Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 390fa990

History | View | Annotate | Download (111.5 kB)

1
News
2
====
3

    
4

    
5
Version 2.8.3
6
-------------
7

    
8
*(Released Thu, 12 Dec 2013)*
9

    
10
- Fixed Luxi daemon socket permissions after master-failover
11
- Improve IP version detection code directly checking for colons rather than
12
  passing the family from the cluster object
13
- Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
14
  opportunistically anymore (Issue 622)
15
- Allow link local IPv6 gateways (Issue 624)
16
- Fix error printing (Issue 616)
17
- Fix a bug in InstanceSetParams concerning names: in case no name is passed in
18
  disk modifications, keep the old one. If name=none then set disk name to
19
  None.
20
- Update build_chroot script to work with the latest hackage packages
21
- Add a packet number limit to "fping" in master-ip-setup (Issue 630)
22
- Fix evacuation out of drained node (Issue 615)
23
- Add default file_driver if missing (Issue 571)
24
- Fix job error message after unclean master shutdown (Issue 618)
25
- Lock group(s) when creating instances (Issue 621)
26
- SetDiskID() before accepting an instance (Issue 633)
27
- Allow the ext template disks to receive arbitrary parameters, both at creation
28
  time and while being modified
29
- Xen handle domain shutdown (future proofing cherry-pick)
30
- Refactor reading live data in htools (future proofing cherry-pick)
31

    
32

    
33
Version 2.8.2
34
-------------
35

    
36
*(Released Thu, 07 Nov 2013)*
37

    
38
- DRBD: ensure peers are UpToDate for dual-primary
39
- Improve error message for replace-disks
40
- More dependency checks at configure time
41
- Placate warnings on ganeti.outils_unittest.py
42

    
43

    
44
Version 2.8.1
45
-------------
46

    
47
*(Released Thu, 17 Oct 2013)*
48

    
49
- Correctly start/stop luxid during gnt-cluster master-failover
50
- Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
51
- Fix path for the job queue serial file
52
- Improved harep man page
53
- Minor documentation improvements
54

    
55

    
56
Version 2.8.0
57
-------------
58

    
59
*(Released Mon, 30 Sep 2013)*
60

    
61
Incompatible/important changes
62
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
63

    
64
- Instance policy can contain multiple instance specs, as described in
65
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
66
  <design-partitioned>`. As a consequence, it's not possible to partially change
67
  or override instance specs. Bounding specs (min and max) can be specified as a
68
  whole using the new option ``--ipolicy-bounds-specs``, while standard
69
  specs use the new option ``--ipolicy-std-specs``.
70
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
71
  gnt-instance is a valid YAML object.
72
- hail now honors network restrictions when allocating nodes. This led to an
73
  update of the IAllocator protocol. See the IAllocator documentation for
74
  details.
75
- confd now only answers static configuration request over the network. luxid
76
  was extracted, listens on the local LUXI socket and responds to live queries.
77
  This allows finer grained permissions if using separate users.
78

    
79
New features
80
~~~~~~~~~~~~
81

    
82
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
83
  to always require authentication, ``--require-authentication``. It can
84
  be specified in ``$sysconfdir/default/ganeti``.
85
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
86
  be used to manage the disk templates to be used by instances in the cluster.
87
  Initially, it will be set to a list that includes plain, drbd, if they were
88
  enabled by specifying a volume group name, and file and sharedfile, if those
89
  were enabled at configure time. Additionally, it will include all disk
90
  templates that are currently used by instances. The order of disk templates
91
  will be based on Ganeti's history of supporting them. In the future, the
92
  first entry of the list will be used as a default disk template on instance
93
  creation.
94
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
95
  configuration back to the previous stable version.
96
- Disk templates in group ipolicy can be restored to the default value.
97
- Initial support for diskless instances and virtual clusters in QA.
98
- More QA and unit tests for instance policies.
99
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
100
  describing why the opcode itself was executed.
101
- The monitoring daemon is now available. It allows users to query the cluster
102
  for obtaining information about the status of the system. The daemon is only
103
  responsible for providing the information over the network: the actual data
104
  gathering is performed by data collectors (currently, only the DRBD status
105
  collector is available).
106
- In order to help developers work on Ganeti, a new script
107
  (``devel/build_chroot``) is provided, for building a chroot that contains all
108
  the required development libraries and tools for compiling Ganeti on a Debian
109
  Squeeze system.
110
- A new tool, ``harep``, for performing self-repair and recreation of instances
111
  in Ganeti has been added.
112
- Split queries are enabled for tags, network, exports, cluster info, groups,
113
  jobs, nodes.
114
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
115
  It prints the command line to set the current policies, to ease
116
  changing them.
117
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
118
  devices for KVM virtio-net interfaces will get created with VNET_HDR
119
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
120
  virtio-net interfaces, which prevents host kernel tainting and log
121
  flooding, when dealing with broken or malicious virtio-net drivers.
122
  It's set to true by default.
123
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
124
  failures.
125
- Support 'viridian' parameter in Xen HVM
126
- Support DSA SSH keys in bootstrap
127
- To simplify the work of packaging frameworks that want to add the needed users
128
  and groups in a split-user setup themselves, at build time three files in
129
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
130
  the groups to be generated, the ``users`` file contains, one per line, the
131
  users to be generated, optionally followed by their primary group, where
132
  important. The ``groupmemberships`` file contains, one per line, additional
133
  user-group membership relations that need to be established. The syntax of
134
  these files will remain stable in all future versions.
135

    
136

    
137
New dependencies
138
~~~~~~~~~~~~~~~~
139
The following new dependencies have been added:
140

