8 *(Released Wed, 5 Mar 2014)*
10 - Fix incorrect invocation of hooks on offline nodes (issue 742)
11 - Fix incorrect exit code of gnt-cluster verify in certain circumstances
14 Inherited from the 2.9 branch:
16 - Fix overflow problem in hbal that caused it to break when waiting for
17 jobs for more than 10 minutes (issue 717)
18 - Make hbal properly handle non-LVM storage
19 - Properly export and import NIC parameters, and do so in a backwards
20 compatible way (issue 716)
21 - Fix net-common script in case of routed mode (issue 728)
22 - Improve documentation (issues 724, 730)
28 *(Released Thu, 20 Feb 2014)*
30 Incompatible/important changes
31 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
33 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
34 default. Specify --no-wait-for-sync to override this behavior.
35 - The Ganeti python code now adheres to a private-module layout. In particular,
36 the module 'ganeti' is no longer in the python search path.
37 - On instance allocation, the iallocator now considers non-LVM storage
38 properly. In particular, actual file storage space information is used
39 when allocating space for a file/sharedfile instance.
40 - When disabling disk templates cluster-wide, the cluster now first
41 checks whether there are instances still using those templates.
42 - 'gnt-node list-storage' now also reports storage information about
43 file-based storage types.
44 - In case of non drbd instances, export \*_SECONDARY environment variables
45 as empty strings (and not "None") during 'instance-migrate' related hooks.
50 - KVM hypervisors can now access RBD storage directly without having to
51 go through a block device.
52 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
53 procedure between two Ganeti versions that are both 2.10 or higher.
54 - The move-instance command can now change disk templates when moving
55 instances, and does not require any node placement options to be
56 specified if the destination cluster has a default iallocator.
57 - Users can now change the soundhw and cpuid settings for XEN hypervisors.
58 - Hail and hbal now have the (optional) capability of accessing average CPU
59 load information through the monitoring deamon, and to use it to dynamically
60 adapt the allocation of instances.
61 - Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
62 so that disk and NIC modifications take effect without the need of actual
63 reboot. There are a couple of constrains currently for this feature:
65 - only KVM hypervisor (versions >= 1.0) supports it,
66 - one can not (yet) hotplug a disk using userspace access mode for RBD
67 - in case of a downgrade instances should suffer a reboot in order to
68 be migratable (due to core change of runtime files)
73 - A new test framework for logical units was introduced and the test
74 coverage for logical units was improved significantly.
75 - Opcodes are entirely generated from Haskell using the tool 'hs2py' and
76 the module 'src/Ganeti/OpCodes.hs'.
77 - Constants are also generated from Haskell using the tool
78 'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
79 exception of socket related constants, which require changing the
80 cluster configuration file, and HVS related constants, because they
81 are part of a port of instance queries to Haskell. As a result, these
82 changes will be part of the next release of Ganeti.
87 The following new dependencies have been added/updated.
91 - The version requirements for ``python-mock`` have increased to at least
92 version 1.0.1. It is still used for testing only.
97 - Fix integer overflow problem in hbal
103 *(Released Wed, 12 Feb 2014)*
105 This was the third RC release of the 2.10 series. Since 2.10.0 rc2:
107 - Improved hotplug robustness
108 - Start Ganeti daemons after ensure-dirs during upgrade
109 - Documentation improvements
111 Inherited from the 2.9 branch:
113 - Fix the RAPI instances-multi-alloc call
114 - assign unique filenames to file-based disks
115 - gracefully handle degraded non-diskless instances with 0 disks (issue 697)
116 - noded now runs with its specified group, which is the default group,
117 defaulting to root (issue 707)
118 - make using UUIDs to identify nodes in gnt-node consistently possible
125 *(Released Fri, 31 Jan 2014)*
127 This was the second RC release of the 2.10 series. Since 2.10.0 rc1:
129 - Documentation improvements
130 - Run drbdsetup syncer only on network attach
131 - Include target node in hooks nodes for migration
133 - Support post-upgrade hooks during cluster upgrades
135 Inherited from the 2.9 branch:
137 - Ensure that all the hypervisors exist in the config file (Issue 640)
138 - Correctly recognise the role as master node (Issue 687)
139 - configure: allow detection of Sphinx 1.2+ (Issue 502)
140 - gnt-instance now honors the KVM path correctly (Issue 691)
142 Inherited from the 2.8 branch:
144 - Change the list separator for the usb_devices parameter from comma to space.
145 Commas could not work because they are already the hypervisor option
146 separator (Issue 649)
147 - Add support for blktap2 file-driver (Issue 638)
148 - Add network tag definitions to the haskell codebase (Issue 641)
149 - Fix RAPI network tag handling
150 - Add the network tags to the tags searched by gnt-cluster search-tags
151 - Fix caching bug preventing jobs from being cancelled
152 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
158 *(Released Tue, 17 Dec 2013)*
160 This was the first RC release of the 2.10 series. Since 2.10.0 beta1:
162 - All known issues in 2.10.0 beta1 have been resolved (see changes from
164 - Improve handling of KVM runtime files from earlier Ganeti versions
165 - Documentation fixes
167 Inherited from the 2.9 branch:
169 - use custom KVM path if set for version checking
170 - SingleNotifyPipeCondition: don't share pollers
172 Inherited from the 2.8 branch:
174 - Fixed Luxi daemon socket permissions after master-failover
175 - Improve IP version detection code directly checking for colons rather than
176 passing the family from the cluster object
177 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
178 opportunistically anymore (Issue 622)
179 - Allow link local IPv6 gateways (Issue 624)
180 - Fix error printing (Issue 616)
181 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
182 disk modifications, keep the old one. If name=none then set disk name to
184 - Update build_chroot script to work with the latest hackage packages
185 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
186 - Fix evacuation out of drained node (Issue 615)
187 - Add default file_driver if missing (Issue 571)
188 - Fix job error message after unclean master shutdown (Issue 618)
189 - Lock group(s) when creating instances (Issue 621)
190 - SetDiskID() before accepting an instance (Issue 633)
191 - Allow the ext template disks to receive arbitrary parameters, both at creation
192 time and while being modified
193 - Xen handle domain shutdown (future proofing cherry-pick)
194 - Refactor reading live data in htools (future proofing cherry-pick)
200 *(Released Wed, 27 Nov 2013)*
202 This was the first beta release of the 2.10 series. All important changes
203 are listed in the latest 2.10 entry.
208 The following issues are known to be present in the beta and will be fixed
211 - Issue 477: Wrong permissions for confd LUXI socket
212 - Issue 621: Instance related opcodes do not aquire network/group locks
213 - Issue 622: Assertion Error: Node locks differ from node resource locks
214 - Issue 623: IPv6 Masterd <-> Luxid communication error
220 *(Released Tue, 25 Feb 2014)*
222 - Fix overflow problem in hbal that caused it to break when waiting for
223 jobs for more than 10 minutes (issue 717)
224 - Make hbal properly handle non-LVM storage
225 - Properly export and import NIC parameters, and do so in a backwards
226 compatible way (issue 716)
227 - Fix net-common script in case of routed mode (issue 728)
228 - Improve documentation (issues 724, 730)
234 *(Released Mon, 10 Feb 2014)*
236 - Fix the RAPI instances-multi-alloc call
237 - assign unique filenames to file-based disks
238 - gracefully handle degraded non-diskless instances with 0 disks (issue 697)
239 - noded now runs with its specified group, which is the default group,
240 defaulting to root (issue 707)
241 - make using UUIDs to identify nodes in gnt-node consistently possible
248 *(Released Mon, 27 Jan 2014)*
250 - Ensure that all the hypervisors exist in the config file (Issue 640)
251 - Correctly recognise the role as master node (Issue 687)
252 - configure: allow detection of Sphinx 1.2+ (Issue 502)
253 - gnt-instance now honors the KVM path correctly (Issue 691)
255 Inherited from the 2.8 branch:
257 - Change the list separator for the usb_devices parameter from comma to space.
258 Commas could not work because they are already the hypervisor option
259 separator (Issue 649)
260 - Add support for blktap2 file-driver (Issue 638)
261 - Add network tag definitions to the haskell codebase (Issue 641)
262 - Fix RAPI network tag handling
263 - Add the network tags to the tags searched by gnt-cluster search-tags
264 - Fix caching bug preventing jobs from being cancelled
265 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
271 *(Released Fri, 13 Dec 2013)*
273 - use custom KVM path if set for version checking
274 - SingleNotifyPipeCondition: don't share pollers
276 Inherited from the 2.8 branch:
278 - Fixed Luxi daemon socket permissions after master-failover
279 - Improve IP version detection code directly checking for colons rather than
280 passing the family from the cluster object
281 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
282 opportunistically anymore (Issue 622)
283 - Allow link local IPv6 gateways (Issue 624)
284 - Fix error printing (Issue 616)
285 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
286 disk modifications, keep the old one. If name=none then set disk name to
288 - Update build_chroot script to work with the latest hackage packages
289 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
290 - Fix evacuation out of drained node (Issue 615)
291 - Add default file_driver if missing (Issue 571)
292 - Fix job error message after unclean master shutdown (Issue 618)
293 - Lock group(s) when creating instances (Issue 621)
294 - SetDiskID() before accepting an instance (Issue 633)
295 - Allow the ext template disks to receive arbitrary parameters, both at creation
296 time and while being modified
297 - Xen handle domain shutdown (future proofing cherry-pick)
298 - Refactor reading live data in htools (future proofing cherry-pick)
304 *(Released Wed, 13 Nov 2013)*
306 - fix bug, that kept nodes offline when readding
307 - when verifying DRBD versions, ignore unavailable nodes
308 - fix bug that made the console unavailable on kvm in split-user
310 - DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
316 *(Released Tue, 5 Nov 2013)*
318 Incompatible/important changes
319 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
321 - hroller now also plans for capacity to move non-redundant instances off
322 any node to be rebooted; the old behavior of completely ignoring any
323 non-redundant instances can be restored by adding the --ignore-non-redundant
325 - The cluster option '--no-lvm-storage' was removed in favor of the new option
326 '--enabled-disk-templates'.
327 - On instance creation, disk templates no longer need to be specified
328 with '-t'. The default disk template will be taken from the list of
329 enabled disk templates.
330 - The monitoring daemon is now running as root, in order to be able to collect
331 information only available to root (such as the state of Xen instances).
332 - The ConfD client is now IPv6 compatible.
333 - File and shared file storage is no longer dis/enabled at configure time,
334 but using the option '--enabled-disk-templates' at cluster initialization and
336 - The default directories for file and shared file storage are not anymore
337 specified at configure time, but taken from the cluster's configuration.
338 They can be set at cluster initialization and modification with
339 '--file-storage-dir' and '--shared-file-storage-dir'.
340 - Cluster verification now includes stricter checks regarding the
341 default file and shared file storage directories. It now checks that
342 the directories are explicitely allowed in the 'file-storage-paths' file and
343 that the directories exist on all nodes.
344 - The list of allowed disk templates in the instance policy and the list
345 of cluster-wide enabled disk templates is now checked for consistency
346 on cluster or group modification. On cluster initialization, the ipolicy
347 disk templates are ensured to be a subset of the cluster-wide enabled
353 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
354 the correct command syntax. It is possible to use different DRBD versions
355 on different nodes as long as they are compatible to each other. This
356 enables rolling upgrades of DRBD with no downtime. As permanent operation
357 of different DRBD versions within a node group is discouraged,
358 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
359 - New "inst-status-xen" data collector for the monitoring daemon, providing
360 information about the state of the xen instances on the nodes.
361 - New "lv" data collector for the monitoring daemon, collecting data about the
362 logical volumes on the nodes, and pairing them with the name of the instances
364 - New "diskstats" data collector, collecting the data from /proc/diskstats and
365 presenting them over the monitoring daemon interface.
366 - The ConfD client is now IPv6 compatible.
370 The following new dependencies have been added.
374 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
375 for the unit tests (and only used for testing).
379 - ``hslogger`` (http://software.complete.org/hslogger) is now always
380 required, even if confd is not enabled.
385 - Correctly start/stop luxid during gnt-cluster master-failover (inherited
387 - Improved error messsages (inherited from stable-2.8)
393 *(Released Tue, 15 Oct 2013)*
395 The third release candidate in the 2.9 series. Since 2.9.0 rc2:
397 - in implicit configuration upgrade, match ipolicy with enabled disk templates
398 - improved harep documentation (inherited from stable-2.8)
404 *(Released Wed, 9 Oct 2013)*
406 The second release candidate in the 2.9 series. Since 2.9.0 rc1:
408 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
409 at least one DRBD instance.
410 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
417 *(Released Tue, 1 Oct 2013)*
419 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
422 - update of the documentation, in particular installation instructions
423 - merging of LD_* constants into DT_* constants
424 - python style changes to be compatible with newer versions of pylint
430 *(Released Thu, 29 Aug 2013)*
432 This was the first beta release of the 2.9 series. All important changes
433 are listed in the latest 2.9 entry.
439 *(Released Thu, 23 Jan 2014)*
441 - Change the list separator for the usb_devices parameter from comma to space.
