7 *(Released Fri, 19 Nov 2010)*
9 A few small bugs fixed, and some improvements to the build system:
11 - Fix documentation regarding conversion to drbd
12 - Fix validation of parameters in cluster modify (``gnt-cluster modify
14 - Fix error handling in node modify with multiple changes
15 - Allow remote imports without checked names
21 *(Released Tue, 19 Oct 2010)*
23 - Disable SSL session ID cache in RPC client
29 *(Released Thu, 14 Oct 2010)*
31 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
32 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
34 - Fix problems with interaction between Curl and Python's HTTP server,
35 resulting in increased speed in many RPC calls
36 - Improve our release script to prevent breakage with older aclocal and
43 *(Released Thu, 7 Oct 2010)*
45 - Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with
46 value retrieved from master
47 - Added support for blacklisted or hidden OS definitions
48 - Added simple lock monitor (accessible via (``gnt-debug locks``)
49 - Added support for -mem-path in KVM hypervisor abstraction layer
50 - Allow overriding instance parameters in tool for inter-cluster
51 instance moves (``tools/move-instance``)
52 - Improved opcode summaries (e.g. in ``gnt-job list``)
53 - Improve consistency of OS listing by sorting it
54 - Documentation updates
60 *(Released Fri, 8 Oct 2010)*
62 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
68 *(Released Mon, 4 Oct 2010)*
70 - Fixed regression in ``gnt-instance rename``
76 *(Released Wed, 22 Sep 2010)*
78 - Fixed OS_VARIANT variable for OS scripts
79 - Fixed cluster tag operations via RAPI
80 - Made ``setup-ssh`` exit with non-zero code if an error occurred
81 - Disabled RAPI CA checks in watcher
87 *(Released Mon, 23 Aug 2010)*
89 - Support DRBD versions of the format "a.b.c.d"
91 - Re-introduce support for usage from multiple threads in RAPI client
92 - Instance renames and modify via RAPI
93 - Work around race condition between processing and archival in job
95 - Mark opcodes following failed one as failed, too
96 - Job field ``lock_status`` was removed due to difficulties making it
97 work with the changed job queue in Ganeti 2.2; a better way to monitor
98 locks is expected for a later 2.2.x release
99 - Fixed dry-run behaviour with many commands
100 - Support ``ssh-agent`` again when adding nodes
101 - Many additional bugfixes
107 *(Released Fri, 30 Jul 2010)*
109 Important change: the internal RPC mechanism between Ganeti nodes has
110 changed from using a home-grown http library (based on the Python base
111 libraries) to use the PycURL library. This requires that PycURL is
112 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
113 against GnuTLS by default. cURL's support for GnuTLS had known issues
114 before cURL 7.21.0 and we recommend using the latest cURL release or
115 linking against OpenSSL. Most other distributions already link PycURL
116 and cURL against OpenSSL. The command::
118 python -c 'import pycurl; print pycurl.version'
120 can be used to determine the libraries PycURL and cURL are linked
123 Other significant changes:
125 - Rewrote much of the internals of the job queue, in order to achieve
126 better parallelism; this decouples job query operations from the job
127 processing, and it should allow much nicer behaviour of the master
128 daemon under load, and it also has uncovered some long-standing bugs
129 related to the job serialisation (now fixed)
130 - Added a default iallocator setting to the cluster parameters,
131 eliminating the need to always pass nodes or an iallocator for
132 operations that require selection of new node(s)
133 - Added experimental support for the LXC virtualization method
134 - Added support for OS parameters, which allows the installation of
135 instances to pass parameter to OS scripts in order to customise the
137 - Added a hypervisor parameter controlling the migration type (live or
138 non-live), since hypervisors have various levels of reliability; this
139 has renamed the 'live' parameter to 'mode'
140 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
141 logical volumes, meaning that cluster verify will ignore them and not
142 flag their presence as errors
143 - The watcher will now reset the error count for failed instances after
144 8 hours, thus allowing self-healing if the problem that caused the
145 instances to be down/fail to start has cleared in the meantime
146 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
147 check for, and warn, if the drbd module parameter ``usermode_helper``
148 is not consistent with the cluster-wide setting; this is needed to
149 make diagnose easier of failed drbd creations
150 - Started adding base IPv6 support, but this is not yet
151 enabled/available for use
152 - Rename operations (cluster, instance) will now return the new name,
153 which is especially useful if a short name was passed in
154 - Added support for instance migration in RAPI
155 - Added a tool to pre-configure nodes for the SSH setup, before joining
156 them to the cluster; this will allow in the future a simplified model
157 for node joining (but not yet fully enabled in 2.2); this needs the
158 paramiko python library
159 - Fixed handling of name-resolving errors
160 - Fixed consistency of job results on the error path
161 - Fixed master-failover race condition when executed multiple times in
163 - Fixed many bugs related to the job queue (mostly introduced during the
164 2.2 development cycle, so not all are impacting 2.1)
165 - Fixed instance migration with missing disk symlinks
166 - Fixed handling of unknown jobs in ``gnt-job archive``
167 - And many other small fixes/improvements
171 - Enhanced both the unittest and the QA coverage
172 - Switched the opcode validation to a generic model, and extended the
173 validation to all opcode parameters
174 - Changed more parts of the code that write shell scripts to use the
176 - Switched the master daemon to use the asyncore library for the Luxi
183 *(Released Thu, 17 Jun 2010)*
185 - Added tool (``move-instance``) and infrastructure to move instances
186 between separate clusters (see :doc:`separate documentation
187 <move-instance>` and :doc:`design document <design-2.2>`)
188 - Added per-request RPC timeout
189 - RAPI now requires a Content-Type header for requests with a body (e.g.
