9 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
15 *(Released Tue, 2 Apr 2013)*
17 - Networks no longer have a "type" slot, since this information was
18 unused in Ganeti: instead of it tags should be used.
19 - Diskless instances are now externally mirrored (Issue 237). This for
20 now has only been tested in conjunction with explicit target nodes for
22 - The rapi client now has a ``target_node`` option to MigrateInstance.
23 - Fix early exit return code for hbal (Issue 386).
24 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
25 - Fix ``rbd showmapped`` output parsing (Issue 312).
26 - Networks are now referenced indexed by UUID, rather than name. This
27 will require running cfgupgrade, from 2.7.0beta1, if networks are in
29 - The OS environment now includes network information.
30 - Deleting of a network is now disallowed if any instance nic is using
31 it, to prevent dangling references.
32 - External storage is now documented in man pages.
33 - The exclusive_storage flag can now only be set at nodegroup level.
34 - Hbal can now submit an explicit priority with its jobs.
35 - Many network related locking fixes.
36 - Bump up the required pylint version to 0.25.1.
37 - Fix the ``no_remember`` option in RAPI client.
38 - Many ipolicy related tests, qa, and fixes.
39 - Many documentation improvements and fixes.
40 - Fix building with ``--disable-file-storage``.
41 - Fix ``-q`` option in htools, which was broken if passed more than
43 - Some haskell/python interaction improvements and fixes.
44 - Fix iallocator in case of missing LVM storage.
45 - Fix confd config load in case of ``--no-lvm-storage``.
46 - The confd/query functionality is now mentioned in the security
53 *(Released Wed, 6 Feb 2013)*
55 - ``gnt-instance batch-create`` has been changed to use the bulk create
56 opcode from Ganeti. This lead to incompatible changes in the format of
57 the JSON file. It's now not a custom dict anymore but a dict
58 compatible with the ``OpInstanceCreate`` opcode.
59 - Parent directories for file storage need to be listed in
60 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
61 write the file automatically based on old configuration values, but it
62 can not distribute it across all nodes and the file contents should be
63 verified. Use ``gnt-cluster copyfile
64 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
65 upgraded. The reason for requiring this list of paths now is that
66 before it would have been possible to inject new paths via RPC,
67 allowing files to be created in arbitrary locations. The RPC protocol
68 is protected using SSL/X.509 certificates, but as a design principle
69 Ganeti does not permit arbitrary paths to be passed.
70 - The parsing of the variants file for OSes (see
71 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
72 lines and comment lines (starting with ``#``) are ignored for better
74 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
75 longer available. ``gnt-node add`` now invokes a new tool on the
76 destination node, named ``prepare-node-join``, to configure the SSH
77 daemon. Paramiko is no longer necessary to configure nodes' SSH
78 daemons via ``gnt-node add``.
79 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
80 for RAPI users. It allows granting permissions to query for
81 information to a specific user without giving
82 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
83 - A new tool named ``node-cleanup`` has been added. It cleans remains of
84 a cluster from a machine by stopping all daemons, removing
85 certificates and ssconf files. Unless the ``--no-backup`` option is
86 given, copies of the certificates are made.
87 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
88 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
89 the flag is not reset after a master failover.
90 - Instance creations now support the use of opportunistic locking,
91 potentially speeding up the (parallel) creation of multiple instances.
92 This feature is currently only available via the :doc:`RAPI
93 <rapi>` interface and when an instance allocator is used. If the
94 ``opportunistic_locking`` parameter is set the opcode will try to
95 acquire as many locks as possible, but will not wait for any locks
96 held by other opcodes. If not enough resources can be found to
97 allocate the instance, the temporary error code
98 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
99 retried thereafter, with or without opportunistic locking.
100 - The functionality for allocating multiple instances at once has been
101 overhauled and is now also available through :doc:`RAPI <rapi>`.
102 - Man pages can now be included when the documentation is built, in
103 which case the output is in ``doc/man-html``. The configure-time
104 option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
106 - A new htool called ``hroller`` helps scheduling parallel shutdown of
107 nodes on clusters where all instances are shut down, but without ever
108 rebooting both primary and secondary node for an instance at the same
110 - Queries not needing locks or RPC access to the node can now be
111 performed by the confd daemon, making them independent from jobs, and
112 thus faster to execute. This is selectable at configure time.
113 - The KVM hypervisor has been updated with many new hypervisor
114 parameters, including a generic one for passing arbitrary command line
115 values. See a complete list in :manpage:`gnt-instance(8)`.
116 - A new tool, called ``mon-collector``, is the stand-alone executor of
117 the data collectors for a monitoring system. As of this version, it
118 just includes the DRBD data collector, that can be executed by calling
119 ``mon-collector`` using the ``drbd`` parameter. See
120 :manpage:`mon-collector(7)`.
121 - New ``exclusive-storage`` node parameter added. When it's set to true,
122 physical disks are assigned in an exclusive fashion to instances, as
123 documented in :doc:`Partitioned Ganeti <design-partitioned>`.
124 Currently, only instances using the ``plain`` disk template are
126 - New network management functionality to support automatic allocation
127 of IP addresses and managing of network parameters. See
128 :manpage:`gnt-network(8)` for more details.
129 - New external storage backend, to allow managing arbitrary storage
130 systems external to the cluster. See
131 :manpage:`ganeti-extstorage-interface(7)`.
132 - Instance renames of LVM-based instances will now update the LV tags
133 (which can be used to recover the instance-to-LV mapping in case of
135 - ``hbal`` will now exit with status 0 if, during job execution over
136 LUXI, early exit has been requested and all jobs are successful;
137 before, exit status 1 was used, which cannot be differentiated from
144 *(Released Fri, 21 Dec 2012)*
146 Important behaviour change: hbal won't rebalance anymore instances which
147 have the ``auto_balance`` attribute set to false. This was the intention
148 all along, but until now it only skipped those from the N+1 memory
149 reservation (DRBD-specific).
151 A significant number of bug fixes in this release:
153 - Fixed disk adoption interaction with ipolicy checks.
154 - Fixed networking issues when instances are started, stopped or
155 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
156 - Fixed the warning in cluster verify for shared storage instances not
158 - Fixed removal of storage directory on shared file storage (issue 262).
159 - Fixed validation of LVM volume group name in OpClusterSetParams
160 (``gnt-cluster modify``) (issue 285).
161 - Fixed runtime memory increases (``gnt-instance modify -m``).
162 - Fixed live migration under Xen's ``xl`` mode.
163 - Fixed ``gnt-instance console`` with ``xl``.
164 - Fixed building with newer Haskell compiler/libraries.
165 - Fixed PID file writing in Haskell daemons (confd); this prevents
166 restart issues if confd was launched manually (outside of
167 ``daemon-util``) while another copy of it was running
168 - Fixed a type error when doing live migrations with KVM (issue 297) and
169 the error messages for failing migrations have been improved.
170 - Fixed opcode validation for the out-of-band commands (``gnt-node
172 - Fixed a type error when unsetting OS hypervisor parameters (issue
173 311); now it's possible to unset all OS-specific hypervisor
175 - Fixed the ``dry-run`` mode for many operations: verification of
176 results was over-zealous but didn't take into account the ``dry-run``
177 operation, resulting in "wrong" failures.
178 - Fixed bash completion in ``gnt-job list`` when the job queue has
179 hundreds of entries; especially with older ``bash`` versions, this
180 results in significant CPU usage.
182 And lastly, a few other improvements have been made:
184 - Added option to force master-failover without voting (issue 282).
185 - Clarified error message on lock conflict (issue 287).
186 - Logging of newly submitted jobs has been improved (issue 290).
187 - Hostname checks have been made uniform between instance rename and
189 - The ``--submit`` option is now supported by ``gnt-debug delay``.
190 - Shutting down the master daemon by sending SIGTERM now stops it from
191 processing jobs waiting for locks; instead, those jobs will be started
192 once again after the master daemon is started the next time (issue
194 - Support for Xen's ``xl`` program has been improved (besides the fixes
196 - Reduced logging noise in the Haskell confd daemon (only show one log
197 entry for each config reload, instead of two).
198 - Several man page updates and typo fixes.
204 *(Released Fri, 12 Oct 2012)*
206 A small bugfix release. Among the bugs fixed:
208 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
209 made the command unusable.
210 - Commands that issue many jobs don't fail anymore just because some jobs
211 take so long that other jobs are archived.
212 - Failures during ``gnt-instance reinstall`` are reflected by the exit
214 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
216 - When ``always_failover`` is set, ``--allow-failover`` is not required
217 in migrate commands anymore.
218 - ``bash_completion`` works even if extglob is disabled.
219 - Fixed bug with locks that made failover for RDB-based instances fail.
220 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
221 a random node instead of one based on the allocator metric.
222 - Support for newer versions of pylint and pep8.
223 - Hail doesn't fail anymore when trying to add an instance of type
224 ``file``, ``sharedfile`` or ``rbd``.
225 - Added new Makefile target to rebuild the whole distribution, so that
226 all files are included.
232 *(Released Fri, 27 Jul 2012)*
235 .. attention:: The ``LUXI`` protocol has been made more consistent
236 regarding its handling of command arguments. This, however, leads to
237 incompatibility issues with previous versions. Please ensure that you
238 restart Ganeti daemons soon after the upgrade, otherwise most
239 ``LUXI`` calls (job submission, setting/resetting the drain flag,
240 pausing/resuming the watcher, cancelling and archiving jobs, querying
241 the cluster configuration) will fail.
250 The current ``admin_up`` field, which used to denote whether an instance
251 should be running or not, has been removed. Instead, ``admin_state`` is
252 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
254 The rational behind this is that an instance being “down” can have
257 - it could be down during a reboot
258 - it could be temporarily be down for a reinstall
259 - or it could be down because it is deprecated and kept just for its
262 The previous Boolean state was making it difficult to do capacity
263 calculations: should Ganeti reserve memory for a down instance? Now, the
264 tri-state field makes it clear:
266 - in ``up`` and ``down`` state, all resources are reserved for the
267 instance, and it can be at any time brought up if it is down
268 - in ``offline`` state, only disk space is reserved for it, but not
271 The field can have an extra use: since the transition between ``up`` and
272 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
273 transition between ``offline`` and ``down`` is done via ``gnt-instance
274 modify``, it is possible to given different rights to users. For
275 example, owners of an instance could be allowed to start/stop it, but
276 not transition it out of the offline state.