442 Commas could not work because they are already the hypervisor option
443 separator (Issue 649)
444 - Add support for blktap2 file-driver (Issue 638)
445 - Add network tag definitions to the haskell codebase (Issue 641)
446 - Fix RAPI network tag handling
447 - Add the network tags to the tags searched by gnt-cluster search-tags
448 - Fix caching bug preventing jobs from being cancelled
449 - Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
455 *(Released Thu, 12 Dec 2013)*
457 - Fixed Luxi daemon socket permissions after master-failover
458 - Improve IP version detection code directly checking for colons rather than
459 passing the family from the cluster object
460 - Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
461 opportunistically anymore (Issue 622)
462 - Allow link local IPv6 gateways (Issue 624)
463 - Fix error printing (Issue 616)
464 - Fix a bug in InstanceSetParams concerning names: in case no name is passed in
465 disk modifications, keep the old one. If name=none then set disk name to
467 - Update build_chroot script to work with the latest hackage packages
468 - Add a packet number limit to "fping" in master-ip-setup (Issue 630)
469 - Fix evacuation out of drained node (Issue 615)
470 - Add default file_driver if missing (Issue 571)
471 - Fix job error message after unclean master shutdown (Issue 618)
472 - Lock group(s) when creating instances (Issue 621)
473 - SetDiskID() before accepting an instance (Issue 633)
474 - Allow the ext template disks to receive arbitrary parameters, both at creation
475 time and while being modified
476 - Xen handle domain shutdown (future proofing cherry-pick)
477 - Refactor reading live data in htools (future proofing cherry-pick)
483 *(Released Thu, 07 Nov 2013)*
485 - DRBD: ensure peers are UpToDate for dual-primary
486 - Improve error message for replace-disks
487 - More dependency checks at configure time
488 - Placate warnings on ganeti.outils_unittest.py
494 *(Released Thu, 17 Oct 2013)*
496 - Correctly start/stop luxid during gnt-cluster master-failover
497 - Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
498 - Fix path for the job queue serial file
499 - Improved harep man page
500 - Minor documentation improvements
506 *(Released Mon, 30 Sep 2013)*
508 Incompatible/important changes
509 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
511 - Instance policy can contain multiple instance specs, as described in
512 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
513 <design-partitioned>`. As a consequence, it's not possible to partially change
514 or override instance specs. Bounding specs (min and max) can be specified as a
515 whole using the new option ``--ipolicy-bounds-specs``, while standard
516 specs use the new option ``--ipolicy-std-specs``.
517 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
518 gnt-instance is a valid YAML object.
519 - hail now honors network restrictions when allocating nodes. This led to an
520 update of the IAllocator protocol. See the IAllocator documentation for
522 - confd now only answers static configuration request over the network. luxid
523 was extracted, listens on the local LUXI socket and responds to live queries.
524 This allows finer grained permissions if using separate users.
529 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
530 to always require authentication, ``--require-authentication``. It can
531 be specified in ``$sysconfdir/default/ganeti``.
532 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
533 be used to manage the disk templates to be used by instances in the cluster.
534 Initially, it will be set to a list that includes plain, drbd, if they were
535 enabled by specifying a volume group name, and file and sharedfile, if those
536 were enabled at configure time. Additionally, it will include all disk
537 templates that are currently used by instances. The order of disk templates
538 will be based on Ganeti's history of supporting them. In the future, the
539 first entry of the list will be used as a default disk template on instance
541 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
542 configuration back to the previous stable version.
543 - Disk templates in group ipolicy can be restored to the default value.
544 - Initial support for diskless instances and virtual clusters in QA.
545 - More QA and unit tests for instance policies.
546 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
547 describing why the opcode itself was executed.
548 - The monitoring daemon is now available. It allows users to query the cluster
549 for obtaining information about the status of the system. The daemon is only
550 responsible for providing the information over the network: the actual data
551 gathering is performed by data collectors (currently, only the DRBD status
552 collector is available).
553 - In order to help developers work on Ganeti, a new script
554 (``devel/build_chroot``) is provided, for building a chroot that contains all
555 the required development libraries and tools for compiling Ganeti on a Debian
557 - A new tool, ``harep``, for performing self-repair and recreation of instances
558 in Ganeti has been added.
559 - Split queries are enabled for tags, network, exports, cluster info, groups,
561 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
562 It prints the command line to set the current policies, to ease
564 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
565 devices for KVM virtio-net interfaces will get created with VNET_HDR
566 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
567 virtio-net interfaces, which prevents host kernel tainting and log
568 flooding, when dealing with broken or malicious virtio-net drivers.
569 It's set to true by default.
570 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
572 - Support 'viridian' parameter in Xen HVM
573 - Support DSA SSH keys in bootstrap
574 - To simplify the work of packaging frameworks that want to add the needed users
575 and groups in a split-user setup themselves, at build time three files in
576 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
577 the groups to be generated, the ``users`` file contains, one per line, the
578 users to be generated, optionally followed by their primary group, where
579 important. The ``groupmemberships`` file contains, one per line, additional
580 user-group membership relations that need to be established. The syntax of
581 these files will remain stable in all future versions.
586 The following new dependencies have been added:
589 - The ``curl`` library is not optional anymore for compiling the Haskell code.
590 - ``snap-server`` library (if monitoring is enabled).
593 - The minimum Python version needed to run Ganeti is now 2.6.
594 - ``yaml`` library (only for running the QA).
598 - Perform proper cleanup on termination of Haskell daemons
599 - Fix corner-case in handling of remaining retry time
605 *(Released Tue, 17 Sep 2013)*
607 - To simplify the work of packaging frameworks that want to add the needed users
608 and groups in a split-user setup themselves, at build time three files in
609 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
610 the groups to be generated, the ``users`` file contains, one per line, the
611 users to be generated, optionally followed by their primary group, where
612 important. The ``groupmemberships`` file contains, one per line, additional
613 user-group membership relations that need to be established. The syntax of
614 these files will remain stable in all future versions.
615 - Add a default to file-driver when unspecified over RAPI (Issue 571)
616 - Mark the DSA host pubkey as optional, and remove it during config downgrade
618 - Some documentation fixes
624 *(Released Tue, 27 Aug 2013)*
626 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
628 - Support 'viridian' parameter in Xen HVM (Issue 233)
629 - Include VCS version in ``gnt-cluster version``
630 - Support DSA SSH keys in bootstrap (Issue 338)
631 - Fix batch creation of instances
632 - Use FQDN to check master node status (Issue 551)
633 - Make the DRBD collector more failure-resilient
639 *(Released Fri, 2 Aug 2013)*
641 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
643 - Fix upgrading/downgrading from 2.7
644 - Increase maximum RAPI message size
645 - Documentation updates
646 - Split ``confd`` between ``luxid`` and ``confd``
647 - Merge 2.7 series up to the 2.7.1 release
648 - Allow the ``modify_etc_hosts`` option to be changed
649 - Add better debugging for ``luxid`` queries
650 - Expose bulk parameter for GetJobs in RAPI client
651 - Expose missing ``network`` fields in RAPI
652 - Add some ``cluster verify`` tests
653 - Some unittest fixes
654 - Fix a malfunction in ``hspace``'s tiered allocation
655 - Fix query compatibility between haskell and python implementations
656 - Add the ``vnet_hdr`` HV parameter for KVM
657 - Add ``--cleanup`` to instance failover
658 - Change the connected groups format in ``gnt-network info`` output; it
659 was previously displayed as a raw list by mistake. (Merged from 2.7)
665 *(Released Mon, 24 Jun 2013)*
667 This was the first beta release of the 2.8 series. All important changes
668 are listed in the latest 2.8 entry.
674 *(Released Thu, 26 Sep 2013)*
676 - Change the connected groups format in ``gnt-network info`` output; it
677 was previously displayed as a raw list by mistake
678 - Check disk template in right dict when copying
679 - Support multi-instance allocs without iallocator
680 - Fix some errors in the documentation
681 - Fix formatting of tuple in an error message
687 *(Released Thu, 25 Jul 2013)*
689 - Add logrotate functionality in daemon-util
690 - Add logrotate example file
691 - Add missing fields to network queries over rapi
692 - Fix network object timestamps
693 - Add support for querying network timestamps
694 - Fix a typo in the example crontab
695 - Fix a documentation typo
701 *(Released Thu, 04 Jul 2013)*
703 Incompatible/important changes
704 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
706 - Instance policies for disk size were documented to be on a per-disk
707 basis, but hail applied them to the sum of all disks. This has been
709 - ``hbal`` will now exit with status 0 if, during job execution over
710 LUXI, early exit has been requested and all jobs are successful;
711 before, exit status 1 was used, which cannot be differentiated from
713 - Compatibility with newer versions of rbd has been fixed
714 - ``gnt-instance batch-create`` has been changed to use the bulk create
715 opcode from Ganeti. This lead to incompatible changes in the format of
716 the JSON file. It's now not a custom dict anymore but a dict
717 compatible with the ``OpInstanceCreate`` opcode.
718 - Parent directories for file storage need to be listed in
719 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
720 write the file automatically based on old configuration values, but it
721 can not distribute it across all nodes and the file contents should be
722 verified. Use ``gnt-cluster copyfile
723 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
724 upgraded. The reason for requiring this list of paths now is that
725 before it would have been possible to inject new paths via RPC,
726 allowing files to be created in arbitrary locations. The RPC protocol
727 is protected using SSL/X.509 certificates, but as a design principle
728 Ganeti does not permit arbitrary paths to be passed.
729 - The parsing of the variants file for OSes (see
730 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
731 lines and comment lines (starting with ``#``) are ignored for better
733 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
734 longer available. ``gnt-node add`` now invokes a new tool on the
735 destination node, named ``prepare-node-join``, to configure the SSH
736 daemon. Paramiko is no longer necessary to configure nodes' SSH
737 daemons via ``gnt-node add``.
738 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
739 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
740 the flag is not reset after a master failover.
741 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
742 is recommended. 2.6 will be mandatory from the 2.8 series.
748 - New network management functionality to support automatic allocation
749 of IP addresses and managing of network parameters. See
750 :manpage:`gnt-network(8)` for more details.
751 - New external storage backend, to allow managing arbitrary storage
752 systems external to the cluster. See
753 :manpage:`ganeti-extstorage-interface(7)`.
754 - New ``exclusive-storage`` node parameter added, restricted to
755 nodegroup level. When it's set to true, physical disks are assigned in
756 an exclusive fashion to instances, as documented in :doc:`Partitioned
757 Ganeti <design-partitioned>`. Currently, only instances using the
758 ``plain`` disk template are supported.
759 - The KVM hypervisor has been updated with many new hypervisor
760 parameters, including a generic one for passing arbitrary command line
761 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
762 compatible up to qemu 1.4.
763 - A new tool, called ``mon-collector``, is the stand-alone executor of
764 the data collectors for a monitoring system. As of this version, it
765 just includes the DRBD data collector, that can be executed by calling
766 ``mon-collector`` using the ``drbd`` parameter. See
767 :manpage:`mon-collector(7)`.
768 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
769 for RAPI users. It allows granting permissions to query for
770 information to a specific user without giving
771 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
772 - A new tool named ``node-cleanup`` has been added. It cleans remains of
773 a cluster from a machine by stopping all daemons, removing
774 certificates and ssconf files. Unless the ``--no-backup`` option is
775 given, copies of the certificates are made.
776 - Instance creations now support the use of opportunistic locking,
777 potentially speeding up the (parallel) creation of multiple instances.
778 This feature is currently only available via the :doc:`RAPI
779 <rapi>` interface and when an instance allocator is used. If the
780 ``opportunistic_locking`` parameter is set the opcode will try to
781 acquire as many locks as possible, but will not wait for any locks
782 held by other opcodes. If not enough resources can be found to
783 allocate the instance, the temporary error code
784 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
785 retried thereafter, with or without opportunistic locking.
786 - New experimental linux-ha resource scripts.
787 - Restricted-commands support: ganeti can now be asked (via command line
788 or rapi) to perform commands on a node. These are passed via ganeti
789 RPC rather than ssh. This functionality is restricted to commands
790 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
791 reasons. The file is not copied automatically.
797 - Diskless instances are now externally mirrored (Issue 237). This for
798 now has only been tested in conjunction with explicit target nodes for
800 - Queries not needing locks or RPC access to the node can now be
801 performed by the confd daemon, making them independent from jobs, and
802 thus faster to execute. This is selectable at configure time.
803 - The functionality for allocating multiple instances at once has been
804 overhauled and is now also available through :doc:`RAPI <rapi>`.
806 There are no significant changes from version 2.7.0~rc3.
812 *(Released Tue, 25 Jun 2013)*
814 - Fix permissions on the confd query socket (Issue 477)
815 - Fix permissions on the job archive dir (Issue 498)
816 - Fix handling of an internal exception in replace-disks (Issue 472)
817 - Fix gnt-node info handling of shortened names (Issue 497)
818 - Fix gnt-instance grow-disk when wiping is enabled
819 - Documentation improvements, and support for newer pandoc
820 - Fix hspace honoring ipolicy for disks (Issue 484)
821 - Improve handling of the ``kvm_extra`` HV parameter
827 *(Released Fri, 24 May 2013)*
829 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
831 - Disks added through ``gnt-instance modify`` or created through
832 ``gnt-instance recreate-disks`` are wiped, if the
833 ``prealloc_wipe_disks`` flag is set.
834 - If wiping newly created disks fails, the disks are removed. Also,
835 partial failures in creating disks through ``gnt-instance modify``
836 triggers a cleanup of the partially-created disks.
