beside the instance specs, there are a few other settings belonging to
388 the instance policy framework. It is possible now to customise, per
389 cluster and node-group:
391 - the list of allowed disk templates
392 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
393 - the maximum ratio of instance to spindles (see below for more
394 information) for local storage
396 All these together should allow all tools that talk to Ganeti to know
397 what are the ranges of allowed values for instances and the
398 over-subscription that is allowed.
400 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
401 instance configuration, and the physical CPU configuration from the
402 node. For the spindle ratios however, we didn't track before these
403 values, so new parameters have been added:
405 - a new node parameter ``spindle_count``, defaults to 1, customisable at
406 node group or node level
407 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
409 Note that spindles in this context doesn't need to mean actual
410 mechanical hard-drives; it's just a relative number for both the node
411 I/O capacity and instance I/O consumption.
413 Instance migration behaviour
414 ++++++++++++++++++++++++++++
416 While live-migration is in general desirable over failover, it is
417 possible that for some workloads it is actually worse, due to the
418 variable time of the “suspend” phase during live migration.
420 To allow the tools to work consistently over such instances (without
421 having to hard-code instance names), a new backend parameter
422 ``always_failover`` has been added to control the migration/failover
423 behaviour. When set to True, all migration requests for an instance will
424 instead fall-back to failover.
426 Instance memory ballooning
427 ++++++++++++++++++++++++++
429 Initial support for memory ballooning has been added. The memory for an
430 instance is no longer fixed (backend parameter ``memory``), but instead
431 can vary between minimum and maximum values (backend parameters
432 ``minmem`` and ``maxmem``). Currently we only change an instance's
435 - live migrating or failing over and instance and the target node
436 doesn't have enough memory
437 - user requests changing the memory via ``gnt-instance modify
443 In order to control the use of specific CPUs by instance, support for
444 controlling CPU pinning has been added for the Xen, HVM and LXC
445 hypervisors. This is controlled by a new hypervisor parameter
446 ``cpu_mask``; details about possible values for this are in the
447 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
448 VCPU-to-CPU mapping) form will work well only when all nodes in your
449 cluster have the same amount of CPUs.
454 Another area in which Ganeti was not customisable were the parameters
455 used for storage configuration, e.g. how many stripes to use for LVM,
456 DRBD resync configuration, etc.
458 To improve this area, we've added disks parameters, which are
459 customisable at cluster and node group level, and which allow to
460 specify various parameters for disks (DRBD has the most parameters
461 currently), for example:
463 - DRBD resync algorithm and parameters (e.g. speed)
464 - the default VG for meta-data volumes for DRBD
465 - number of stripes for LVM (plain disk template)
468 These parameters can be modified via ``gnt-cluster modify -D …`` and
469 ``gnt-group modify -D …``, and are used at either instance creation (in
470 case of LVM stripes, for example) or at disk “activation” time
473 Rados block device support
474 ++++++++++++++++++++++++++
476 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
477 denoted by the ``rbd`` disk template type. This is considered
478 experimental, feedback is welcome. For details on configuring it, see
479 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
484 The existing master IP functionality works well only in simple setups (a
485 single network shared by all nodes); however, if nodes belong to
486 different networks, then the ``/32`` setup and lack of routing
487 information is not enough.
489 To allow the master IP to function well in more complex cases, the
490 system was reworked as follows:
492 - a master IP netmask setting has been added
493 - the master IP activation/turn-down code was moved from the node daemon
495 - whether to run the Ganeti-supplied master IP script or a user-supplied
496 on is a ``gnt-cluster init`` setting
498 Details about the location of the standard and custom setup scripts are
499 in the man page :manpage:`gnt-cluster(8)`; for information about the
500 setup script protocol, look at the Ganeti-supplied script.
505 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
508 It is now possible to use TLS-protected connections, and when renewing
509 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
510 it is now possible to specify spice or spice CA certificates. Also, it
511 is possible to configure a password for SPICE sessions via the
512 hypervisor parameter ``spice_password_file``.
514 There are also new parameters to control the compression and streaming
515 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
516 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
517 for the spice parameters.
519 Lastly, it is now possible to see the SPICE connection information via
520 ``gnt-instance console``.
525 A new tool (``tools/ovfconverter``) has been added that supports
526 conversion between Ganeti and the `Open Virtualization Format
527 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
530 This relies on the ``qemu-img`` tool to convert the disk formats, so the
531 actual compatibility with other virtualization solutions depends on it.
536 The configuration query daemon (``ganeti-confd``) is now optional, and
537 has been rewritten in Haskell; whether to use the daemon at all, use the
538 Python (default) or the Haskell version is selectable at configure time
539 via the ``--enable-confd`` parameter, which can take one of the
540 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
541 daemon will result in a smaller footprint; for larger systems, we
542 welcome feedback on the Haskell version which might become the default
545 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
546 daemon running. The Python version doesn't implement the new call.
549 User interface changes
550 ~~~~~~~~~~~~~~~~~~~~~~
552 We have replaced the ``--disks`` option of ``gnt-instance
553 replace-disks`` with a more flexible ``--disk`` option, which allows
554 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
555 disk size and mode can be changed upon recreation (via ``gnt-instance
556 recreate-disks``, which accepts the same ``--disk`` option).
558 As many people are used to a ``show`` command, we have added that as an
559 alias to ``info`` on all ``gnt-*`` commands.
561 The ``gnt-instance grow-disk`` command has a new mode in which it can
562 accept the target size of the disk, instead of the delta; this can be
563 more safe since two runs in absolute mode will be idempotent, and
564 sometimes it's also easier to specify the desired size directly.
566 Also the handling of instances with regard to offline secondaries has
567 been improved. Instance operations should not fail because one of it's
568 secondary nodes is offline, even though it's safe to proceed.
570 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
571 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
572 minors to instance name.
577 RAPI coverage has improved, with (for example) new resources for
578 recreate-disks, node power-cycle, etc.
583 There is partial support for ``xl`` in the Xen hypervisor; feedback is
586 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
587 whether to still support Python 2.4 or move to Python 2.6 as minimum
590 Support for Fedora has been slightly improved; the provided example
591 init.d script should work better on it and the INSTALL file should
592 document the needed dependencies.
597 The deprecated ``QueryLocks`` LUXI request has been removed. Use
598 ``Query(what=QR_LOCK, ...)`` instead.
600 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
601 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
602 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
603 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
604 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
606 RAPI client: ``CertificateError`` now derives from
607 ``GanetiApiError``. This should make it more easy to handle Ganeti
610 Deprecation warnings due to PyCrypto/paramiko import in
611 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
612 make sure to run an up-to-date paramiko version, if you use this tool.
614 The QA scripts now depend on Python 2.5 or above (the main code base
615 still works with Python 2.4).
617 The configuration file (``config.data``) is now written without
618 indentation for performance reasons; if you want to edit it, it can be
619 re-formatted via ``tools/fmtjson``.