190 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
191 :rfc:`2616` (HTTP/1.1), section 7.2.1)
192 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
194 - Implemented initial support for running Ganeti daemons as separate
195 users, see configure-time flags ``--with-user-prefix`` and
196 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
198 - Instances can be removed after export (``gnt-backup export
200 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
201 key (instead of 1024 bit)
202 - Added new cluster configuration file for cluster domain secret
203 - Import/export now use SSL instead of SSH
204 - Added support for showing estimated time when exporting an instance,
205 see the ``ganeti-os-interface(7)`` manpage and look for
212 *(Released Tue, 16 Nov 2010)*
214 Some more bugfixes. Unless critical bugs occur, this will be the last
217 - Fix case of MAC special-values
218 - Fix mac checker regex
219 - backend: Fix typo causing “out of range” error
220 - Add missing --units in gnt-instance list man page
226 *(Released Tue, 24 Aug 2010)*
229 - Don't ignore secondary node silently on non-mirrored disk templates
231 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
232 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
233 - Properly document the usb_mouse parameter
234 - Fix path in ganeti-rapi(8) (issue 116)
235 - Adjust error message when the ganeti user's .ssh directory is
237 - Add same-node-check when changing the disk template to drbd
243 *(Released Fri, 16 Jul 2010)*
246 - Add an option to only select some reboot types during qa/burnin.
247 (on some hypervisors consequent reboots are not supported)
248 - Fix infrequent race condition in master failover. Sometimes the old
249 master ip address would be still detected as up for a short time
250 after it was removed, causing failover to fail.
251 - Decrease mlockall warnings when the ctypes module is missing. On
252 Python 2.4 we support running even if no ctypes module is installed,
253 but we were too verbose about this issue.
254 - Fix building on old distributions, on which man doesn't have a
256 - Fix RAPI not to ignore the MAC address on instance creation
257 - Implement the old instance creation format in the RAPI client.
263 *(Released Thu, 01 Jul 2010)*
265 A small bugfix release:
266 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
267 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
268 a non-existing option
269 - Fix instance create: the --force-variant option was ignored
270 - Improve pylint 0.21 compatibility and warnings with Python 2.6
271 - Fix modify node storage with non-FQDN arguments
272 - Fix RAPI client to authenticate under Python 2.6 when used
273 for more than 5 requests needing authentication
274 - Fix gnt-instance modify -t (storage) giving a wrong error message
275 when converting a non-shutdown drbd instance to plain
281 *(Released Fri, 18 Jun 2010)*
283 A small bugfix release:
285 - Fix live migration of KVM instances started with older Ganeti
286 versions which had fewer hypervisor parameters
287 - Fix gnt-instance grow-disk on down instances
288 - Fix an error-reporting bug during instance migration
289 - Better checking of the ``--net`` and ``--disk`` values, to avoid
290 silently ignoring broken ones
291 - Fix an RPC error reporting bug affecting, for example, RAPI client
293 - Fix bug triggered by different API version os-es on different nodes
294 - Fix a bug in instance startup with custom hvparams: OS level
295 parameters would fail to be applied.
296 - Fix the RAPI client under Python 2.6 (but more work is needed to
297 make it work completely well with OpenSSL)
298 - Fix handling of errors when resolving names from DNS
304 *(Released Thu, 3 Jun 2010)*
306 A medium sized development cycle. Some new features, and some
307 fixes/small improvements/cleanups.
312 The node deamon now tries to mlock itself into memory, unless the
313 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
314 its logs, and falls back to console logging. This allows emergency
315 features such as ``gnt-node powercycle`` to work even in the event of a
316 broken node disk (tested offlining the disk hosting the node's
317 filesystem and dropping its memory caches; don't try this at home)
319 KVM: add vhost-net acceleration support. It can be tested with a new
320 enough version of the kernel and of qemu-kvm.
322 KVM: Add instance chrooting feature. If you use privilege dropping for
323 your VMs you can also now force them to chroot to an empty directory,
324 before starting the emulated guest.