837 - Removing the master IP address doesn't fail if the address has been
839 - Fix ownership of the OS log dir
840 - Workaround missing SO_PEERCRED constant (Issue 191)
846 *(Released Fri, 3 May 2013)*
848 This was the first release candidate of the 2.7 series. Since beta3:
850 - Fix kvm compatibility with qemu 1.4 (Issue 389)
851 - Documentation updates (admin guide, upgrade notes, install
852 instructions) (Issue 372)
853 - Fix gnt-group list nodes and instances count (Issue 436)
854 - Fix compilation without non-mandatory libraries (Issue 441)
855 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
856 - Make confd logging more verbose at INFO level (Issue 435)
857 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
858 - Fix failure path for instance storage type conversion (Issue 229)
859 - Update htools text backend documentation
860 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
861 - Disable inter-cluster instance move for file-based instances, because
862 it is dependant on instance export, which is not supported for
863 file-based instances. (Issue 414)
864 - Fix gnt-job crashes on non-ascii characters (Issue 427)
865 - Fix volume group checks on non-vm-capable nodes (Issue 432)
871 *(Released Mon, 22 Apr 2013)*
873 This was the third beta release of the 2.7 series. Since beta2:
875 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
876 - Fix data loss on wrong usage of ``gnt-instance move``
877 - Properly export errors in confd-based job queries
878 - Add ``users-setup`` tool
879 - Fix iallocator protocol to report 0 as a disk size for diskless
880 instances. This avoids hail breaking when a diskless instance is
882 - Fix job queue directory permission problem that made confd job queries
883 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
884 for access to archived jobs (Issue 406).
885 - Limit the sizes of networks supported by ``gnt-network`` to something
886 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
887 - Fix bugs in instance disk template conversion
888 - Fix GHC 7 compatibility
889 - Fix ``burnin`` install path (Issue 426).
890 - Allow very small disk grows (Issue 347).
891 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
892 that noded doesn't import masterd code (Issue 419).
893 - Make sure the default metavg at cluster init is the same as the vg, if
894 unspecified (Issue 358).
895 - Fix cleanup of partially created disks (part of Issue 416)
901 *(Released Tue, 2 Apr 2013)*
903 This was the second beta release of the 2.7 series. Since beta1:
905 - Networks no longer have a "type" slot, since this information was
906 unused in Ganeti: instead of it tags should be used.
907 - The rapi client now has a ``target_node`` option to MigrateInstance.
908 - Fix early exit return code for hbal (Issue 386).
909 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
910 - Fix ``rbd showmapped`` output parsing (Issue 312).
911 - Networks are now referenced indexed by UUID, rather than name. This
912 will require running cfgupgrade, from 2.7.0beta1, if networks are in
914 - The OS environment now includes network information.
915 - Deleting of a network is now disallowed if any instance nic is using
916 it, to prevent dangling references.
917 - External storage is now documented in man pages.
918 - The exclusive_storage flag can now only be set at nodegroup level.
919 - Hbal can now submit an explicit priority with its jobs.
920 - Many network related locking fixes.
921 - Bump up the required pylint version to 0.25.1.
922 - Fix the ``no_remember`` option in RAPI client.
923 - Many ipolicy related tests, qa, and fixes.
924 - Many documentation improvements and fixes.
925 - Fix building with ``--disable-file-storage``.
926 - Fix ``-q`` option in htools, which was broken if passed more than
928 - Some haskell/python interaction improvements and fixes.
929 - Fix iallocator in case of missing LVM storage.
930 - Fix confd config load in case of ``--no-lvm-storage``.
931 - The confd/query functionality is now mentioned in the security
938 *(Released Wed, 6 Feb 2013)*
940 This was the first beta release of the 2.7 series. All important changes
941 are listed in the latest 2.7 entry.
947 *(Released Fri, 21 Dec 2012)*
949 Important behaviour change: hbal won't rebalance anymore instances which
950 have the ``auto_balance`` attribute set to false. This was the intention
951 all along, but until now it only skipped those from the N+1 memory
952 reservation (DRBD-specific).
954 A significant number of bug fixes in this release:
956 - Fixed disk adoption interaction with ipolicy checks.
957 - Fixed networking issues when instances are started, stopped or
958 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
959 - Fixed the warning in cluster verify for shared storage instances not
961 - Fixed removal of storage directory on shared file storage (issue 262).
962 - Fixed validation of LVM volume group name in OpClusterSetParams
963 (``gnt-cluster modify``) (issue 285).
964 - Fixed runtime memory increases (``gnt-instance modify -m``).
965 - Fixed live migration under Xen's ``xl`` mode.
966 - Fixed ``gnt-instance console`` with ``xl``.
967 - Fixed building with newer Haskell compiler/libraries.
968 - Fixed PID file writing in Haskell daemons (confd); this prevents
969 restart issues if confd was launched manually (outside of
970 ``daemon-util``) while another copy of it was running
971 - Fixed a type error when doing live migrations with KVM (issue 297) and
972 the error messages for failing migrations have been improved.
973 - Fixed opcode validation for the out-of-band commands (``gnt-node
975 - Fixed a type error when unsetting OS hypervisor parameters (issue
976 311); now it's possible to unset all OS-specific hypervisor
978 - Fixed the ``dry-run`` mode for many operations: verification of
979 results was over-zealous but didn't take into account the ``dry-run``
980 operation, resulting in "wrong" failures.
981 - Fixed bash completion in ``gnt-job list`` when the job queue has
982 hundreds of entries; especially with older ``bash`` versions, this
983 results in significant CPU usage.
985 And lastly, a few other improvements have been made:
987 - Added option to force master-failover without voting (issue 282).
988 - Clarified error message on lock conflict (issue 287).
989 - Logging of newly submitted jobs has been improved (issue 290).
990 - Hostname checks have been made uniform between instance rename and
992 - The ``--submit`` option is now supported by ``gnt-debug delay``.
993 - Shutting down the master daemon by sending SIGTERM now stops it from
994 processing jobs waiting for locks; instead, those jobs will be started
995 once again after the master daemon is started the next time (issue
997 - Support for Xen's ``xl`` program has been improved (besides the fixes
999 - Reduced logging noise in the Haskell confd daemon (only show one log
1000 entry for each config reload, instead of two).
1001 - Several man page updates and typo fixes.
1007 *(Released Fri, 12 Oct 2012)*
1009 A small bugfix release. Among the bugs fixed:
1011 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
1012 made the command unusable.
1013 - Commands that issue many jobs don't fail anymore just because some jobs
1014 take so long that other jobs are archived.
1015 - Failures during ``gnt-instance reinstall`` are reflected by the exit
1017 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
1019 - When ``always_failover`` is set, ``--allow-failover`` is not required
1020 in migrate commands anymore.
1021 - ``bash_completion`` works even if extglob is disabled.
1022 - Fixed bug with locks that made failover for RDB-based instances fail.
1023 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
1024 a random node instead of one based on the allocator metric.
1025 - Support for newer versions of pylint and pep8.
1026 - Hail doesn't fail anymore when trying to add an instance of type
1027 ``file``, ``sharedfile`` or ``rbd``.
1028 - Added new Makefile target to rebuild the whole distribution, so that
1029 all files are included.
1035 *(Released Fri, 27 Jul 2012)*
1038 .. attention:: The ``LUXI`` protocol has been made more consistent
1039 regarding its handling of command arguments. This, however, leads to
1040 incompatibility issues with previous versions. Please ensure that you
1041 restart Ganeti daemons soon after the upgrade, otherwise most
1042 ``LUXI`` calls (job submission, setting/resetting the drain flag,
1043 pausing/resuming the watcher, cancelling and archiving jobs, querying
1044 the cluster configuration) will fail.
1053 The current ``admin_up`` field, which used to denote whether an instance
1054 should be running or not, has been removed. Instead, ``admin_state`` is
1055 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
1057 The rational behind this is that an instance being “down” can have
1060 - it could be down during a reboot
1061 - it could be temporarily be down for a reinstall
1062 - or it could be down because it is deprecated and kept just for its
1065 The previous Boolean state was making it difficult to do capacity
1066 calculations: should Ganeti reserve memory for a down instance? Now, the
1067 tri-state field makes it clear:
1069 - in ``up`` and ``down`` state, all resources are reserved for the
1070 instance, and it can be at any time brought up if it is down
1071 - in ``offline`` state, only disk space is reserved for it, but not
1074 The field can have an extra use: since the transition between ``up`` and
1075 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
1076 transition between ``offline`` and ``down`` is done via ``gnt-instance
1077 modify``, it is possible to given different rights to users. For
1078 example, owners of an instance could be allowed to start/stop it, but
1079 not transition it out of the offline state.
1081 Instance policies and specs
1082 +++++++++++++++++++++++++++
1084 In previous Ganeti versions, an instance creation request was not
1085 limited on the minimum size and on the maximum size just by the cluster
1086 resources. As such, any policy could be implemented only in third-party
1087 clients (RAPI clients, or shell wrappers over ``gnt-*``
1088 tools). Furthermore, calculating cluster capacity via ``hspace`` again
1089 required external input with regards to instance sizes.
1091 In order to improve these workflows and to allow for example better
1092 per-node group differentiation, we introduced instance specs, which
1095 - minimum instance disk size, disk count, memory size, cpu count
1096 - maximum values for the above metrics
1097 - and “standard” values (used in ``hspace`` to calculate the standard
1100 The minimum/maximum values can be also customised at node-group level,
1101 for example allowing more powerful hardware to support bigger instance
1104 Beside the instance specs, there are a few other settings belonging to
1105 the instance policy framework. It is possible now to customise, per
1106 cluster and node-group:
1108 - the list of allowed disk templates
1109 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
1110 - the maximum ratio of instance to spindles (see below for more
1111 information) for local storage
1113 All these together should allow all tools that talk to Ganeti to know
1114 what are the ranges of allowed values for instances and the
1115 over-subscription that is allowed.
1117 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
1118 instance configuration, and the physical CPU configuration from the
1119 node. For the spindle ratios however, we didn't track before these
1120 values, so new parameters have been added:
1122 - a new node parameter ``spindle_count``, defaults to 1, customisable at
1123 node group or node level
1124 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
1126 Note that spindles in this context doesn't need to mean actual
1127 mechanical hard-drives; it's just a relative number for both the node
1128 I/O capacity and instance I/O consumption.
1130 Instance migration behaviour
1131 ++++++++++++++++++++++++++++
1133 While live-migration is in general desirable over failover, it is
1134 possible that for some workloads it is actually worse, due to the
1135 variable time of the “suspend” phase during live migration.
1137 To allow the tools to work consistently over such instances (without
1138 having to hard-code instance names), a new backend parameter
1139 ``always_failover`` has been added to control the migration/failover
1140 behaviour. When set to True, all migration requests for an instance will
1141 instead fall-back to failover.
1143 Instance memory ballooning
1144 ++++++++++++++++++++++++++
1146 Initial support for memory ballooning has been added. The memory for an
1147 instance is no longer fixed (backend parameter ``memory``), but instead
1148 can vary between minimum and maximum values (backend parameters
1149 ``minmem`` and ``maxmem``). Currently we only change an instance's
1152 - live migrating or failing over and instance and the target node
1153 doesn't have enough memory
1154 - user requests changing the memory via ``gnt-instance modify
1157 Instance CPU pinning
1158 ++++++++++++++++++++
1160 In order to control the use of specific CPUs by instance, support for
1161 controlling CPU pinning has been added for the Xen, HVM and LXC
1162 hypervisors. This is controlled by a new hypervisor parameter
1163 ``cpu_mask``; details about possible values for this are in the
1164 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
1165 VCPU-to-CPU mapping) form will work well only when all nodes in your
1166 cluster have the same amount of CPUs.
1171 Another area in which Ganeti was not customisable were the parameters
1172 used for storage configuration, e.g. how many stripes to use for LVM,
1173 DRBD resync configuration, etc.
1175 To improve this area, we've added disks parameters, which are
1176 customisable at cluster and node group level, and which allow to
1177 specify various parameters for disks (DRBD has the most parameters
1178 currently), for example:
1180 - DRBD resync algorithm and parameters (e.g. speed)
1181 - the default VG for meta-data volumes for DRBD
1182 - number of stripes for LVM (plain disk template)
1185 These parameters can be modified via ``gnt-cluster modify -D …`` and
1186 ``gnt-group modify -D …``, and are used at either instance creation (in
1187 case of LVM stripes, for example) or at disk “activation” time
1188 (e.g. resync speed).
1190 Rados block device support
1191 ++++++++++++++++++++++++++
1193 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
1194 denoted by the ``rbd`` disk template type. This is considered
1195 experimental, feedback is welcome. For details on configuring it, see
1196 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
1201 The existing master IP functionality works well only in simple setups (a
1202 single network shared by all nodes); however, if nodes belong to
1203 different networks, then the ``/32`` setup and lack of routing
1204 information is not enough.
1206 To allow the master IP to function well in more complex cases, the
1207 system was reworked as follows:
1209 - a master IP netmask setting has been added
1210 - the master IP activation/turn-down code was moved from the node daemon
1211 to a separate script
1212 - whether to run the Ganeti-supplied master IP script or a user-supplied
1213 on is a ``gnt-cluster init`` setting
1215 Details about the location of the standard and custom setup scripts are
1216 in the man page :manpage:`gnt-cluster(8)`; for information about the
1217 setup script protocol, look at the Ganeti-supplied script.