278 Instance policies and specs
279 +++++++++++++++++++++++++++
281 In previous Ganeti versions, an instance creation request was not
282 limited on the minimum size and on the maximum size just by the cluster
283 resources. As such, any policy could be implemented only in third-party
284 clients (RAPI clients, or shell wrappers over ``gnt-*``
285 tools). Furthermore, calculating cluster capacity via ``hspace`` again
286 required external input with regards to instance sizes.
288 In order to improve these workflows and to allow for example better
289 per-node group differentiation, we introduced instance specs, which
292 - minimum instance disk size, disk count, memory size, cpu count
293 - maximum values for the above metrics
294 - and “standard” values (used in ``hspace`` to calculate the standard
297 The minimum/maximum values can be also customised at node-group level,
298 for example allowing more powerful hardware to support bigger instance
301 Beside the instance specs, there are a few other settings belonging to
302 the instance policy framework. It is possible now to customise, per
303 cluster and node-group:
305 - the list of allowed disk templates
306 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
307 - the maximum ratio of instance to spindles (see below for more
308 information) for local storage
310 All these together should allow all tools that talk to Ganeti to know
311 what are the ranges of allowed values for instances and the
312 over-subscription that is allowed.
314 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
315 instance configuration, and the physical CPU configuration from the
316 node. For the spindle ratios however, we didn't track before these
317 values, so new parameters have been added:
319 - a new node parameter ``spindle_count``, defaults to 1, customisable at
320 node group or node level
321 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
323 Note that spindles in this context doesn't need to mean actual
324 mechanical hard-drives; it's just a relative number for both the node
325 I/O capacity and instance I/O consumption.
327 Instance migration behaviour
328 ++++++++++++++++++++++++++++
330 While live-migration is in general desirable over failover, it is
331 possible that for some workloads it is actually worse, due to the
332 variable time of the “suspend” phase during live migration.
334 To allow the tools to work consistently over such instances (without
335 having to hard-code instance names), a new backend parameter
336 ``always_failover`` has been added to control the migration/failover
337 behaviour. When set to True, all migration requests for an instance will
338 instead fall-back to failover.
340 Instance memory ballooning
341 ++++++++++++++++++++++++++
343 Initial support for memory ballooning has been added. The memory for an
344 instance is no longer fixed (backend parameter ``memory``), but instead
345 can vary between minimum and maximum values (backend parameters
346 ``minmem`` and ``maxmem``). Currently we only change an instance's
349 - live migrating or failing over and instance and the target node
350 doesn't have enough memory
351 - user requests changing the memory via ``gnt-instance modify
357 In order to control the use of specific CPUs by instance, support for
358 controlling CPU pinning has been added for the Xen, HVM and LXC
359 hypervisors. This is controlled by a new hypervisor parameter
360 ``cpu_mask``; details about possible values for this are in the
361 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
362 VCPU-to-CPU mapping) form will work well only when all nodes in your
363 cluster have the same amount of CPUs.
368 Another area in which Ganeti was not customisable were the parameters
369 used for storage configuration, e.g. how many stripes to use for LVM,
370 DRBD resync configuration, etc.
372 To improve this area, we've added disks parameters, which are
373 customisable at cluster and node group level, and which allow to
374 specify various parameters for disks (DRBD has the most parameters
375 currently), for example:
377 - DRBD resync algorithm and parameters (e.g. speed)
378 - the default VG for meta-data volumes for DRBD
379 - number of stripes for LVM (plain disk template)
382 These parameters can be modified via ``gnt-cluster modify -D …`` and
383 ``gnt-group modify -D …``, and are used at either instance creation (in
384 case of LVM stripes, for example) or at disk “activation” time
387 Rados block device support
388 ++++++++++++++++++++++++++
390 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
391 denoted by the ``rbd`` disk template type. This is considered
392 experimental, feedback is welcome. For details on configuring it, see
393 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
398 The existing master IP functionality works well only in simple setups (a
399 single network shared by all nodes); however, if nodes belong to
400 different networks, then the ``/32`` setup and lack of routing
401 information is not enough.
403 To allow the master IP to function well in more complex cases, the
404 system was reworked as follows:
406 - a master IP netmask setting has been added
407 - the master IP activation/turn-down code was moved from the node daemon
409 - whether to run the Ganeti-supplied master IP script or a user-supplied
410 on is a ``gnt-cluster init`` setting
412 Details about the location of the standard and custom setup scripts are
413 in the man page :manpage:`gnt-cluster(8)`; for information about the
414 setup script protocol, look at the Ganeti-supplied script.
419 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
422 It is now possible to use TLS-protected connections, and when renewing
423 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
424 it is now possible to specify spice or spice CA certificates. Also, it
425 is possible to configure a password for SPICE sessions via the
426 hypervisor parameter ``spice_password_file``.
428 There are also new parameters to control the compression and streaming
429 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
430 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
431 for the spice parameters.
433 Lastly, it is now possible to see the SPICE connection information via
434 ``gnt-instance console``.
439 A new tool (``tools/ovfconverter``) has been added that supports
440 conversion between Ganeti and the `Open Virtualization Format
441 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
444 This relies on the ``qemu-img`` tool to convert the disk formats, so the
445 actual compatibility with other virtualization solutions depends on it.
450 The configuration query daemon (``ganeti-confd``) is now optional, and
451 has been rewritten in Haskell; whether to use the daemon at all, use the
452 Python (default) or the Haskell version is selectable at configure time
453 via the ``--enable-confd`` parameter, which can take one of the
454 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
455 daemon will result in a smaller footprint; for larger systems, we
456 welcome feedback on the Haskell version which might become the default
459 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
460 daemon running. The Python version doesn't implement the new call.
463 User interface changes
464 ~~~~~~~~~~~~~~~~~~~~~~
466 We have replaced the ``--disks`` option of ``gnt-instance
467 replace-disks`` with a more flexible ``--disk`` option, which allows
468 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
469 disk size and mode can be changed upon recreation (via ``gnt-instance
470 recreate-disks``, which accepts the same ``--disk`` option).
472 As many people are used to a ``show`` command, we have added that as an
473 alias to ``info`` on all ``gnt-*`` commands.
475 The ``gnt-instance grow-disk`` command has a new mode in which it can
476 accept the target size of the disk, instead of the delta; this can be
477 more safe since two runs in absolute mode will be idempotent, and
478 sometimes it's also easier to specify the desired size directly.
480 Also the handling of instances with regard to offline secondaries has
481 been improved. Instance operations should not fail because one of it's
482 secondary nodes is offline, even though it's safe to proceed.
484 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
485 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
486 minors to instance name.
491 RAPI coverage has improved, with (for example) new resources for
492 recreate-disks, node power-cycle, etc.
497 There is partial support for ``xl`` in the Xen hypervisor; feedback is
500 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
501 whether to still support Python 2.4 or move to Python 2.6 as minimum
504 Support for Fedora has been slightly improved; the provided example
505 init.d script should work better on it and the INSTALL file should
506 document the needed dependencies.
511 The deprecated ``QueryLocks`` LUXI request has been removed. Use
512 ``Query(what=QR_LOCK, ...)`` instead.
514 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
515 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
516 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
517 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
518 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
520 RAPI client: ``CertificateError`` now derives from
521 ``GanetiApiError``. This should make it more easy to handle Ganeti
524 Deprecation warnings due to PyCrypto/paramiko import in
525 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
526 make sure to run an up-to-date paramiko version, if you use this tool.
528 The QA scripts now depend on Python 2.5 or above (the main code base
529 still works with Python 2.4).
531 The configuration file (``config.data``) is now written without
532 indentation for performance reasons; if you want to edit it, it can be
533 re-formatted via ``tools/fmtjson``.
535 A number of bugs has been fixed in the cluster merge tool.
537 ``x509`` certification verification (used in import-export) has been
538 changed to allow the same clock skew as permitted by the cluster
539 verification. This will remove some rare but hard to diagnose errors in
546 *(Released Thu, 19 Jul 2012)*
548 Very few changes from rc4 to the final release, only bugfixes:
550 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
551 instance, fix CDROM booting for KVM instances)
552 - fixed node group modification of node parameters
553 - fixed issue in LUClusterVerifyGroup with multi-group clusters
554 - fixed generation of bash completion to ensure a stable ordering
561 *(Released Fri, 13 Jul 2012)*
563 Third release candidate for 2.6. The following changes were done from
566 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
567 - Fixed an inconsistency in the LUXI protocol with the provided
568 arguments (NOT backwards compatible)
569 - Fixed a bug with node groups ipolicy where ``min`` was greater than
570 the cluster ``std`` value
571 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
572 easier instance debugging on nodes (requires ``hconfd`` to work)
578 *(Released Tue, 03 Jul 2012)*
580 Second release candidate for 2.6. The following changes were done from
583 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
585 - Fixed a RAPI regression on missing beparams/memory
586 - Fixed redistribution of files on offline nodes
587 - Added possibility to run activate-disks even though secondaries are
588 offline. With this change it relaxes also the strictness on some other
589 commands which use activate disks internally:
590 * ``gnt-instance start|reboot|rename|backup|export``
591 - Made it possible to remove safely an instance if its secondaries are
593 - Made it possible to reinstall even though secondaries are offline
599 *(Released Mon, 25 Jun 2012)*
601 First release candidate for 2.6. The following changes were done from
604 - Fixed bugs with disk parameters and ``rbd`` templates as well as
606 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
607 behaviour. It supports now the modify operation
608 - ``hcheck`` implemented to analyze cluster health and possibility of
609 improving health by rebalance
610 - ``hbal`` has been improved in dealing with split instances
616 *(Released Mon, 11 Jun 2012)*
618 Second beta release of 2.6. The following changes were done from beta2
621 - Fixed ``daemon-util`` with non-root user models
622 - Fixed creation of plain instances with ``--no-wait-for-sync``
623 - Fix wrong iv_names when running ``cfgupgrade``
624 - Export more information in RAPI group queries
625 - Fixed bug when changing instance network interfaces
626 - Extended burnin to do NIC changes
627 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
628 - Changed default for DRBD barriers
629 - Fixed DRBD error reporting for syncer rate
630 - Verify the options on disk parameters
632 And of course various fixes to documentation and improved unittests and
639 *(Released Wed, 23 May 2012)*
641 First beta release of 2.6. The following changes were done from beta1 to
644 - integrated patch for distributions without ``start-stop-daemon``
645 - adapted example init.d script to work on Fedora
646 - fixed log handling in Haskell daemons
647 - adapted checks in the watcher for pycurl linked against libnss
648 - add partial support for ``xl`` instead of ``xm`` for Xen
649 - fixed a type issue in cluster verification
650 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
653 Plus integrated fixes from the 2.5 branch:
655 - fixed ``kvm-ifup`` to use ``/bin/bash``
656 - fixed parallel build failures
657 - KVM live migration when using a custom keymap
663 *(Released Tue, 24 Jul 2012)*
665 A small bugfix release, with no new features:
667 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
668 different default shell (e.g. Debian, Ubuntu)
669 - fixed KVM startup and live migration with a custom keymap (fixes Issue
670 243 and Debian bug #650664)
671 - fixed compatibility with KVM versions that don't support multiple boot
672 devices (fixes Issue 230 and Debian bug #624256)