326 KVM: Add maximum migration bandwith and maximum downtime tweaking
327 support (requires a new-enough version of qemu-kvm).
329 Cluster verify will now warn if the master node doesn't have the master
332 Add a new (incompatible) instance creation request format to RAPI which
333 supports all parameters (previously only a subset was supported, and it
334 wasn't possible to extend the old format to accomodate all the new
335 features. The old format is still supported, and a client can check for
336 this feature, before using it, by checking for its presence in the
337 ``features`` RAPI resource.
339 Now with ancient latin support. Try it passing the ``--roman`` option to
340 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
341 (requires the python-roman module to be installed, in order to work).
346 As usual many internal code refactorings, documentation updates, and
349 - Lots of improvements and cleanups to the experimental Remote API
350 (RAPI) client library.
351 - A new unit test suite for the core daemon libraries.
352 - A fix to creating missing directories makes sure the umask is not
353 applied anymore. This enforces the same directory permissions
355 - Better handling terminating daemons with ctrl+c (used when running
356 them in debugging mode).
357 - Fix a race condition in live migrating a KVM instance, when stat()
358 on the old proc status file returned EINVAL, which is an unexpected
360 - Fixed manpage checking with newer man and utf-8 charachters. But now
361 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
367 *(Released Fri, 7 May 2010)*
369 Fix a bug which prevented untagged KVM instances from starting.
375 *(Released Fri, 7 May 2010)*
377 Another release with a long development cycle, during which many
378 different features were added.
383 The KVM hypervisor now can run the individual instances as non-root, to
384 reduce the impact of a VM being hijacked due to bugs in the
385 hypervisor. It is possible to run all instances as a single (non-root)
386 user, to manually specify a user for each instance, or to dynamically
387 allocate a user out of a cluster-wide pool to each instance, with the
388 guarantee that no two instances will run under the same user ID on any
391 An experimental RAPI client library, that can be used standalone
392 (without the other Ganeti libraries), is provided in the source tree as
393 ``lib/rapi/client.py``. Note this client might change its interface in
394 the future, as we iterate on its capabilities.
396 A new command, ``gnt-cluster renew-crypto`` has been added to easily
397 replace the cluster's certificates and crypto keys. This might help in
398 case they have been compromised, or have simply expired.
400 A new disk option for instance creation has been added that allows one
401 to "adopt" currently existing logical volumes, with data
402 preservation. This should allow easier migration to Ganeti from
403 unmanaged (or managed via other software) instances.
405 Another disk improvement is the possibility to convert between redundant
406 (DRBD) and plain (LVM) disk configuration for an instance. This should
407 allow better scalability (starting with one node and growing the
408 cluster, or shrinking a two-node cluster to one node).
410 A new feature that could help with automated node failovers has been
411 implemented: if a node sees itself as offline (by querying the master
412 candidates), it will try to shutdown (hard) all instances and any active
413 DRBD devices. This reduces the risk of duplicate instances if an
414 external script automatically failovers the instances on such nodes. To
415 enable this, the cluster parameter ``maintain_node_health`` should be
416 enabled; in the future this option (per the name) will enable other
417 automatic maintenance features.
419 Instance export/import now will reuse the original instance
420 specifications for all parameters; that means exporting an instance,
421 deleting it and the importing it back should give an almost identical
422 instance. Note that the default import behaviour has changed from
423 before, where it created only one NIC; now it recreates the original
426 Cluster verify has added a few new checks: SSL certificates validity,
427 /etc/hosts consistency across the cluster, etc.
432 As usual, many internal changes were done, documentation fixes,
435 - Fixed cluster initialization with disabled cluster storage (regression
437 - File-based storage supports growing the disks
438 - Fixed behaviour of node role changes
439 - Fixed cluster verify for some corner cases, plus a general rewrite of
440 cluster verify to allow future extension with more checks
441 - Fixed log spamming by watcher and node daemon (regression introduced
443 - Fixed possible validation issues when changing the list of enabled
445 - Fixed cleanup of /etc/hosts during node removal
446 - Fixed RAPI response for invalid methods
447 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
448 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
449 from ide disks, etc.)
450 - Allow OS changes without re-installation (to record a changed OS
451 outside of Ganeti, or to allow OS renames)
452 - Allow instance creation without OS installation (useful for example if
453 the OS will be installed manually, or restored from a backup not in
455 - Implemented option to make cluster ``copyfile`` use the replication
457 - Added list of enabled hypervisors to ssconf (possibly useful for
459 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
461 - A partial form of node re-IP is possible via node readd, which now
462 allows changed node primary IP
463 - Command line utilities now show an informational message if the job is
465 - The logs of the master daemon now show the PID/UID/GID of the
472 *(Released Fri, 12 Mar 2010)*
474 During the 2.1.0 long release candidate cycle, a lot of improvements and
475 changes have accumulated with were released later as 2.1.1.