    
141
For Haskell:
142
- The ``curl`` library is not optional anymore for compiling the Haskell code.
143
- ``snap-server`` library (if monitoring is enabled).
144

    
145
For Python:
146
- The minimum Python version needed to run Ganeti is now 2.6.
147
- ``yaml`` library (only for running the QA).
148

    
149
Since 2.8.0 rc3
150
~~~~~~~~~~~~~~~
151
- Perform proper cleanup on termination of Haskell daemons
152
- Fix corner-case in handling of remaining retry time
153

    
154

    
155
Version 2.8.0 rc3
156
-----------------
157

    
158
*(Released Tue, 17 Sep 2013)*
159

    
160
- To simplify the work of packaging frameworks that want to add the needed users
161
  and groups in a split-user setup themselves, at build time three files in
162
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
163
  the groups to be generated, the ``users`` file contains, one per line, the
164
  users to be generated, optionally followed by their primary group, where
165
  important. The ``groupmemberships`` file contains, one per line, additional
166
  user-group membership relations that need to be established. The syntax of
167
  these files will remain stable in all future versions.
168
- Add a default to file-driver when unspecified over RAPI (Issue 571)
169
- Mark the DSA host pubkey as optional, and remove it during config downgrade
170
  (Issue 560)
171
- Some documentation fixes
172

    
173

    
174
Version 2.8.0 rc2
175
-----------------
176

    
177
*(Released Tue, 27 Aug 2013)*
178

    
179
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
180

    
181
- Support 'viridian' parameter in Xen HVM (Issue 233)
182
- Include VCS version in ``gnt-cluster version``
183
- Support DSA SSH keys in bootstrap (Issue 338)
184
- Fix batch creation of instances
185
- Use FQDN to check master node status (Issue 551)
186
- Make the DRBD collector more failure-resilient
187

    
188

    
189
Version 2.8.0 rc1
190
-----------------
191

    
192
*(Released Fri, 2 Aug 2013)*
193

    
194
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
195

    
196
- Fix upgrading/downgrading from 2.7
197
- Increase maximum RAPI message size
198
- Documentation updates
199
- Split ``confd`` between ``luxid`` and ``confd``
200
- Merge 2.7 series up to the 2.7.1 release
201
- Allow the ``modify_etc_hosts`` option to be changed
202
- Add better debugging for ``luxid`` queries
203
- Expose bulk parameter for GetJobs in RAPI client
204
- Expose missing ``network`` fields in RAPI
205
- Add some ``cluster verify`` tests
206
- Some unittest fixes
207
- Fix a malfunction in ``hspace``'s tiered allocation
208
- Fix query compatibility between haskell and python implementations
209
- Add the ``vnet_hdr`` HV parameter for KVM
210
- Add ``--cleanup`` to instance failover
211
- Change the connected groups format in ``gnt-network info`` output; it
212
  was previously displayed as a raw list by mistake. (Merged from 2.7)
213

    
214

    
215
Version 2.8.0 beta1
216
-------------------
217

    
218
*(Released Mon, 24 Jun 2013)*
219

    
220
This was the first beta release of the 2.8 series. All important changes
221
are listed in the latest 2.8 entry.
222

    
223

    
224
Version 2.7.2
225
-------------
226

    
227
*(Released Thu, 26 Sep 2013)*
228

    
229
- Change the connected groups format in ``gnt-network info`` output; it
230
  was previously displayed as a raw list by mistake
231
- Check disk template in right dict when copying
232
- Support multi-instance allocs without iallocator
233
- Fix some errors in the documentation
234
- Fix formatting of tuple in an error message
235

    
236

    
237
Version 2.7.1
238
-------------
239

    
240
*(Released Thu, 25 Jul 2013)*
241

    
242
- Add logrotate functionality in daemon-util
243
- Add logrotate example file
244
- Add missing fields to network queries over rapi
245
- Fix network object timestamps
246
- Add support for querying network timestamps
247
- Fix a typo in the example crontab
248
- Fix a documentation typo
249

    
250

    
251
Version 2.7.0
252
-------------
253

    
254
*(Released Thu, 04 Jul 2013)*
255

    
256
Incompatible/important changes
257
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
258

    
259
- Instance policies for disk size were documented to be on a per-disk
260
  basis, but hail applied them to the sum of all disks. This has been
261
  fixed.
262
- ``hbal`` will now exit with status 0 if, during job execution over
263
  LUXI, early exit has been requested and all jobs are successful;
264
  before, exit status 1 was used, which cannot be differentiated from
265
  "job error" case
266
- Compatibility with newer versions of rbd has been fixed
267
- ``gnt-instance batch-create`` has been changed to use the bulk create
268
  opcode from Ganeti. This lead to incompatible changes in the format of
269
  the JSON file. It's now not a custom dict anymore but a dict
270
  compatible with the ``OpInstanceCreate`` opcode.
271
- Parent directories for file storage need to be listed in
272
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
273
  write the file automatically based on old configuration values, but it
274
  can not distribute it across all nodes and the file contents should be
275
  verified. Use ``gnt-cluster copyfile
276
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
277
  upgraded. The reason for requiring this list of paths now is that
278
  before it would have been possible to inject new paths via RPC,
279
  allowing files to be created in arbitrary locations. The RPC protocol
280
  is protected using SSL/X.509 certificates, but as a design principle
281
  Ganeti does not permit arbitrary paths to be passed.
282
- The parsing of the variants file for OSes (see
283
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
284
  lines and comment lines (starting with ``#``) are ignored for better
285
  readability.
286
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
287
  longer available. ``gnt-node add`` now invokes a new tool on the
288
  destination node, named ``prepare-node-join``, to configure the SSH
289
  daemon. Paramiko is no longer necessary to configure nodes' SSH
290
  daemons via ``gnt-node add``.
291
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
292
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
293
  the flag is not reset after a master failover.
294
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
295
  is recommended. 2.6 will be mandatory from the 2.8 series.
296

    
297

    
298
New features
299
~~~~~~~~~~~~
300