674 Additionally, a few fixes were done to the build system (fixed parallel
675 build failures) and to the unittests (fixed race condition in test for
676 FileID functions, and the default enable/disable mode for QA test is now
683 *(Released Fri, 11 May 2012)*
685 A small bugfix release.
687 The main issues solved are on the topic of compatibility with newer LVM
690 - fixed parsing of ``lv_attr`` field
691 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
692 the ``--force`` flag is needed
694 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
695 to accept kernel 3.x too (was hardcoded to 2.6.*).
697 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
698 of hook results and that also made display of other errors suboptimal
699 was fixed; the code behaves now like 2.4 and earlier.
701 Another change in 2.5, the cleanup of the OS scripts environment, is too
702 aggressive: it removed even the ``PATH`` variable, which requires the OS
703 scripts to *always* need to export it. Since this is a bit too strict,
704 we now export a minimal PATH, the same that we export for hooks.
706 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
707 integrated into this release.
709 Finally, a few other miscellaneous changes were done (no new features,
710 just small improvements):
712 - Fix ``gnt-group --help`` display
713 - Fix hardcoded Xen kernel path
714 - Fix grow-disk handling of invalid units
715 - Update synopsis for ``gnt-cluster repair-disk-sizes``
716 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
722 *(Released Thu, 12 Apr 2012)*
724 Incompatible/important changes and bugfixes
725 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
727 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
728 resource's ``ip_check`` parameter changed from ``True`` to ``False``
729 to match the underlying LUXI interface.
730 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
731 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
732 not maintain backwards-compatibility as the underlying operation
733 changed in an incompatible way. The RAPI client can talk to old
734 servers, but it needs to be told so as the return value changed.
735 - When creating file-based instances via RAPI, the ``file_driver``
736 parameter no longer defaults to ``loop`` and must be specified.
737 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
739 - Support for the undocumented and deprecated RAPI instance creation
740 request format version 0 has been dropped. Use version 1, supported
741 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
742 - Pyparsing 1.4.6 or above is required, see :doc:`installation
743 documentation <install>`.
744 - The "cluster-verify" hooks are now executed per group by the
745 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
746 if you just run ``gnt-cluster verify``, which generates one opcode per
748 - The environment as passed to the OS scripts is cleared, and thus no
749 environment variables defined in the node daemon's environment will be
750 inherited by the scripts.
751 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
753 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
754 support operations involving multiple node groups.
755 - Offline nodes are ignored when failing over an instance.
756 - Support for KVM version 1.0, which changed the version reporting format
758 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
760 - ``Makefile`` is now compatible with Automake 1.11.2
761 - Includes all bugfixes made in the 2.4 series
766 - The ganeti-htools project has been merged into the ganeti-core source
767 tree and will be built as part of Ganeti (see :doc:`install-quick`).
768 - Implemented support for :doc:`shared storage <design-shared-storage>`.
769 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
770 GPT-style partition tables (requires `parted
771 <http://www.gnu.org/s/parted/>`_).
772 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
773 - Allowed adding tags on instance creation.
774 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
775 - Allow instances to be started in a paused state, enabling the user to
776 see the complete console output on boot using the console.
777 - Added new hypervisor flag to control default reboot behaviour
778 (``reboot_behavior``).
779 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
780 - Improved out-of-band management support:
782 - Added ``gnt-node health`` command reporting the health status of
784 - Added ``gnt-node power`` command to manage power status of nodes.
785 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
787 - Instance migration can fall back to failover if instance is not
789 - Filters can be used when listing nodes, instances, groups and locks;
790 see :manpage:`ganeti(7)` manpage.
791 - Added post-execution status as variables to :doc:`hooks <hooks>`
793 - Instance tags are exported/imported together with the instance.
794 - When given an explicit job ID, ``gnt-job info`` will work for archived
796 - Jobs can define dependencies on other jobs (not yet supported via
797 RAPI or command line, but used by internal commands and usable via
800 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
803 - Instance failover is now available as a RAPI resource
804 (``/2/instances/[instance_name]/failover``).
805 - ``gnt-instance info`` defaults to static information if primary node
807 - Opcodes have a new ``comment`` attribute.
808 - Added basic SPICE support to KVM hypervisor.
809 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
811 Node group improvements
812 ~~~~~~~~~~~~~~~~~~~~~~~
814 - ``gnt-cluster verify`` has been modified to check groups separately,
815 thereby improving performance.
816 - Node group support has been added to ``gnt-cluster verify-disks``,
817 which now operates per node group.
818 - Watcher has been changed to work better with node groups.
820 - One process and state file per node group.
821 - Slow watcher in one group doesn't block other group's watcher.
823 - Added new command, ``gnt-group evacuate``, to move all instances in a
824 node group to other groups.
825 - Added ``gnt-instance change-group`` to move an instance to another
827 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
828 per-group operations.
829 - Node groups can be tagged.
830 - Some operations switch from an exclusive to a shared lock as soon as
832 - Instance's primary and secondary nodes' groups are now available as
833 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
834 and ``snodes.group.uuid``).
839 - Numerous updates to documentation and manpages.
841 - :doc:`RAPI <rapi>` documentation now has detailed parameter
843 - Some opcode/job results are now also documented, see :doc:`RAPI
846 - A lockset's internal lock is now also visible in lock monitor.
847 - Log messages from job queue workers now contain information about the
848 opcode they're processing.
849 - ``gnt-instance console`` no longer requires the instance lock.
850 - A short delay when waiting for job changes reduces the number of LUXI
851 requests significantly.
852 - DRBD metadata volumes are overwritten with zeros during disk creation.
853 - Out-of-band commands no longer acquire the cluster lock in exclusive
855 - ``devel/upload`` now uses correct permissions for directories.