480 The node evacuate command (``gnt-node evacuate``) was significantly
481 rewritten, and as such the IAllocator protocol was changed - a new
482 request type has been added. This unfortunate change during a stable
483 series is designed to improve performance of node evacuations; on
484 clusters with more than about five nodes and which are well-balanced,
485 evacuation should proceed in parallel for all instances of the node
486 being evacuated. As such, any existing IAllocator scripts need to be
487 updated, otherwise the above command will fail due to the unknown
488 request. The provided "dumb" allocator has not been updated; but the
489 ganeti-htools package supports the new protocol since version 0.2.4.
491 Another important change is increased validation of node and instance
492 names. This might create problems in special cases, if invalid host
493 names are being used.
495 Also, a new layer of hypervisor parameters has been added, that sits at
496 OS level between the cluster defaults and the instance ones. This allows
497 customisation of virtualization parameters depending on the installed
498 OS. For example instances with OS 'X' may have a different KVM kernel
499 (or any other parameter) than the cluster defaults. This is intended to
500 help managing a multiple OSes on the same cluster, without manual
501 modification of each instance's parameters.
503 A tool for merging clusters, ``cluster-merge``, has been added in the
509 - Improved the int/float conversions that should make the code more
510 robust in face of errors from the node daemons
511 - Fixed the remove node code in case of internal configuration errors
512 - Fixed the node daemon behaviour in face of inconsistent queue
513 directory (e.g. read-only file-system where we can't open the files
515 - Fixed the behaviour of gnt-node modify for master candidate demotion;
516 now it either aborts cleanly or, if given the new “auto_promote”
517 parameter, will automatically promote other nodes as needed
518 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
519 completely prevent Ganeti from working
520 - Fixed bug for instance export when not all disks were successfully
522 - Fixed behaviour of node add when the new node is slow in starting up
524 - Fixed handling of signals in the LUXI client, which should improve
525 behaviour of command-line scripts
526 - Added checks for invalid node/instance names in the configuration (now
527 flagged during cluster verify)
528 - Fixed watcher behaviour for disk activation errors
529 - Fixed two potentially endless loops in http library, which led to the
530 RAPI daemon hanging and consuming 100% CPU in some cases
531 - Fixed bug in RAPI daemon related to hashed passwords
532 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
534 - Enhanced compatibility with non-Debian OSes, but not using absolute
535 path in some commands and allowing customisation of the ssh
536 configuration directory
537 - Fixed possible future issue with new Python versions by abiding to the
538 proper use of ``__slots__`` attribute on classes
539 - Added checks that should prevent directory traversal attacks
540 - Many documentation fixes based on feedback from users
545 - Added an “early_release” more for instance replace disks and node
546 evacuate, where we release locks earlier and thus allow higher
547 parallelism within the cluster
548 - Added watcher hooks, intended to allow the watcher to restart other
549 daemons (e.g. from the ganeti-nbma project), but they can be used of
550 course for any other purpose
551 - Added a compile-time disable for DRBD barriers, to increase
552 performance if the administrator trusts the power supply or the
553 storage system to not lose writes
554 - Added the option of using syslog for logging instead of, or in
555 addition to, Ganeti's own log files
556 - Removed boot restriction for paravirtual NICs for KVM, recent versions
557 can indeed boot from a paravirtual NIC
558 - Added a generic debug level for many operations; while this is not
559 used widely yet, it allows one to pass the debug value all the way to
561 - Enhanced the hooks environment for instance moves (failovers,
562 migrations) where the primary/secondary nodes changed during the
563 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
564 - Enhanced data validations for many user-supplied values; one important
565 item is the restrictions imposed on instance and node names, which
566 might reject some (invalid) host names
567 - Add a configure-time option to disable file-based storage, if it's not
568 needed; this allows greater security separation between the master
569 node and the other nodes from the point of view of the inter-node RPC
571 - Added user notification in interactive tools if job is waiting in the
572 job queue or trying to acquire locks
573 - Added log messages when a job is waiting for locks
574 - Added filtering by node tags in instance operations which admit
575 multiple instances (start, stop, reboot, reinstall)
576 - Added a new tool for cluster mergers, ``cluster-merge``
577 - Parameters from command line which are of the form ``a=b,c=d`` can now
578 use backslash escapes to pass in values which contain commas,
579 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
581 - For KVM, the instance name is the first parameter passed to KVM, so
582 that it's more visible in the process list
588 *(Released Tue, 2 Mar 2010)*
590 Ganeti 2.1 brings many improvements with it. Major changes:
592 - Added infrastructure to ease automated disk repairs
593 - Added new daemon to export configuration data in a cheaper way than
595 - Instance NICs can now be routed instead of being associated with a