    
301
- New network management functionality to support automatic allocation
302
  of IP addresses and managing of network parameters. See
303
  :manpage:`gnt-network(8)` for more details.
304
- New external storage backend, to allow managing arbitrary storage
305
  systems external to the cluster. See
306
  :manpage:`ganeti-extstorage-interface(7)`.
307
- New ``exclusive-storage`` node parameter added, restricted to
308
  nodegroup level. When it's set to true, physical disks are assigned in
309
  an exclusive fashion to instances, as documented in :doc:`Partitioned
310
  Ganeti <design-partitioned>`.  Currently, only instances using the
311
  ``plain`` disk template are supported.
312
- The KVM hypervisor has been updated with many new hypervisor
313
  parameters, including a generic one for passing arbitrary command line
314
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
315
  compatible up to qemu 1.4.
316
- A new tool, called ``mon-collector``, is the stand-alone executor of
317
  the data collectors for a monitoring system. As of this version, it
318
  just includes the DRBD data collector, that can be executed by calling
319
  ``mon-collector`` using the ``drbd`` parameter. See
320
  :manpage:`mon-collector(7)`.
321
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
322
  for RAPI users. It allows granting permissions to query for
323
  information to a specific user without giving
324
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
325
- A new tool named ``node-cleanup`` has been added. It cleans remains of
326
  a cluster from a machine by stopping all daemons, removing
327
  certificates and ssconf files. Unless the ``--no-backup`` option is
328
  given, copies of the certificates are made.
329
- Instance creations now support the use of opportunistic locking,
330
  potentially speeding up the (parallel) creation of multiple instances.
331
  This feature is currently only available via the :doc:`RAPI
332
  <rapi>` interface and when an instance allocator is used. If the
333
  ``opportunistic_locking`` parameter is set the opcode will try to
334
  acquire as many locks as possible, but will not wait for any locks
335
  held by other opcodes. If not enough resources can be found to
336
  allocate the instance, the temporary error code
337
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
338
  retried thereafter, with or without opportunistic locking.
339
- New experimental linux-ha resource scripts.
340
- Restricted-commands support: ganeti can now be asked (via command line
341
  or rapi) to perform commands on a node. These are passed via ganeti
342
  RPC rather than ssh. This functionality is restricted to commands
343
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
344
  reasons. The file is not copied automatically.
345