621 A number of bugs has been fixed in the cluster merge tool.
623 ``x509`` certification verification (used in import-export) has been
624 changed to allow the same clock skew as permitted by the cluster
625 verification. This will remove some rare but hard to diagnose errors in
632 *(Released Thu, 19 Jul 2012)*
634 Very few changes from rc4 to the final release, only bugfixes:
636 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
637 instance, fix CDROM booting for KVM instances)
638 - fixed node group modification of node parameters
639 - fixed issue in LUClusterVerifyGroup with multi-group clusters
640 - fixed generation of bash completion to ensure a stable ordering
647 *(Released Fri, 13 Jul 2012)*
649 Third release candidate for 2.6. The following changes were done from
652 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
653 - Fixed an inconsistency in the LUXI protocol with the provided
654 arguments (NOT backwards compatible)
655 - Fixed a bug with node groups ipolicy where ``min`` was greater than
656 the cluster ``std`` value
657 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
658 easier instance debugging on nodes (requires ``hconfd`` to work)
664 *(Released Tue, 03 Jul 2012)*
666 Second release candidate for 2.6. The following changes were done from
669 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
671 - Fixed a RAPI regression on missing beparams/memory
672 - Fixed redistribution of files on offline nodes
673 - Added possibility to run activate-disks even though secondaries are
674 offline. With this change it relaxes also the strictness on some other
675 commands which use activate disks internally:
676 * ``gnt-instance start|reboot|rename|backup|export``
677 - Made it possible to remove safely an instance if its secondaries are
679 - Made it possible to reinstall even though secondaries are offline
685 *(Released Mon, 25 Jun 2012)*
687 First release candidate for 2.6. The following changes were done from
690 - Fixed bugs with disk parameters and ``rbd`` templates as well as
692 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
693 behaviour. It supports now the modify operation
694 - ``hcheck`` implemented to analyze cluster health and possibility of
695 improving health by rebalance
696 - ``hbal`` has been improved in dealing with split instances
702 *(Released Mon, 11 Jun 2012)*
704 Second beta release of 2.6. The following changes were done from beta2
707 - Fixed ``daemon-util`` with non-root user models
708 - Fixed creation of plain instances with ``--no-wait-for-sync``
709 - Fix wrong iv_names when running ``cfgupgrade``
710 - Export more information in RAPI group queries
711 - Fixed bug when changing instance network interfaces
712 - Extended burnin to do NIC changes
713 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
714 - Changed default for DRBD barriers
715 - Fixed DRBD error reporting for syncer rate
716 - Verify the options on disk parameters
718 And of course various fixes to documentation and improved unittests and
725 *(Released Wed, 23 May 2012)*
727 First beta release of 2.6. The following changes were done from beta1 to
730 - integrated patch for distributions without ``start-stop-daemon``
731 - adapted example init.d script to work on Fedora
732 - fixed log handling in Haskell daemons
733 - adapted checks in the watcher for pycurl linked against libnss
734 - add partial support for ``xl`` instead of ``xm`` for Xen
735 - fixed a type issue in cluster verification
736 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
739 Plus integrated fixes from the 2.5 branch:
741 - fixed ``kvm-ifup`` to use ``/bin/bash``
742 - fixed parallel build failures
743 - KVM live migration when using a custom keymap
749 *(Released Tue, 24 Jul 2012)*
751 A small bugfix release, with no new features:
753 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
754 different default shell (e.g. Debian, Ubuntu)
755 - fixed KVM startup and live migration with a custom keymap (fixes Issue
756 243 and Debian bug #650664)
757 - fixed compatibility with KVM versions that don't support multiple boot
758 devices (fixes Issue 230 and Debian bug #624256)
760 Additionally, a few fixes were done to the build system (fixed parallel
761 build failures) and to the unittests (fixed race condition in test for
762 FileID functions, and the default enable/disable mode for QA test is now
769 *(Released Fri, 11 May 2012)*
771 A small bugfix release.
773 The main issues solved are on the topic of compatibility with newer LVM
776 - fixed parsing of ``lv_attr`` field
777 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
778 the ``--force`` flag is needed
780 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
781 to accept kernel 3.x too (was hardcoded to 2.6.*).
783 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
784 of hook results and that also made display of other errors suboptimal
785 was fixed; the code behaves now like 2.4 and earlier.
787 Another change in 2.5, the cleanup of the OS scripts environment, is too
788 aggressive: it removed even the ``PATH`` variable, which requires the OS
789 scripts to *always* need to export it. Since this is a bit too strict,
790 we now export a minimal PATH, the same that we export for hooks.
792 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
793 integrated into this release.
795 Finally, a few other miscellaneous changes were done (no new features,
796 just small improvements):
798 - Fix ``gnt-group --help`` display
799 - Fix hardcoded Xen kernel path
800 - Fix grow-disk handling of invalid units
801 - Update synopsis for ``gnt-cluster repair-disk-sizes``
802 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
808 *(Released Thu, 12 Apr 2012)*
810 Incompatible/important changes and bugfixes
811 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
813 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
814 resource's ``ip_check`` parameter changed from ``True`` to ``False``
815 to match the underlying LUXI interface.
816 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
817 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
818 not maintain backwards-compatibility as the underlying operation
819 changed in an incompatible way. The RAPI client can talk to old
820 servers, but it needs to be told so as the return value changed.
821 - When creating file-based instances via RAPI, the ``file_driver``
822 parameter no longer defaults to ``loop`` and must be specified.
823 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
825 - Support for the undocumented and deprecated RAPI instance creation
826 request format version 0 has been dropped. Use version 1, supported
827 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
828 - Pyparsing 1.4.6 or above is required, see :doc:`installation
829 documentation <install>`.
830 - The "cluster-verify" hooks are now executed per group by the
831 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
832 if you just run ``gnt-cluster verify``, which generates one opcode per
834 - The environment as passed to the OS scripts is cleared, and thus no
835 environment variables defined in the node daemon's environment will be
836 inherited by the scripts.
837 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
839 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
840 support operations involving multiple node groups.
841 - Offline nodes are ignored when failing over an instance.
842 - Support for KVM version 1.0, which changed the version reporting format
844 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
846 - ``Makefile`` is now compatible with Automake 1.11.2
847 - Includes all bugfixes made in the 2.4 series
852 - The ganeti-htools project has been merged into the ganeti-core source
853 tree and will be built as part of Ganeti (see :doc:`install-quick`).
854 - Implemented support for :doc:`shared storage <design-shared-storage>`.
855 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
856 GPT-style partition tables (requires `parted
857 <http://www.gnu.org/s/parted/>`_).
858 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
859 - Allowed adding tags on instance creation.
860 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
861 - Allow instances to be started in a paused state, enabling the user to
862 see the complete console output on boot using the console.
863 - Added new hypervisor flag to control default reboot behaviour
864 (``reboot_behavior``).
865 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
866 - Improved out-of-band management support:
868 - Added ``gnt-node health`` command reporting the health status of
870 - Added ``gnt-node power`` command to manage power status of nodes.
871 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
873 - Instance migration can fall back to failover if instance is not
875 - Filters can be used when listing nodes, instances, groups and locks;
876 see :manpage:`ganeti(7)` manpage.
877 - Added post-execution status as variables to :doc:`hooks <hooks>`
879 - Instance tags are exported/imported together with the instance.
880 - When given an explicit job ID, ``gnt-job info`` will work for archived
882 - Jobs can define dependencies on other jobs (not yet supported via
883 RAPI or command line, but used by internal commands and usable via
886 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
889 - Instance failover is now available as a RAPI resource
890 (``/2/instances/[instance_name]/failover``).
891 - ``gnt-instance info`` defaults to static information if primary node
893 - Opcodes have a new ``comment`` attribute.
894 - Added basic SPICE support to KVM hypervisor.
895 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
897 Node group improvements
898 ~~~~~~~~~~~~~~~~~~~~~~~
900 - ``gnt-cluster verify`` has been modified to check groups separately,
901 thereby improving performance.
902 - Node group support has been added to ``gnt-cluster verify-disks``,
903 which now operates per node group.
904 - Watcher has been changed to work better with node groups.
906 - One process and state file per node group.
907 - Slow watcher in one group doesn't block other group's watcher.
909 - Added new command, ``gnt-group evacuate``, to move all instances in a
910 node group to other groups.
911 - Added ``gnt-instance change-group`` to move an instance to another
913 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
914 per-group operations.
915 - Node groups can be tagged.
916 - Some operations switch from an exclusive to a shared lock as soon as
918 - Instance's primary and secondary nodes' groups are now available as
919 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
920 and ``snodes.group.uuid``).
925 - Numerous updates to documentation and manpages.
927 - :doc:`RAPI <rapi>` documentation now has detailed parameter
929 - Some opcode/job results are now also documented, see :doc:`RAPI
932 - A lockset's internal lock is now also visible in lock monitor.
933 - Log messages from job queue workers now contain information about the
934 opcode they're processing.
935 - ``gnt-instance console`` no longer requires the instance lock.
936 - A short delay when waiting for job changes reduces the number of LUXI
937 requests significantly.
938 - DRBD metadata volumes are overwritten with zeros during disk creation.
939 - Out-of-band commands no longer acquire the cluster lock in exclusive
941 - ``devel/upload`` now uses correct permissions for directories.