597 - Improved job locking logic to reduce impact of jobs acquiring multiple
598 locks waiting for other long-running jobs
600 In-depth implementation details can be found in the Ganeti 2.1 design
606 - Added chroot hypervisor
607 - Added more options to xen-hvm hypervisor (``kernel_path`` and
609 - Added more options to xen-pvm hypervisor (``use_bootloader``,
610 ``bootloader_path`` and ``bootloader_args``)
611 - Added the ``use_localtime`` option for the xen-hvm and kvm
612 hypervisors, and the default value for this has changed to false (in
613 2.0 xen-hvm always enabled it)
614 - Added luxi call to submit multiple jobs in one go
615 - Added cluster initialization option to not modify ``/etc/hosts``
617 - Added network interface parameters
618 - Added dry run mode to some LUs
619 - Added RAPI resources:
621 - ``/2/instances/[instance_name]/info``
622 - ``/2/instances/[instance_name]/replace-disks``
623 - ``/2/nodes/[node_name]/evacuate``
624 - ``/2/nodes/[node_name]/migrate``
625 - ``/2/nodes/[node_name]/role``
626 - ``/2/nodes/[node_name]/storage``
627 - ``/2/nodes/[node_name]/storage/modify``
628 - ``/2/nodes/[node_name]/storage/repair``
630 - Added OpCodes to evacuate or migrate all instances on a node
631 - Added new command to list storage elements on nodes (``gnt-node
632 list-storage``) and modify them (``gnt-node modify-storage``)
633 - Added new ssconf files with master candidate IP address
634 (``ssconf_master_candidates_ips``), node primary IP address
635 (``ssconf_node_primary_ips``) and node secondary IP address
636 (``ssconf_node_secondary_ips``)
637 - Added ``ganeti-confd`` and a client library to query the Ganeti
638 configuration via UDP
639 - Added ability to run hooks after cluster initialization and before
641 - Added automatic mode for disk replace (``gnt-instance replace-disks
643 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
644 after catastrophic data-loss
645 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
647 - Added ``gnt-instance move`` command to move instances
648 - Added ``gnt-cluster watcher`` command to control watcher
649 - Added ``gnt-node powercycle`` command to powercycle nodes
650 - Added new job status field ``lock_status``
651 - Added parseable error codes to cluster verification (``gnt-cluster
652 verify --error-codes``) and made output less verbose (use
653 ``--verbose`` to restore previous behaviour)
654 - Added UUIDs to the main config entities (cluster, nodes, instances)
655 - Added support for OS variants
656 - Added support for hashed passwords in the Ganeti remote API users file
658 - Added option to specify maximum timeout on instance shutdown
659 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
660 - Added new helper script to start and stop Ganeti daemons
661 (``daemon-util``), with the intent to reduce the work necessary to
662 adjust Ganeti for non-Debian distributions and to start/stop daemons
664 - Added more unittests
665 - Fixed critical bug in ganeti-masterd startup
666 - Removed the configure-time ``kvm-migration-port`` parameter, this is
667 now customisable at the cluster level for both the KVM and Xen
668 hypervisors using the new ``migration_port`` parameter
669 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
670 reinstalling an instance
671 - Allowed ``@`` in tag names
672 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
673 - Many documentation updates
674 - Distribute hypervisor files on ``gnt-cluster redist-conf``
675 - ``gnt-instance reinstall`` can now reinstall multiple instances
676 - Updated many command line parameters
677 - Introduced new OS API version 15
678 - No longer support a default hypervisor
679 - Treat virtual LVs as inexistent
680 - Improved job locking logic to reduce lock contention
681 - Match instance and node names case insensitively
682 - Reimplemented bash completion script to be more complete
689 *(Released Thu, 4 Feb 2010)*
691 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
692 - Fix a string formatting bug
693 - Improve safety of the code in some error paths
694 - Improve data validation in the master of values returned from nodes
700 *(Released Thu, 17 Dec 2009)*
702 - Fix security issue due to missing validation of iallocator names; this
703 allows local and remote execution of arbitrary executables
704 - Fix failure of gnt-node list during instance removal
705 - Ship the RAPI documentation in the archive
711 *(Released Wed, 30 Sep 2009)*
713 - Fixed many wrong messages
714 - Fixed a few bugs related to the locking library
715 - Fixed MAC checking at instance creation time
716 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
717 - Fixed a few issues related to signal handling in both daemons and
719 - Fixed the example startup script provided
720 - Fixed insserv dependencies in the example startup script (patch from
722 - Fixed handling of drained nodes in the iallocator framework
723 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
725 - Fixed error related to invalid job IDs in job polling
726 - Fixed job/opcode persistence on unclean master shutdown
727 - Fixed handling of partial job processing after unclean master
729 - Fixed error reporting from LUs, previously all errors were converted
730 into execution errors
731 - Fixed error reporting from burnin
732 - Decreased significantly the memory usage of the job queue
733 - Optimised slightly multi-job submission
734 - Optimised slightly opcode loading
735 - Backported the multi-job submit framework from the development
736 branch; multi-instance start and stop should be faster
737 - Added script to clean archived jobs after 21 days; this will reduce