    
346

    
347
Misc changes
348
~~~~~~~~~~~~
349

    
350
- Diskless instances are now externally mirrored (Issue 237). This for
351
  now has only been tested in conjunction with explicit target nodes for
352
  migration/failover.
353
- Queries not needing locks or RPC access to the node can now be
354
  performed by the confd daemon, making them independent from jobs, and
355
  thus faster to execute. This is selectable at configure time.
356
- The functionality for allocating multiple instances at once has been
357
  overhauled and is now also available through :doc:`RAPI <rapi>`.
358

    
359
There are no significant changes from version 2.7.0~rc3.
360

    
361

    
362
Version 2.7.0 rc3
363
-----------------
364

    
365
*(Released Tue, 25 Jun 2013)*
366

    
367
- Fix permissions on the confd query socket (Issue 477)
368
- Fix permissions on the job archive dir (Issue 498)
369
- Fix handling of an internal exception in replace-disks (Issue 472)
370
- Fix gnt-node info handling of shortened names (Issue 497)
371
- Fix gnt-instance grow-disk when wiping is enabled
372
- Documentation improvements, and support for newer pandoc
373
- Fix hspace honoring ipolicy for disks (Issue 484)
374
- Improve handling of the ``kvm_extra`` HV parameter
375

    
376

    
377
Version 2.7.0 rc2
378
-----------------
379

    
380
*(Released Fri, 24 May 2013)*
381

    
382
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
383
  symlink.
384
- Disks added through ``gnt-instance modify`` or created through
385
  ``gnt-instance recreate-disks`` are wiped, if the
386
  ``prealloc_wipe_disks`` flag is set.
387
- If wiping newly created disks fails, the disks are removed. Also,
388
  partial failures in creating disks through ``gnt-instance modify``
389
  triggers a cleanup of the partially-created disks.
390
- Removing the master IP address doesn't fail if the address has been
391
  already removed.
392
- Fix ownership of the OS log dir
393
- Workaround missing SO_PEERCRED constant (Issue 191)
394

    
395

    
396
Version 2.7.0 rc1
397
-----------------
398

    
399
*(Released Fri, 3 May 2013)*
400

    
401
This was the first release candidate of the 2.7 series. Since beta3:
402

    
403
- Fix kvm compatibility with qemu 1.4 (Issue 389)
404
- Documentation updates (admin guide, upgrade notes, install
405
  instructions) (Issue 372)
406
- Fix gnt-group list nodes and instances count (Issue 436)
407
- Fix compilation without non-mandatory libraries (Issue 441)
408
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
409
- Make confd logging more verbose at INFO level (Issue 435)
410
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
411
- Fix failure path for instance storage type conversion (Issue 229)
412
- Update htools text backend documentation
413
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
414
- Disable inter-cluster instance move for file-based instances, because
415
  it is dependant on instance export, which is not supported for
416
  file-based instances. (Issue 414)
417
- Fix gnt-job crashes on non-ascii characters (Issue 427)
418
- Fix volume group checks on non-vm-capable nodes (Issue 432)
419

    
420

    
421
Version 2.7.0 beta3
422
-------------------
423

    
424
*(Released Mon, 22 Apr 2013)*
425

    
426
This was the third beta release of the 2.7 series. Since beta2:
427

    
428
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
429
- Fix data loss on wrong usage of ``gnt-instance move``
430
- Properly export errors in confd-based job queries
431
- Add ``users-setup`` tool
432
- Fix iallocator protocol to report 0 as a disk size for diskless
433
  instances. This avoids hail breaking when a diskless instance is
434
  present.
435
- Fix job queue directory permission problem that made confd job queries
436
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
437
  for access to archived jobs (Issue 406).
438
- Limit the sizes of networks supported by ``gnt-network`` to something
439
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
440
- Fix bugs in instance disk template conversion
441
- Fix GHC 7 compatibility
442
- Fix ``burnin`` install path (Issue 426).
443
- Allow very small disk grows (Issue 347).
444
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
445
  that noded doesn't import masterd code (Issue 419).
446
- Make sure the default metavg at cluster init is the same as the vg, if
447
  unspecified (Issue 358).
448
- Fix cleanup of partially created disks (part of Issue 416)
449

    
450

    
451
Version 2.7.0 beta2
452
-------------------
453

    
454
*(Released Tue, 2 Apr 2013)*
455

    
456
This was the second beta release of the 2.7 series. Since beta1:
457

    
458
- Networks no longer have a "type" slot, since this information was
459
  unused in Ganeti: instead of it tags should be used.
460
- The rapi client now has a ``target_node`` option to MigrateInstance.
461
- Fix early exit return code for hbal (Issue 386).
462
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
463
- Fix ``rbd showmapped`` output parsing (Issue 312).
464
- Networks are now referenced indexed by UUID, rather than name. This
465
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
466
  use.
467
- The OS environment now includes network information.
468
- Deleting of a network is now disallowed if any instance nic is using
469
  it, to prevent dangling references.
470
- External storage is now documented in man pages.
471
- The exclusive_storage flag can now only be set at nodegroup level.
472
- Hbal can now submit an explicit priority with its jobs.
473
- Many network related locking fixes.
474
- Bump up the required pylint version to 0.25.1.
475
- Fix the ``no_remember`` option in RAPI client.
476
- Many ipolicy related tests, qa, and fixes.
477
- Many documentation improvements and fixes.
478
- Fix building with ``--disable-file-storage``.
479
- Fix ``-q`` option in htools, which was broken if passed more than
480
  once.
481
- Some haskell/python interaction improvements and fixes.
482
- Fix iallocator in case of missing LVM storage.
483
- Fix confd config load in case of ``--no-lvm-storage``.
484
- The confd/query functionality is now mentioned in the security
485
  documentation.
486

    
487

    
488
Version 2.7.0 beta1
489
-------------------
490

    
491
*(Released Wed, 6 Feb 2013)*
492

    
493
This was the first beta release of the 2.7 series. All important changes
494
are listed in the latest 2.7 entry.
495

    
496

    
497
Version 2.6.2
498
-------------
499

    
500
*(Released Fri, 21 Dec 2012)*
501

    
502
Important behaviour change: hbal won't rebalance anymore instances which
503
have the ``auto_balance`` attribute set to false. This was the intention
504
all along, but until now it only skipped those from the N+1 memory
505
reservation (DRBD-specific).
506

    
507
A significant number of bug fixes in this release:
508

    
509
- Fixed disk adoption interaction with ipolicy checks.
510
- Fixed networking issues when instances are started, stopped or
511
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
512
- Fixed the warning in cluster verify for shared storage instances not
513
  being redundant.
514
- Fixed removal of storage directory on shared file storage (issue 262).
515
- Fixed validation of LVM volume group name in OpClusterSetParams
516
  (``gnt-cluster modify``) (issue 285).
517
- Fixed runtime memory increases (``gnt-instance modify -m``).
518
- Fixed live migration under Xen's ``xl`` mode.
519
- Fixed ``gnt-instance console`` with ``xl``.
520
- Fixed building with newer Haskell compiler/libraries.
521
- Fixed PID file writing in Haskell daemons (confd); this prevents
522
  restart issues if confd was launched manually (outside of
523
  ``daemon-util``) while another copy of it was running
524
- Fixed a type error when doing live migrations with KVM (issue 297) and
525
  the error messages for failing migrations have been improved.
526
- Fixed opcode validation for the out-of-band commands (``gnt-node
527
  power``).
528
- Fixed a type error when unsetting OS hypervisor parameters (issue
529
  311); now it's possible to unset all OS-specific hypervisor
530
  parameters.
531
- Fixed the ``dry-run`` mode for many operations: verification of
532
  results was over-zealous but didn't take into account the ``dry-run``
533
  operation, resulting in "wrong" failures.
534
- Fixed bash completion in ``gnt-job list`` when the job queue has
535
  hundreds of entries; especially with older ``bash`` versions, this
536
  results in significant CPU usage.
537

    
538
And lastly, a few other improvements have been made:
539

    
540
- Added option to force master-failover without voting (issue 282).
541
- Clarified error message on lock conflict (issue 287).
542
- Logging of newly submitted jobs has been improved (issue 290).
543
- Hostname checks have been made uniform between instance rename and
544
  create (issue 291).
545
- The ``--submit`` option is now supported by ``gnt-debug delay``.
546
- Shutting down the master daemon by sending SIGTERM now stops it from
547
  processing jobs waiting for locks; instead, those jobs will be started
548
  once again after the master daemon is started the next time (issue
549
  296).
550
- Support for Xen's ``xl`` program has been improved (besides the fixes
551
  above).
552
- Reduced logging noise in the Haskell confd daemon (only show one log
553
  entry for each config reload, instead of two).
554
- Several man page updates and typo fixes.
555

    
556

    
557
Version 2.6.1
558
-------------
559

    
560
*(Released Fri, 12 Oct 2012)*
561

    
562
A small bugfix release. Among the bugs fixed:
563

    
564
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
565
  made the command unusable.
566
- Commands that issue many jobs don't fail anymore just because some jobs
567
  take so long that other jobs are archived.
568
- Failures during ``gnt-instance reinstall`` are reflected by the exit
569
  status.
570
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
571
  DRBD is enabled.
572
- When ``always_failover`` is set, ``--allow-failover`` is not required
573
  in migrate commands anymore.
574
- ``bash_completion`` works even if extglob is disabled.
575
- Fixed bug with locks that made failover for RDB-based instances fail.
576
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
577
  a random node instead of one based on the allocator metric.
578
- Support for newer versions of pylint and pep8.
579
- Hail doesn't fail anymore when trying to add an instance of type
580
  ``file``, ``sharedfile`` or ``rbd``.
581
- Added new Makefile target to rebuild the whole distribution, so that
582
  all files are included.
583

    
584

    
585
Version 2.6.0
586
-------------
587

    
588
*(Released Fri, 27 Jul 2012)*
589

    
590

    
591
.. attention:: The ``LUXI`` protocol has been made more consistent
592
   regarding its handling of command arguments. This, however, leads to
593
   incompatibility issues with previous versions. Please ensure that you
594
   restart Ganeti daemons soon after the upgrade, otherwise most
595
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
596
   pausing/resuming the watcher, cancelling and archiving jobs, querying
597
   the cluster configuration) will fail.
598

    
599

    
600
New features
601
~~~~~~~~~~~~
602