947 *(Released Fri, 23 Mar 2012)*
949 This was the sixth release candidate of the 2.5 series.
955 *(Released Mon, 9 Jan 2012)*
957 This was the fifth release candidate of the 2.5 series.
963 *(Released Thu, 27 Oct 2011)*
965 This was the fourth release candidate of the 2.5 series.
971 *(Released Wed, 26 Oct 2011)*
973 This was the third release candidate of the 2.5 series.
979 *(Released Tue, 18 Oct 2011)*
981 This was the second release candidate of the 2.5 series.
987 *(Released Tue, 4 Oct 2011)*
989 This was the first release candidate of the 2.5 series.
995 *(Released Wed, 31 Aug 2011)*
997 This was the third beta release of the 2.5 series.
1003 *(Released Mon, 22 Aug 2011)*
1005 This was the second beta release of the 2.5 series.
1011 *(Released Fri, 12 Aug 2011)*
1013 This was the first beta release of the 2.5 series.
1019 *(Released Thu, 27 Oct 2011)*
1021 - Fixed bug when parsing command line parameter values ending in
1023 - Fixed assertion error after unclean master shutdown
1024 - Disable HTTP client pool for RPC, significantly reducing memory usage
1026 - Fixed queue archive creation with wrong permissions
1032 *(Released Tue, 23 Aug 2011)*
1036 - Fixed documentation for importing with ``--src-dir`` option
1037 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1038 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1044 *(Released Fri, 5 Aug 2011)*
1046 Many bug-fixes and a few small features:
1048 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1049 issues when you tried to add an instance with two MAC addresses in one
1051 - KVM: fixed per-instance stored UID value
1052 - KVM: configure bridged NICs at migration start
1053 - KVM: Fix a bug where instance will not start with never KVM versions
1055 - Added OS search path to ``gnt-cluster info``
1056 - Fixed an issue with ``file_storage_dir`` where you were forced to
1057 provide an absolute path, but the documentation states it is a
1058 relative path, the documentation was right
1059 - Added a new parameter to instance stop/start called ``--no-remember``
1060 that will make the state change to not be remembered
1061 - Implemented ``no_remember`` at RAPI level
1062 - Improved the documentation
1063 - Node evacuation: don't call IAllocator if node is already empty
1064 - Fixed bug in DRBD8 replace disks on current nodes
1065 - Fixed bug in recreate-disks for DRBD instances
1066 - Moved assertion checking locks in ``gnt-instance replace-disks``
1067 causing it to abort with not owning the right locks for some situation
1068 - Job queue: Fixed potential race condition when cancelling queued jobs
1069 - Fixed off-by-one bug in job serial generation
1070 - ``gnt-node volumes``: Fix instance names
1071 - Fixed aliases in bash completion
1072 - Fixed a bug in reopening log files after being sent a SIGHUP
1073 - Added a flag to burnin to allow specifying VCPU count
1074 - Bugfixes to non-root Ganeti configuration
1080 *(Released Thu, 12 May 2011)*
1082 Many bug-fixes and a few new small features:
1084 - Fixed a bug related to log opening failures
1085 - Fixed a bug in instance listing with orphan instances
1086 - Fixed a bug which prevented resetting the cluster-level node parameter
1087 ``oob_program`` to the default
1088 - Many fixes related to the ``cluster-merge`` tool
1089 - Fixed a race condition in the lock monitor, which caused failures
1090 during (at least) creation of many instances in parallel
1091 - Improved output for gnt-job info
1092 - Removed the quiet flag on some ssh calls which prevented debugging
1094 - Improved the N+1 failure messages in cluster verify by actually
1095 showing the memory values (needed and available)
1096 - Increased lock attempt timeouts so that when executing long operations
1097 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1098 too early and thus prevent the use of the 'fair' mechanism
1099 - Changed instance query data (``gnt-instance info``) to not acquire
1100 locks unless needed, thus allowing its use on locked instance if only
1101 static information is asked for
1102 - Improved behaviour with filesystems that do not support rename on an
1104 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1105 kept locks on all nodes during the wipe, which is unneeded
1106 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1107 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1108 10GiB) which caused the wipe to fail right at the end in some cases
1109 - Fixed master IP activation when doing master failover with no-voting
1110 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1111 the master node itself
1112 - Fixed potential data-loss in under disk full conditions, where Ganeti
1113 wouldn't check correctly the return code and would consider
1114 partially-written files 'correct'
1115 - Fixed bug related to multiple VGs and DRBD disk replacing
1116 - Added new disk parameter ``metavg`` that allows placement of the meta
1117 device for DRBD in a different volume group
1118 - Fixed error handling in the node daemon when the system libc doesn't
1119 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1120 - Fixed lock release during replace-disks, which kept cluster-wide locks
1121 when doing disk replaces with an iallocator script
1122 - Added check for missing bridges in cluster verify
1123 - Handle EPIPE errors while writing to the terminal better, so that
1124 piping the output to e.g. ``less`` doesn't cause a backtrace
1125 - Fixed rare case where a ^C during Luxi calls could have been
1126 interpreted as server errors, instead of simply terminating
1127 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1129 - Added a few more parameters to the KVM hypervisor, allowing a second
1130 CDROM, custom disk type for CDROMs and a floppy image
1131 - Removed redundant message in instance rename when the name is given
1133 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1134 disks on new nodes, allowing recreation when the original instance
1135 nodes are completely gone
1136 - Added option when converting disk templates to DRBD to skip waiting
1137 for the resync, in order to make the instance available sooner
1138 - Added two new variables to the OS scripts environment (containing the
1140 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1141 to allow use of ``pvgrub`` as bootloader
1142 - Changed the instance memory modifications to only check out-of-memory
1143 conditions on memory increases, and turned the secondary node warnings
1144 into errors (they can still be overridden via ``--force``)
1145 - Fixed the handling of a corner case when the Python installation gets
1146 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1147 to execute a command that doesn't exist
1148 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1149 node of the instance returned failures during instance shutdown; this
1150 adds the option ``--ignore-consistency`` to gnt-instance move
1152 And as usual, various improvements to the error messages, documentation
1159 *(Released Wed, 09 Mar 2011)*
1161 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1162 the RAPI users file if run twice (even with ``--dry-run``).
1164 The release fixes that bug (nothing else changed).
1170 *(Released Mon, 07 Mar 2011)*
1172 Final 2.4.0 release. Just a few small fixes:
1174 - Fixed RAPI node evacuate
1175 - Fixed the kvm-ifup script
1176 - Fixed internal error handling for special job cases
1177 - Updated man page to specify the escaping feature for options
1183 *(Released Mon, 28 Feb 2011)*
1185 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1186 that this feature wiped the disks of the wrong instance, leading to loss
1191 - Fixed title of query field containing instance name
1192 - Expanded the glossary in the documentation
1193 - Fixed one unittest (internal issue)
1199 *(Released Mon, 21 Feb 2011)*
1201 A number of bug fixes plus just a couple functionality changes.
1203 On the user-visible side, the ``gnt-* list`` command output has changed
1204 with respect to "special" field states. The current rc1 style of display
1205 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1206 the default output mode special fields states are displayed as follows:
1208 - Offline resource: ``*``
1209 - Unavailable/not applicable: ``-``
1210 - Data missing (RPC failure): ``?``
1211 - Unknown field: ``??``
1213 Another user-visible change is the addition of ``--force-join`` to
1218 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1219 - Fixed regression in RAPI/instance reinstall where all parameters were
1220 required (instead of optional)
1221 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1222 - Fixed iallocator usage (offline nodes were not considered offline)
1223 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1224 - Fixed hypervisor and OS parameter validation with respect to
1225 non-vm_capable nodes
1226 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1228 - Fixed ``tools/listrunner`` with respect to agent-based usage
1234 *(Released Fri, 4 Feb 2011)*
1236 Many changes and fixes since the beta1 release. While there were some
1237 internal changes, the code has been mostly stabilised for the RC
1240 Note: the dumb allocator was removed in this release, as it was not kept
1241 up-to-date with the IAllocator protocol changes. It is recommended to
1242 use the ``hail`` command from the ganeti-htools package.
1244 Note: the 2.4 and up versions of Ganeti are not compatible with the
1245 0.2.x branch of ganeti-htools. You need to upgrade to
1246 ganeti-htools-0.3.0 (or later).
1248 Regressions fixed from 2.3
1249 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1251 - Fixed the ``gnt-cluster verify-disks`` command
1252 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1254 - Fixed disk adoption breakage
1255 - Fixed wrong headers in instance listing for field aliases
1260 - Fixed corner case in KVM handling of NICs
1261 - Fixed many cases of wrong handling of non-vm_capable nodes
1262 - Fixed a bug where a missing instance symlink was not possible to
1263 recreate with any ``gnt-*`` command (now ``gnt-instance
1264 activate-disks`` does it)
1265 - Fixed the volume group name as reported by ``gnt-cluster
1267 - Increased timeouts for the import-export code, hopefully leading to
1268 fewer aborts due network or instance timeouts
1269 - Fixed bug in ``gnt-node list-storage``
1270 - Fixed bug where not all daemons were started on cluster
1271 initialisation, but only at the first watcher run
1272 - Fixed many bugs in the OOB implementation
1273 - Fixed watcher behaviour in presence of instances with offline
1275 - Fixed instance list output for instances running on the wrong node
1276 - a few fixes to the cluster-merge tool, but it still cannot merge
1277 multi-node groups (currently it is not recommended to use this tool)
1283 - Improved network configuration for the KVM hypervisor
1284 - Added e1000 as a supported NIC for Xen-HVM
1285 - Improved the lvmstrap tool to also be able to use partitions, as
1286 opposed to full disks
1287 - Improved speed of disk wiping (the cluster parameter
1288 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1289 of instance creations
1290 - Added documentation for the OS parameters
1291 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1292 hypervisor is not responding
1293 - Added display of blacklisted and hidden OS information in
1294 ``gnt-cluster info``
1295 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1296 NIC and node parameters, which might create problems with currently
1297 invalid (but undetected) configuration files, but prevents validation
1298 failures when unrelated parameters are modified
1299 - Changed cluster initialisation to wait for the master daemon to become
1301 - Expanded the RAPI interface:
1303 - Added config redistribution resource
1304 - Added activation/deactivation of instance disks
1305 - Added export of console information
1307 - Implemented log file reopening on SIGHUP, which allows using
1308 logrotate(8) for the Ganeti log files
1309 - Added a basic OOB helper script as an example
1315 *(Released Fri, 14 Jan 2011)*
1320 - Fixed timezone issues when formatting timestamps
1321 - Added support for node groups, available via ``gnt-group`` and other
1323 - Added out-of-band framework and management, see :doc:`design
1324 document <design-oob>`
1325 - Removed support for roman numbers from ``gnt-node list`` and
1326 ``gnt-instance list``.