1222 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
1225 It is now possible to use TLS-protected connections, and when renewing
1226 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
1227 it is now possible to specify spice or spice CA certificates. Also, it
1228 is possible to configure a password for SPICE sessions via the
1229 hypervisor parameter ``spice_password_file``.
1231 There are also new parameters to control the compression and streaming
1232 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1233 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1234 for the spice parameters.
1236 Lastly, it is now possible to see the SPICE connection information via
1237 ``gnt-instance console``.
1242 A new tool (``tools/ovfconverter``) has been added that supports
1243 conversion between Ganeti and the `Open Virtualization Format
1244 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1247 This relies on the ``qemu-img`` tool to convert the disk formats, so the
1248 actual compatibility with other virtualization solutions depends on it.
1250 Confd daemon changes
1251 ++++++++++++++++++++
1253 The configuration query daemon (``ganeti-confd``) is now optional, and
1254 has been rewritten in Haskell; whether to use the daemon at all, use the
1255 Python (default) or the Haskell version is selectable at configure time
1256 via the ``--enable-confd`` parameter, which can take one of the
1257 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
1258 daemon will result in a smaller footprint; for larger systems, we
1259 welcome feedback on the Haskell version which might become the default
1262 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1263 daemon running. The Python version doesn't implement the new call.
1266 User interface changes
1267 ~~~~~~~~~~~~~~~~~~~~~~
1269 We have replaced the ``--disks`` option of ``gnt-instance
1270 replace-disks`` with a more flexible ``--disk`` option, which allows
1271 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1272 disk size and mode can be changed upon recreation (via ``gnt-instance
1273 recreate-disks``, which accepts the same ``--disk`` option).
1275 As many people are used to a ``show`` command, we have added that as an
1276 alias to ``info`` on all ``gnt-*`` commands.
1278 The ``gnt-instance grow-disk`` command has a new mode in which it can
1279 accept the target size of the disk, instead of the delta; this can be
1280 more safe since two runs in absolute mode will be idempotent, and
1281 sometimes it's also easier to specify the desired size directly.
1283 Also the handling of instances with regard to offline secondaries has
1284 been improved. Instance operations should not fail because one of it's
1285 secondary nodes is offline, even though it's safe to proceed.
1287 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1288 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1289 minors to instance name.
1294 RAPI coverage has improved, with (for example) new resources for
1295 recreate-disks, node power-cycle, etc.
1300 There is partial support for ``xl`` in the Xen hypervisor; feedback is
1303 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1304 whether to still support Python 2.4 or move to Python 2.6 as minimum
1307 Support for Fedora has been slightly improved; the provided example
1308 init.d script should work better on it and the INSTALL file should
1309 document the needed dependencies.
1314 The deprecated ``QueryLocks`` LUXI request has been removed. Use
1315 ``Query(what=QR_LOCK, ...)`` instead.
1317 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1318 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1319 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1320 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1321 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1323 RAPI client: ``CertificateError`` now derives from
1324 ``GanetiApiError``. This should make it more easy to handle Ganeti
1327 Deprecation warnings due to PyCrypto/paramiko import in
1328 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
1329 make sure to run an up-to-date paramiko version, if you use this tool.
1331 The QA scripts now depend on Python 2.5 or above (the main code base
1332 still works with Python 2.4).
1334 The configuration file (``config.data``) is now written without
1335 indentation for performance reasons; if you want to edit it, it can be
1336 re-formatted via ``tools/fmtjson``.
1338 A number of bugs has been fixed in the cluster merge tool.
1340 ``x509`` certification verification (used in import-export) has been
1341 changed to allow the same clock skew as permitted by the cluster
1342 verification. This will remove some rare but hard to diagnose errors in
1349 *(Released Thu, 19 Jul 2012)*
1351 Very few changes from rc4 to the final release, only bugfixes:
1353 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
1354 instance, fix CDROM booting for KVM instances)
1355 - fixed node group modification of node parameters
1356 - fixed issue in LUClusterVerifyGroup with multi-group clusters
1357 - fixed generation of bash completion to ensure a stable ordering
1364 *(Released Fri, 13 Jul 2012)*
1366 Third release candidate for 2.6. The following changes were done from
1369 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1370 - Fixed an inconsistency in the LUXI protocol with the provided
1371 arguments (NOT backwards compatible)
1372 - Fixed a bug with node groups ipolicy where ``min`` was greater than
1373 the cluster ``std`` value
1374 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1375 easier instance debugging on nodes (requires ``hconfd`` to work)
1381 *(Released Tue, 03 Jul 2012)*
1383 Second release candidate for 2.6. The following changes were done from
1386 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1388 - Fixed a RAPI regression on missing beparams/memory
1389 - Fixed redistribution of files on offline nodes
1390 - Added possibility to run activate-disks even though secondaries are
1391 offline. With this change it relaxes also the strictness on some other
1392 commands which use activate disks internally:
1393 * ``gnt-instance start|reboot|rename|backup|export``
1394 - Made it possible to remove safely an instance if its secondaries are
1396 - Made it possible to reinstall even though secondaries are offline
1402 *(Released Mon, 25 Jun 2012)*
1404 First release candidate for 2.6. The following changes were done from
1407 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1409 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1410 behaviour. It supports now the modify operation
1411 - ``hcheck`` implemented to analyze cluster health and possibility of
1412 improving health by rebalance
1413 - ``hbal`` has been improved in dealing with split instances
1419 *(Released Mon, 11 Jun 2012)*
1421 Second beta release of 2.6. The following changes were done from beta2
1424 - Fixed ``daemon-util`` with non-root user models
1425 - Fixed creation of plain instances with ``--no-wait-for-sync``
1426 - Fix wrong iv_names when running ``cfgupgrade``
1427 - Export more information in RAPI group queries
1428 - Fixed bug when changing instance network interfaces
1429 - Extended burnin to do NIC changes
1430 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1431 - Changed default for DRBD barriers
1432 - Fixed DRBD error reporting for syncer rate
1433 - Verify the options on disk parameters
1435 And of course various fixes to documentation and improved unittests and
1442 *(Released Wed, 23 May 2012)*
1444 First beta release of 2.6. The following changes were done from beta1 to
1447 - integrated patch for distributions without ``start-stop-daemon``
1448 - adapted example init.d script to work on Fedora
1449 - fixed log handling in Haskell daemons
1450 - adapted checks in the watcher for pycurl linked against libnss
1451 - add partial support for ``xl`` instead of ``xm`` for Xen
1452 - fixed a type issue in cluster verification
1453 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1456 Plus integrated fixes from the 2.5 branch:
1458 - fixed ``kvm-ifup`` to use ``/bin/bash``
1459 - fixed parallel build failures
1460 - KVM live migration when using a custom keymap
1466 *(Released Tue, 24 Jul 2012)*
1468 A small bugfix release, with no new features:
1470 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1471 different default shell (e.g. Debian, Ubuntu)
1472 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1473 243 and Debian bug #650664)
1474 - fixed compatibility with KVM versions that don't support multiple boot
1475 devices (fixes Issue 230 and Debian bug #624256)
1477 Additionally, a few fixes were done to the build system (fixed parallel
1478 build failures) and to the unittests (fixed race condition in test for
1479 FileID functions, and the default enable/disable mode for QA test is now
1486 *(Released Fri, 11 May 2012)*
1488 A small bugfix release.
1490 The main issues solved are on the topic of compatibility with newer LVM
1493 - fixed parsing of ``lv_attr`` field
1494 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1495 the ``--force`` flag is needed
1497 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1498 to accept kernel 3.x too (was hardcoded to 2.6.*).
1500 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1501 of hook results and that also made display of other errors suboptimal
1502 was fixed; the code behaves now like 2.4 and earlier.
1504 Another change in 2.5, the cleanup of the OS scripts environment, is too
1505 aggressive: it removed even the ``PATH`` variable, which requires the OS
1506 scripts to *always* need to export it. Since this is a bit too strict,
1507 we now export a minimal PATH, the same that we export for hooks.
1509 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1510 integrated into this release.
1512 Finally, a few other miscellaneous changes were done (no new features,
1513 just small improvements):
1515 - Fix ``gnt-group --help`` display
1516 - Fix hardcoded Xen kernel path
1517 - Fix grow-disk handling of invalid units
1518 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1519 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1525 *(Released Thu, 12 Apr 2012)*
1527 Incompatible/important changes and bugfixes
1528 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1530 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1531 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1532 to match the underlying LUXI interface.
1533 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1534 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1535 not maintain backwards-compatibility as the underlying operation
1536 changed in an incompatible way. The RAPI client can talk to old
1537 servers, but it needs to be told so as the return value changed.
1538 - When creating file-based instances via RAPI, the ``file_driver``
1539 parameter no longer defaults to ``loop`` and must be specified.
1540 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1542 - Support for the undocumented and deprecated RAPI instance creation
1543 request format version 0 has been dropped. Use version 1, supported
1544 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1545 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1546 documentation <install>`.
1547 - The "cluster-verify" hooks are now executed per group by the
1548 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1549 if you just run ``gnt-cluster verify``, which generates one opcode per
1551 - The environment as passed to the OS scripts is cleared, and thus no
1552 environment variables defined in the node daemon's environment will be
1553 inherited by the scripts.
1554 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1556 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1557 support operations involving multiple node groups.
1558 - Offline nodes are ignored when failing over an instance.
1559 - Support for KVM version 1.0, which changed the version reporting format
1561 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1563 - ``Makefile`` is now compatible with Automake 1.11.2
1564 - Includes all bugfixes made in the 2.4 series
1569 - The ganeti-htools project has been merged into the ganeti-core source
1570 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1571 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1572 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1573 GPT-style partition tables (requires `parted
1574 <http://www.gnu.org/s/parted/>`_).
1575 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1576 - Allowed adding tags on instance creation.
1577 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1578 - Allow instances to be started in a paused state, enabling the user to
1579 see the complete console output on boot using the console.
1580 - Added new hypervisor flag to control default reboot behaviour
1581 (``reboot_behavior``).
1582 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1583 - Improved out-of-band management support:
1585 - Added ``gnt-node health`` command reporting the health status of
1587 - Added ``gnt-node power`` command to manage power status of nodes.
1588 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1590 - Instance migration can fall back to failover if instance is not
1592 - Filters can be used when listing nodes, instances, groups and locks;
1593 see :manpage:`ganeti(7)` manpage.
1594 - Added post-execution status as variables to :doc:`hooks <hooks>`
1596 - Instance tags are exported/imported together with the instance.
1597 - When given an explicit job ID, ``gnt-job info`` will work for archived
1599 - Jobs can define dependencies on other jobs (not yet supported via
1600 RAPI or command line, but used by internal commands and usable via
1603 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1606 - Instance failover is now available as a RAPI resource
1607 (``/2/instances/[instance_name]/failover``).
1608 - ``gnt-instance info`` defaults to static information if primary node
1610 - Opcodes have a new ``comment`` attribute.
1611 - Added basic SPICE support to KVM hypervisor.
1612 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1614 Node group improvements
1615 ~~~~~~~~~~~~~~~~~~~~~~~
1617 - ``gnt-cluster verify`` has been modified to check groups separately,
1618 thereby improving performance.
1619 - Node group support has been added to ``gnt-cluster verify-disks``,
1620 which now operates per node group.
1621 - Watcher has been changed to work better with node groups.
1623 - One process and state file per node group.
1624 - Slow watcher in one group doesn't block other group's watcher.
1626 - Added new command, ``gnt-group evacuate``, to move all instances in a
1627 node group to other groups.
1628 - Added ``gnt-instance change-group`` to move an instance to another
1630 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1631 per-group operations.
1632 - Node groups can be tagged.
1633 - Some operations switch from an exclusive to a shared lock as soon as
1635 - Instance's primary and secondary nodes' groups are now available as
1636 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1637 and ``snodes.group.uuid``).
1642 - Numerous updates to documentation and manpages.
1644 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1646 - Some opcode/job results are now also documented, see :doc:`RAPI
1649 - A lockset's internal lock is now also visible in lock monitor.
1650 - Log messages from job queue workers now contain information about the
1651 opcode they're processing.
1652 - ``gnt-instance console`` no longer requires the instance lock.
1653 - A short delay when waiting for job changes reduces the number of LUXI
1654 requests significantly.
1655 - DRBD metadata volumes are overwritten with zeros during disk creation.
1656 - Out-of-band commands no longer acquire the cluster lock in exclusive
1658 - ``devel/upload`` now uses correct permissions for directories.