861 *(Released Fri, 23 Mar 2012)*
863 This was the sixth release candidate of the 2.5 series.
869 *(Released Mon, 9 Jan 2012)*
871 This was the fifth release candidate of the 2.5 series.
877 *(Released Thu, 27 Oct 2011)*
879 This was the fourth release candidate of the 2.5 series.
885 *(Released Wed, 26 Oct 2011)*
887 This was the third release candidate of the 2.5 series.
893 *(Released Tue, 18 Oct 2011)*
895 This was the second release candidate of the 2.5 series.
901 *(Released Tue, 4 Oct 2011)*
903 This was the first release candidate of the 2.5 series.
909 *(Released Wed, 31 Aug 2011)*
911 This was the third beta release of the 2.5 series.
917 *(Released Mon, 22 Aug 2011)*
919 This was the second beta release of the 2.5 series.
925 *(Released Fri, 12 Aug 2011)*
927 This was the first beta release of the 2.5 series.
933 *(Released Thu, 27 Oct 2011)*
935 - Fixed bug when parsing command line parameter values ending in
937 - Fixed assertion error after unclean master shutdown
938 - Disable HTTP client pool for RPC, significantly reducing memory usage
940 - Fixed queue archive creation with wrong permissions
946 *(Released Tue, 23 Aug 2011)*
950 - Fixed documentation for importing with ``--src-dir`` option
951 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
952 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
958 *(Released Fri, 5 Aug 2011)*
960 Many bug-fixes and a few small features:
962 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
963 issues when you tried to add an instance with two MAC addresses in one
965 - KVM: fixed per-instance stored UID value
966 - KVM: configure bridged NICs at migration start
967 - KVM: Fix a bug where instance will not start with never KVM versions
969 - Added OS search path to ``gnt-cluster info``
970 - Fixed an issue with ``file_storage_dir`` where you were forced to
971 provide an absolute path, but the documentation states it is a
972 relative path, the documentation was right
973 - Added a new parameter to instance stop/start called ``--no-remember``
974 that will make the state change to not be remembered
975 - Implemented ``no_remember`` at RAPI level
976 - Improved the documentation
977 - Node evacuation: don't call IAllocator if node is already empty
978 - Fixed bug in DRBD8 replace disks on current nodes
979 - Fixed bug in recreate-disks for DRBD instances
980 - Moved assertion checking locks in ``gnt-instance replace-disks``
981 causing it to abort with not owning the right locks for some situation
982 - Job queue: Fixed potential race condition when cancelling queued jobs
983 - Fixed off-by-one bug in job serial generation
984 - ``gnt-node volumes``: Fix instance names
985 - Fixed aliases in bash completion
986 - Fixed a bug in reopening log files after being sent a SIGHUP
987 - Added a flag to burnin to allow specifying VCPU count
988 - Bugfixes to non-root Ganeti configuration
994 *(Released Thu, 12 May 2011)*
996 Many bug-fixes and a few new small features:
998 - Fixed a bug related to log opening failures
999 - Fixed a bug in instance listing with orphan instances
1000 - Fixed a bug which prevented resetting the cluster-level node parameter
1001 ``oob_program`` to the default
1002 - Many fixes related to the ``cluster-merge`` tool
1003 - Fixed a race condition in the lock monitor, which caused failures
1004 during (at least) creation of many instances in parallel
1005 - Improved output for gnt-job info
1006 - Removed the quiet flag on some ssh calls which prevented debugging
1008 - Improved the N+1 failure messages in cluster verify by actually
1009 showing the memory values (needed and available)
1010 - Increased lock attempt timeouts so that when executing long operations
1011 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1012 too early and thus prevent the use of the 'fair' mechanism
1013 - Changed instance query data (``gnt-instance info``) to not acquire
1014 locks unless needed, thus allowing its use on locked instance if only
1015 static information is asked for
1016 - Improved behaviour with filesystems that do not support rename on an
1018 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1019 kept locks on all nodes during the wipe, which is unneeded
1020 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1021 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1022 10GiB) which caused the wipe to fail right at the end in some cases
1023 - Fixed master IP activation when doing master failover with no-voting
1024 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1025 the master node itself
1026 - Fixed potential data-loss in under disk full conditions, where Ganeti
1027 wouldn't check correctly the return code and would consider
1028 partially-written files 'correct'
1029 - Fixed bug related to multiple VGs and DRBD disk replacing
1030 - Added new disk parameter ``metavg`` that allows placement of the meta
1031 device for DRBD in a different volume group
1032 - Fixed error handling in the node daemon when the system libc doesn't
1033 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1034 - Fixed lock release during replace-disks, which kept cluster-wide locks
1035 when doing disk replaces with an iallocator script
1036 - Added check for missing bridges in cluster verify
1037 - Handle EPIPE errors while writing to the terminal better, so that
1038 piping the output to e.g. ``less`` doesn't cause a backtrace
1039 - Fixed rare case where a ^C during Luxi calls could have been
1040 interpreted as server errors, instead of simply terminating
1041 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1043 - Added a few more parameters to the KVM hypervisor, allowing a second
1044 CDROM, custom disk type for CDROMs and a floppy image
1045 - Removed redundant message in instance rename when the name is given
1047 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1048 disks on new nodes, allowing recreation when the original instance
1049 nodes are completely gone
1050 - Added option when converting disk templates to DRBD to skip waiting
1051 for the resync, in order to make the instance available sooner
1052 - Added two new variables to the OS scripts environment (containing the
1054 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1055 to allow use of ``pvgrub`` as bootloader
1056 - Changed the instance memory modifications to only check out-of-memory
1057 conditions on memory increases, and turned the secondary node warnings
1058 into errors (they can still be overridden via ``--force``)
1059 - Fixed the handling of a corner case when the Python installation gets
1060 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1061 to execute a command that doesn't exist
1062 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1063 node of the instance returned failures during instance shutdown; this
1064 adds the option ``--ignore-consistency`` to gnt-instance move
1066 And as usual, various improvements to the error messages, documentation
1073 *(Released Wed, 09 Mar 2011)*
1075 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1076 the RAPI users file if run twice (even with ``--dry-run``).
1078 The release fixes that bug (nothing else changed).
1084 *(Released Mon, 07 Mar 2011)*
1086 Final 2.4.0 release. Just a few small fixes:
1088 - Fixed RAPI node evacuate
1089 - Fixed the kvm-ifup script
1090 - Fixed internal error handling for special job cases
1091 - Updated man page to specify the escaping feature for options
1097 *(Released Mon, 28 Feb 2011)*
1099 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1100 that this feature wiped the disks of the wrong instance, leading to loss
1105 - Fixed title of query field containing instance name
1106 - Expanded the glossary in the documentation
1107 - Fixed one unittest (internal issue)
1113 *(Released Mon, 21 Feb 2011)*
1115 A number of bug fixes plus just a couple functionality changes.
1117 On the user-visible side, the ``gnt-* list`` command output has changed
1118 with respect to "special" field states. The current rc1 style of display
1119 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1120 the default output mode special fields states are displayed as follows:
1122 - Offline resource: ``*``
1123 - Unavailable/not applicable: ``-``
1124 - Data missing (RPC failure): ``?``
1125 - Unknown field: ``??``
1127 Another user-visible change is the addition of ``--force-join`` to
1132 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1133 - Fixed regression in RAPI/instance reinstall where all parameters were
1134 required (instead of optional)
1135 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1136 - Fixed iallocator usage (offline nodes were not considered offline)
1137 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1138 - Fixed hypervisor and OS parameter validation with respect to
1139 non-vm_capable nodes
1140 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1142 - Fixed ``tools/listrunner`` with respect to agent-based usage
1148 *(Released Fri, 4 Feb 2011)*
1150 Many changes and fixes since the beta1 release. While there were some
1151 internal changes, the code has been mostly stabilised for the RC
1154 Note: the dumb allocator was removed in this release, as it was not kept
1155 up-to-date with the IAllocator protocol changes. It is recommended to
1156 use the ``hail`` command from the ganeti-htools package.
1158 Note: the 2.4 and up versions of Ganeti are not compatible with the
1159 0.2.x branch of ganeti-htools. You need to upgrade to
1160 ganeti-htools-0.3.0 (or later).
1162 Regressions fixed from 2.3
1163 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1165 - Fixed the ``gnt-cluster verify-disks`` command
1166 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1168 - Fixed disk adoption breakage
1169 - Fixed wrong headers in instance listing for field aliases
1174 - Fixed corner case in KVM handling of NICs
1175 - Fixed many cases of wrong handling of non-vm_capable nodes
1176 - Fixed a bug where a missing instance symlink was not possible to
1177 recreate with any ``gnt-*`` command (now ``gnt-instance
1178 activate-disks`` does it)
1179 - Fixed the volume group name as reported by ``gnt-cluster
1181 - Increased timeouts for the import-export code, hopefully leading to
1182 fewer aborts due network or instance timeouts
1183 - Fixed bug in ``gnt-node list-storage``
1184 - Fixed bug where not all daemons were started on cluster
1185 initialisation, but only at the first watcher run
1186 - Fixed many bugs in the OOB implementation
1187 - Fixed watcher behaviour in presence of instances with offline
1189 - Fixed instance list output for instances running on the wrong node
1190 - a few fixes to the cluster-merge tool, but it still cannot merge
1191 multi-node groups (currently it is not recommended to use this tool)
1197 - Improved network configuration for the KVM hypervisor
1198 - Added e1000 as a supported NIC for Xen-HVM
1199 - Improved the lvmstrap tool to also be able to use partitions, as
1200 opposed to full disks
1201 - Improved speed of disk wiping (the cluster parameter
1202 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1203 of instance creations
1204 - Added documentation for the OS parameters
1205 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1206 hypervisor is not responding
1207 - Added display of blacklisted and hidden OS information in
1208 ``gnt-cluster info``
1209 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1210 NIC and node parameters, which might create problems with currently
1211 invalid (but undetected) configuration files, but prevents validation
1212 failures when unrelated parameters are modified
1213 - Changed cluster initialisation to wait for the master daemon to become
1215 - Expanded the RAPI interface:
1217 - Added config redistribution resource
1218 - Added activation/deactivation of instance disks
1219 - Added export of console information
1221 - Implemented log file reopening on SIGHUP, which allows using
1222 logrotate(8) for the Ganeti log files
1223 - Added a basic OOB helper script as an example
1229 *(Released Fri, 14 Jan 2011)*
1234 - Fixed timezone issues when formatting timestamps
1235 - Added support for node groups, available via ``gnt-group`` and other
1237 - Added out-of-band framework and management, see :doc:`design
1238 document <design-oob>`
1239 - Removed support for roman numbers from ``gnt-node list`` and
1240 ``gnt-instance list``.
1241 - Allowed modification of master network interface via ``gnt-cluster
1242 modify --master-netdev``
1243 - Accept offline secondaries while shutting down instance disks
1244 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1245 - Added support for multiple LVM volume groups
1246 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1248 - Added commands to list available fields:
1250 - ``gnt-node list-fields``
1251 - ``gnt-group list-fields``
1252 - ``gnt-instance list-fields``
1254 - Updated documentation and man pages
1259 - Moved ``rapi_users`` file into separate directory, now named
1260 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1262 - Added new tool for running commands on many machines,
1263 ``tools/ganeti-listrunner``
1264 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1265 improving the ``gnt-instance console`` output
1266 - Allowed customisation of disk index separator at ``configure`` time
1267 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1268 - Added support for non-partitioned md disks in ``lvmstrap``
1269 - Added script to gracefully power off KVM instances
1270 - Split ``utils`` module into smaller parts
1271 - Changed query operations to return more detailed information, e.g.
1272 whether an information is unavailable due to an offline node. To use
1273 this new functionality, the LUXI call ``Query`` must be used. Field
1274 information is now stored by the master daemon and can be retrieved
1275 using ``QueryFields``. Instances, nodes and groups can also be queried
1276 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1277 exposed via RAPI). The following commands make use of this
1278 infrastructure change:
1280 - ``gnt-group list``
1281 - ``gnt-group list-fields``
1283 - ``gnt-node list-fields``
1284 - ``gnt-instance list``
1285 - ``gnt-instance list-fields``
1286 - ``gnt-debug locks``
1291 - New RAPI resources (see :doc:`rapi`):
1295 - ``/2/groups/[group_name]``
1296 - ``/2/groups/[group_name]/assign-nodes``
1297 - ``/2/groups/[group_name]/modify``
1298 - ``/2/groups/[group_name]/rename``
1299 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1303 - Implemented ``no_install`` for instance creation
1304 - Implemented OS parameters for instance reinstallation, allowing
1305 use of special settings on reinstallation (e.g. for preserving data)