1327 - Allowed modification of master network interface via ``gnt-cluster
1328 modify --master-netdev``
1329 - Accept offline secondaries while shutting down instance disks
1330 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1331 - Added support for multiple LVM volume groups
1332 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1334 - Added commands to list available fields:
1336 - ``gnt-node list-fields``
1337 - ``gnt-group list-fields``
1338 - ``gnt-instance list-fields``
1340 - Updated documentation and man pages
1345 - Moved ``rapi_users`` file into separate directory, now named
1346 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1348 - Added new tool for running commands on many machines,
1349 ``tools/ganeti-listrunner``
1350 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1351 improving the ``gnt-instance console`` output
1352 - Allowed customisation of disk index separator at ``configure`` time
1353 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1354 - Added support for non-partitioned md disks in ``lvmstrap``
1355 - Added script to gracefully power off KVM instances
1356 - Split ``utils`` module into smaller parts
1357 - Changed query operations to return more detailed information, e.g.
1358 whether an information is unavailable due to an offline node. To use
1359 this new functionality, the LUXI call ``Query`` must be used. Field
1360 information is now stored by the master daemon and can be retrieved
1361 using ``QueryFields``. Instances, nodes and groups can also be queried
1362 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1363 exposed via RAPI). The following commands make use of this
1364 infrastructure change:
1366 - ``gnt-group list``
1367 - ``gnt-group list-fields``
1369 - ``gnt-node list-fields``
1370 - ``gnt-instance list``
1371 - ``gnt-instance list-fields``
1372 - ``gnt-debug locks``
1377 - New RAPI resources (see :doc:`rapi`):
1381 - ``/2/groups/[group_name]``
1382 - ``/2/groups/[group_name]/assign-nodes``
1383 - ``/2/groups/[group_name]/modify``
1384 - ``/2/groups/[group_name]/rename``
1385 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1389 - Implemented ``no_install`` for instance creation
1390 - Implemented OS parameters for instance reinstallation, allowing
1391 use of special settings on reinstallation (e.g. for preserving data)
1396 - Added IPv6 support in import/export
1397 - Pause DRBD synchronization while wiping disks on instance creation
1398 - Updated unittests and QA scripts
1399 - Improved network parameters passed to KVM
1400 - Converted man pages from docbook to reStructuredText
1406 *(Released Mon, 20 Dec 2010)*
1408 Released version 2.3.1~rc1 without any changes.
1414 *(Released Wed, 1 Dec 2010)*
1416 - impexpd: Disable OpenSSL compression in socat if possible (backport
1417 from master, commit e90739d625b, see :doc:`installation guide
1418 <install-quick>` for details)
1419 - Changed unittest coverage report to exclude test scripts
1420 - Added script to check version format
1426 *(Released Wed, 1 Dec 2010)*
1428 Released version 2.3.0~rc1 without any changes.
1434 *(Released Fri, 19 Nov 2010)*
1436 A number of bugfixes and documentation updates:
1438 - Update ganeti-os-interface documentation
1439 - Fixed a bug related to duplicate MACs or similar items which should be
1441 - Fix breakage in OS state modify
1442 - Reinstall instance: disallow offline secondaries (fixes bug related to
1443 OS changing but reinstall failing)
1444 - plus all the other fixes between 2.2.1 and 2.2.2
1450 *(Released Tue, 2 Nov 2010)*
1452 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1453 - Fixed master failover race with watcher
1454 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1456 - Accept previously stopped instance for export with instance removal
1457 - Simplify and extend the environment variables for instance OS scripts
1458 - Added new node flags, ``master_capable`` and ``vm_capable``
1459 - Added optional instance disk wiping prior during allocation. This is a
1460 cluster-wide option and can be set/modified using
1461 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1462 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1463 :doc:`install-quick`
1464 - Added a new watcher option (``--ignore-pause``)
1465 - Added option to ignore offline node on instance start/stop
1466 (``--ignore-offline``)
1467 - Allow overriding OS parameters with ``gnt-instance reinstall``
1468 - Added ability to change node's secondary IP address using ``gnt-node
1470 - Implemented privilege separation for all daemons except
1471 ``ganeti-noded``, see ``configure`` options
1472 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1474 - Implemented job priorities (see ``ganeti(7)`` manpage)
1475 - Ignore failures while shutting down instances during failover from
1477 - Exit daemon's bootstrap process only once daemon is ready
1478 - Export more information via ``LUInstanceQuery``/remote API
1479 - Improved documentation, QA and unittests
1480 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1481 restart if the file was created or changed
1482 - Added LUXI protocol version sent with each request and response,
1483 allowing detection of server/client mismatches
1484 - Moved the Python scripts among gnt-* and ganeti-* into modules
1485 - Moved all code related to setting up SSH to an external script,
1487 - Infrastructure changes for node group support in future versions
1493 *(Released Fri, 19 Nov 2010)*
1495 A few small bugs fixed, and some improvements to the build system:
1497 - Fix documentation regarding conversion to drbd
1498 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1500 - Fix error handling in node modify with multiple changes
1501 - Allow remote imports without checked names
1507 *(Released Tue, 19 Oct 2010)*
1509 - Disable SSL session ID cache in RPC client
1515 *(Released Thu, 14 Oct 2010)*
1517 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1518 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1520 - Fix problems with interaction between Curl and Python's HTTP server,
1521 resulting in increased speed in many RPC calls
1522 - Improve our release script to prevent breakage with older aclocal and
1529 *(Released Thu, 7 Oct 2010)*
1531 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1532 value retrieved from master
1533 - Added support for blacklisted or hidden OS definitions
1534 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1535 - Added support for -mem-path in KVM hypervisor abstraction layer
1536 - Allow overriding instance parameters in tool for inter-cluster
1537 instance moves (``tools/move-instance``)
1538 - Improved opcode summaries (e.g. in ``gnt-job list``)
1539 - Improve consistency of OS listing by sorting it
1540 - Documentation updates
1546 *(Released Fri, 8 Oct 2010)*
1548 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1554 *(Released Mon, 4 Oct 2010)*
1556 - Fixed regression in ``gnt-instance rename``
1562 *(Released Wed, 22 Sep 2010)*
1564 - Fixed OS_VARIANT variable for OS scripts
1565 - Fixed cluster tag operations via RAPI
1566 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1567 - Disabled RAPI CA checks in watcher
1573 *(Released Mon, 23 Aug 2010)*
1575 - Support DRBD versions of the format "a.b.c.d"
1577 - Re-introduce support for usage from multiple threads in RAPI client
1578 - Instance renames and modify via RAPI
1579 - Work around race condition between processing and archival in job
1581 - Mark opcodes following failed one as failed, too
1582 - Job field ``lock_status`` was removed due to difficulties making it
1583 work with the changed job queue in Ganeti 2.2; a better way to monitor
1584 locks is expected for a later 2.2.x release
1585 - Fixed dry-run behaviour with many commands
1586 - Support ``ssh-agent`` again when adding nodes
1587 - Many additional bugfixes
1593 *(Released Fri, 30 Jul 2010)*
1595 Important change: the internal RPC mechanism between Ganeti nodes has
1596 changed from using a home-grown http library (based on the Python base
1597 libraries) to use the PycURL library. This requires that PycURL is
1598 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1599 against GnuTLS by default. cURL's support for GnuTLS had known issues
1600 before cURL 7.21.0 and we recommend using the latest cURL release or
1601 linking against OpenSSL. Most other distributions already link PycURL
1602 and cURL against OpenSSL. The command::
1604 python -c 'import pycurl; print pycurl.version'
1606 can be used to determine the libraries PycURL and cURL are linked
1609 Other significant changes:
1611 - Rewrote much of the internals of the job queue, in order to achieve
1612 better parallelism; this decouples job query operations from the job
1613 processing, and it should allow much nicer behaviour of the master
1614 daemon under load, and it also has uncovered some long-standing bugs
1615 related to the job serialisation (now fixed)
1616 - Added a default iallocator setting to the cluster parameters,
1617 eliminating the need to always pass nodes or an iallocator for
1618 operations that require selection of new node(s)
1619 - Added experimental support for the LXC virtualization method
1620 - Added support for OS parameters, which allows the installation of
1621 instances to pass parameter to OS scripts in order to customise the
1623 - Added a hypervisor parameter controlling the migration type (live or
1624 non-live), since hypervisors have various levels of reliability; this
1625 has renamed the 'live' parameter to 'mode'
1626 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1627 logical volumes, meaning that cluster verify will ignore them and not
1628 flag their presence as errors
1629 - The watcher will now reset the error count for failed instances after
1630 8 hours, thus allowing self-healing if the problem that caused the
1631 instances to be down/fail to start has cleared in the meantime
1632 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1633 check for, and warn, if the drbd module parameter ``usermode_helper``
1634 is not consistent with the cluster-wide setting; this is needed to
1635 make diagnose easier of failed drbd creations
1636 - Started adding base IPv6 support, but this is not yet
1637 enabled/available for use
1638 - Rename operations (cluster, instance) will now return the new name,
1639 which is especially useful if a short name was passed in
1640 - Added support for instance migration in RAPI
1641 - Added a tool to pre-configure nodes for the SSH setup, before joining
1642 them to the cluster; this will allow in the future a simplified model
1643 for node joining (but not yet fully enabled in 2.2); this needs the
1644 paramiko python library
1645 - Fixed handling of name-resolving errors
1646 - Fixed consistency of job results on the error path
1647 - Fixed master-failover race condition when executed multiple times in
1649 - Fixed many bugs related to the job queue (mostly introduced during the
1650 2.2 development cycle, so not all are impacting 2.1)
1651 - Fixed instance migration with missing disk symlinks
1652 - Fixed handling of unknown jobs in ``gnt-job archive``
1653 - And many other small fixes/improvements
1657 - Enhanced both the unittest and the QA coverage
1658 - Switched the opcode validation to a generic model, and extended the
1659 validation to all opcode parameters
1660 - Changed more parts of the code that write shell scripts to use the
1662 - Switched the master daemon to use the asyncore library for the Luxi
1669 *(Released Thu, 17 Jun 2010)*
1671 - Added tool (``move-instance``) and infrastructure to move instances
1672 between separate clusters (see :doc:`separate documentation
1673 <move-instance>` and :doc:`design document <design-2.2>`)
1674 - Added per-request RPC timeout
1675 - RAPI now requires a Content-Type header for requests with a body (e.g.
