8 Ganeti 2.1 brings many improvements with it. Major changes:
10 - Added infrastructure to ease automated disk repairs
11 - Added new daemon to export configuration data in a cheaper way than
13 - Instance NICs can now be routed instead of being associated with a
15 - Improved job locking logic to reduce impact of jobs acquiring multiple
16 locks waiting for other long-running jobs
18 Detailed implementation details can be found in the Ganeti 2.1 design
24 - Added chroot hypervisor
25 - Added more options to xen-hvm hypervisor (``kernel_path`` and
27 - Added more options to xen-pvm hypervisor (``use_bootloader``,
28 ``bootloader_path`` and ``bootloader_args``)
29 - Added luxi call to submit multiple jobs in one go
30 - Added cluster initialization time option to not modify ``/etc/hosts``
32 - Added network interface parameters
33 - Added dry run mode to some LUs
34 - Added RAPI resources:
36 - ``/2/instances/[instance_name]/info``
37 - ``/2/instances/[instance_name]/replace-disks``
38 - ``/2/nodes/[node_name]/evacuate``
39 - ``/2/nodes/[node_name]/migrate``
40 - ``/2/nodes/[node_name]/role``
41 - ``/2/nodes/[node_name]/storage``
42 - ``/2/nodes/[node_name]/storage/modify``
43 - ``/2/nodes/[node_name]/storage/repair``
45 - Added OpCodes to evacuate or migrate all instances on a node
46 - Added new command to list storage elements on nodes (``gnt-node
47 list-storage``) and modify them (``gnt-node modify-storage``)
48 - Added new ssconf files with master candidate IP address
49 (``ssconf_master_candidates_ips``), node primary IP address
50 (``ssconf_node_primary_ips``) and node secondary IP address
51 (``ssconf_node_secondary_ips``)
52 - Added ``ganeti-confd`` and a client library to query the Ganeti
54 - Added ability to run hooks after cluster initialization and before
56 - Added automatic mode for disk replace (``gnt-instance replace-disks
58 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
59 after catastrophic data-loss
60 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
62 - Added ``gnt-instance move`` command to move instances
63 - Added ``gnt-cluster watcher`` command to control watcher
64 - Added ``gnt-node powercycle`` command to powercycle nodes
65 - Added new job status field ``lock_status``
66 - Added parseable error codes to cluster verification (``gnt-cluster
67 verify --error-codes``) and made output less verbose (use
68 ``--verbose`` to restore previous behaviour)
69 - Added UUIDs to the main config entities (cluster, nodes, instances)
70 - Added support for OS variants
71 - Added support for hashed passwords in the Ganeti remote API users file
73 - Added option to specify maximum timeout on instance shutdown
74 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
75 - Added more unittests
76 - Fixed critical bug in ganeti-masterd startup
77 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
78 reinstalling an instance
79 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
80 - Many documentation updates
81 - Distribute hypervisor files on ``gnt-cluster redist-conf``
82 - ``gnt-instance reinstall`` can now reinstall multiple instances
83 - Updated many command line parameters
84 - Introduced new OS API version 15
85 - No longer support a default hypervisor
86 - Treat virtual LVs as inexistent
87 - Improved job locking logic to reduce lock contention
88 - Match instance and node names case insensitively
89 - Reimplemented bash completion script to be more complete
96 - Fixed many wrong messages
97 - Fixed a few bugs related to the locking library
98 - Fixed MAC checking at instance creation time
99 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
100 - Fixed a few issues related to signal handling in both daemons and
102 - Fixed the example startup script provided
103 - Fixed insserv dependencies in the example startup script (patch from
105 - Fixed handling of drained nodes in the iallocator framework
106 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
108 - Fixed error related to invalid job IDs in job polling
109 - Fixed job/opcode persistence on unclean master shutdown
110 - Fixed handling of partial job processing after unclean master
112 - Fixed error reporting from LUs, previously all errors were converted
113 into execution errors
114 - Fixed error reporting from burnin
115 - Decreased significantly the memory usage of the job queue
116 - Optimised slightly multi-job submission
117 - Optimised slightly opcode loading
118 - Backported the multi-job submit framework from the development
119 branch; multi-instance start and stop should be faster
120 - Added script to clean archived jobs after 21 days; this will reduce
121 the size of the queue directory
122 - Added some extra checks in disk size tracking
123 - Added an example ethers hook script
124 - Added a cluster parameter that prevents Ganeti from modifying of
126 - Added more node information to RAPI responses
127 - Added a ``gnt-job watch`` command that allows following the ouput of a
129 - Added a bind-address option to ganeti-rapi
130 - Added more checks to the configuration verify
131 - Enhanced the burnin script such that some operations can be retried