1664 *(Released Fri, 23 Mar 2012)*
1666 This was the sixth release candidate of the 2.5 series.
1672 *(Released Mon, 9 Jan 2012)*
1674 This was the fifth release candidate of the 2.5 series.
1680 *(Released Thu, 27 Oct 2011)*
1682 This was the fourth release candidate of the 2.5 series.
1688 *(Released Wed, 26 Oct 2011)*
1690 This was the third release candidate of the 2.5 series.
1696 *(Released Tue, 18 Oct 2011)*
1698 This was the second release candidate of the 2.5 series.
1704 *(Released Tue, 4 Oct 2011)*
1706 This was the first release candidate of the 2.5 series.
1712 *(Released Wed, 31 Aug 2011)*
1714 This was the third beta release of the 2.5 series.
1720 *(Released Mon, 22 Aug 2011)*
1722 This was the second beta release of the 2.5 series.
1728 *(Released Fri, 12 Aug 2011)*
1730 This was the first beta release of the 2.5 series.
1736 *(Released Thu, 27 Oct 2011)*
1738 - Fixed bug when parsing command line parameter values ending in
1740 - Fixed assertion error after unclean master shutdown
1741 - Disable HTTP client pool for RPC, significantly reducing memory usage
1743 - Fixed queue archive creation with wrong permissions
1749 *(Released Tue, 23 Aug 2011)*
1753 - Fixed documentation for importing with ``--src-dir`` option
1754 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1755 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1761 *(Released Fri, 5 Aug 2011)*
1763 Many bug-fixes and a few small features:
1765 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1766 issues when you tried to add an instance with two MAC addresses in one
1768 - KVM: fixed per-instance stored UID value
1769 - KVM: configure bridged NICs at migration start
1770 - KVM: Fix a bug where instance will not start with never KVM versions
1772 - Added OS search path to ``gnt-cluster info``
1773 - Fixed an issue with ``file_storage_dir`` where you were forced to
1774 provide an absolute path, but the documentation states it is a
1775 relative path, the documentation was right
1776 - Added a new parameter to instance stop/start called ``--no-remember``
1777 that will make the state change to not be remembered
1778 - Implemented ``no_remember`` at RAPI level
1779 - Improved the documentation
1780 - Node evacuation: don't call IAllocator if node is already empty
1781 - Fixed bug in DRBD8 replace disks on current nodes
1782 - Fixed bug in recreate-disks for DRBD instances
1783 - Moved assertion checking locks in ``gnt-instance replace-disks``
1784 causing it to abort with not owning the right locks for some situation
1785 - Job queue: Fixed potential race condition when cancelling queued jobs
1786 - Fixed off-by-one bug in job serial generation
1787 - ``gnt-node volumes``: Fix instance names
1788 - Fixed aliases in bash completion
1789 - Fixed a bug in reopening log files after being sent a SIGHUP
1790 - Added a flag to burnin to allow specifying VCPU count
1791 - Bugfixes to non-root Ganeti configuration
1797 *(Released Thu, 12 May 2011)*
1799 Many bug-fixes and a few new small features:
1801 - Fixed a bug related to log opening failures
1802 - Fixed a bug in instance listing with orphan instances
1803 - Fixed a bug which prevented resetting the cluster-level node parameter
1804 ``oob_program`` to the default
1805 - Many fixes related to the ``cluster-merge`` tool
1806 - Fixed a race condition in the lock monitor, which caused failures
1807 during (at least) creation of many instances in parallel
1808 - Improved output for gnt-job info
1809 - Removed the quiet flag on some ssh calls which prevented debugging
1811 - Improved the N+1 failure messages in cluster verify by actually
1812 showing the memory values (needed and available)
1813 - Increased lock attempt timeouts so that when executing long operations
1814 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1815 too early and thus prevent the use of the 'fair' mechanism
1816 - Changed instance query data (``gnt-instance info``) to not acquire
1817 locks unless needed, thus allowing its use on locked instance if only
1818 static information is asked for
1819 - Improved behaviour with filesystems that do not support rename on an
1821 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1822 kept locks on all nodes during the wipe, which is unneeded
1823 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1824 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1825 10GiB) which caused the wipe to fail right at the end in some cases
1826 - Fixed master IP activation when doing master failover with no-voting
1827 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1828 the master node itself
1829 - Fixed potential data-loss in under disk full conditions, where Ganeti
1830 wouldn't check correctly the return code and would consider
1831 partially-written files 'correct'
1832 - Fixed bug related to multiple VGs and DRBD disk replacing
1833 - Added new disk parameter ``metavg`` that allows placement of the meta
1834 device for DRBD in a different volume group
1835 - Fixed error handling in the node daemon when the system libc doesn't
1836 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1837 - Fixed lock release during replace-disks, which kept cluster-wide locks
1838 when doing disk replaces with an iallocator script
1839 - Added check for missing bridges in cluster verify
1840 - Handle EPIPE errors while writing to the terminal better, so that
1841 piping the output to e.g. ``less`` doesn't cause a backtrace
1842 - Fixed rare case where a ^C during Luxi calls could have been
1843 interpreted as server errors, instead of simply terminating
1844 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1846 - Added a few more parameters to the KVM hypervisor, allowing a second
1847 CDROM, custom disk type for CDROMs and a floppy image
1848 - Removed redundant message in instance rename when the name is given
1850 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1851 disks on new nodes, allowing recreation when the original instance
1852 nodes are completely gone
1853 - Added option when converting disk templates to DRBD to skip waiting
1854 for the resync, in order to make the instance available sooner
1855 - Added two new variables to the OS scripts environment (containing the
1857 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1858 to allow use of ``pvgrub`` as bootloader
1859 - Changed the instance memory modifications to only check out-of-memory
1860 conditions on memory increases, and turned the secondary node warnings
1861 into errors (they can still be overridden via ``--force``)
1862 - Fixed the handling of a corner case when the Python installation gets
1863 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1864 to execute a command that doesn't exist
1865 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1866 node of the instance returned failures during instance shutdown; this
1867 adds the option ``--ignore-consistency`` to gnt-instance move
1869 And as usual, various improvements to the error messages, documentation
1876 *(Released Wed, 09 Mar 2011)*
1878 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1879 the RAPI users file if run twice (even with ``--dry-run``).
1881 The release fixes that bug (nothing else changed).
1887 *(Released Mon, 07 Mar 2011)*
1889 Final 2.4.0 release. Just a few small fixes:
1891 - Fixed RAPI node evacuate
1892 - Fixed the kvm-ifup script
1893 - Fixed internal error handling for special job cases
1894 - Updated man page to specify the escaping feature for options
1900 *(Released Mon, 28 Feb 2011)*
1902 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1903 that this feature wiped the disks of the wrong instance, leading to loss
1908 - Fixed title of query field containing instance name
1909 - Expanded the glossary in the documentation
1910 - Fixed one unittest (internal issue)
1916 *(Released Mon, 21 Feb 2011)*
1918 A number of bug fixes plus just a couple functionality changes.
1920 On the user-visible side, the ``gnt-* list`` command output has changed
1921 with respect to "special" field states. The current rc1 style of display
1922 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1923 the default output mode special fields states are displayed as follows:
1925 - Offline resource: ``*``
1926 - Unavailable/not applicable: ``-``
1927 - Data missing (RPC failure): ``?``
1928 - Unknown field: ``??``
1930 Another user-visible change is the addition of ``--force-join`` to
1935 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1936 - Fixed regression in RAPI/instance reinstall where all parameters were
1937 required (instead of optional)
1938 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1939 - Fixed iallocator usage (offline nodes were not considered offline)
1940 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1941 - Fixed hypervisor and OS parameter validation with respect to
1942 non-vm_capable nodes
1943 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1945 - Fixed ``tools/listrunner`` with respect to agent-based usage
1951 *(Released Fri, 4 Feb 2011)*
1953 Many changes and fixes since the beta1 release. While there were some
1954 internal changes, the code has been mostly stabilised for the RC
1957 Note: the dumb allocator was removed in this release, as it was not kept
1958 up-to-date with the IAllocator protocol changes. It is recommended to
1959 use the ``hail`` command from the ganeti-htools package.
1961 Note: the 2.4 and up versions of Ganeti are not compatible with the
1962 0.2.x branch of ganeti-htools. You need to upgrade to
1963 ganeti-htools-0.3.0 (or later).
1965 Regressions fixed from 2.3
1966 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1968 - Fixed the ``gnt-cluster verify-disks`` command
1969 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1971 - Fixed disk adoption breakage
1972 - Fixed wrong headers in instance listing for field aliases
1977 - Fixed corner case in KVM handling of NICs
1978 - Fixed many cases of wrong handling of non-vm_capable nodes
1979 - Fixed a bug where a missing instance symlink was not possible to
1980 recreate with any ``gnt-*`` command (now ``gnt-instance
1981 activate-disks`` does it)
1982 - Fixed the volume group name as reported by ``gnt-cluster
1984 - Increased timeouts for the import-export code, hopefully leading to
1985 fewer aborts due network or instance timeouts
1986 - Fixed bug in ``gnt-node list-storage``
1987 - Fixed bug where not all daemons were started on cluster
1988 initialisation, but only at the first watcher run
1989 - Fixed many bugs in the OOB implementation
1990 - Fixed watcher behaviour in presence of instances with offline
1992 - Fixed instance list output for instances running on the wrong node
1993 - a few fixes to the cluster-merge tool, but it still cannot merge
1994 multi-node groups (currently it is not recommended to use this tool)
2000 - Improved network configuration for the KVM hypervisor
2001 - Added e1000 as a supported NIC for Xen-HVM
2002 - Improved the lvmstrap tool to also be able to use partitions, as
2003 opposed to full disks
2004 - Improved speed of disk wiping (the cluster parameter
2005 ``prealloc_wipe_disks``, so that it has a low impact on the total time
2006 of instance creations
2007 - Added documentation for the OS parameters
2008 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
2009 hypervisor is not responding
2010 - Added display of blacklisted and hidden OS information in
2011 ``gnt-cluster info``
2012 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
2013 NIC and node parameters, which might create problems with currently
2014 invalid (but undetected) configuration files, but prevents validation
2015 failures when unrelated parameters are modified
2016 - Changed cluster initialisation to wait for the master daemon to become
2018 - Expanded the RAPI interface:
2020 - Added config redistribution resource
2021 - Added activation/deactivation of instance disks
2022 - Added export of console information
2024 - Implemented log file reopening on SIGHUP, which allows using
2025 logrotate(8) for the Ganeti log files
2026 - Added a basic OOB helper script as an example
2032 *(Released Fri, 14 Jan 2011)*
2037 - Fixed timezone issues when formatting timestamps
2038 - Added support for node groups, available via ``gnt-group`` and other
2040 - Added out-of-band framework and management, see :doc:`design
2041 document <design-oob>`
2042 - Removed support for roman numbers from ``gnt-node list`` and
2043 ``gnt-instance list``.
2044 - Allowed modification of master network interface via ``gnt-cluster
2045 modify --master-netdev``
2046 - Accept offline secondaries while shutting down instance disks
2047 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
2048 - Added support for multiple LVM volume groups
2049 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
2051 - Added commands to list available fields:
2053 - ``gnt-node list-fields``
2054 - ``gnt-group list-fields``
2055 - ``gnt-instance list-fields``
2057 - Updated documentation and man pages
2062 - Moved ``rapi_users`` file into separate directory, now named
2063 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
2065 - Added new tool for running commands on many machines,
2066 ``tools/ganeti-listrunner``
2067 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
2068 improving the ``gnt-instance console`` output
2069 - Allowed customisation of disk index separator at ``configure`` time
2070 - Export node group allocation policy to :doc:`iallocator <iallocator>`
2071 - Added support for non-partitioned md disks in ``lvmstrap``
2072 - Added script to gracefully power off KVM instances
2073 - Split ``utils`` module into smaller parts
2074 - Changed query operations to return more detailed information, e.g.
2075 whether an information is unavailable due to an offline node. To use
2076 this new functionality, the LUXI call ``Query`` must be used. Field
2077 information is now stored by the master daemon and can be retrieved
2078 using ``QueryFields``. Instances, nodes and groups can also be queried
2079 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
2080 exposed via RAPI). The following commands make use of this
2081 infrastructure change:
2083 - ``gnt-group list``
2084 - ``gnt-group list-fields``
2086 - ``gnt-node list-fields``
2087 - ``gnt-instance list``
2088 - ``gnt-instance list-fields``
2089 - ``gnt-debug locks``
2094 - New RAPI resources (see :doc:`rapi`):
2098 - ``/2/groups/[group_name]``
2099 - ``/2/groups/[group_name]/assign-nodes``
2100 - ``/2/groups/[group_name]/modify``
2101 - ``/2/groups/[group_name]/rename``
2102 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
2106 - Implemented ``no_install`` for instance creation
2107 - Implemented OS parameters for instance reinstallation, allowing
2108 use of special settings on reinstallation (e.g. for preserving data)
2113 - Added IPv6 support in import/export
2114 - Pause DRBD synchronization while wiping disks on instance creation
2115 - Updated unittests and QA scripts
2116 - Improved network parameters passed to KVM
2117 - Converted man pages from docbook to reStructuredText
2123 *(Released Mon, 20 Dec 2010)*
2125 Released version 2.3.1~rc1 without any changes.
2131 *(Released Wed, 1 Dec 2010)*
2133 - impexpd: Disable OpenSSL compression in socat if possible (backport
2134 from master, commit e90739d625b, see :doc:`installation guide
2135 <install-quick>` for details)