1676 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1677 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1678 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1680 - Implemented initial support for running Ganeti daemons as separate
1681 users, see configure-time flags ``--with-user-prefix`` and
1682 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1684 - Instances can be removed after export (``gnt-backup export
1685 --remove-instance``)
1686 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1687 key (instead of 1024 bit)
1688 - Added new cluster configuration file for cluster domain secret
1689 - Import/export now use SSL instead of SSH
1690 - Added support for showing estimated time when exporting an instance,
1691 see the ``ganeti-os-interface(7)`` manpage and look for
1698 *(Released Tue, 16 Nov 2010)*
1700 Some more bugfixes. Unless critical bugs occur, this will be the last
1703 - Fix case of MAC special-values
1704 - Fix mac checker regex
1705 - backend: Fix typo causing "out of range" error
1706 - Add missing --units in gnt-instance list man page
1712 *(Released Tue, 24 Aug 2010)*
1715 - Don't ignore secondary node silently on non-mirrored disk templates
1717 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1718 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1719 - Properly document the usb_mouse parameter
1720 - Fix path in ganeti-rapi(8) (issue 116)
1721 - Adjust error message when the ganeti user's .ssh directory is
1723 - Add same-node-check when changing the disk template to drbd
1729 *(Released Fri, 16 Jul 2010)*
1732 - Add an option to only select some reboot types during qa/burnin.
1733 (on some hypervisors consequent reboots are not supported)
1734 - Fix infrequent race condition in master failover. Sometimes the old
1735 master ip address would be still detected as up for a short time
1736 after it was removed, causing failover to fail.
1737 - Decrease mlockall warnings when the ctypes module is missing. On
1738 Python 2.4 we support running even if no ctypes module is installed,
1739 but we were too verbose about this issue.
1740 - Fix building on old distributions, on which man doesn't have a
1742 - Fix RAPI not to ignore the MAC address on instance creation
1743 - Implement the old instance creation format in the RAPI client.
1749 *(Released Thu, 01 Jul 2010)*
1751 A small bugfix release:
1752 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1753 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1754 a non-existing option
1755 - Fix instance create: the --force-variant option was ignored
1756 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1757 - Fix modify node storage with non-FQDN arguments
1758 - Fix RAPI client to authenticate under Python 2.6 when used
1759 for more than 5 requests needing authentication
1760 - Fix gnt-instance modify -t (storage) giving a wrong error message
1761 when converting a non-shutdown drbd instance to plain
1767 *(Released Fri, 18 Jun 2010)*
1769 A small bugfix release:
1771 - Fix live migration of KVM instances started with older Ganeti
1772 versions which had fewer hypervisor parameters
1773 - Fix gnt-instance grow-disk on down instances
1774 - Fix an error-reporting bug during instance migration
1775 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1776 silently ignoring broken ones
1777 - Fix an RPC error reporting bug affecting, for example, RAPI client
1779 - Fix bug triggered by different API version os-es on different nodes
1780 - Fix a bug in instance startup with custom hvparams: OS level
1781 parameters would fail to be applied.
1782 - Fix the RAPI client under Python 2.6 (but more work is needed to
1783 make it work completely well with OpenSSL)
1784 - Fix handling of errors when resolving names from DNS
1790 *(Released Thu, 3 Jun 2010)*
1792 A medium sized development cycle. Some new features, and some
1793 fixes/small improvements/cleanups.
1795 Significant features
1796 ~~~~~~~~~~~~~~~~~~~~
1798 The node deamon now tries to mlock itself into memory, unless the
1799 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1800 its logs, and falls back to console logging. This allows emergency
1801 features such as ``gnt-node powercycle`` to work even in the event of a
1802 broken node disk (tested offlining the disk hosting the node's
1803 filesystem and dropping its memory caches; don't try this at home)
1805 KVM: add vhost-net acceleration support. It can be tested with a new
1806 enough version of the kernel and of qemu-kvm.
1808 KVM: Add instance chrooting feature. If you use privilege dropping for
1809 your VMs you can also now force them to chroot to an empty directory,
1810 before starting the emulated guest.
1812 KVM: Add maximum migration bandwith and maximum downtime tweaking
1813 support (requires a new-enough version of qemu-kvm).
1815 Cluster verify will now warn if the master node doesn't have the master
1816 ip configured on it.
1818 Add a new (incompatible) instance creation request format to RAPI which
1819 supports all parameters (previously only a subset was supported, and it
1820 wasn't possible to extend the old format to accomodate all the new
1821 features. The old format is still supported, and a client can check for
1822 this feature, before using it, by checking for its presence in the
1823 ``features`` RAPI resource.
1825 Now with ancient latin support. Try it passing the ``--roman`` option to
1826 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1827 (requires the python-roman module to be installed, in order to work).
1832 As usual many internal code refactorings, documentation updates, and
1835 - Lots of improvements and cleanups to the experimental Remote API
1836 (RAPI) client library.
1837 - A new unit test suite for the core daemon libraries.
1838 - A fix to creating missing directories makes sure the umask is not
1839 applied anymore. This enforces the same directory permissions
1841 - Better handling terminating daemons with ctrl+c (used when running
1842 them in debugging mode).
1843 - Fix a race condition in live migrating a KVM instance, when stat()
1844 on the old proc status file returned EINVAL, which is an unexpected
1846 - Fixed manpage checking with newer man and utf-8 charachters. But now
1847 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1853 *(Released Fri, 7 May 2010)*
1855 Fix a bug which prevented untagged KVM instances from starting.
1861 *(Released Fri, 7 May 2010)*
1863 Another release with a long development cycle, during which many
1864 different features were added.
1866 Significant features
1867 ~~~~~~~~~~~~~~~~~~~~
1869 The KVM hypervisor now can run the individual instances as non-root, to
1870 reduce the impact of a VM being hijacked due to bugs in the
1871 hypervisor. It is possible to run all instances as a single (non-root)
1872 user, to manually specify a user for each instance, or to dynamically
1873 allocate a user out of a cluster-wide pool to each instance, with the
1874 guarantee that no two instances will run under the same user ID on any
1877 An experimental RAPI client library, that can be used standalone
1878 (without the other Ganeti libraries), is provided in the source tree as
1879 ``lib/rapi/client.py``. Note this client might change its interface in
1880 the future, as we iterate on its capabilities.
1882 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1883 replace the cluster's certificates and crypto keys. This might help in
1884 case they have been compromised, or have simply expired.