1310 - Added IPv6 support in import/export
1311 - Pause DRBD synchronization while wiping disks on instance creation
1312 - Updated unittests and QA scripts
1313 - Improved network parameters passed to KVM
1314 - Converted man pages from docbook to reStructuredText
1320 *(Released Mon, 20 Dec 2010)*
1322 Released version 2.3.1~rc1 without any changes.
1328 *(Released Wed, 1 Dec 2010)*
1330 - impexpd: Disable OpenSSL compression in socat if possible (backport
1331 from master, commit e90739d625b, see :doc:`installation guide
1332 <install-quick>` for details)
1333 - Changed unittest coverage report to exclude test scripts
1334 - Added script to check version format
1340 *(Released Wed, 1 Dec 2010)*
1342 Released version 2.3.0~rc1 without any changes.
1348 *(Released Fri, 19 Nov 2010)*
1350 A number of bugfixes and documentation updates:
1352 - Update ganeti-os-interface documentation
1353 - Fixed a bug related to duplicate MACs or similar items which should be
1355 - Fix breakage in OS state modify
1356 - Reinstall instance: disallow offline secondaries (fixes bug related to
1357 OS changing but reinstall failing)
1358 - plus all the other fixes between 2.2.1 and 2.2.2
1364 *(Released Tue, 2 Nov 2010)*
1366 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1367 - Fixed master failover race with watcher
1368 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1370 - Accept previously stopped instance for export with instance removal
1371 - Simplify and extend the environment variables for instance OS scripts
1372 - Added new node flags, ``master_capable`` and ``vm_capable``
1373 - Added optional instance disk wiping prior during allocation. This is a
1374 cluster-wide option and can be set/modified using
1375 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1376 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1377 :doc:`install-quick`
1378 - Added a new watcher option (``--ignore-pause``)
1379 - Added option to ignore offline node on instance start/stop
1380 (``--ignore-offline``)
1381 - Allow overriding OS parameters with ``gnt-instance reinstall``
1382 - Added ability to change node's secondary IP address using ``gnt-node
1384 - Implemented privilege separation for all daemons except
1385 ``ganeti-noded``, see ``configure`` options
1386 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1388 - Implemented job priorities (see ``ganeti(7)`` manpage)
1389 - Ignore failures while shutting down instances during failover from
1391 - Exit daemon's bootstrap process only once daemon is ready
1392 - Export more information via ``LUInstanceQuery``/remote API
1393 - Improved documentation, QA and unittests
1394 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1395 restart if the file was created or changed
1396 - Added LUXI protocol version sent with each request and response,
1397 allowing detection of server/client mismatches
1398 - Moved the Python scripts among gnt-* and ganeti-* into modules
1399 - Moved all code related to setting up SSH to an external script,
1401 - Infrastructure changes for node group support in future versions
1407 *(Released Fri, 19 Nov 2010)*
1409 A few small bugs fixed, and some improvements to the build system:
1411 - Fix documentation regarding conversion to drbd
1412 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1414 - Fix error handling in node modify with multiple changes
1415 - Allow remote imports without checked names
1421 *(Released Tue, 19 Oct 2010)*
1423 - Disable SSL session ID cache in RPC client
1429 *(Released Thu, 14 Oct 2010)*
1431 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1432 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1434 - Fix problems with interaction between Curl and Python's HTTP server,
1435 resulting in increased speed in many RPC calls
1436 - Improve our release script to prevent breakage with older aclocal and
1443 *(Released Thu, 7 Oct 2010)*
1445 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1446 value retrieved from master
1447 - Added support for blacklisted or hidden OS definitions
1448 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1449 - Added support for -mem-path in KVM hypervisor abstraction layer
1450 - Allow overriding instance parameters in tool for inter-cluster
1451 instance moves (``tools/move-instance``)
1452 - Improved opcode summaries (e.g. in ``gnt-job list``)
1453 - Improve consistency of OS listing by sorting it
1454 - Documentation updates
1460 *(Released Fri, 8 Oct 2010)*
1462 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1468 *(Released Mon, 4 Oct 2010)*
1470 - Fixed regression in ``gnt-instance rename``
1476 *(Released Wed, 22 Sep 2010)*
1478 - Fixed OS_VARIANT variable for OS scripts
1479 - Fixed cluster tag operations via RAPI
1480 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1481 - Disabled RAPI CA checks in watcher
1487 *(Released Mon, 23 Aug 2010)*
1489 - Support DRBD versions of the format "a.b.c.d"
1491 - Re-introduce support for usage from multiple threads in RAPI client
1492 - Instance renames and modify via RAPI
1493 - Work around race condition between processing and archival in job
1495 - Mark opcodes following failed one as failed, too
1496 - Job field ``lock_status`` was removed due to difficulties making it
1497 work with the changed job queue in Ganeti 2.2; a better way to monitor
1498 locks is expected for a later 2.2.x release
1499 - Fixed dry-run behaviour with many commands
1500 - Support ``ssh-agent`` again when adding nodes
1501 - Many additional bugfixes
1507 *(Released Fri, 30 Jul 2010)*
1509 Important change: the internal RPC mechanism between Ganeti nodes has
1510 changed from using a home-grown http library (based on the Python base
1511 libraries) to use the PycURL library. This requires that PycURL is
1512 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1513 against GnuTLS by default. cURL's support for GnuTLS had known issues
1514 before cURL 7.21.0 and we recommend using the latest cURL release or
1515 linking against OpenSSL. Most other distributions already link PycURL
1516 and cURL against OpenSSL. The command::
1518 python -c 'import pycurl; print pycurl.version'
1520 can be used to determine the libraries PycURL and cURL are linked
1523 Other significant changes:
1525 - Rewrote much of the internals of the job queue, in order to achieve
1526 better parallelism; this decouples job query operations from the job
1527 processing, and it should allow much nicer behaviour of the master
1528 daemon under load, and it also has uncovered some long-standing bugs
1529 related to the job serialisation (now fixed)
1530 - Added a default iallocator setting to the cluster parameters,
1531 eliminating the need to always pass nodes or an iallocator for
1532 operations that require selection of new node(s)
1533 - Added experimental support for the LXC virtualization method
1534 - Added support for OS parameters, which allows the installation of
1535 instances to pass parameter to OS scripts in order to customise the
1537 - Added a hypervisor parameter controlling the migration type (live or
1538 non-live), since hypervisors have various levels of reliability; this
1539 has renamed the 'live' parameter to 'mode'
1540 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1541 logical volumes, meaning that cluster verify will ignore them and not
1542 flag their presence as errors
1543 - The watcher will now reset the error count for failed instances after
1544 8 hours, thus allowing self-healing if the problem that caused the
1545 instances to be down/fail to start has cleared in the meantime
1546 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1547 check for, and warn, if the drbd module parameter ``usermode_helper``
1548 is not consistent with the cluster-wide setting; this is needed to
1549 make diagnose easier of failed drbd creations
1550 - Started adding base IPv6 support, but this is not yet
1551 enabled/available for use
1552 - Rename operations (cluster, instance) will now return the new name,
1553 which is especially useful if a short name was passed in
1554 - Added support for instance migration in RAPI
1555 - Added a tool to pre-configure nodes for the SSH setup, before joining
1556 them to the cluster; this will allow in the future a simplified model
1557 for node joining (but not yet fully enabled in 2.2); this needs the
1558 paramiko python library
1559 - Fixed handling of name-resolving errors
1560 - Fixed consistency of job results on the error path
1561 - Fixed master-failover race condition when executed multiple times in
1563 - Fixed many bugs related to the job queue (mostly introduced during the
1564 2.2 development cycle, so not all are impacting 2.1)
1565 - Fixed instance migration with missing disk symlinks
1566 - Fixed handling of unknown jobs in ``gnt-job archive``
1567 - And many other small fixes/improvements
1571 - Enhanced both the unittest and the QA coverage
1572 - Switched the opcode validation to a generic model, and extended the
1573 validation to all opcode parameters
1574 - Changed more parts of the code that write shell scripts to use the
1576 - Switched the master daemon to use the asyncore library for the Luxi
1583 *(Released Thu, 17 Jun 2010)*
1585 - Added tool (``move-instance``) and infrastructure to move instances
1586 between separate clusters (see :doc:`separate documentation
1587 <move-instance>` and :doc:`design document <design-2.2>`)
1588 - Added per-request RPC timeout
1589 - RAPI now requires a Content-Type header for requests with a body (e.g.
1590 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1591 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1592 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1594 - Implemented initial support for running Ganeti daemons as separate
1595 users, see configure-time flags ``--with-user-prefix`` and
1596 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1598 - Instances can be removed after export (``gnt-backup export
1599 --remove-instance``)
1600 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1601 key (instead of 1024 bit)
1602 - Added new cluster configuration file for cluster domain secret
1603 - Import/export now use SSL instead of SSH
1604 - Added support for showing estimated time when exporting an instance,
1605 see the ``ganeti-os-interface(7)`` manpage and look for
1612 *(Released Tue, 16 Nov 2010)*
1614 Some more bugfixes. Unless critical bugs occur, this will be the last
1617 - Fix case of MAC special-values
1618 - Fix mac checker regex
1619 - backend: Fix typo causing "out of range" error
1620 - Add missing --units in gnt-instance list man page
1626 *(Released Tue, 24 Aug 2010)*
1629 - Don't ignore secondary node silently on non-mirrored disk templates
1631 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1632 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1633 - Properly document the usb_mouse parameter
1634 - Fix path in ganeti-rapi(8) (issue 116)
1635 - Adjust error message when the ganeti user's .ssh directory is
1637 - Add same-node-check when changing the disk template to drbd
1643 *(Released Fri, 16 Jul 2010)*
1646 - Add an option to only select some reboot types during qa/burnin.
1647 (on some hypervisors consequent reboots are not supported)
1648 - Fix infrequent race condition in master failover. Sometimes the old
1649 master ip address would be still detected as up for a short time
1650 after it was removed, causing failover to fail.
1651 - Decrease mlockall warnings when the ctypes module is missing. On
1652 Python 2.4 we support running even if no ctypes module is installed,
1653 but we were too verbose about this issue.
1654 - Fix building on old distributions, on which man doesn't have a
1656 - Fix RAPI not to ignore the MAC address on instance creation
1657 - Implement the old instance creation format in the RAPI client.