2136 - Changed unittest coverage report to exclude test scripts
2137 - Added script to check version format
2143 *(Released Wed, 1 Dec 2010)*
2145 Released version 2.3.0~rc1 without any changes.
2151 *(Released Fri, 19 Nov 2010)*
2153 A number of bugfixes and documentation updates:
2155 - Update ganeti-os-interface documentation
2156 - Fixed a bug related to duplicate MACs or similar items which should be
2158 - Fix breakage in OS state modify
2159 - Reinstall instance: disallow offline secondaries (fixes bug related to
2160 OS changing but reinstall failing)
2161 - plus all the other fixes between 2.2.1 and 2.2.2
2167 *(Released Tue, 2 Nov 2010)*
2169 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
2170 - Fixed master failover race with watcher
2171 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
2173 - Accept previously stopped instance for export with instance removal
2174 - Simplify and extend the environment variables for instance OS scripts
2175 - Added new node flags, ``master_capable`` and ``vm_capable``
2176 - Added optional instance disk wiping prior during allocation. This is a
2177 cluster-wide option and can be set/modified using
2178 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
2179 - Added IPv6 support, see :doc:`design document <design-2.3>` and
2180 :doc:`install-quick`
2181 - Added a new watcher option (``--ignore-pause``)
2182 - Added option to ignore offline node on instance start/stop
2183 (``--ignore-offline``)
2184 - Allow overriding OS parameters with ``gnt-instance reinstall``
2185 - Added ability to change node's secondary IP address using ``gnt-node
2187 - Implemented privilege separation for all daemons except
2188 ``ganeti-noded``, see ``configure`` options
2189 - Complain if an instance's disk is marked faulty in ``gnt-cluster
2191 - Implemented job priorities (see ``ganeti(7)`` manpage)
2192 - Ignore failures while shutting down instances during failover from
2194 - Exit daemon's bootstrap process only once daemon is ready
2195 - Export more information via ``LUInstanceQuery``/remote API
2196 - Improved documentation, QA and unittests
2197 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
2198 restart if the file was created or changed
2199 - Added LUXI protocol version sent with each request and response,
2200 allowing detection of server/client mismatches
2201 - Moved the Python scripts among gnt-* and ganeti-* into modules
2202 - Moved all code related to setting up SSH to an external script,
2204 - Infrastructure changes for node group support in future versions
2210 *(Released Fri, 19 Nov 2010)*
2212 A few small bugs fixed, and some improvements to the build system:
2214 - Fix documentation regarding conversion to drbd
2215 - Fix validation of parameters in cluster modify (``gnt-cluster modify
2217 - Fix error handling in node modify with multiple changes
2218 - Allow remote imports without checked names
2224 *(Released Tue, 19 Oct 2010)*
2226 - Disable SSL session ID cache in RPC client
2232 *(Released Thu, 14 Oct 2010)*
2234 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
2235 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2237 - Fix problems with interaction between Curl and Python's HTTP server,
2238 resulting in increased speed in many RPC calls
2239 - Improve our release script to prevent breakage with older aclocal and
2246 *(Released Thu, 7 Oct 2010)*
2248 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2249 value retrieved from master
2250 - Added support for blacklisted or hidden OS definitions
2251 - Added simple lock monitor (accessible via (``gnt-debug locks``)
2252 - Added support for -mem-path in KVM hypervisor abstraction layer
2253 - Allow overriding instance parameters in tool for inter-cluster
2254 instance moves (``tools/move-instance``)
2255 - Improved opcode summaries (e.g. in ``gnt-job list``)
2256 - Improve consistency of OS listing by sorting it
2257 - Documentation updates
2263 *(Released Fri, 8 Oct 2010)*
2265 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
2271 *(Released Mon, 4 Oct 2010)*
2273 - Fixed regression in ``gnt-instance rename``
2279 *(Released Wed, 22 Sep 2010)*
2281 - Fixed OS_VARIANT variable for OS scripts
2282 - Fixed cluster tag operations via RAPI
2283 - Made ``setup-ssh`` exit with non-zero code if an error occurred
2284 - Disabled RAPI CA checks in watcher
2290 *(Released Mon, 23 Aug 2010)*
2292 - Support DRBD versions of the format "a.b.c.d"
2294 - Re-introduce support for usage from multiple threads in RAPI client
2295 - Instance renames and modify via RAPI
2296 - Work around race condition between processing and archival in job
2298 - Mark opcodes following failed one as failed, too
2299 - Job field ``lock_status`` was removed due to difficulties making it
2300 work with the changed job queue in Ganeti 2.2; a better way to monitor
2301 locks is expected for a later 2.2.x release
2302 - Fixed dry-run behaviour with many commands
2303 - Support ``ssh-agent`` again when adding nodes
2304 - Many additional bugfixes
2310 *(Released Fri, 30 Jul 2010)*
2312 Important change: the internal RPC mechanism between Ganeti nodes has
2313 changed from using a home-grown http library (based on the Python base
2314 libraries) to use the PycURL library. This requires that PycURL is
2315 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2316 against GnuTLS by default. cURL's support for GnuTLS had known issues
2317 before cURL 7.21.0 and we recommend using the latest cURL release or
2318 linking against OpenSSL. Most other distributions already link PycURL
2319 and cURL against OpenSSL. The command::
2321 python -c 'import pycurl; print pycurl.version'
2323 can be used to determine the libraries PycURL and cURL are linked
2326 Other significant changes:
2328 - Rewrote much of the internals of the job queue, in order to achieve
2329 better parallelism; this decouples job query operations from the job
2330 processing, and it should allow much nicer behaviour of the master
2331 daemon under load, and it also has uncovered some long-standing bugs
2332 related to the job serialisation (now fixed)
2333 - Added a default iallocator setting to the cluster parameters,
2334 eliminating the need to always pass nodes or an iallocator for
2335 operations that require selection of new node(s)
2336 - Added experimental support for the LXC virtualization method
2337 - Added support for OS parameters, which allows the installation of
2338 instances to pass parameter to OS scripts in order to customise the
2340 - Added a hypervisor parameter controlling the migration type (live or
2341 non-live), since hypervisors have various levels of reliability; this
2342 has renamed the 'live' parameter to 'mode'
2343 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
2344 logical volumes, meaning that cluster verify will ignore them and not
2345 flag their presence as errors
2346 - The watcher will now reset the error count for failed instances after
2347 8 hours, thus allowing self-healing if the problem that caused the
2348 instances to be down/fail to start has cleared in the meantime
2349 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2350 check for, and warn, if the drbd module parameter ``usermode_helper``
2351 is not consistent with the cluster-wide setting; this is needed to
2352 make diagnose easier of failed drbd creations
2353 - Started adding base IPv6 support, but this is not yet
2354 enabled/available for use
2355 - Rename operations (cluster, instance) will now return the new name,
2356 which is especially useful if a short name was passed in
2357 - Added support for instance migration in RAPI
2358 - Added a tool to pre-configure nodes for the SSH setup, before joining
2359 them to the cluster; this will allow in the future a simplified model
2360 for node joining (but not yet fully enabled in 2.2); this needs the
2361 paramiko python library
2362 - Fixed handling of name-resolving errors
2363 - Fixed consistency of job results on the error path
2364 - Fixed master-failover race condition when executed multiple times in
2366 - Fixed many bugs related to the job queue (mostly introduced during the
2367 2.2 development cycle, so not all are impacting 2.1)
2368 - Fixed instance migration with missing disk symlinks
2369 - Fixed handling of unknown jobs in ``gnt-job archive``
2370 - And many other small fixes/improvements
2374 - Enhanced both the unittest and the QA coverage
2375 - Switched the opcode validation to a generic model, and extended the
2376 validation to all opcode parameters
2377 - Changed more parts of the code that write shell scripts to use the
2379 - Switched the master daemon to use the asyncore library for the Luxi
2386 *(Released Thu, 17 Jun 2010)*
2388 - Added tool (``move-instance``) and infrastructure to move instances
2389 between separate clusters (see :doc:`separate documentation
2390 <move-instance>` and :doc:`design document <design-2.2>`)
2391 - Added per-request RPC timeout
2392 - RAPI now requires a Content-Type header for requests with a body (e.g.
2393 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2394 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2395 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2397 - Implemented initial support for running Ganeti daemons as separate
2398 users, see configure-time flags ``--with-user-prefix`` and
2399 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2401 - Instances can be removed after export (``gnt-backup export
2402 --remove-instance``)
2403 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2404 key (instead of 1024 bit)
2405 - Added new cluster configuration file for cluster domain secret
2406 - Import/export now use SSL instead of SSH
2407 - Added support for showing estimated time when exporting an instance,
2408 see the ``ganeti-os-interface(7)`` manpage and look for
2415 *(Released Tue, 16 Nov 2010)*
2417 Some more bugfixes. Unless critical bugs occur, this will be the last
2420 - Fix case of MAC special-values
2421 - Fix mac checker regex
2422 - backend: Fix typo causing "out of range" error
2423 - Add missing --units in gnt-instance list man page
2429 *(Released Tue, 24 Aug 2010)*
2432 - Don't ignore secondary node silently on non-mirrored disk templates
2434 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2435 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2436 - Properly document the usb_mouse parameter
2437 - Fix path in ganeti-rapi(8) (issue 116)
2438 - Adjust error message when the ganeti user's .ssh directory is
2440 - Add same-node-check when changing the disk template to drbd
2446 *(Released Fri, 16 Jul 2010)*
2449 - Add an option to only select some reboot types during qa/burnin.
2450 (on some hypervisors consequent reboots are not supported)
2451 - Fix infrequent race condition in master failover. Sometimes the old
2452 master ip address would be still detected as up for a short time
2453 after it was removed, causing failover to fail.
2454 - Decrease mlockall warnings when the ctypes module is missing. On
2455 Python 2.4 we support running even if no ctypes module is installed,
2456 but we were too verbose about this issue.
2457 - Fix building on old distributions, on which man doesn't have a
2459 - Fix RAPI not to ignore the MAC address on instance creation
2460 - Implement the old instance creation format in the RAPI client.
2466 *(Released Thu, 01 Jul 2010)*
2468 A small bugfix release:
2469 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2470 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2471 a non-existing option
2472 - Fix instance create: the --force-variant option was ignored
2473 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2474 - Fix modify node storage with non-FQDN arguments
2475 - Fix RAPI client to authenticate under Python 2.6 when used
2476 for more than 5 requests needing authentication
2477 - Fix gnt-instance modify -t (storage) giving a wrong error message
2478 when converting a non-shutdown drbd instance to plain
2484 *(Released Fri, 18 Jun 2010)*
2486 A small bugfix release:
2488 - Fix live migration of KVM instances started with older Ganeti
2489 versions which had fewer hypervisor parameters
2490 - Fix gnt-instance grow-disk on down instances
2491 - Fix an error-reporting bug during instance migration
2492 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2493 silently ignoring broken ones
2494 - Fix an RPC error reporting bug affecting, for example, RAPI client
2496 - Fix bug triggered by different API version os-es on different nodes
2497 - Fix a bug in instance startup with custom hvparams: OS level
2498 parameters would fail to be applied.
2499 - Fix the RAPI client under Python 2.6 (but more work is needed to
2500 make it work completely well with OpenSSL)
2501 - Fix handling of errors when resolving names from DNS
2507 *(Released Thu, 3 Jun 2010)*
2509 A medium sized development cycle. Some new features, and some
2510 fixes/small improvements/cleanups.
2512 Significant features
2513 ~~~~~~~~~~~~~~~~~~~~
2515 The node deamon now tries to mlock itself into memory, unless the
2516 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2517 its logs, and falls back to console logging. This allows emergency
2518 features such as ``gnt-node powercycle`` to work even in the event of a
2519 broken node disk (tested offlining the disk hosting the node's
2520 filesystem and dropping its memory caches; don't try this at home)
2522 KVM: add vhost-net acceleration support. It can be tested with a new
2523 enough version of the kernel and of qemu-kvm.
2525 KVM: Add instance chrooting feature. If you use privilege dropping for
2526 your VMs you can also now force them to chroot to an empty directory,
2527 before starting the emulated guest.
2529 KVM: Add maximum migration bandwith and maximum downtime tweaking
2530 support (requires a new-enough version of qemu-kvm).
2532 Cluster verify will now warn if the master node doesn't have the master
2533 ip configured on it.
2535 Add a new (incompatible) instance creation request format to RAPI which
2536 supports all parameters (previously only a subset was supported, and it
2537 wasn't possible to extend the old format to accomodate all the new
2538 features. The old format is still supported, and a client can check for
2539 this feature, before using it, by checking for its presence in the
2540 ``features`` RAPI resource.
2542 Now with ancient latin support. Try it passing the ``--roman`` option to
2543 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2544 (requires the python-roman module to be installed, in order to work).
2549 As usual many internal code refactorings, documentation updates, and
2552 - Lots of improvements and cleanups to the experimental Remote API
2553 (RAPI) client library.
2554 - A new unit test suite for the core daemon libraries.
2555 - A fix to creating missing directories makes sure the umask is not
2556 applied anymore. This enforces the same directory permissions
2558 - Better handling terminating daemons with ctrl+c (used when running
2559 them in debugging mode).
2560 - Fix a race condition in live migrating a KVM instance, when stat()
2561 on the old proc status file returned EINVAL, which is an unexpected
2563 - Fixed manpage checking with newer man and utf-8 charachters. But now
2564 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2570 *(Released Fri, 7 May 2010)*
2572 Fix a bug which prevented untagged KVM instances from starting.
2578 *(Released Fri, 7 May 2010)*
2580 Another release with a long development cycle, during which many
2581 different features were added.
2583 Significant features
2584 ~~~~~~~~~~~~~~~~~~~~
2586 The KVM hypervisor now can run the individual instances as non-root, to
2587 reduce the impact of a VM being hijacked due to bugs in the
2588 hypervisor. It is possible to run all instances as a single (non-root)
2589 user, to manually specify a user for each instance, or to dynamically
2590 allocate a user out of a cluster-wide pool to each instance, with the
2591 guarantee that no two instances will run under the same user ID on any
2594 An experimental RAPI client library, that can be used standalone
2595 (without the other Ganeti libraries), is provided in the source tree as
2596 ``lib/rapi/client.py``. Note this client might change its interface in
2597 the future, as we iterate on its capabilities.