1886 A new disk option for instance creation has been added that allows one
1887 to "adopt" currently existing logical volumes, with data
1888 preservation. This should allow easier migration to Ganeti from
1889 unmanaged (or managed via other software) instances.
1891 Another disk improvement is the possibility to convert between redundant
1892 (DRBD) and plain (LVM) disk configuration for an instance. This should
1893 allow better scalability (starting with one node and growing the
1894 cluster, or shrinking a two-node cluster to one node).
1896 A new feature that could help with automated node failovers has been
1897 implemented: if a node sees itself as offline (by querying the master
1898 candidates), it will try to shutdown (hard) all instances and any active
1899 DRBD devices. This reduces the risk of duplicate instances if an
1900 external script automatically failovers the instances on such nodes. To
1901 enable this, the cluster parameter ``maintain_node_health`` should be
1902 enabled; in the future this option (per the name) will enable other
1903 automatic maintenance features.
1905 Instance export/import now will reuse the original instance
1906 specifications for all parameters; that means exporting an instance,
1907 deleting it and the importing it back should give an almost identical
1908 instance. Note that the default import behaviour has changed from
1909 before, where it created only one NIC; now it recreates the original
1912 Cluster verify has added a few new checks: SSL certificates validity,
1913 /etc/hosts consistency across the cluster, etc.
1918 As usual, many internal changes were done, documentation fixes,
1921 - Fixed cluster initialization with disabled cluster storage (regression
1922 introduced in 2.1.1)
1923 - File-based storage supports growing the disks
1924 - Fixed behaviour of node role changes
1925 - Fixed cluster verify for some corner cases, plus a general rewrite of
1926 cluster verify to allow future extension with more checks
1927 - Fixed log spamming by watcher and node daemon (regression introduced
1929 - Fixed possible validation issues when changing the list of enabled
1931 - Fixed cleanup of /etc/hosts during node removal
1932 - Fixed RAPI response for invalid methods
1933 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1934 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1935 from ide disks, etc.)
1936 - Allow OS changes without re-installation (to record a changed OS
1937 outside of Ganeti, or to allow OS renames)
1938 - Allow instance creation without OS installation (useful for example if
1939 the OS will be installed manually, or restored from a backup not in
1941 - Implemented option to make cluster ``copyfile`` use the replication
1943 - Added list of enabled hypervisors to ssconf (possibly useful for
1945 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1947 - A partial form of node re-IP is possible via node readd, which now
1948 allows changed node primary IP
1949 - Command line utilities now show an informational message if the job is
1951 - The logs of the master daemon now show the PID/UID/GID of the
1958 *(Released Fri, 12 Mar 2010)*
1960 During the 2.1.0 long release candidate cycle, a lot of improvements and
1961 changes have accumulated with were released later as 2.1.1.
1966 The node evacuate command (``gnt-node evacuate``) was significantly
1967 rewritten, and as such the IAllocator protocol was changed - a new
1968 request type has been added. This unfortunate change during a stable
1969 series is designed to improve performance of node evacuations; on
1970 clusters with more than about five nodes and which are well-balanced,
1971 evacuation should proceed in parallel for all instances of the node
1972 being evacuated. As such, any existing IAllocator scripts need to be
1973 updated, otherwise the above command will fail due to the unknown
1974 request. The provided "dumb" allocator has not been updated; but the
1975 ganeti-htools package supports the new protocol since version 0.2.4.
1977 Another important change is increased validation of node and instance
1978 names. This might create problems in special cases, if invalid host
1979 names are being used.
1981 Also, a new layer of hypervisor parameters has been added, that sits at
1982 OS level between the cluster defaults and the instance ones. This allows
1983 customisation of virtualization parameters depending on the installed
1984 OS. For example instances with OS 'X' may have a different KVM kernel
1985 (or any other parameter) than the cluster defaults. This is intended to
1986 help managing a multiple OSes on the same cluster, without manual
1987 modification of each instance's parameters.
1989 A tool for merging clusters, ``cluster-merge``, has been added in the
1990 tools sub-directory.
1995 - Improved the int/float conversions that should make the code more
1996 robust in face of errors from the node daemons
1997 - Fixed the remove node code in case of internal configuration errors
1998 - Fixed the node daemon behaviour in face of inconsistent queue
1999 directory (e.g. read-only file-system where we can't open the files
2001 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2002 now it either aborts cleanly or, if given the new "auto_promote"
2003 parameter, will automatically promote other nodes as needed
2004 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2005 completely prevent Ganeti from working
2006 - Fixed bug for instance export when not all disks were successfully
2008 - Fixed behaviour of node add when the new node is slow in starting up
2010 - Fixed handling of signals in the LUXI client, which should improve
2011 behaviour of command-line scripts
2012 - Added checks for invalid node/instance names in the configuration (now
2013 flagged during cluster verify)
2014 - Fixed watcher behaviour for disk activation errors
2015 - Fixed two potentially endless loops in http library, which led to the
2016 RAPI daemon hanging and consuming 100% CPU in some cases
2017 - Fixed bug in RAPI daemon related to hashed passwords
2018 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2020 - Enhanced compatibility with non-Debian OSes, but not using absolute
2021 path in some commands and allowing customisation of the ssh
2022 configuration directory
2023 - Fixed possible future issue with new Python versions by abiding to the
2024 proper use of ``__slots__`` attribute on classes
2025 - Added checks that should prevent directory traversal attacks
2026 - Many documentation fixes based on feedback from users
2031 - Added an "early_release" more for instance replace disks and node
2032 evacuate, where we release locks earlier and thus allow higher
2033 parallelism within the cluster
2034 - Added watcher hooks, intended to allow the watcher to restart other
2035 daemons (e.g. from the ganeti-nbma project), but they can be used of
2036 course for any other purpose
2037 - Added a compile-time disable for DRBD barriers, to increase
2038 performance if the administrator trusts the power supply or the
2039 storage system to not lose writes
2040 - Added the option of using syslog for logging instead of, or in
2041 addition to, Ganeti's own log files
2042 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2043 can indeed boot from a paravirtual NIC
2044 - Added a generic debug level for many operations; while this is not
2045 used widely yet, it allows one to pass the debug value all the way to
2047 - Enhanced the hooks environment for instance moves (failovers,
2048 migrations) where the primary/secondary nodes changed during the
2049 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2050 - Enhanced data validations for many user-supplied values; one important
2051 item is the restrictions imposed on instance and node names, which
2052 might reject some (invalid) host names
2053 - Add a configure-time option to disable file-based storage, if it's not
2054 needed; this allows greater security separation between the master
2055 node and the other nodes from the point of view of the inter-node RPC
2057 - Added user notification in interactive tools if job is waiting in the
2058 job queue or trying to acquire locks
2059 - Added log messages when a job is waiting for locks
2060 - Added filtering by node tags in instance operations which admit
2061 multiple instances (start, stop, reboot, reinstall)
2062 - Added a new tool for cluster mergers, ``cluster-merge``
2063 - Parameters from command line which are of the form ``a=b,c=d`` can now
2064 use backslash escapes to pass in values which contain commas,
2065 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2067 - For KVM, the instance name is the first parameter passed to KVM, so
2068 that it's more visible in the process list
2074 *(Released Tue, 2 Mar 2010)*
2076 Ganeti 2.1 brings many improvements with it. Major changes:
2078 - Added infrastructure to ease automated disk repairs
2079 - Added new daemon to export configuration data in a cheaper way than
2080 using the remote API
2081 - Instance NICs can now be routed instead of being associated with a
2083 - Improved job locking logic to reduce impact of jobs acquiring multiple
2084 locks waiting for other long-running jobs
2086 In-depth implementation details can be found in the Ganeti 2.1 design
2092 - Added chroot hypervisor
2093 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2095 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2096 ``bootloader_path`` and ``bootloader_args``)
2097 - Added the ``use_localtime`` option for the xen-hvm and kvm
2098 hypervisors, and the default value for this has changed to false (in
2099 2.0 xen-hvm always enabled it)
2100 - Added luxi call to submit multiple jobs in one go
2101 - Added cluster initialization option to not modify ``/etc/hosts``
2103 - Added network interface parameters
2104 - Added dry run mode to some LUs
2105 - Added RAPI resources:
2107 - ``/2/instances/[instance_name]/info``
2108 - ``/2/instances/[instance_name]/replace-disks``
2109 - ``/2/nodes/[node_name]/evacuate``
2110 - ``/2/nodes/[node_name]/migrate``
2111 - ``/2/nodes/[node_name]/role``
2112 - ``/2/nodes/[node_name]/storage``
2113 - ``/2/nodes/[node_name]/storage/modify``
2114 - ``/2/nodes/[node_name]/storage/repair``
2116 - Added OpCodes to evacuate or migrate all instances on a node
2117 - Added new command to list storage elements on nodes (``gnt-node
2118 list-storage``) and modify them (``gnt-node modify-storage``)
2119 - Added new ssconf files with master candidate IP address
2120 (``ssconf_master_candidates_ips``), node primary IP address
2121 (``ssconf_node_primary_ips``) and node secondary IP address
2122 (``ssconf_node_secondary_ips``)
2123 - Added ``ganeti-confd`` and a client library to query the Ganeti
2124 configuration via UDP
2125 - Added ability to run hooks after cluster initialization and before
2127 - Added automatic mode for disk replace (``gnt-instance replace-disks
2129 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2130 after catastrophic data-loss
2131 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2133 - Added ``gnt-instance move`` command to move instances
2134 - Added ``gnt-cluster watcher`` command to control watcher
2135 - Added ``gnt-node powercycle`` command to powercycle nodes
2136 - Added new job status field ``lock_status``
2137 - Added parseable error codes to cluster verification (``gnt-cluster
2138 verify --error-codes``) and made output less verbose (use
2139 ``--verbose`` to restore previous behaviour)
2140 - Added UUIDs to the main config entities (cluster, nodes, instances)
2141 - Added support for OS variants