1663 *(Released Thu, 01 Jul 2010)*
1665 A small bugfix release:
1666 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1667 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1668 a non-existing option
1669 - Fix instance create: the --force-variant option was ignored
1670 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1671 - Fix modify node storage with non-FQDN arguments
1672 - Fix RAPI client to authenticate under Python 2.6 when used
1673 for more than 5 requests needing authentication
1674 - Fix gnt-instance modify -t (storage) giving a wrong error message
1675 when converting a non-shutdown drbd instance to plain
1681 *(Released Fri, 18 Jun 2010)*
1683 A small bugfix release:
1685 - Fix live migration of KVM instances started with older Ganeti
1686 versions which had fewer hypervisor parameters
1687 - Fix gnt-instance grow-disk on down instances
1688 - Fix an error-reporting bug during instance migration
1689 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1690 silently ignoring broken ones
1691 - Fix an RPC error reporting bug affecting, for example, RAPI client
1693 - Fix bug triggered by different API version os-es on different nodes
1694 - Fix a bug in instance startup with custom hvparams: OS level
1695 parameters would fail to be applied.
1696 - Fix the RAPI client under Python 2.6 (but more work is needed to
1697 make it work completely well with OpenSSL)
1698 - Fix handling of errors when resolving names from DNS
1704 *(Released Thu, 3 Jun 2010)*
1706 A medium sized development cycle. Some new features, and some
1707 fixes/small improvements/cleanups.
1709 Significant features
1710 ~~~~~~~~~~~~~~~~~~~~
1712 The node deamon now tries to mlock itself into memory, unless the
1713 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1714 its logs, and falls back to console logging. This allows emergency
1715 features such as ``gnt-node powercycle`` to work even in the event of a
1716 broken node disk (tested offlining the disk hosting the node's
1717 filesystem and dropping its memory caches; don't try this at home)
1719 KVM: add vhost-net acceleration support. It can be tested with a new
1720 enough version of the kernel and of qemu-kvm.
1722 KVM: Add instance chrooting feature. If you use privilege dropping for
1723 your VMs you can also now force them to chroot to an empty directory,
1724 before starting the emulated guest.
1726 KVM: Add maximum migration bandwith and maximum downtime tweaking
1727 support (requires a new-enough version of qemu-kvm).
1729 Cluster verify will now warn if the master node doesn't have the master
1730 ip configured on it.
1732 Add a new (incompatible) instance creation request format to RAPI which
1733 supports all parameters (previously only a subset was supported, and it
1734 wasn't possible to extend the old format to accomodate all the new
1735 features. The old format is still supported, and a client can check for
1736 this feature, before using it, by checking for its presence in the
1737 ``features`` RAPI resource.
1739 Now with ancient latin support. Try it passing the ``--roman`` option to
1740 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1741 (requires the python-roman module to be installed, in order to work).
1746 As usual many internal code refactorings, documentation updates, and
1749 - Lots of improvements and cleanups to the experimental Remote API
1750 (RAPI) client library.
1751 - A new unit test suite for the core daemon libraries.
1752 - A fix to creating missing directories makes sure the umask is not
1753 applied anymore. This enforces the same directory permissions
1755 - Better handling terminating daemons with ctrl+c (used when running
1756 them in debugging mode).
1757 - Fix a race condition in live migrating a KVM instance, when stat()
1758 on the old proc status file returned EINVAL, which is an unexpected
1760 - Fixed manpage checking with newer man and utf-8 charachters. But now
1761 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1767 *(Released Fri, 7 May 2010)*
1769 Fix a bug which prevented untagged KVM instances from starting.
1775 *(Released Fri, 7 May 2010)*
1777 Another release with a long development cycle, during which many
1778 different features were added.
1780 Significant features
1781 ~~~~~~~~~~~~~~~~~~~~
1783 The KVM hypervisor now can run the individual instances as non-root, to
1784 reduce the impact of a VM being hijacked due to bugs in the
1785 hypervisor. It is possible to run all instances as a single (non-root)
1786 user, to manually specify a user for each instance, or to dynamically
1787 allocate a user out of a cluster-wide pool to each instance, with the
1788 guarantee that no two instances will run under the same user ID on any
1791 An experimental RAPI client library, that can be used standalone
1792 (without the other Ganeti libraries), is provided in the source tree as
1793 ``lib/rapi/client.py``. Note this client might change its interface in
1794 the future, as we iterate on its capabilities.
1796 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1797 replace the cluster's certificates and crypto keys. This might help in
1798 case they have been compromised, or have simply expired.
1800 A new disk option for instance creation has been added that allows one
1801 to "adopt" currently existing logical volumes, with data
1802 preservation. This should allow easier migration to Ganeti from
1803 unmanaged (or managed via other software) instances.
1805 Another disk improvement is the possibility to convert between redundant
1806 (DRBD) and plain (LVM) disk configuration for an instance. This should
1807 allow better scalability (starting with one node and growing the
1808 cluster, or shrinking a two-node cluster to one node).
1810 A new feature that could help with automated node failovers has been
1811 implemented: if a node sees itself as offline (by querying the master
1812 candidates), it will try to shutdown (hard) all instances and any active
1813 DRBD devices. This reduces the risk of duplicate instances if an
1814 external script automatically failovers the instances on such nodes. To
1815 enable this, the cluster parameter ``maintain_node_health`` should be
1816 enabled; in the future this option (per the name) will enable other
1817 automatic maintenance features.
1819 Instance export/import now will reuse the original instance
1820 specifications for all parameters; that means exporting an instance,
1821 deleting it and the importing it back should give an almost identical
1822 instance. Note that the default import behaviour has changed from
1823 before, where it created only one NIC; now it recreates the original
1826 Cluster verify has added a few new checks: SSL certificates validity,
1827 /etc/hosts consistency across the cluster, etc.
1832 As usual, many internal changes were done, documentation fixes,
1835 - Fixed cluster initialization with disabled cluster storage (regression
1836 introduced in 2.1.1)
1837 - File-based storage supports growing the disks
1838 - Fixed behaviour of node role changes
1839 - Fixed cluster verify for some corner cases, plus a general rewrite of
1840 cluster verify to allow future extension with more checks
1841 - Fixed log spamming by watcher and node daemon (regression introduced
1843 - Fixed possible validation issues when changing the list of enabled
1845 - Fixed cleanup of /etc/hosts during node removal
1846 - Fixed RAPI response for invalid methods
1847 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1848 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1849 from ide disks, etc.)
1850 - Allow OS changes without re-installation (to record a changed OS
1851 outside of Ganeti, or to allow OS renames)
1852 - Allow instance creation without OS installation (useful for example if
1853 the OS will be installed manually, or restored from a backup not in
1855 - Implemented option to make cluster ``copyfile`` use the replication
1857 - Added list of enabled hypervisors to ssconf (possibly useful for
1859 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1861 - A partial form of node re-IP is possible via node readd, which now
1862 allows changed node primary IP
1863 - Command line utilities now show an informational message if the job is
1865 - The logs of the master daemon now show the PID/UID/GID of the
1872 *(Released Fri, 12 Mar 2010)*
1874 During the 2.1.0 long release candidate cycle, a lot of improvements and
1875 changes have accumulated with were released later as 2.1.1.
1880 The node evacuate command (``gnt-node evacuate``) was significantly
1881 rewritten, and as such the IAllocator protocol was changed - a new
1882 request type has been added. This unfortunate change during a stable
1883 series is designed to improve performance of node evacuations; on
1884 clusters with more than about five nodes and which are well-balanced,
1885 evacuation should proceed in parallel for all instances of the node
1886 being evacuated. As such, any existing IAllocator scripts need to be
1887 updated, otherwise the above command will fail due to the unknown
1888 request. The provided "dumb" allocator has not been updated; but the
1889 ganeti-htools package supports the new protocol since version 0.2.4.
1891 Another important change is increased validation of node and instance
1892 names. This might create problems in special cases, if invalid host
1893 names are being used.
1895 Also, a new layer of hypervisor parameters has been added, that sits at
1896 OS level between the cluster defaults and the instance ones. This allows
1897 customisation of virtualization parameters depending on the installed
1898 OS. For example instances with OS 'X' may have a different KVM kernel
1899 (or any other parameter) than the cluster defaults. This is intended to
1900 help managing a multiple OSes on the same cluster, without manual
1901 modification of each instance's parameters.
1903 A tool for merging clusters, ``cluster-merge``, has been added in the
1904 tools sub-directory.