2599 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2600 replace the cluster's certificates and crypto keys. This might help in
2601 case they have been compromised, or have simply expired.
2603 A new disk option for instance creation has been added that allows one
2604 to "adopt" currently existing logical volumes, with data
2605 preservation. This should allow easier migration to Ganeti from
2606 unmanaged (or managed via other software) instances.
2608 Another disk improvement is the possibility to convert between redundant
2609 (DRBD) and plain (LVM) disk configuration for an instance. This should
2610 allow better scalability (starting with one node and growing the
2611 cluster, or shrinking a two-node cluster to one node).
2613 A new feature that could help with automated node failovers has been
2614 implemented: if a node sees itself as offline (by querying the master
2615 candidates), it will try to shutdown (hard) all instances and any active
2616 DRBD devices. This reduces the risk of duplicate instances if an
2617 external script automatically failovers the instances on such nodes. To
2618 enable this, the cluster parameter ``maintain_node_health`` should be
2619 enabled; in the future this option (per the name) will enable other
2620 automatic maintenance features.
2622 Instance export/import now will reuse the original instance
2623 specifications for all parameters; that means exporting an instance,
2624 deleting it and the importing it back should give an almost identical
2625 instance. Note that the default import behaviour has changed from
2626 before, where it created only one NIC; now it recreates the original
2629 Cluster verify has added a few new checks: SSL certificates validity,
2630 /etc/hosts consistency across the cluster, etc.
2635 As usual, many internal changes were done, documentation fixes,
2638 - Fixed cluster initialization with disabled cluster storage (regression
2639 introduced in 2.1.1)
2640 - File-based storage supports growing the disks
2641 - Fixed behaviour of node role changes
2642 - Fixed cluster verify for some corner cases, plus a general rewrite of
2643 cluster verify to allow future extension with more checks
2644 - Fixed log spamming by watcher and node daemon (regression introduced
2646 - Fixed possible validation issues when changing the list of enabled
2648 - Fixed cleanup of /etc/hosts during node removal
2649 - Fixed RAPI response for invalid methods
2650 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2651 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2652 from ide disks, etc.)
2653 - Allow OS changes without re-installation (to record a changed OS
2654 outside of Ganeti, or to allow OS renames)
2655 - Allow instance creation without OS installation (useful for example if
2656 the OS will be installed manually, or restored from a backup not in
2658 - Implemented option to make cluster ``copyfile`` use the replication
2660 - Added list of enabled hypervisors to ssconf (possibly useful for
2662 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2664 - A partial form of node re-IP is possible via node readd, which now
2665 allows changed node primary IP
2666 - Command line utilities now show an informational message if the job is
2668 - The logs of the master daemon now show the PID/UID/GID of the
2675 *(Released Fri, 12 Mar 2010)*
2677 During the 2.1.0 long release candidate cycle, a lot of improvements and
2678 changes have accumulated with were released later as 2.1.1.
2683 The node evacuate command (``gnt-node evacuate``) was significantly
2684 rewritten, and as such the IAllocator protocol was changed - a new
2685 request type has been added. This unfortunate change during a stable
2686 series is designed to improve performance of node evacuations; on
2687 clusters with more than about five nodes and which are well-balanced,
2688 evacuation should proceed in parallel for all instances of the node
2689 being evacuated. As such, any existing IAllocator scripts need to be
2690 updated, otherwise the above command will fail due to the unknown
2691 request. The provided "dumb" allocator has not been updated; but the
2692 ganeti-htools package supports the new protocol since version 0.2.4.
2694 Another important change is increased validation of node and instance
2695 names. This might create problems in special cases, if invalid host
2696 names are being used.
2698 Also, a new layer of hypervisor parameters has been added, that sits at
2699 OS level between the cluster defaults and the instance ones. This allows
2700 customisation of virtualization parameters depending on the installed
2701 OS. For example instances with OS 'X' may have a different KVM kernel
2702 (or any other parameter) than the cluster defaults. This is intended to
2703 help managing a multiple OSes on the same cluster, without manual
2704 modification of each instance's parameters.
2706 A tool for merging clusters, ``cluster-merge``, has been added in the
2707 tools sub-directory.
2712 - Improved the int/float conversions that should make the code more
2713 robust in face of errors from the node daemons
2714 - Fixed the remove node code in case of internal configuration errors
2715 - Fixed the node daemon behaviour in face of inconsistent queue
2716 directory (e.g. read-only file-system where we can't open the files
2718 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2719 now it either aborts cleanly or, if given the new "auto_promote"
2720 parameter, will automatically promote other nodes as needed
2721 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2722 completely prevent Ganeti from working
2723 - Fixed bug for instance export when not all disks were successfully
2725 - Fixed behaviour of node add when the new node is slow in starting up
2727 - Fixed handling of signals in the LUXI client, which should improve
2728 behaviour of command-line scripts
2729 - Added checks for invalid node/instance names in the configuration (now
2730 flagged during cluster verify)
2731 - Fixed watcher behaviour for disk activation errors
2732 - Fixed two potentially endless loops in http library, which led to the
2733 RAPI daemon hanging and consuming 100% CPU in some cases
2734 - Fixed bug in RAPI daemon related to hashed passwords
2735 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2737 - Enhanced compatibility with non-Debian OSes, but not using absolute
2738 path in some commands and allowing customisation of the ssh
2739 configuration directory
2740 - Fixed possible future issue with new Python versions by abiding to the
2741 proper use of ``__slots__`` attribute on classes
2742 - Added checks that should prevent directory traversal attacks
2743 - Many documentation fixes based on feedback from users
2748 - Added an "early_release" more for instance replace disks and node
2749 evacuate, where we release locks earlier and thus allow higher
2750 parallelism within the cluster
2751 - Added watcher hooks, intended to allow the watcher to restart other
2752 daemons (e.g. from the ganeti-nbma project), but they can be used of
2753 course for any other purpose
2754 - Added a compile-time disable for DRBD barriers, to increase
2755 performance if the administrator trusts the power supply or the
2756 storage system to not lose writes
2757 - Added the option of using syslog for logging instead of, or in
2758 addition to, Ganeti's own log files
2759 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2760 can indeed boot from a paravirtual NIC
2761 - Added a generic debug level for many operations; while this is not
2762 used widely yet, it allows one to pass the debug value all the way to
2764 - Enhanced the hooks environment for instance moves (failovers,
2765 migrations) where the primary/secondary nodes changed during the
2766 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2767 - Enhanced data validations for many user-supplied values; one important
2768 item is the restrictions imposed on instance and node names, which
2769 might reject some (invalid) host names
2770 - Add a configure-time option to disable file-based storage, if it's not
2771 needed; this allows greater security separation between the master
2772 node and the other nodes from the point of view of the inter-node RPC
2774 - Added user notification in interactive tools if job is waiting in the
2775 job queue or trying to acquire locks
2776 - Added log messages when a job is waiting for locks
2777 - Added filtering by node tags in instance operations which admit
2778 multiple instances (start, stop, reboot, reinstall)
2779 - Added a new tool for cluster mergers, ``cluster-merge``
2780 - Parameters from command line which are of the form ``a=b,c=d`` can now
2781 use backslash escapes to pass in values which contain commas,
2782 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2784 - For KVM, the instance name is the first parameter passed to KVM, so
2785 that it's more visible in the process list
2791 *(Released Tue, 2 Mar 2010)*
2793 Ganeti 2.1 brings many improvements with it. Major changes:
2795 - Added infrastructure to ease automated disk repairs
2796 - Added new daemon to export configuration data in a cheaper way than
2797 using the remote API
2798 - Instance NICs can now be routed instead of being associated with a
2800 - Improved job locking logic to reduce impact of jobs acquiring multiple
2801 locks waiting for other long-running jobs
2803 In-depth implementation details can be found in the Ganeti 2.1 design
2809 - Added chroot hypervisor
2810 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2812 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2813 ``bootloader_path`` and ``bootloader_args``)
2814 - Added the ``use_localtime`` option for the xen-hvm and kvm
2815 hypervisors, and the default value for this has changed to false (in
2816 2.0 xen-hvm always enabled it)
2817 - Added luxi call to submit multiple jobs in one go
2818 - Added cluster initialization option to not modify ``/etc/hosts``
2820 - Added network interface parameters
2821 - Added dry run mode to some LUs
2822 - Added RAPI resources:
2824 - ``/2/instances/[instance_name]/info``
2825 - ``/2/instances/[instance_name]/replace-disks``
2826 - ``/2/nodes/[node_name]/evacuate``
2827 - ``/2/nodes/[node_name]/migrate``
2828 - ``/2/nodes/[node_name]/role``
2829 - ``/2/nodes/[node_name]/storage``
2830 - ``/2/nodes/[node_name]/storage/modify``
2831 - ``/2/nodes/[node_name]/storage/repair``
2833 - Added OpCodes to evacuate or migrate all instances on a node
2834 - Added new command to list storage elements on nodes (``gnt-node
2835 list-storage``) and modify them (``gnt-node modify-storage``)
2836 - Added new ssconf files with master candidate IP address
2837 (``ssconf_master_candidates_ips``), node primary IP address
2838 (``ssconf_node_primary_ips``) and node secondary IP address
2839 (``ssconf_node_secondary_ips``)
2840 - Added ``ganeti-confd`` and a client library to query the Ganeti
2841 configuration via UDP
2842 - Added ability to run hooks after cluster initialization and before
2844 - Added automatic mode for disk replace (``gnt-instance replace-disks
2846 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2847 after catastrophic data-loss
2848 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2850 - Added ``gnt-instance move`` command to move instances
2851 - Added ``gnt-cluster watcher`` command to control watcher
2852 - Added ``gnt-node powercycle`` command to powercycle nodes
2853 - Added new job status field ``lock_status``
2854 - Added parseable error codes to cluster verification (``gnt-cluster
2855 verify --error-codes``) and made output less verbose (use
2856 ``--verbose`` to restore previous behaviour)
2857 - Added UUIDs to the main config entities (cluster, nodes, instances)
2858 - Added support for OS variants
2859 - Added support for hashed passwords in the Ganeti remote API users file
2861 - Added option to specify maximum timeout on instance shutdown
2862 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2863 - Added new helper script to start and stop Ganeti daemons
2864 (``daemon-util``), with the intent to reduce the work necessary to
2865 adjust Ganeti for non-Debian distributions and to start/stop daemons
2867 - Added more unittests
2868 - Fixed critical bug in ganeti-masterd startup
2869 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2870 now customisable at the cluster level for both the KVM and Xen
2871 hypervisors using the new ``migration_port`` parameter
2872 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2873 reinstalling an instance
2874 - Allowed ``@`` in tag names
2875 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2876 - Many documentation updates
2877 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2878 - ``gnt-instance reinstall`` can now reinstall multiple instances
2879 - Updated many command line parameters
2880 - Introduced new OS API version 15
2881 - No longer support a default hypervisor
2882 - Treat virtual LVs as inexistent
2883 - Improved job locking logic to reduce lock contention
2884 - Match instance and node names case insensitively
2885 - Reimplemented bash completion script to be more complete
2892 *(Released Thu, 4 Feb 2010)*
2894 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2895 - Fix a string formatting bug
2896 - Improve safety of the code in some error paths
2897 - Improve data validation in the master of values returned from nodes
2903 *(Released Thu, 17 Dec 2009)*
2905 - Fix security issue due to missing validation of iallocator names; this
2906 allows local and remote execution of arbitrary executables
2907 - Fix failure of gnt-node list during instance removal
2908 - Ship the RAPI documentation in the archive
2914 *(Released Wed, 30 Sep 2009)*
2916 - Fixed many wrong messages
2917 - Fixed a few bugs related to the locking library
2918 - Fixed MAC checking at instance creation time
2919 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2920 - Fixed a few issues related to signal handling in both daemons and
2922 - Fixed the example startup script provided
2923 - Fixed insserv dependencies in the example startup script (patch from
2925 - Fixed handling of drained nodes in the iallocator framework
2926 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2928 - Fixed error related to invalid job IDs in job polling
2929 - Fixed job/opcode persistence on unclean master shutdown
2930 - Fixed handling of partial job processing after unclean master
2932 - Fixed error reporting from LUs, previously all errors were converted
2933 into execution errors
2934 - Fixed error reporting from burnin
2935 - Decreased significantly the memory usage of the job queue
2936 - Optimised slightly multi-job submission
2937 - Optimised slightly opcode loading
2938 - Backported the multi-job submit framework from the development
2939 branch; multi-instance start and stop should be faster
2940 - Added script to clean archived jobs after 21 days; this will reduce
2941 the size of the queue directory
2942 - Added some extra checks in disk size tracking
2943 - Added an example ethers hook script
2944 - Added a cluster parameter that prevents Ganeti from modifying of
2946 - Added more node information to RAPI responses
2947 - Added a ``gnt-job watch`` command that allows following the ouput of a
2949 - Added a bind-address option to ganeti-rapi
2950 - Added more checks to the configuration verify
2951 - Enhanced the burnin script such that some operations can be retried
2953 - Converted instance reinstall to multi-instance model
2959 *(Released Fri, 7 Aug 2009)*
2961 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2962 to allow using the pre-2.0.2 behaviour in activation, if any existing
2963 instances have mismatched disk sizes in the configuration
2964 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2965 any configuration mismatches for disk sizes