2142 - Added support for hashed passwords in the Ganeti remote API users file
2144 - Added option to specify maximum timeout on instance shutdown
2145 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2146 - Added new helper script to start and stop Ganeti daemons
2147 (``daemon-util``), with the intent to reduce the work necessary to
2148 adjust Ganeti for non-Debian distributions and to start/stop daemons
2150 - Added more unittests
2151 - Fixed critical bug in ganeti-masterd startup
2152 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2153 now customisable at the cluster level for both the KVM and Xen
2154 hypervisors using the new ``migration_port`` parameter
2155 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2156 reinstalling an instance
2157 - Allowed ``@`` in tag names
2158 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2159 - Many documentation updates
2160 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2161 - ``gnt-instance reinstall`` can now reinstall multiple instances
2162 - Updated many command line parameters
2163 - Introduced new OS API version 15
2164 - No longer support a default hypervisor
2165 - Treat virtual LVs as inexistent
2166 - Improved job locking logic to reduce lock contention
2167 - Match instance and node names case insensitively
2168 - Reimplemented bash completion script to be more complete
2175 *(Released Thu, 4 Feb 2010)*
2177 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2178 - Fix a string formatting bug
2179 - Improve safety of the code in some error paths
2180 - Improve data validation in the master of values returned from nodes
2186 *(Released Thu, 17 Dec 2009)*
2188 - Fix security issue due to missing validation of iallocator names; this
2189 allows local and remote execution of arbitrary executables
2190 - Fix failure of gnt-node list during instance removal
2191 - Ship the RAPI documentation in the archive
2197 *(Released Wed, 30 Sep 2009)*
2199 - Fixed many wrong messages
2200 - Fixed a few bugs related to the locking library
2201 - Fixed MAC checking at instance creation time
2202 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2203 - Fixed a few issues related to signal handling in both daemons and
2205 - Fixed the example startup script provided
2206 - Fixed insserv dependencies in the example startup script (patch from
2208 - Fixed handling of drained nodes in the iallocator framework
2209 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2211 - Fixed error related to invalid job IDs in job polling
2212 - Fixed job/opcode persistence on unclean master shutdown
2213 - Fixed handling of partial job processing after unclean master
2215 - Fixed error reporting from LUs, previously all errors were converted
2216 into execution errors
2217 - Fixed error reporting from burnin
2218 - Decreased significantly the memory usage of the job queue
2219 - Optimised slightly multi-job submission
2220 - Optimised slightly opcode loading
2221 - Backported the multi-job submit framework from the development
2222 branch; multi-instance start and stop should be faster
2223 - Added script to clean archived jobs after 21 days; this will reduce
2224 the size of the queue directory
2225 - Added some extra checks in disk size tracking
2226 - Added an example ethers hook script
2227 - Added a cluster parameter that prevents Ganeti from modifying of
2229 - Added more node information to RAPI responses
2230 - Added a ``gnt-job watch`` command that allows following the ouput of a
2232 - Added a bind-address option to ganeti-rapi
2233 - Added more checks to the configuration verify
2234 - Enhanced the burnin script such that some operations can be retried
2236 - Converted instance reinstall to multi-instance model
2242 *(Released Fri, 7 Aug 2009)*
2244 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2245 to allow using the pre-2.0.2 behaviour in activation, if any existing
2246 instances have mismatched disk sizes in the configuration
2247 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2248 any configuration mismatches for disk sizes
2249 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2250 failover to work on two-node clusters
2251 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2253 - Fixed detection of OS script errors in ``gnt-backup export``
2254 - Fixed exit code of ``gnt-backup export``
2260 *(Released Fri, 17 Jul 2009)*
2262 - Added experimental support for stripped logical volumes; this should
2263 enhance performance but comes with a higher complexity in the block
2264 device handling; stripping is only enabled when passing
2265 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2266 affected even in the non-stripped mode
2267 - Improved resiliency against transient failures at the end of DRBD
2268 resyncs, and in general of DRBD resync checks
2269 - Fixed a couple of issues with exports and snapshot errors
2270 - Fixed a couple of issues in instance listing
2271 - Added display of the disk size in ``gnt-instance info``
2272 - Fixed checking for valid OSes in instance creation
2273 - Fixed handling of the "vcpus" parameter in instance listing and in
2274 general of invalid parameters
2275 - Fixed http server library, and thus RAPI, to handle invalid
2276 username/password combinations correctly; this means that now they
2277 report unauthorized for queries too, not only for modifications,
2278 allowing earlier detect of configuration problems
2279 - Added a new "role" node list field, equivalent to the master/master
2280 candidate/drained/offline flags combinations
2281 - Fixed cluster modify and changes of candidate pool size
2282 - Fixed cluster verify error messages for wrong files on regular nodes
2283 - Fixed a couple of issues with node demotion from master candidate role
2284 - Fixed node readd issues
2285 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2286 - Added a new ``--no-voting`` option for masterfailover to fix failover
2287 on two-nodes clusters when the former master node is unreachable
2288 - Added instance reinstall over RAPI
2294 *(Released Tue, 16 Jun 2009)*
2296 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2297 allow re-adding the start in single-user option (regression from 1.2)
2298 - the watcher writes the instance status to a file, to allow monitoring
2299 to report the instance status (from the master) based on cached
2300 results of the watcher's queries; while this can get stale if the
2301 watcher is being locked due to other work on the cluster, this is
2302 still an improvement
2303 - the watcher now also restarts the node daemon and the rapi daemon if
2305 - fixed the watcher to handle full and drained queue cases
2306 - hooks export more instance data in the environment, which helps if
2307 hook scripts need to take action based on the instance's properties
2308 (no longer need to query back into ganeti)
2309 - instance failovers when the instance is stopped do not check for free
2310 RAM, so that failing over a stopped instance is possible in low memory
2312 - rapi uses queries for tags instead of jobs (for less job traffic), and
2313 for cluster tags it won't talk to masterd at all but read them from
2315 - a couple of error handling fixes in RAPI
2316 - drbd handling: improved the error handling of inconsistent disks after
2317 resync to reduce the frequency of "there are some degraded disks for
2318 this instance" messages
2319 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2320 error handling path called a wrong function name)
2326 *(Released Wed, 27 May 2009)*
2328 - no changes from rc5
2334 *(Released Wed, 20 May 2009)*
2336 - fix a couple of bugs (validation, argument checks)
2337 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2338 - some small improvements to RAPI and IAllocator
2339 - make watcher automatically start the master daemon if down
2345 *(Released Mon, 27 Apr 2009)*
2347 - change the OS list to not require locks; this helps with big clusters
2348 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2349 volume group is broken
2350 - ``gnt-instance info``, without any arguments, doesn't run for all
2351 instances anymore; either pass ``--all`` or pass the desired
2352 instances; this helps against mistakes on big clusters where listing
2353 the information for all instances takes a long time
2354 - miscellaneous doc and man pages fixes
2360 *(Released Wed, 8 Apr 2009)*
2362 - Change the internal locking model of some ``gnt-node`` commands, in
2363 order to reduce contention (and blocking of master daemon) when
2364 batching many creation/reinstall jobs
2365 - Fixes to Xen soft reboot
2366 - No longer build documentation at build time, instead distribute it in
2367 the archive, in order to reduce the need for the whole docbook/rst
2374 *(Released Fri, 27 Mar 2009)*
2376 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2377 - Fix watcher startup sequence, improves the behaviour of busy clusters
2378 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2379 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2380 bind address and other places
2381 - Some documentation fixes and updates
2387 *(Released Mon, 2 Mar 2009)*
2389 - More documentation updates, now all docs should be more-or-less
2391 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2393 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2395 - Fix an issue related to $libdir/run/ganeti and cluster creation
2401 *(Released Thu, 19 Feb 2009)*
2403 - Xen PVM and KVM have switched the default value for the instance root
2404 disk to the first partition on the first drive, instead of the whole
2405 drive; this means that the OS installation scripts must be changed
2407 - Man pages have been updated
2408 - RAPI has been switched by default to HTTPS, and the exported functions
2409 should all work correctly
2410 - RAPI v1 has been removed
2411 - Many improvements to the KVM hypervisor
2412 - Block device errors are now better reported
2413 - Many other bugfixes and small improvements
2419 *(Released Mon, 26 Jan 2009)*
2421 - Version 2 is a general rewrite of the code and therefore the
2422 differences are too many to list, see the design document for 2.0 in
2423 the ``doc/`` subdirectory for more details
2424 - In this beta version there is not yet a migration path from 1.2 (there
2425 will be one in the final 2.0 release)
2426 - A few significant changes are:
2428 - all commands are executed by a daemon (``ganeti-masterd``) and the
2429 various ``gnt-*`` commands are just front-ends to it
2430 - all the commands are entered into, and executed from a job queue,
2431 see the ``gnt-job(8)`` manpage
2432 - the RAPI daemon supports read-write operations, secured by basic
2433 HTTP authentication on top of HTTPS
2434 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2435 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2436 to migrate to DRBD 8 first while still running Ganeti 1.2)
2437 - DRBD devices are using statically allocated minor numbers, which
2438 will be assigned to existing instances during the migration process
2439 - there is support for both Xen PVM and Xen HVM instances running on
2441 - KVM virtualization is supported too
2442 - file-based storage has been implemented, which means that it is
2443 possible to run the cluster without LVM and DRBD storage, for
2444 example using a shared filesystem exported from shared storage (and
2445 still have live migration)
2451 *(Released Tue, 13 Jan 2009)*
2453 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2454 - Reuse the old instance mac address by default on instance import, if
2455 the instance name is the same.