738 the size of the queue directory
739 - Added some extra checks in disk size tracking
740 - Added an example ethers hook script
741 - Added a cluster parameter that prevents Ganeti from modifying of
743 - Added more node information to RAPI responses
744 - Added a ``gnt-job watch`` command that allows following the ouput of a
746 - Added a bind-address option to ganeti-rapi
747 - Added more checks to the configuration verify
748 - Enhanced the burnin script such that some operations can be retried
750 - Converted instance reinstall to multi-instance model
756 *(Released Fri, 7 Aug 2009)*
758 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
759 to allow using the pre-2.0.2 behaviour in activation, if any existing
760 instances have mismatched disk sizes in the configuration
761 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
762 any configuration mismatches for disk sizes
763 - Added ``gnt-master cluste-failover --no-voting`` to allow master
764 failover to work on two-node clusters
765 - Fixed the ``--net`` option of ``gnt-backup import``, which was
767 - Fixed detection of OS script errors in ``gnt-backup export``
768 - Fixed exit code of ``gnt-backup export``
774 *(Released Fri, 17 Jul 2009)*
776 - Added experimental support for stripped logical volumes; this should
777 enhance performance but comes with a higher complexity in the block
778 device handling; stripping is only enabled when passing
779 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
780 affected even in the non-stripped mode
781 - Improved resiliency against transient failures at the end of DRBD
782 resyncs, and in general of DRBD resync checks
783 - Fixed a couple of issues with exports and snapshot errors
784 - Fixed a couple of issues in instance listing
785 - Added display of the disk size in ``gnt-instance info``
786 - Fixed checking for valid OSes in instance creation
787 - Fixed handling of the "vcpus" parameter in instance listing and in
788 general of invalid parameters
789 - Fixed http server library, and thus RAPI, to handle invalid
790 username/password combinations correctly; this means that now they
791 report unauthorized for queries too, not only for modifications,
792 allowing earlier detect of configuration problems
793 - Added a new "role" node list field, equivalent to the master/master
794 candidate/drained/offline flags combinations
795 - Fixed cluster modify and changes of candidate pool size
796 - Fixed cluster verify error messages for wrong files on regular nodes
797 - Fixed a couple of issues with node demotion from master candidate role
798 - Fixed node readd issues
799 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
800 - Added a new ``--no-voting`` option for masterfailover to fix failover
801 on two-nodes clusters when the former master node is unreachable
802 - Added instance reinstall over RAPI
808 *(Released Tue, 16 Jun 2009)*
810 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
811 allow re-adding the start in single-user option (regression from 1.2)
812 - the watcher writes the instance status to a file, to allow monitoring
813 to report the instance status (from the master) based on cached
814 results of the watcher's queries; while this can get stale if the
815 watcher is being locked due to other work on the cluster, this is
817 - the watcher now also restarts the node daemon and the rapi daemon if
819 - fixed the watcher to handle full and drained queue cases
820 - hooks export more instance data in the environment, which helps if
821 hook scripts need to take action based on the instance's properties
822 (no longer need to query back into ganeti)
823 - instance failovers when the instance is stopped do not check for free
824 RAM, so that failing over a stopped instance is possible in low memory
826 - rapi uses queries for tags instead of jobs (for less job traffic), and
827 for cluster tags it won't talk to masterd at all but read them from
829 - a couple of error handling fixes in RAPI
830 - drbd handling: improved the error handling of inconsistent disks after
831 resync to reduce the frequency of "there are some degraded disks for
832 this instance" messages
833 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
834 error handling path called a wrong function name)
840 *(Released Wed, 27 May 2009)*
842 - no changes from rc5
845 Version 2.0 release candidate 5
846 -------------------------------
848 *(Released Wed, 20 May 2009)*
850 - fix a couple of bugs (validation, argument checks)
851 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
852 - some small improvements to RAPI and IAllocator
853 - make watcher automatically start the master daemon if down
856 Version 2.0 release candidate 4
857 -------------------------------
859 *(Released Mon, 27 Apr 2009)*
861 - change the OS list to not require locks; this helps with big clusters
862 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
863 volume group is broken
864 - ``gnt-instance info``, without any arguments, doesn't run for all
865 instances anymore; either pass ``--all`` or pass the desired
866 instances; this helps against mistakes on big clusters where listing
867 the information for all instances takes a long time
868 - miscellaneous doc and man pages fixes
871 Version 2.0 release candidate 3
872 -------------------------------
874 *(Released Wed, 8 Apr 2009)*
876 - Change the internal locking model of some ``gnt-node`` commands, in
877 order to reduce contention (and blocking of master daemon) when
878 batching many creation/reinstall jobs
879 - Fixes to Xen soft reboot
880 - No longer build documentation at build time, instead distribute it in