133 - Converted instance reinstall to multi-instance model
139 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
140 to allow using the pre-2.0.2 behaviour in activation, if any existing
141 instances have mismatched disk sizes in the configuration
142 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
143 any configuration mismatches for disk sizes
144 - Added ``gnt-master cluste-failover --no-voting`` to allow master
145 failover to work on two-node clusters
146 - Fixed the ``--net`` option of ``gnt-backup import``, which was
148 - Fixed detection of OS script errors in ``gnt-backup export``
149 - Fixed exit code of ``gnt-backup export``
155 - Added experimental support for stripped logical volumes; this should
156 enhance performance but comes with a higher complexity in the block
157 device handling; stripping is only enabled when passing
158 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
159 affected even in the non-stripped mode
160 - Improved resiliency against transient failures at the end of DRBD
161 resyncs, and in general of DRBD resync checks
162 - Fixed a couple of issues with exports and snapshot errors
163 - Fixed a couple of issues in instance listing
164 - Added display of the disk size in ``gnt-instance info``
165 - Fixed checking for valid OSes in instance creation
166 - Fixed handling of the "vcpus" parameter in instance listing and in
167 general of invalid parameters
168 - Fixed http server library, and thus RAPI, to handle invalid
169 username/password combinations correctly; this means that now they
170 report unauthorized for queries too, not only for modifications,
171 allowing earlier detect of configuration problems
172 - Added a new "role" node list field, equivalent to the master/master
173 candidate/drained/offline flags combinations
174 - Fixed cluster modify and changes of candidate pool size
175 - Fixed cluster verify error messages for wrong files on regular nodes
176 - Fixed a couple of issues with node demotion from master candidate role
177 - Fixed node readd issues
178 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
179 - Added a new ``--no-voting`` option for masterfailover to fix failover
180 on two-nodes clusters when the former master node is unreachable
181 - Added instance reinstall over RAPI
187 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
188 allow re-adding the start in single-user option (regression from 1.2)
189 - the watcher writes the instance status to a file, to allow monitoring
190 to report the instance status (from the master) based on cached
191 results of the watcher's queries; while this can get stale if the
192 watcher is being locked due to other work on the cluster, this is
194 - the watcher now also restarts the node daemon and the rapi daemon if
196 - fixed the watcher to handle full and drained queue cases
197 - hooks export more instance data in the environment, which helps if
198 hook scripts need to take action based on the instance's properties
199 (no longer need to query back into ganeti)
200 - instance failovers when the instance is stopped do not check for free
201 RAM, so that failing over a stopped instance is possible in low memory
203 - rapi uses queries for tags instead of jobs (for less job traffic), and
204 for cluster tags it won't talk to masterd at all but read them from
206 - a couple of error handling fixes in RAPI
207 - drbd handling: improved the error handling of inconsistent disks after
208 resync to reduce the frequency of "there are some degraded disks for
209 this instance" messages
210 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
211 error handling path called a wrong function name)
217 - no changes from rc5
220 Version 2.0 release candidate 5
221 -------------------------------
223 - fix a couple of bugs (validation, argument checks)
224 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
225 - some small improvements to RAPI and IAllocator
226 - make watcher automatically start the master daemon if down
229 Version 2.0 release candidate 4
230 -------------------------------
232 - change the OS list to not require locks; this helps with big clusters
233 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
234 volume group is broken
235 - ``gnt-instance info``, without any arguments, doesn't run for all
236 instances anymore; either pass ``--all`` or pass the desired
237 instances; this helps against mistakes on big clusters where listing
238 the information for all instances takes a long time
239 - miscellaneous doc and man pages fixes
242 Version 2.0 release candidate 3
243 -------------------------------
245 - Change the internal locking model of some ``gnt-node`` commands, in
246 order to reduce contention (and blocking of master daemon) when
247 batching many creation/reinstall jobs
248 - Fixes to Xen soft reboot
249 - No longer build documentation at build time, instead distribute it in
250 the archive, in order to reduce the need for the whole docbook/rst
254 Version 2.0 release candidate 2
255 -------------------------------
257 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
258 - Fix watcher startup sequence, improves the behaviour of busy clusters