    
603
Instance run status
604
+++++++++++++++++++
605

    
606
The current ``admin_up`` field, which used to denote whether an instance
607
should be running or not, has been removed. Instead, ``admin_state`` is
608
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
609

    
610
The rational behind this is that an instance being “down” can have
611
different meanings:
612

    
613
- it could be down during a reboot
614
- it could be temporarily be down for a reinstall
615
- or it could be down because it is deprecated and kept just for its
616
  disk
617

    
618
The previous Boolean state was making it difficult to do capacity
619
calculations: should Ganeti reserve memory for a down instance? Now, the
620
tri-state field makes it clear:
621

    
622
- in ``up`` and ``down`` state, all resources are reserved for the
623
  instance, and it can be at any time brought up if it is down
624
- in ``offline`` state, only disk space is reserved for it, but not
625
  memory or CPUs
626

    
627
The field can have an extra use: since the transition between ``up`` and
628
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
629
transition between ``offline`` and ``down`` is done via ``gnt-instance
630
modify``, it is possible to given different rights to users. For
631
example, owners of an instance could be allowed to start/stop it, but
632
not transition it out of the offline state.
633

    
634
Instance policies and specs
635
+++++++++++++++++++++++++++
636

    
637
In previous Ganeti versions, an instance creation request was not
638
limited on the minimum size and on the maximum size just by the cluster
639
resources. As such, any policy could be implemented only in third-party
640
clients (RAPI clients, or shell wrappers over ``gnt-*``
641
tools). Furthermore, calculating cluster capacity via ``hspace`` again
642
required external input with regards to instance sizes.
643

    
644
In order to improve these workflows and to allow for example better
645
per-node group differentiation, we introduced instance specs, which
646
allow declaring:
647

    
648
- minimum instance disk size, disk count, memory size, cpu count
649
- maximum values for the above metrics
650
- and “standard” values (used in ``hspace`` to calculate the standard
651
  sized instances)
652

    
653
The minimum/maximum values can be also customised at node-group level,
654
for example allowing more powerful hardware to support bigger instance
655
memory sizes.
656

    
657
Beside the instance specs, there are a few other settings belonging to
658
the instance policy framework. It is possible now to customise, per
659
cluster and node-group:
660

    
661
- the list of allowed disk templates
662
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
663
- the maximum ratio of instance to spindles (see below for more
664
  information) for local storage
665

    
666
All these together should allow all tools that talk to Ganeti to know
667
what are the ranges of allowed values for instances and the
668
over-subscription that is allowed.
669

    
670
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
671
instance configuration, and the physical CPU configuration from the
672
node. For the spindle ratios however, we didn't track before these
673
values, so new parameters have been added:
674

    
675
- a new node parameter ``spindle_count``, defaults to 1, customisable at
676
  node group or node level
677
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
678

    
679
Note that spindles in this context doesn't need to mean actual
680
mechanical hard-drives; it's just a relative number for both the node
681
I/O capacity and instance I/O consumption.
682

    
683
Instance migration behaviour
684
++++++++++++++++++++++++++++
685

    
686
While live-migration is in general desirable over failover, it is
687
possible that for some workloads it is actually worse, due to the
688
variable time of the “suspend” phase during live migration.
689

    
690
To allow the tools to work consistently over such instances (without
691
having to hard-code instance names), a new backend parameter
692
``always_failover`` has been added to control the migration/failover
693
behaviour. When set to True, all migration requests for an instance will
694
instead fall-back to failover.
695