2966 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2967 failover to work on two-node clusters
2968 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2970 - Fixed detection of OS script errors in ``gnt-backup export``
2971 - Fixed exit code of ``gnt-backup export``
2977 *(Released Fri, 17 Jul 2009)*
2979 - Added experimental support for stripped logical volumes; this should
2980 enhance performance but comes with a higher complexity in the block
2981 device handling; stripping is only enabled when passing
2982 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2983 affected even in the non-stripped mode
2984 - Improved resiliency against transient failures at the end of DRBD
2985 resyncs, and in general of DRBD resync checks
2986 - Fixed a couple of issues with exports and snapshot errors
2987 - Fixed a couple of issues in instance listing
2988 - Added display of the disk size in ``gnt-instance info``
2989 - Fixed checking for valid OSes in instance creation
2990 - Fixed handling of the "vcpus" parameter in instance listing and in
2991 general of invalid parameters
2992 - Fixed http server library, and thus RAPI, to handle invalid
2993 username/password combinations correctly; this means that now they
2994 report unauthorized for queries too, not only for modifications,
2995 allowing earlier detect of configuration problems
2996 - Added a new "role" node list field, equivalent to the master/master
2997 candidate/drained/offline flags combinations
2998 - Fixed cluster modify and changes of candidate pool size
2999 - Fixed cluster verify error messages for wrong files on regular nodes
3000 - Fixed a couple of issues with node demotion from master candidate role
3001 - Fixed node readd issues
3002 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
3003 - Added a new ``--no-voting`` option for masterfailover to fix failover
3004 on two-nodes clusters when the former master node is unreachable
3005 - Added instance reinstall over RAPI
3011 *(Released Tue, 16 Jun 2009)*
3013 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
3014 allow re-adding the start in single-user option (regression from 1.2)
3015 - the watcher writes the instance status to a file, to allow monitoring
3016 to report the instance status (from the master) based on cached
3017 results of the watcher's queries; while this can get stale if the
3018 watcher is being locked due to other work on the cluster, this is
3019 still an improvement
3020 - the watcher now also restarts the node daemon and the rapi daemon if
3022 - fixed the watcher to handle full and drained queue cases
3023 - hooks export more instance data in the environment, which helps if
3024 hook scripts need to take action based on the instance's properties
3025 (no longer need to query back into ganeti)
3026 - instance failovers when the instance is stopped do not check for free
3027 RAM, so that failing over a stopped instance is possible in low memory
3029 - rapi uses queries for tags instead of jobs (for less job traffic), and
3030 for cluster tags it won't talk to masterd at all but read them from
3032 - a couple of error handling fixes in RAPI
3033 - drbd handling: improved the error handling of inconsistent disks after
3034 resync to reduce the frequency of "there are some degraded disks for
3035 this instance" messages
3036 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
3037 error handling path called a wrong function name)
3043 *(Released Wed, 27 May 2009)*
3045 - no changes from rc5
3051 *(Released Wed, 20 May 2009)*
3053 - fix a couple of bugs (validation, argument checks)
3054 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
3055 - some small improvements to RAPI and IAllocator
3056 - make watcher automatically start the master daemon if down
3062 *(Released Mon, 27 Apr 2009)*
3064 - change the OS list to not require locks; this helps with big clusters
3065 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
3066 volume group is broken
3067 - ``gnt-instance info``, without any arguments, doesn't run for all
3068 instances anymore; either pass ``--all`` or pass the desired
3069 instances; this helps against mistakes on big clusters where listing
3070 the information for all instances takes a long time
3071 - miscellaneous doc and man pages fixes
3077 *(Released Wed, 8 Apr 2009)*
3079 - Change the internal locking model of some ``gnt-node`` commands, in
3080 order to reduce contention (and blocking of master daemon) when
3081 batching many creation/reinstall jobs
3082 - Fixes to Xen soft reboot
3083 - No longer build documentation at build time, instead distribute it in
3084 the archive, in order to reduce the need for the whole docbook/rst
3091 *(Released Fri, 27 Mar 2009)*
3093 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
3094 - Fix watcher startup sequence, improves the behaviour of busy clusters
3095 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
3096 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
3097 bind address and other places
3098 - Some documentation fixes and updates
3104 *(Released Mon, 2 Mar 2009)*
3106 - More documentation updates, now all docs should be more-or-less
3108 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
3110 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
3112 - Fix an issue related to $libdir/run/ganeti and cluster creation
3118 *(Released Thu, 19 Feb 2009)*
3120 - Xen PVM and KVM have switched the default value for the instance root
3121 disk to the first partition on the first drive, instead of the whole
3122 drive; this means that the OS installation scripts must be changed
3124 - Man pages have been updated
3125 - RAPI has been switched by default to HTTPS, and the exported functions
3126 should all work correctly
3127 - RAPI v1 has been removed
3128 - Many improvements to the KVM hypervisor
3129 - Block device errors are now better reported
3130 - Many other bugfixes and small improvements
3136 *(Released Mon, 26 Jan 2009)*
3138 - Version 2 is a general rewrite of the code and therefore the
3139 differences are too many to list, see the design document for 2.0 in
3140 the ``doc/`` subdirectory for more details
3141 - In this beta version there is not yet a migration path from 1.2 (there
3142 will be one in the final 2.0 release)
3143 - A few significant changes are:
3145 - all commands are executed by a daemon (``ganeti-masterd``) and the
3146 various ``gnt-*`` commands are just front-ends to it
3147 - all the commands are entered into, and executed from a job queue,
3148 see the ``gnt-job(8)`` manpage
3149 - the RAPI daemon supports read-write operations, secured by basic
3150 HTTP authentication on top of HTTPS
3151 - DRBD version 0.7 support has been removed, DRBD 8 is the only
3152 supported version (when migrating from Ganeti 1.2 to 2.0, you need
3153 to migrate to DRBD 8 first while still running Ganeti 1.2)
3154 - DRBD devices are using statically allocated minor numbers, which
3155 will be assigned to existing instances during the migration process
3156 - there is support for both Xen PVM and Xen HVM instances running on
3158 - KVM virtualization is supported too
3159 - file-based storage has been implemented, which means that it is
3160 possible to run the cluster without LVM and DRBD storage, for
3161 example using a shared filesystem exported from shared storage (and
3162 still have live migration)
3168 *(Released Tue, 13 Jan 2009)*
3170 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
3171 - Reuse the old instance mac address by default on instance import, if
3172 the instance name is the same.
3173 - Handle situations in which the node info rpc returns incomplete
3175 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
3176 - Improved version of batcher:
3178 - state file support
3179 - instance mac address support
3180 - support for HVM clusters/instances
3182 - Add an option to show the number of cpu sockets and nodes in
3184 - Support OSes that handle more than one version of the OS api (but do
3185 not change the current API in any other way)
3186 - Fix ``gnt-node migrate``
3187 - ``gnt-debug`` man page
3188 - Fixes various more typos and small issues
3189 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
3195 *(Released Wed, 24 Sep 2008)*
3197 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
3198 type of disk exported to fully virtualized instances.
3199 - provide access to the serial console of HVM instances
3200 - instance auto_balance flag, set by default. If turned off it will
3201 avoid warnings on cluster verify if there is not enough memory to fail
3202 over an instance. in the future it will prevent automatically failing
3203 it over when we will support that.
3204 - batcher tool for instance creation, see ``tools/README.batcher``
3205 - ``gnt-instance reinstall --select-os`` to interactively select a new
3206 operating system when reinstalling an instance.
3207 - when changing the memory amount on instance modify a check has been
3208 added that the instance will be able to start. also warnings are
3209 emitted if the instance will not be able to fail over, if auto_balance
3211 - documentation fixes
3212 - sync fields between ``gnt-instance list/modify/add/import``
3213 - fix a race condition in drbd when the sync speed was set after giving
3214 the device a remote peer.
3220 *(Released Tue, 22 Jul 2008)*
3222 - note: the allowed size and number of tags per object were reduced
3223 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
3224 - fixed twisted 8.x compatibility
3225 - fixed ``gnt-instance replace-disks`` with iallocator
3226 - add TCP keepalives on twisted connections to detect restarted nodes
3227 - disk increase support, see ``gnt-instance grow-disk``
3228 - implement bulk node/instance query for RAPI
3229 - add tags in node/instance listing (optional)
3230 - experimental migration (and live migration) support, read the man page
3231 for ``gnt-instance migrate``
3232 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3233 has some small improvements in handling its state file
3239 *(Released Fri, 13 Jun 2008)*
3241 - Experimental readonly, REST-based remote API implementation;
3242 automatically started on master node, TCP port 5080, if enabled by
3243 ``--enable-rapi`` parameter to configure script.
3244 - Instance allocator support. Add and import instance accept a
3245 ``--iallocator`` parameter, and call that instance allocator to decide
3246 which node to use for the instance. The iallocator document describes
3247 what's expected from an allocator script.
3248 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3249 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3250 node is lost there is still enough memory to fail over the instances
3252 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3253 ``gnt-cluster verify``, to check for site-specific compliance. All the
3254 hooks will run, and their output, if any, will be displayed. Any
3255 failing hook will make the verification return an error value.
3256 - ``gnt-cluster verify`` now checks that its peers are reachable on the
3257 primary and secondary interfaces
3258 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3259 that is still declared as part of the cluster and has failed.
3260 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
3261 specifying output fields, that just adds the chosen fields to the
3263 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
3264 export from the filesystem.
3265 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3266 have been added. Using them you can enable/disable acpi and pae
3267 support, and specify a path for a cd image to be exported to the
3268 instance. These parameters as the name suggest only work on HVM
3270 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3271 PAE support will be set to the previously hardcoded values, but the
3272 (previously hardcoded) path to the CDROM ISO image will be unset and
3273 if required, needs to be set manually with ``gnt-instance modify``
3275 - The address to which an instance's VNC console is bound is now
3276 selectable per-instance, rather than being cluster wide. Of course
3277 this only applies to instances controlled via VNC, so currently just
3278 applies to HVM clusters.
3284 *(Released Mon, 18 Feb 2008)*
3286 - more tweaks to the disk activation code (especially helpful for DRBD)
3287 - change the default ``gnt-instance list`` output format, now there is
3288 one combined status field (see the manpage for the exact values this
3290 - some more fixes for the mac export to hooks change
3291 - make Ganeti not break with DRBD 8.2.x (which changed the version
3292 format in ``/proc/drbd``) (issue 24)
3293 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3294 template, allowing migration from DRBD0.7+MD to DRBD8
3300 *(Released Wed, 30 Jan 2008)*
3302 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3304 - add command aliases infrastructure and a few aliases
3305 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
3306 man pages and the ``--help`` option of ``gnt-node
3307 list``/``gnt-instance list``
3308 - fix ``gnt-backup list`` with down nodes (issue 21)
3309 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3310 - fix the dist archive and add a check for including svn/git files in
3312 - some developer-related changes: improve the burnin and the QA suite,
3313 add an upload script for testing during development
3319 *(Released Wed, 16 Jan 2008)*
3321 - experimental HVM support, read the install document, section
3322 "Initializing the cluster"
3323 - allow for the PVM hypervisor per-instance kernel and initrd paths
3324 - add a new command ``gnt-cluster verify-disks`` which uses a new
3325 algorithm to improve the reconnection of the DRBD pairs if the device
3326 on the secondary node has gone away
3327 - make logical volume code auto-activate LVs at disk activation time
3328 - slightly improve the speed of activating disks
3329 - allow specification of the MAC address at instance creation time, and
3330 changing it later via ``gnt-instance modify``
3331 - fix handling of external commands that generate lots of output on
3333 - update documentation with regard to minimum version of DRBD8 supported
3339 *(Released Tue, 4 Dec 2007)*
3341 - Log the ``xm create`` output to the node daemon log on failure (to
3342 help diagnosing the error)
3343 - In debug mode, log all external commands output if failed to the logs
3344 - Change parsing of lvm commands to ignore stderr
3350 *(Released Wed, 28 Nov 2007)*
3352 - Another round of updates to the DRBD 8 code to deal with more failures
3353 in the replace secondary node operation
3354 - Some more logging of failures in disk operations (lvm, drbd)
3355 - A few documentation updates
3362 *(Released Tue, 13 Nov 2007)*
3364 - Change configuration file format from Python's Pickle to JSON.
3365 Upgrading is possible using the cfgupgrade utility.
3366 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3367 faster replace disks and is more stable (DRBD 8 has many improvements
3368 compared to DRBD 0.7)
3369 - Added command line tags support (see man pages for ``gnt-instance``,
3370 ``gnt-node``, ``gnt-cluster``)
3371 - Added instance rename support
3372 - Added multi-instance startup/shutdown
3373 - Added cluster rename support
3374 - Added ``gnt-node evacuate`` to simplify some node operations
3375 - Added instance reboot operation that can speedup reboot as compared to
3377 - Soften the requirement that hostnames are in FQDN format
3378 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3380 - Removed dependency on debian's patched fping that uses the
3381 non-standard ``-S`` option
3382 - Now the OS definitions are searched for in multiple, configurable
3383 paths (easier for distros to package)
3384 - Some changes to the hooks infrastructure (especially the new
3385 post-configuration update hook)
3386 - Other small bugfixes
3388 .. vim: set textwidth=72 syntax=rst :