1909 - Improved the int/float conversions that should make the code more
1910 robust in face of errors from the node daemons
1911 - Fixed the remove node code in case of internal configuration errors
1912 - Fixed the node daemon behaviour in face of inconsistent queue
1913 directory (e.g. read-only file-system where we can't open the files
1915 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1916 now it either aborts cleanly or, if given the new "auto_promote"
1917 parameter, will automatically promote other nodes as needed
1918 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1919 completely prevent Ganeti from working
1920 - Fixed bug for instance export when not all disks were successfully
1922 - Fixed behaviour of node add when the new node is slow in starting up
1924 - Fixed handling of signals in the LUXI client, which should improve
1925 behaviour of command-line scripts
1926 - Added checks for invalid node/instance names in the configuration (now
1927 flagged during cluster verify)
1928 - Fixed watcher behaviour for disk activation errors
1929 - Fixed two potentially endless loops in http library, which led to the
1930 RAPI daemon hanging and consuming 100% CPU in some cases
1931 - Fixed bug in RAPI daemon related to hashed passwords
1932 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1934 - Enhanced compatibility with non-Debian OSes, but not using absolute
1935 path in some commands and allowing customisation of the ssh
1936 configuration directory
1937 - Fixed possible future issue with new Python versions by abiding to the
1938 proper use of ``__slots__`` attribute on classes
1939 - Added checks that should prevent directory traversal attacks
1940 - Many documentation fixes based on feedback from users
1945 - Added an "early_release" more for instance replace disks and node
1946 evacuate, where we release locks earlier and thus allow higher
1947 parallelism within the cluster
1948 - Added watcher hooks, intended to allow the watcher to restart other
1949 daemons (e.g. from the ganeti-nbma project), but they can be used of
1950 course for any other purpose
1951 - Added a compile-time disable for DRBD barriers, to increase
1952 performance if the administrator trusts the power supply or the
1953 storage system to not lose writes
1954 - Added the option of using syslog for logging instead of, or in
1955 addition to, Ganeti's own log files
1956 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1957 can indeed boot from a paravirtual NIC
1958 - Added a generic debug level for many operations; while this is not
1959 used widely yet, it allows one to pass the debug value all the way to
1961 - Enhanced the hooks environment for instance moves (failovers,
1962 migrations) where the primary/secondary nodes changed during the
1963 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1964 - Enhanced data validations for many user-supplied values; one important
1965 item is the restrictions imposed on instance and node names, which
1966 might reject some (invalid) host names
1967 - Add a configure-time option to disable file-based storage, if it's not
1968 needed; this allows greater security separation between the master
1969 node and the other nodes from the point of view of the inter-node RPC
1971 - Added user notification in interactive tools if job is waiting in the
1972 job queue or trying to acquire locks
1973 - Added log messages when a job is waiting for locks
1974 - Added filtering by node tags in instance operations which admit
1975 multiple instances (start, stop, reboot, reinstall)
1976 - Added a new tool for cluster mergers, ``cluster-merge``
1977 - Parameters from command line which are of the form ``a=b,c=d`` can now
1978 use backslash escapes to pass in values which contain commas,
1979 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1981 - For KVM, the instance name is the first parameter passed to KVM, so
1982 that it's more visible in the process list
1988 *(Released Tue, 2 Mar 2010)*
1990 Ganeti 2.1 brings many improvements with it. Major changes:
1992 - Added infrastructure to ease automated disk repairs
1993 - Added new daemon to export configuration data in a cheaper way than
1994 using the remote API
1995 - Instance NICs can now be routed instead of being associated with a
1997 - Improved job locking logic to reduce impact of jobs acquiring multiple
1998 locks waiting for other long-running jobs
2000 In-depth implementation details can be found in the Ganeti 2.1 design
2006 - Added chroot hypervisor
2007 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2009 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2010 ``bootloader_path`` and ``bootloader_args``)
2011 - Added the ``use_localtime`` option for the xen-hvm and kvm
2012 hypervisors, and the default value for this has changed to false (in
2013 2.0 xen-hvm always enabled it)
2014 - Added luxi call to submit multiple jobs in one go
2015 - Added cluster initialization option to not modify ``/etc/hosts``
2017 - Added network interface parameters
2018 - Added dry run mode to some LUs
2019 - Added RAPI resources:
2021 - ``/2/instances/[instance_name]/info``
2022 - ``/2/instances/[instance_name]/replace-disks``
2023 - ``/2/nodes/[node_name]/evacuate``
2024 - ``/2/nodes/[node_name]/migrate``
2025 - ``/2/nodes/[node_name]/role``
2026 - ``/2/nodes/[node_name]/storage``
2027 - ``/2/nodes/[node_name]/storage/modify``
2028 - ``/2/nodes/[node_name]/storage/repair``
2030 - Added OpCodes to evacuate or migrate all instances on a node
2031 - Added new command to list storage elements on nodes (``gnt-node
2032 list-storage``) and modify them (``gnt-node modify-storage``)
2033 - Added new ssconf files with master candidate IP address
2034 (``ssconf_master_candidates_ips``), node primary IP address
2035 (``ssconf_node_primary_ips``) and node secondary IP address
2036 (``ssconf_node_secondary_ips``)
2037 - Added ``ganeti-confd`` and a client library to query the Ganeti
2038 configuration via UDP
2039 - Added ability to run hooks after cluster initialization and before
2041 - Added automatic mode for disk replace (``gnt-instance replace-disks
2043 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2044 after catastrophic data-loss
2045 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2047 - Added ``gnt-instance move`` command to move instances
2048 - Added ``gnt-cluster watcher`` command to control watcher
2049 - Added ``gnt-node powercycle`` command to powercycle nodes
2050 - Added new job status field ``lock_status``
2051 - Added parseable error codes to cluster verification (``gnt-cluster
2052 verify --error-codes``) and made output less verbose (use
2053 ``--verbose`` to restore previous behaviour)
2054 - Added UUIDs to the main config entities (cluster, nodes, instances)
2055 - Added support for OS variants
2056 - Added support for hashed passwords in the Ganeti remote API users file
2058 - Added option to specify maximum timeout on instance shutdown
2059 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2060 - Added new helper script to start and stop Ganeti daemons
2061 (``daemon-util``), with the intent to reduce the work necessary to
2062 adjust Ganeti for non-Debian distributions and to start/stop daemons
2064 - Added more unittests
2065 - Fixed critical bug in ganeti-masterd startup
2066 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2067 now customisable at the cluster level for both the KVM and Xen
2068 hypervisors using the new ``migration_port`` parameter
2069 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2070 reinstalling an instance
2071 - Allowed ``@`` in tag names
2072 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2073 - Many documentation updates
2074 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2075 - ``gnt-instance reinstall`` can now reinstall multiple instances
2076 - Updated many command line parameters
2077 - Introduced new OS API version 15
2078 - No longer support a default hypervisor
2079 - Treat virtual LVs as inexistent
2080 - Improved job locking logic to reduce lock contention
2081 - Match instance and node names case insensitively
2082 - Reimplemented bash completion script to be more complete
2089 *(Released Thu, 4 Feb 2010)*
2091 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2092 - Fix a string formatting bug
2093 - Improve safety of the code in some error paths
2094 - Improve data validation in the master of values returned from nodes
2100 *(Released Thu, 17 Dec 2009)*
2102 - Fix security issue due to missing validation of iallocator names; this
2103 allows local and remote execution of arbitrary executables
2104 - Fix failure of gnt-node list during instance removal
2105 - Ship the RAPI documentation in the archive
2111 *(Released Wed, 30 Sep 2009)*
2113 - Fixed many wrong messages
2114 - Fixed a few bugs related to the locking library
2115 - Fixed MAC checking at instance creation time
2116 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2117 - Fixed a few issues related to signal handling in both daemons and
2119 - Fixed the example startup script provided
2120 - Fixed insserv dependencies in the example startup script (patch from
2122 - Fixed handling of drained nodes in the iallocator framework
2123 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2125 - Fixed error related to invalid job IDs in job polling
2126 - Fixed job/opcode persistence on unclean master shutdown
2127 - Fixed handling of partial job processing after unclean master
2129 - Fixed error reporting from LUs, previously all errors were converted
2130 into execution errors
2131 - Fixed error reporting from burnin
2132 - Decreased significantly the memory usage of the job queue
2133 - Optimised slightly multi-job submission
2134 - Optimised slightly opcode loading
2135 - Backported the multi-job submit framework from the development
2136 branch; multi-instance start and stop should be faster
2137 - Added script to clean archived jobs after 21 days; this will reduce
2138 the size of the queue directory
2139 - Added some extra checks in disk size tracking
2140 - Added an example ethers hook script
2141 - Added a cluster parameter that prevents Ganeti from modifying of
2143 - Added more node information to RAPI responses
2144 - Added a ``gnt-job watch`` command that allows following the ouput of a
2146 - Added a bind-address option to ganeti-rapi
2147 - Added more checks to the configuration verify
2148 - Enhanced the burnin script such that some operations can be retried
2150 - Converted instance reinstall to multi-instance model
2156 *(Released Fri, 7 Aug 2009)*
2158 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2159 to allow using the pre-2.0.2 behaviour in activation, if any existing
2160 instances have mismatched disk sizes in the configuration
2161 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2162 any configuration mismatches for disk sizes
2163 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2164 failover to work on two-node clusters
2165 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2167 - Fixed detection of OS script errors in ``gnt-backup export``
2168 - Fixed exit code of ``gnt-backup export``
2174 *(Released Fri, 17 Jul 2009)*
2176 - Added experimental support for stripped logical volumes; this should
2177 enhance performance but comes with a higher complexity in the block
2178 device handling; stripping is only enabled when passing
2179 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2180 affected even in the non-stripped mode
2181 - Improved resiliency against transient failures at the end of DRBD
2182 resyncs, and in general of DRBD resync checks
2183 - Fixed a couple of issues with exports and snapshot errors
2184 - Fixed a couple of issues in instance listing
2185 - Added display of the disk size in ``gnt-instance info``
2186 - Fixed checking for valid OSes in instance creation
2187 - Fixed handling of the "vcpus" parameter in instance listing and in
2188 general of invalid parameters
2189 - Fixed http server library, and thus RAPI, to handle invalid
2190 username/password combinations correctly; this means that now they
2191 report unauthorized for queries too, not only for modifications,
2192 allowing earlier detect of configuration problems
2193 - Added a new "role" node list field, equivalent to the master/master
2194 candidate/drained/offline flags combinations
2195 - Fixed cluster modify and changes of candidate pool size
2196 - Fixed cluster verify error messages for wrong files on regular nodes
2197 - Fixed a couple of issues with node demotion from master candidate role
2198 - Fixed node readd issues
2199 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2200 - Added a new ``--no-voting`` option for masterfailover to fix failover
2201 on two-nodes clusters when the former master node is unreachable
2202 - Added instance reinstall over RAPI
2208 *(Released Tue, 16 Jun 2009)*
2210 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2211 allow re-adding the start in single-user option (regression from 1.2)
2212 - the watcher writes the instance status to a file, to allow monitoring
2213 to report the instance status (from the master) based on cached
2214 results of the watcher's queries; while this can get stale if the
2215 watcher is being locked due to other work on the cluster, this is
2216 still an improvement
2217 - the watcher now also restarts the node daemon and the rapi daemon if