    
696
Instance memory ballooning
697
++++++++++++++++++++++++++
698

    
699
Initial support for memory ballooning has been added. The memory for an
700
instance is no longer fixed (backend parameter ``memory``), but instead
701
can vary between minimum and maximum values (backend parameters
702
``minmem`` and ``maxmem``). Currently we only change an instance's
703
memory when:
704

    
705
- live migrating or failing over and instance and the target node
706
  doesn't have enough memory
707
- user requests changing the memory via ``gnt-instance modify
708
  --runtime-memory``
709

    
710
Instance CPU pinning
711
++++++++++++++++++++
712

    
713
In order to control the use of specific CPUs by instance, support for
714
controlling CPU pinning has been added for the Xen, HVM and LXC
715
hypervisors. This is controlled by a new hypervisor parameter
716
``cpu_mask``; details about possible values for this are in the
717
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
718
VCPU-to-CPU mapping) form will work well only when all nodes in your
719
cluster have the same amount of CPUs.
720

    
721
Disk parameters
722
+++++++++++++++
723

    
724
Another area in which Ganeti was not customisable were the parameters
725
used for storage configuration, e.g. how many stripes to use for LVM,
726
DRBD resync configuration, etc.
727

    
728
To improve this area, we've added disks parameters, which are
729
customisable at cluster and node group level, and which allow to
730
specify various parameters for disks (DRBD has the most parameters
731
currently), for example:
732

    
733
- DRBD resync algorithm and parameters (e.g. speed)
734
- the default VG for meta-data volumes for DRBD
735
- number of stripes for LVM (plain disk template)
736
- the RBD pool
737

    
738
These parameters can be modified via ``gnt-cluster modify -D …`` and
739
``gnt-group modify -D …``, and are used at either instance creation (in
740
case of LVM stripes, for example) or at disk “activation” time
741
(e.g. resync speed).
742

    
743
Rados block device support
744
++++++++++++++++++++++++++
745

    
746
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
747
denoted by the ``rbd`` disk template type. This is considered
748
experimental, feedback is welcome. For details on configuring it, see
749
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
750

    
751
Master IP setup
752
+++++++++++++++
753

    
754
The existing master IP functionality works well only in simple setups (a
755
single network shared by all nodes); however, if nodes belong to
756
different networks, then the ``/32`` setup and lack of routing
757
information is not enough.
758

    
759
To allow the master IP to function well in more complex cases, the
760
system was reworked as follows:
761

    
762
- a master IP netmask setting has been added
763
- the master IP activation/turn-down code was moved from the node daemon
764
  to a separate script
765
- whether to run the Ganeti-supplied master IP script or a user-supplied
766
  on is a ``gnt-cluster init`` setting
767

    
768
Details about the location of the standard and custom setup scripts are
769
in the man page :manpage:`gnt-cluster(8)`; for information about the
770
setup script protocol, look at the Ganeti-supplied script.
771

    
772
SPICE support
773
+++++++++++++
774

    
775
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
776
improved.
777

    
778
It is now possible to use TLS-protected connections, and when renewing
779
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
780
it is now possible to specify spice or spice CA certificates. Also, it
781
is possible to configure a password for SPICE sessions via the
782
hypervisor parameter ``spice_password_file``.
783

    
784
There are also new parameters to control the compression and streaming
785
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
786
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
787
for the spice parameters.
788

    
789
Lastly, it is now possible to see the SPICE connection information via
790
``gnt-instance console``.
791

    
792
OVF converter
793
+++++++++++++
794

    
795
A new tool (``tools/ovfconverter``) has been added that supports
796
conversion between Ganeti and the `Open Virtualization Format
797
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
798
from).
799

    
800
This relies on the ``qemu-img`` tool to convert the disk formats, so the
801
actual compatibility with other virtualization solutions depends on it.
802

    
803
Confd daemon changes
804
++++++++++++++++++++
805

    
806
The configuration query daemon (``ganeti-confd``) is now optional, and
807
has been rewritten in Haskell; whether to use the daemon at all, use the
808
Python (default) or the Haskell version is selectable at configure time
809
via the ``--enable-confd`` parameter, which can take one of the
810
``haskell``, ``python`` or ``no`` values. If not used, disabling the
811
daemon will result in a smaller footprint; for larger systems, we
812
welcome feedback on the Haskell version which might become the default
813
in future versions.
814

    
815
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
816
daemon running. The Python version doesn't implement the new call.
817

    
818

    
819
User interface changes
820
~~~~~~~~~~~~~~~~~~~~~~
821

    
822
We have replaced the ``--disks`` option of ``gnt-instance
823
replace-disks`` with a more flexible ``--disk`` option, which allows
824
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
825
disk size and mode can be changed upon recreation (via ``gnt-instance
826
recreate-disks``, which accepts the same ``--disk`` option).
827

    
828
As many people are used to a ``show`` command, we have added that as an
829
alias to ``info`` on all ``gnt-*`` commands.
830

    
831
The ``gnt-instance grow-disk`` command has a new mode in which it can
832
accept the target size of the disk, instead of the delta; this can be
833
more safe since two runs in absolute mode will be idempotent, and
834
sometimes it's also easier to specify the desired size directly.
835

    
836
Also the handling of instances with regard to offline secondaries has
837
been improved. Instance operations should not fail because one of it's
838
secondary nodes is offline, even though it's safe to proceed.
839

    
840
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
841
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
842
minors to instance name.
843

    
844
API changes
845
~~~~~~~~~~~
846