2456 - Handle situations in which the node info rpc returns incomplete
2458 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2459 - Improved version of batcher:
2461 - state file support
2462 - instance mac address support
2463 - support for HVM clusters/instances
2465 - Add an option to show the number of cpu sockets and nodes in
2467 - Support OSes that handle more than one version of the OS api (but do
2468 not change the current API in any other way)
2469 - Fix ``gnt-node migrate``
2470 - ``gnt-debug`` man page
2471 - Fixes various more typos and small issues
2472 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2478 *(Released Wed, 24 Sep 2008)*
2480 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2481 type of disk exported to fully virtualized instances.
2482 - provide access to the serial console of HVM instances
2483 - instance auto_balance flag, set by default. If turned off it will
2484 avoid warnings on cluster verify if there is not enough memory to fail
2485 over an instance. in the future it will prevent automatically failing
2486 it over when we will support that.
2487 - batcher tool for instance creation, see ``tools/README.batcher``
2488 - ``gnt-instance reinstall --select-os`` to interactively select a new
2489 operating system when reinstalling an instance.
2490 - when changing the memory amount on instance modify a check has been
2491 added that the instance will be able to start. also warnings are
2492 emitted if the instance will not be able to fail over, if auto_balance
2494 - documentation fixes
2495 - sync fields between ``gnt-instance list/modify/add/import``
2496 - fix a race condition in drbd when the sync speed was set after giving
2497 the device a remote peer.
2503 *(Released Tue, 22 Jul 2008)*
2505 - note: the allowed size and number of tags per object were reduced
2506 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2507 - fixed twisted 8.x compatibility
2508 - fixed ``gnt-instance replace-disks`` with iallocator
2509 - add TCP keepalives on twisted connections to detect restarted nodes
2510 - disk increase support, see ``gnt-instance grow-disk``
2511 - implement bulk node/instance query for RAPI
2512 - add tags in node/instance listing (optional)
2513 - experimental migration (and live migration) support, read the man page
2514 for ``gnt-instance migrate``
2515 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2516 has some small improvements in handling its state file
2522 *(Released Fri, 13 Jun 2008)*
2524 - Experimental readonly, REST-based remote API implementation;
2525 automatically started on master node, TCP port 5080, if enabled by
2526 ``--enable-rapi`` parameter to configure script.
2527 - Instance allocator support. Add and import instance accept a
2528 ``--iallocator`` parameter, and call that instance allocator to decide
2529 which node to use for the instance. The iallocator document describes
2530 what's expected from an allocator script.
2531 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2532 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2533 node is lost there is still enough memory to fail over the instances
2535 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2536 ``gnt-cluster verify``, to check for site-specific compliance. All the
2537 hooks will run, and their output, if any, will be displayed. Any
2538 failing hook will make the verification return an error value.
2539 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2540 primary and secondary interfaces
2541 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2542 that is still declared as part of the cluster and has failed.
2543 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2544 specifying output fields, that just adds the chosen fields to the
2546 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2547 export from the filesystem.
2548 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2549 have been added. Using them you can enable/disable acpi and pae
2550 support, and specify a path for a cd image to be exported to the
2551 instance. These parameters as the name suggest only work on HVM
2553 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2554 PAE support will be set to the previously hardcoded values, but the
2555 (previously hardcoded) path to the CDROM ISO image will be unset and
2556 if required, needs to be set manually with ``gnt-instance modify``
2558 - The address to which an instance's VNC console is bound is now
2559 selectable per-instance, rather than being cluster wide. Of course
2560 this only applies to instances controlled via VNC, so currently just
2561 applies to HVM clusters.
2567 *(Released Mon, 18 Feb 2008)*
2569 - more tweaks to the disk activation code (especially helpful for DRBD)
2570 - change the default ``gnt-instance list`` output format, now there is
2571 one combined status field (see the manpage for the exact values this
2573 - some more fixes for the mac export to hooks change
2574 - make Ganeti not break with DRBD 8.2.x (which changed the version
2575 format in ``/proc/drbd``) (issue 24)
2576 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2577 template, allowing migration from DRBD0.7+MD to DRBD8
2583 *(Released Wed, 30 Jan 2008)*
2585 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2587 - add command aliases infrastructure and a few aliases
2588 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2589 man pages and the ``--help`` option of ``gnt-node
2590 list``/``gnt-instance list``
2591 - fix ``gnt-backup list`` with down nodes (issue 21)
2592 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2593 - fix the dist archive and add a check for including svn/git files in
2595 - some developer-related changes: improve the burnin and the QA suite,
2596 add an upload script for testing during development
2602 *(Released Wed, 16 Jan 2008)*
2604 - experimental HVM support, read the install document, section
2605 "Initializing the cluster"
2606 - allow for the PVM hypervisor per-instance kernel and initrd paths
2607 - add a new command ``gnt-cluster verify-disks`` which uses a new
2608 algorithm to improve the reconnection of the DRBD pairs if the device
2609 on the secondary node has gone away
2610 - make logical volume code auto-activate LVs at disk activation time
2611 - slightly improve the speed of activating disks
2612 - allow specification of the MAC address at instance creation time, and
2613 changing it later via ``gnt-instance modify``
2614 - fix handling of external commands that generate lots of output on
2616 - update documentation with regard to minimum version of DRBD8 supported
2622 *(Released Tue, 4 Dec 2007)*
2624 - Log the ``xm create`` output to the node daemon log on failure (to
2625 help diagnosing the error)
2626 - In debug mode, log all external commands output if failed to the logs
2627 - Change parsing of lvm commands to ignore stderr
2633 *(Released Wed, 28 Nov 2007)*
2635 - Another round of updates to the DRBD 8 code to deal with more failures
2636 in the replace secondary node operation
2637 - Some more logging of failures in disk operations (lvm, drbd)
2638 - A few documentation updates
2645 *(Released Tue, 13 Nov 2007)*
2647 - Change configuration file format from Python's Pickle to JSON.
2648 Upgrading is possible using the cfgupgrade utility.
2649 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2650 faster replace disks and is more stable (DRBD 8 has many improvements
2651 compared to DRBD 0.7)
2652 - Added command line tags support (see man pages for ``gnt-instance``,
2653 ``gnt-node``, ``gnt-cluster``)
2654 - Added instance rename support
2655 - Added multi-instance startup/shutdown
2656 - Added cluster rename support
2657 - Added ``gnt-node evacuate`` to simplify some node operations
2658 - Added instance reboot operation that can speedup reboot as compared to
2660 - Soften the requirement that hostnames are in FQDN format
2661 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2663 - Removed dependency on debian's patched fping that uses the
2664 non-standard ``-S`` option
2665 - Now the OS definitions are searched for in multiple, configurable
2666 paths (easier for distros to package)
2667 - Some changes to the hooks infrastructure (especially the new
2668 post-configuration update hook)
2669 - Other small bugfixes
2671 .. vim: set textwidth=72 syntax=rst :