259 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
260 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
261 bind address and other places
262 - Some documentation fixes and updates
265 Version 2.0 release candidate 1
266 -------------------------------
268 - More documentation updates, now all docs should be more-or-less
270 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
272 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
274 - Fix an issue related to $libdir/run/ganeti and cluster creation
280 - Xen PVM and KVM have switched the default value for the instance root
281 disk to the first partition on the first drive, instead of the whole
282 drive; this means that the OS installation scripts must be changed
284 - Man pages have been updated
285 - RAPI has been switched by default to HTTPS, and the exported functions
286 should all work correctly
287 - RAPI v1 has been removed
288 - Many improvements to the KVM hypervisor
289 - Block device errors are now better reported
290 - Many other bugfixes and small improvements
296 - Version 2 is a general rewrite of the code and therefore the
297 differences are too many to list, see the design document for 2.0 in
298 the ``doc/`` subdirectory for more details
299 - In this beta version there is not yet a migration path from 1.2 (there
300 will be one in the final 2.0 release)
301 - A few significant changes are:
303 - all commands are executed by a daemon (``ganeti-masterd``) and the
304 various ``gnt-*`` commands are just front-ends to it
305 - all the commands are entered into, and executed from a job queue,
306 see the ``gnt-job(8)`` manpage
307 - the RAPI daemon supports read-write operations, secured by basic
308 HTTP authentication on top of HTTPS
309 - DRBD version 0.7 support has been removed, DRBD 8 is the only
310 supported version (when migrating from Ganeti 1.2 to 2.0, you need
311 to migrate to DRBD 8 first while still running Ganeti 1.2)
312 - DRBD devices are using statically allocated minor numbers, which
313 will be assigned to existing instances during the migration process
314 - there is support for both Xen PVM and Xen HVM instances running on
316 - KVM virtualization is supported too
317 - file-based storage has been implemented, which means that it is
318 possible to run the cluster without LVM and DRBD storage, for
319 example using a shared filesystem exported from shared storage (and
320 still have live migration)
326 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
327 - Reuse the old instance mac address by default on instance import, if
328 the instance name is the same.
329 - Handle situations in which the node info rpc returns incomplete
331 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
332 - Improved version of batcher:
335 - instance mac address support
336 - support for HVM clusters/instances
338 - Add an option to show the number of cpu sockets and nodes in
340 - Support OSes that handle more than one version of the OS api (but do
341 not change the current API in any other way)
342 - Fix ``gnt-node migrate``
343 - ``gnt-debug`` man page
344 - Fixes various more typos and small issues
345 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
351 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
352 type of disk exported to fully virtualized instances.
353 - provide access to the serial console of HVM instances
354 - instance auto_balance flag, set by default. If turned off it will
355 avoid warnings on cluster verify if there is not enough memory to fail
356 over an instance. in the future it will prevent automatically failing
357 it over when we will support that.
358 - batcher tool for instance creation, see ``tools/README.batcher``
359 - ``gnt-instance reinstall --select-os`` to interactively select a new
360 operating system when reinstalling an instance.
361 - when changing the memory amount on instance modify a check has been
362 added that the instance will be able to start. also warnings are
363 emitted if the instance will not be able to fail over, if auto_balance
365 - documentation fixes
366 - sync fields between ``gnt-instance list/modify/add/import``
367 - fix a race condition in drbd when the sync speed was set after giving
368 the device a remote peer.
374 - note: the allowed size and number of tags per object were reduced
375 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
376 - fixed twisted 8.x compatibility
377 - fixed ``gnt-instance replace-disks`` with iallocator
378 - add TCP keepalives on twisted connections to detect restarted nodes
379 - disk increase support, see ``gnt-instance grow-disk``
380 - implement bulk node/instance query for RAPI
381 - add tags in node/instance listing (optional)
382 - experimental migration (and live migration) support, read the man page
383 for ``gnt-instance migrate``
384 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
385 has some small improvements in handling its state file
391 - Experimental readonly, REST-based remote API implementation;
392 automatically started on master node, TCP port 5080, if enabled by
393 ``--enable-rapi`` parameter to configure script.