    
847
RAPI coverage has improved, with (for example) new resources for
848
recreate-disks, node power-cycle, etc.
849

    
850
Compatibility
851
~~~~~~~~~~~~~
852

    
853
There is partial support for ``xl`` in the Xen hypervisor; feedback is
854
welcome.
855

    
856
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
857
whether to still support Python 2.4 or move to Python 2.6 as minimum
858
required version.
859

    
860
Support for Fedora has been slightly improved; the provided example
861
init.d script should work better on it and the INSTALL file should
862
document the needed dependencies.
863

    
864
Internal changes
865
~~~~~~~~~~~~~~~~
866

    
867
The deprecated ``QueryLocks`` LUXI request has been removed. Use
868
``Query(what=QR_LOCK, ...)`` instead.
869

    
870
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
871
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
872
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
873
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
874
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
875

    
876
RAPI client: ``CertificateError`` now derives from
877
``GanetiApiError``. This should make it more easy to handle Ganeti
878
errors.
879

    
880
Deprecation warnings due to PyCrypto/paramiko import in
881
``tools/setup-ssh`` have been silenced, as usually they are safe; please
882
make sure to run an up-to-date paramiko version, if you use this tool.
883

    
884
The QA scripts now depend on Python 2.5 or above (the main code base
885
still works with Python 2.4).
886

    
887
The configuration file (``config.data``) is now written without
888
indentation for performance reasons; if you want to edit it, it can be
889
re-formatted via ``tools/fmtjson``.
890

    
891
A number of bugs has been fixed in the cluster merge tool.
892

    
893
``x509`` certification verification (used in import-export) has been
894
changed to allow the same clock skew as permitted by the cluster
895
verification. This will remove some rare but hard to diagnose errors in
896
import-export.
897

    
898

    
899
Version 2.6.0 rc4
900
-----------------
901

    
902
*(Released Thu, 19 Jul 2012)*
903

    
904
Very few changes from rc4 to the final release, only bugfixes:
905

    
906
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
907
  instance, fix CDROM booting for KVM instances)
908
- fixed node group modification of node parameters
909
- fixed issue in LUClusterVerifyGroup with multi-group clusters
910
- fixed generation of bash completion to ensure a stable ordering
911
- fixed a few typos
912

    
913

    
914
Version 2.6.0 rc3
915
-----------------
916

    
917
*(Released Fri, 13 Jul 2012)*
918

    
919
Third release candidate for 2.6. The following changes were done from
920
rc3 to rc4:
921

    
922
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
923
- Fixed an inconsistency in the LUXI protocol with the provided
924
  arguments (NOT backwards compatible)
925
- Fixed a bug with node groups ipolicy where ``min`` was greater than
926
  the cluster ``std`` value
927
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
928
  easier instance debugging on nodes (requires ``hconfd`` to work)
929

    
930

    
931
Version 2.6.0 rc2
932
-----------------
933

    
934
*(Released Tue, 03 Jul 2012)*
935

    
936
Second release candidate for 2.6. The following changes were done from
937
rc2 to rc3:
938

    
939
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
940
  master candidates
941
- Fixed a RAPI regression on missing beparams/memory
942
- Fixed redistribution of files on offline nodes
943
- Added possibility to run activate-disks even though secondaries are
944
  offline. With this change it relaxes also the strictness on some other
945
  commands which use activate disks internally:
946
  * ``gnt-instance start|reboot|rename|backup|export``
947
- Made it possible to remove safely an instance if its secondaries are
948
  offline
949
- Made it possible to reinstall even though secondaries are offline
950

    
951

    
952
Version 2.6.0 rc1
953
-----------------
954

    
955
*(Released Mon, 25 Jun 2012)*
956

    
957
First release candidate for 2.6. The following changes were done from
958
rc1 to rc2:
959

    
960
- Fixed bugs with disk parameters and ``rbd`` templates as well as
961
  ``instance_os_add``
962
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
963
  behaviour. It supports now the modify operation
964
- ``hcheck`` implemented to analyze cluster health and possibility of
965
  improving health by rebalance
966
- ``hbal`` has been improved in dealing with split instances
967

    
968

    
969
Version 2.6.0 beta2
970
-------------------
971

    
972
*(Released Mon, 11 Jun 2012)*
973

    
974
Second beta release of 2.6. The following changes were done from beta2
975
to rc1:
976

    
977
- Fixed ``daemon-util`` with non-root user models
978
- Fixed creation of plain instances with ``--no-wait-for-sync``
979
- Fix wrong iv_names when running ``cfgupgrade``
980
- Export more information in RAPI group queries
981
- Fixed bug when changing instance network interfaces
982
- Extended burnin to do NIC changes
983
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
984
- Changed default for DRBD barriers
985
- Fixed DRBD error reporting for syncer rate
986
- Verify the options on disk parameters
987

    
988
And of course various fixes to documentation and improved unittests and
989
QA.
990

    
991

    
992
Version 2.6.0 beta1
993
-------------------
994

    
995
*(Released Wed, 23 May 2012)*
996

    
997
First beta release of 2.6. The following changes were done from beta1 to
998
beta2:
999

    
1000
- integrated patch for distributions without ``start-stop-daemon``
1001
- adapted example init.d script to work on Fedora
1002
- fixed log handling in Haskell daemons
1003
- adapted checks in the watcher for pycurl linked against libnss
1004
- add partial support for ``xl`` instead of ``xm`` for Xen
1005
- fixed a type issue in cluster verification
1006
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1007
  clusters)
1008

    
1009
Plus integrated fixes from the 2.5 branch:
1010

    
1011
- fixed ``kvm-ifup`` to use ``/bin/bash``
1012
- fixed parallel build failures
1013
- KVM live migration when using a custom keymap
1014