881 the archive, in order to reduce the need for the whole docbook/rst
885 Version 2.0 release candidate 2
886 -------------------------------
888 *(Released Fri, 27 Mar 2009)*
890 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
891 - Fix watcher startup sequence, improves the behaviour of busy clusters
892 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
893 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
894 bind address and other places
895 - Some documentation fixes and updates
898 Version 2.0 release candidate 1
899 -------------------------------
901 *(Released Mon, 2 Mar 2009)*
903 - More documentation updates, now all docs should be more-or-less
905 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
907 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
909 - Fix an issue related to $libdir/run/ganeti and cluster creation
915 *(Released Thu, 19 Feb 2009)*
917 - Xen PVM and KVM have switched the default value for the instance root
918 disk to the first partition on the first drive, instead of the whole
919 drive; this means that the OS installation scripts must be changed
921 - Man pages have been updated
922 - RAPI has been switched by default to HTTPS, and the exported functions
923 should all work correctly
924 - RAPI v1 has been removed
925 - Many improvements to the KVM hypervisor
926 - Block device errors are now better reported
927 - Many other bugfixes and small improvements
933 *(Released Mon, 26 Jan 2009)*
935 - Version 2 is a general rewrite of the code and therefore the
936 differences are too many to list, see the design document for 2.0 in
937 the ``doc/`` subdirectory for more details
938 - In this beta version there is not yet a migration path from 1.2 (there
939 will be one in the final 2.0 release)
940 - A few significant changes are:
942 - all commands are executed by a daemon (``ganeti-masterd``) and the
943 various ``gnt-*`` commands are just front-ends to it
944 - all the commands are entered into, and executed from a job queue,
945 see the ``gnt-job(8)`` manpage
946 - the RAPI daemon supports read-write operations, secured by basic
947 HTTP authentication on top of HTTPS
948 - DRBD version 0.7 support has been removed, DRBD 8 is the only
949 supported version (when migrating from Ganeti 1.2 to 2.0, you need
950 to migrate to DRBD 8 first while still running Ganeti 1.2)
951 - DRBD devices are using statically allocated minor numbers, which
952 will be assigned to existing instances during the migration process
953 - there is support for both Xen PVM and Xen HVM instances running on
955 - KVM virtualization is supported too
956 - file-based storage has been implemented, which means that it is
957 possible to run the cluster without LVM and DRBD storage, for
958 example using a shared filesystem exported from shared storage (and
959 still have live migration)
965 *(Released Tue, 13 Jan 2009)*
967 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
968 - Reuse the old instance mac address by default on instance import, if
969 the instance name is the same.
970 - Handle situations in which the node info rpc returns incomplete
972 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
973 - Improved version of batcher:
976 - instance mac address support
977 - support for HVM clusters/instances
979 - Add an option to show the number of cpu sockets and nodes in
981 - Support OSes that handle more than one version of the OS api (but do
982 not change the current API in any other way)
983 - Fix ``gnt-node migrate``
984 - ``gnt-debug`` man page
985 - Fixes various more typos and small issues
986 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
992 *(Released Wed, 24 Sep 2008)*
994 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
995 type of disk exported to fully virtualized instances.
996 - provide access to the serial console of HVM instances
997 - instance auto_balance flag, set by default. If turned off it will
998 avoid warnings on cluster verify if there is not enough memory to fail
999 over an instance. in the future it will prevent automatically failing
1000 it over when we will support that.
1001 - batcher tool for instance creation, see ``tools/README.batcher``
1002 - ``gnt-instance reinstall --select-os`` to interactively select a new
1003 operating system when reinstalling an instance.
1004 - when changing the memory amount on instance modify a check has been
1005 added that the instance will be able to start. also warnings are
1006 emitted if the instance will not be able to fail over, if auto_balance
1008 - documentation fixes
1009 - sync fields between ``gnt-instance list/modify/add/import``
1010 - fix a race condition in drbd when the sync speed was set after giving
1011 the device a remote peer.
1017 *(Released Tue, 22 Jul 2008)*
1019 - note: the allowed size and number of tags per object were reduced
1020 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1021 - fixed twisted 8.x compatibility
1022 - fixed ``gnt-instance replace-disks`` with iallocator
1023 - add TCP keepalives on twisted connections to detect restarted nodes
1024 - disk increase support, see ``gnt-instance grow-disk``
1025 - implement bulk node/instance query for RAPI
1026 - add tags in node/instance listing (optional)
1027 - experimental migration (and live migration) support, read the man page
1028 for ``gnt-instance migrate``
1029 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1030 has some small improvements in handling its state file
1036 *(Released Fri, 13 Jun 2008)*
1038 - Experimental readonly, REST-based remote API implementation;
1039 automatically started on master node, TCP port 5080, if enabled by
1040 ``--enable-rapi`` parameter to configure script.