394 - Instance allocator support. Add and import instance accept a
395 ``--iallocator`` parameter, and call that instance allocator to decide
396 which node to use for the instance. The iallocator document describes
397 what's expected from an allocator script.
398 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
399 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
400 node is lost there is still enough memory to fail over the instances
402 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
403 ``gnt-cluster verify``, to check for site-specific compliance. All the
404 hooks will run, and their output, if any, will be displayed. Any
405 failing hook will make the verification return an error value.
406 - ``gnt-cluster verify`` now checks that its peers are reachable on the
407 primary and secondary interfaces
408 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
409 that is still declared as part of the cluster and has failed.
410 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
411 specifying output fields, that just adds the chosen fields to the
413 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
414 export from the filesystem.
415 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
416 have been added. Using them you can enable/disable acpi and pae
417 support, and specify a path for a cd image to be exported to the
418 instance. These parameters as the name suggest only work on HVM
420 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
421 PAE support will be set to the previously hardcoded values, but the
422 (previously hardcoded) path to the CDROM ISO image will be unset and
423 if required, needs to be set manually with ``gnt-instance modify``
425 - The address to which an instance's VNC console is bound is now
426 selectable per-instance, rather than being cluster wide. Of course
427 this only applies to instances controlled via VNC, so currently just
428 applies to HVM clusters.
434 - more tweaks to the disk activation code (especially helpful for DRBD)
435 - change the default ``gnt-instance list`` output format, now there is
436 one combined status field (see the manpage for the exact values this
438 - some more fixes for the mac export to hooks change
439 - make Ganeti not break with DRBD 8.2.x (which changed the version
440 format in ``/proc/drbd``) (issue 24)
441 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
442 template, allowing migration from DRBD0.7+MD to DRBD8
448 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
450 - add command aliases infrastructure and a few aliases
451 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
452 man pages and the ``--help`` option of ``gnt-node
453 list``/``gnt-instance list``
454 - fix ``gnt-backup list`` with down nodes (issue 21)
455 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
456 - fix the dist archive and add a check for including svn/git files in
458 - some developer-related changes: improve the burnin and the QA suite,
459 add an upload script for testing during development
465 - experimental HVM support, read the install document, section
466 "Initializing the cluster"
467 - allow for the PVM hypervisor per-instance kernel and initrd paths
468 - add a new command ``gnt-cluster verify-disks`` which uses a new
469 algorithm to improve the reconnection of the DRBD pairs if the device
470 on the secondary node has gone away
471 - make logical volume code auto-activate LVs at disk activation time
472 - slightly improve the speed of activating disks
473 - allow specification of the MAC address at instance creation time, and
474 changing it later via ``gnt-instance modify``
475 - fix handling of external commands that generate lots of output on
477 - update documentation with regard to minimum version of DRBD8 supported
483 - Log the ``xm create`` output to the node daemon log on failure (to
484 help diagnosing the error)
485 - In debug mode, log all external commands output if failed to the logs
486 - Change parsing of lvm commands to ignore stderr
492 - Another round of updates to the DRBD 8 code to deal with more failures
493 in the replace secondary node operation
494 - Some more logging of failures in disk operations (lvm, drbd)
495 - A few documentation updates
502 - Change configuration file format from Python's Pickle to JSON.
503 Upgrading is possible using the cfgupgrade utility.
504 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
505 faster replace disks and is more stable (DRBD 8 has many improvements
506 compared to DRBD 0.7)
507 - Added command line tags support (see man pages for ``gnt-instance``,
508 ``gnt-node``, ``gnt-cluster``)
509 - Added instance rename support
510 - Added multi-instance startup/shutdown
511 - Added cluster rename support
512 - Added ``gnt-node evacuate`` to simplify some node operations
513 - Added instance reboot operation that can speedup reboot as compared to
515 - Soften the requirement that hostnames are in FQDN format
516 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
518 - Removed dependency on debian's patched fping that uses the
519 non-standard ``-S`` option
520 - Now the OS definitions are searched for in multiple, configurable
521 paths (easier for distros to package)
522 - Some changes to the hooks infrastructure (especially the new
523 post-configuration update hook)
524 - Other small bugfixes
526 .. vim: set textwidth=72 :