2219 - fixed the watcher to handle full and drained queue cases
2220 - hooks export more instance data in the environment, which helps if
2221 hook scripts need to take action based on the instance's properties
2222 (no longer need to query back into ganeti)
2223 - instance failovers when the instance is stopped do not check for free
2224 RAM, so that failing over a stopped instance is possible in low memory
2226 - rapi uses queries for tags instead of jobs (for less job traffic), and
2227 for cluster tags it won't talk to masterd at all but read them from
2229 - a couple of error handling fixes in RAPI
2230 - drbd handling: improved the error handling of inconsistent disks after
2231 resync to reduce the frequency of "there are some degraded disks for
2232 this instance" messages
2233 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2234 error handling path called a wrong function name)
2240 *(Released Wed, 27 May 2009)*
2242 - no changes from rc5
2248 *(Released Wed, 20 May 2009)*
2250 - fix a couple of bugs (validation, argument checks)
2251 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2252 - some small improvements to RAPI and IAllocator
2253 - make watcher automatically start the master daemon if down
2259 *(Released Mon, 27 Apr 2009)*
2261 - change the OS list to not require locks; this helps with big clusters
2262 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2263 volume group is broken
2264 - ``gnt-instance info``, without any arguments, doesn't run for all
2265 instances anymore; either pass ``--all`` or pass the desired
2266 instances; this helps against mistakes on big clusters where listing
2267 the information for all instances takes a long time
2268 - miscellaneous doc and man pages fixes
2274 *(Released Wed, 8 Apr 2009)*
2276 - Change the internal locking model of some ``gnt-node`` commands, in
2277 order to reduce contention (and blocking of master daemon) when
2278 batching many creation/reinstall jobs
2279 - Fixes to Xen soft reboot
2280 - No longer build documentation at build time, instead distribute it in
2281 the archive, in order to reduce the need for the whole docbook/rst
2288 *(Released Fri, 27 Mar 2009)*
2290 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2291 - Fix watcher startup sequence, improves the behaviour of busy clusters
2292 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2293 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2294 bind address and other places
2295 - Some documentation fixes and updates
2301 *(Released Mon, 2 Mar 2009)*
2303 - More documentation updates, now all docs should be more-or-less
2305 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2307 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2309 - Fix an issue related to $libdir/run/ganeti and cluster creation
2315 *(Released Thu, 19 Feb 2009)*
2317 - Xen PVM and KVM have switched the default value for the instance root
2318 disk to the first partition on the first drive, instead of the whole
2319 drive; this means that the OS installation scripts must be changed
2321 - Man pages have been updated
2322 - RAPI has been switched by default to HTTPS, and the exported functions
2323 should all work correctly
2324 - RAPI v1 has been removed
2325 - Many improvements to the KVM hypervisor
2326 - Block device errors are now better reported
2327 - Many other bugfixes and small improvements
2333 *(Released Mon, 26 Jan 2009)*
2335 - Version 2 is a general rewrite of the code and therefore the
2336 differences are too many to list, see the design document for 2.0 in
2337 the ``doc/`` subdirectory for more details
2338 - In this beta version there is not yet a migration path from 1.2 (there
2339 will be one in the final 2.0 release)
2340 - A few significant changes are:
2342 - all commands are executed by a daemon (``ganeti-masterd``) and the
2343 various ``gnt-*`` commands are just front-ends to it
2344 - all the commands are entered into, and executed from a job queue,
2345 see the ``gnt-job(8)`` manpage
2346 - the RAPI daemon supports read-write operations, secured by basic
2347 HTTP authentication on top of HTTPS
2348 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2349 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2350 to migrate to DRBD 8 first while still running Ganeti 1.2)
2351 - DRBD devices are using statically allocated minor numbers, which
2352 will be assigned to existing instances during the migration process
2353 - there is support for both Xen PVM and Xen HVM instances running on
2355 - KVM virtualization is supported too
2356 - file-based storage has been implemented, which means that it is
2357 possible to run the cluster without LVM and DRBD storage, for
2358 example using a shared filesystem exported from shared storage (and
2359 still have live migration)
2365 *(Released Tue, 13 Jan 2009)*
2367 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2368 - Reuse the old instance mac address by default on instance import, if
2369 the instance name is the same.
2370 - Handle situations in which the node info rpc returns incomplete
2372 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2373 - Improved version of batcher:
2375 - state file support
2376 - instance mac address support
2377 - support for HVM clusters/instances
2379 - Add an option to show the number of cpu sockets and nodes in
2381 - Support OSes that handle more than one version of the OS api (but do
2382 not change the current API in any other way)
2383 - Fix ``gnt-node migrate``
2384 - ``gnt-debug`` man page
2385 - Fixes various more typos and small issues
2386 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2392 *(Released Wed, 24 Sep 2008)*
2394 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2395 type of disk exported to fully virtualized instances.
2396 - provide access to the serial console of HVM instances
2397 - instance auto_balance flag, set by default. If turned off it will
2398 avoid warnings on cluster verify if there is not enough memory to fail
2399 over an instance. in the future it will prevent automatically failing
2400 it over when we will support that.
2401 - batcher tool for instance creation, see ``tools/README.batcher``
2402 - ``gnt-instance reinstall --select-os`` to interactively select a new
2403 operating system when reinstalling an instance.
2404 - when changing the memory amount on instance modify a check has been
2405 added that the instance will be able to start. also warnings are
2406 emitted if the instance will not be able to fail over, if auto_balance
2408 - documentation fixes
2409 - sync fields between ``gnt-instance list/modify/add/import``
2410 - fix a race condition in drbd when the sync speed was set after giving
2411 the device a remote peer.
2417 *(Released Tue, 22 Jul 2008)*
2419 - note: the allowed size and number of tags per object were reduced
2420 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2421 - fixed twisted 8.x compatibility
2422 - fixed ``gnt-instance replace-disks`` with iallocator
2423 - add TCP keepalives on twisted connections to detect restarted nodes
2424 - disk increase support, see ``gnt-instance grow-disk``
2425 - implement bulk node/instance query for RAPI
2426 - add tags in node/instance listing (optional)
2427 - experimental migration (and live migration) support, read the man page
2428 for ``gnt-instance migrate``
2429 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2430 has some small improvements in handling its state file
2436 *(Released Fri, 13 Jun 2008)*
2438 - Experimental readonly, REST-based remote API implementation;
2439 automatically started on master node, TCP port 5080, if enabled by
2440 ``--enable-rapi`` parameter to configure script.
2441 - Instance allocator support. Add and import instance accept a
2442 ``--iallocator`` parameter, and call that instance allocator to decide
2443 which node to use for the instance. The iallocator document describes
2444 what's expected from an allocator script.
2445 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2446 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2447 node is lost there is still enough memory to fail over the instances
2449 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2450 ``gnt-cluster verify``, to check for site-specific compliance. All the
2451 hooks will run, and their output, if any, will be displayed. Any
2452 failing hook will make the verification return an error value.
2453 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2454 primary and secondary interfaces
2455 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2456 that is still declared as part of the cluster and has failed.
2457 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2458 specifying output fields, that just adds the chosen fields to the
2460 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2461 export from the filesystem.
2462 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2463 have been added. Using them you can enable/disable acpi and pae
2464 support, and specify a path for a cd image to be exported to the
2465 instance. These parameters as the name suggest only work on HVM
2467 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2468 PAE support will be set to the previously hardcoded values, but the
2469 (previously hardcoded) path to the CDROM ISO image will be unset and
2470 if required, needs to be set manually with ``gnt-instance modify``
2472 - The address to which an instance's VNC console is bound is now
2473 selectable per-instance, rather than being cluster wide. Of course
2474 this only applies to instances controlled via VNC, so currently just
2475 applies to HVM clusters.
2481 *(Released Mon, 18 Feb 2008)*
2483 - more tweaks to the disk activation code (especially helpful for DRBD)
2484 - change the default ``gnt-instance list`` output format, now there is
2485 one combined status field (see the manpage for the exact values this
2487 - some more fixes for the mac export to hooks change
2488 - make Ganeti not break with DRBD 8.2.x (which changed the version
2489 format in ``/proc/drbd``) (issue 24)
2490 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2491 template, allowing migration from DRBD0.7+MD to DRBD8
2497 *(Released Wed, 30 Jan 2008)*
2499 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2501 - add command aliases infrastructure and a few aliases
2502 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2503 man pages and the ``--help`` option of ``gnt-node
2504 list``/``gnt-instance list``
2505 - fix ``gnt-backup list`` with down nodes (issue 21)
2506 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2507 - fix the dist archive and add a check for including svn/git files in
2509 - some developer-related changes: improve the burnin and the QA suite,
2510 add an upload script for testing during development
2516 *(Released Wed, 16 Jan 2008)*
2518 - experimental HVM support, read the install document, section
2519 "Initializing the cluster"
2520 - allow for the PVM hypervisor per-instance kernel and initrd paths
2521 - add a new command ``gnt-cluster verify-disks`` which uses a new
2522 algorithm to improve the reconnection of the DRBD pairs if the device
2523 on the secondary node has gone away
2524 - make logical volume code auto-activate LVs at disk activation time
2525 - slightly improve the speed of activating disks
2526 - allow specification of the MAC address at instance creation time, and
2527 changing it later via ``gnt-instance modify``
2528 - fix handling of external commands that generate lots of output on
2530 - update documentation with regard to minimum version of DRBD8 supported
2536 *(Released Tue, 4 Dec 2007)*
2538 - Log the ``xm create`` output to the node daemon log on failure (to
2539 help diagnosing the error)
2540 - In debug mode, log all external commands output if failed to the logs
2541 - Change parsing of lvm commands to ignore stderr
2547 *(Released Wed, 28 Nov 2007)*
2549 - Another round of updates to the DRBD 8 code to deal with more failures
2550 in the replace secondary node operation
2551 - Some more logging of failures in disk operations (lvm, drbd)
2552 - A few documentation updates
2559 *(Released Tue, 13 Nov 2007)*
2561 - Change configuration file format from Python's Pickle to JSON.
2562 Upgrading is possible using the cfgupgrade utility.
2563 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2564 faster replace disks and is more stable (DRBD 8 has many improvements
2565 compared to DRBD 0.7)
2566 - Added command line tags support (see man pages for ``gnt-instance``,
2567 ``gnt-node``, ``gnt-cluster``)
2568 - Added instance rename support
2569 - Added multi-instance startup/shutdown
2570 - Added cluster rename support
2571 - Added ``gnt-node evacuate`` to simplify some node operations
2572 - Added instance reboot operation that can speedup reboot as compared to
2574 - Soften the requirement that hostnames are in FQDN format
2575 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2577 - Removed dependency on debian's patched fping that uses the
2578 non-standard ``-S`` option
2579 - Now the OS definitions are searched for in multiple, configurable
2580 paths (easier for distros to package)
2581 - Some changes to the hooks infrastructure (especially the new
2582 post-configuration update hook)
2583 - Other small bugfixes
2585 .. vim: set textwidth=72 syntax=rst :