1041 - Instance allocator support. Add and import instance accept a
1042 ``--iallocator`` parameter, and call that instance allocator to decide
1043 which node to use for the instance. The iallocator document describes
1044 what's expected from an allocator script.
1045 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1046 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1047 node is lost there is still enough memory to fail over the instances
1049 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1050 ``gnt-cluster verify``, to check for site-specific compliance. All the
1051 hooks will run, and their output, if any, will be displayed. Any
1052 failing hook will make the verification return an error value.
1053 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1054 primary and secondary interfaces
1055 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1056 that is still declared as part of the cluster and has failed.
1057 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1058 specifying output fields, that just adds the chosen fields to the
1060 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1061 export from the filesystem.
1062 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1063 have been added. Using them you can enable/disable acpi and pae
1064 support, and specify a path for a cd image to be exported to the
1065 instance. These parameters as the name suggest only work on HVM
1067 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1068 PAE support will be set to the previously hardcoded values, but the
1069 (previously hardcoded) path to the CDROM ISO image will be unset and
1070 if required, needs to be set manually with ``gnt-instance modify``
1072 - The address to which an instance's VNC console is bound is now
1073 selectable per-instance, rather than being cluster wide. Of course
1074 this only applies to instances controlled via VNC, so currently just
1075 applies to HVM clusters.
1081 *(Released Mon, 18 Feb 2008)*
1083 - more tweaks to the disk activation code (especially helpful for DRBD)
1084 - change the default ``gnt-instance list`` output format, now there is
1085 one combined status field (see the manpage for the exact values this
1087 - some more fixes for the mac export to hooks change
1088 - make Ganeti not break with DRBD 8.2.x (which changed the version
1089 format in ``/proc/drbd``) (issue 24)
1090 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1091 template, allowing migration from DRBD0.7+MD to DRBD8
1097 *(Released Wed, 30 Jan 2008)*
1099 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1101 - add command aliases infrastructure and a few aliases
1102 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1103 man pages and the ``--help`` option of ``gnt-node
1104 list``/``gnt-instance list``
1105 - fix ``gnt-backup list`` with down nodes (issue 21)
1106 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1107 - fix the dist archive and add a check for including svn/git files in
1109 - some developer-related changes: improve the burnin and the QA suite,
1110 add an upload script for testing during development
1116 *(Released Wed, 16 Jan 2008)*
1118 - experimental HVM support, read the install document, section
1119 "Initializing the cluster"
1120 - allow for the PVM hypervisor per-instance kernel and initrd paths
1121 - add a new command ``gnt-cluster verify-disks`` which uses a new
1122 algorithm to improve the reconnection of the DRBD pairs if the device
1123 on the secondary node has gone away
1124 - make logical volume code auto-activate LVs at disk activation time
1125 - slightly improve the speed of activating disks
1126 - allow specification of the MAC address at instance creation time, and
1127 changing it later via ``gnt-instance modify``
1128 - fix handling of external commands that generate lots of output on
1130 - update documentation with regard to minimum version of DRBD8 supported
1136 *(Released Tue, 4 Dec 2007)*
1138 - Log the ``xm create`` output to the node daemon log on failure (to
1139 help diagnosing the error)
1140 - In debug mode, log all external commands output if failed to the logs
1141 - Change parsing of lvm commands to ignore stderr
1147 *(Released Wed, 28 Nov 2007)*
1149 - Another round of updates to the DRBD 8 code to deal with more failures
1150 in the replace secondary node operation
1151 - Some more logging of failures in disk operations (lvm, drbd)
1152 - A few documentation updates
1159 *(Released Tue, 13 Nov 2007)*
1161 - Change configuration file format from Python's Pickle to JSON.
1162 Upgrading is possible using the cfgupgrade utility.
1163 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1164 faster replace disks and is more stable (DRBD 8 has many improvements
1165 compared to DRBD 0.7)
1166 - Added command line tags support (see man pages for ``gnt-instance``,
1167 ``gnt-node``, ``gnt-cluster``)
1168 - Added instance rename support
1169 - Added multi-instance startup/shutdown
1170 - Added cluster rename support
1171 - Added ``gnt-node evacuate`` to simplify some node operations
1172 - Added instance reboot operation that can speedup reboot as compared to
1174 - Soften the requirement that hostnames are in FQDN format
1175 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1177 - Removed dependency on debian's patched fping that uses the
1178 non-standard ``-S`` option
1179 - Now the OS definitions are searched for in multiple, configurable
1180 paths (easier for distros to package)
1181 - Some changes to the hooks infrastructure (especially the new
1182 post-configuration update hook)
1183 - Other small bugfixes
1185 .. vim: set textwidth=72 :