    
1015

    
1016
Version 2.5.2
1017
-------------
1018

    
1019
*(Released Tue, 24 Jul 2012)*
1020

    
1021
A small bugfix release, with no new features:
1022

    
1023
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1024
  different default shell (e.g. Debian, Ubuntu)
1025
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1026
  243 and Debian bug #650664)
1027
- fixed compatibility with KVM versions that don't support multiple boot
1028
  devices (fixes Issue 230 and Debian bug #624256)
1029

    
1030
Additionally, a few fixes were done to the build system (fixed parallel
1031
build failures) and to the unittests (fixed race condition in test for
1032
FileID functions, and the default enable/disable mode for QA test is now
1033
customisable).
1034

    
1035

    
1036
Version 2.5.1
1037
-------------
1038

    
1039
*(Released Fri, 11 May 2012)*
1040

    
1041
A small bugfix release.
1042

    
1043
The main issues solved are on the topic of compatibility with newer LVM
1044
releases:
1045

    
1046
- fixed parsing of ``lv_attr`` field
1047
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1048
  the ``--force`` flag is needed
1049

    
1050
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1051
to accept kernel 3.x too (was hardcoded to 2.6.*).
1052

    
1053
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1054
of hook results and that also made display of other errors suboptimal
1055
was fixed; the code behaves now like 2.4 and earlier.
1056

    
1057
Another change in 2.5, the cleanup of the OS scripts environment, is too
1058
aggressive: it removed even the ``PATH`` variable, which requires the OS
1059
scripts to *always* need to export it. Since this is a bit too strict,
1060
we now export a minimal PATH, the same that we export for hooks.
1061

    
1062
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1063
integrated into this release.
1064

    
1065
Finally, a few other miscellaneous changes were done (no new features,
1066
just small improvements):
1067

    
1068
- Fix ``gnt-group --help`` display
1069
- Fix hardcoded Xen kernel path
1070
- Fix grow-disk handling of invalid units
1071
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1072
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1073

    
1074

    
1075
Version 2.5.0
1076
-------------
1077

    
1078
*(Released Thu, 12 Apr 2012)*
1079

    
1080
Incompatible/important changes and bugfixes
1081
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1082

    
1083
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1084
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1085
  to match the underlying LUXI interface.
1086
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1087
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1088
  not maintain backwards-compatibility as the underlying operation
1089
  changed in an incompatible way. The RAPI client can talk to old
1090
  servers, but it needs to be told so as the return value changed.
1091
- When creating file-based instances via RAPI, the ``file_driver``
1092
  parameter no longer defaults to ``loop`` and must be specified.
1093
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1094
  ``link`` instead.
1095
- Support for the undocumented and deprecated RAPI instance creation
1096
  request format version 0 has been dropped. Use version 1, supported
1097
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1098
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1099
  documentation <install>`.
1100
- The "cluster-verify" hooks are now executed per group by the
1101
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1102
  if you just run ``gnt-cluster verify``, which generates one opcode per
1103
  group.
1104
- The environment as passed to the OS scripts is cleared, and thus no
1105
  environment variables defined in the node daemon's environment will be
1106
  inherited by the scripts.
1107
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1108
  deprecated.
1109
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1110
  support operations involving multiple node groups.
1111
- Offline nodes are ignored when failing over an instance.
1112
- Support for KVM version 1.0, which changed the version reporting format
1113
  from 3 to 2 digits.
1114
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1115
  removal.
1116
- ``Makefile`` is now compatible with Automake 1.11.2
1117
- Includes all bugfixes made in the 2.4 series
1118

    
1119
New features
1120
~~~~~~~~~~~~
1121

    
1122
- The ganeti-htools project has been merged into the ganeti-core source
1123
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1124
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1125
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1126
  GPT-style partition tables (requires `parted
1127
  <http://www.gnu.org/s/parted/>`_).
1128
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1129
- Allowed adding tags on instance creation.
1130
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1131
- Allow instances to be started in a paused state, enabling the user to
1132
  see the complete console output on boot using the console.
1133
- Added new hypervisor flag to control default reboot behaviour
1134
  (``reboot_behavior``).
1135
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1136
- Improved out-of-band management support:
1137

    
1138
  - Added ``gnt-node health`` command reporting the health status of
1139
    nodes.
1140
  - Added ``gnt-node power`` command to manage power status of nodes.
1141
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1142

    
1143
- Instance migration can fall back to failover if instance is not
1144
  running.
1145
- Filters can be used when listing nodes, instances, groups and locks;
1146
  see :manpage:`ganeti(7)` manpage.
1147
- Added post-execution status as variables to :doc:`hooks <hooks>`
1148
  environment.
1149
- Instance tags are exported/imported together with the instance.
1150
- When given an explicit job ID, ``gnt-job info`` will work for archived
1151
  jobs.
1152
- Jobs can define dependencies on other jobs (not yet supported via
1153
  RAPI or command line, but used by internal commands and usable via
1154
  LUXI).
1155

    
1156
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1157
    dependencies.
1158

    
1159
- Instance failover is now available as a RAPI resource
1160
  (``/2/instances/[instance_name]/failover``).
1161
- ``gnt-instance info`` defaults to static information if primary node
1162
  is offline.
1163
- Opcodes have a new ``comment`` attribute.
1164
- Added basic SPICE support to KVM hypervisor.
1165
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1166

    
1167
Node group improvements
1168
~~~~~~~~~~~~~~~~~~~~~~~
1169

    
1170
- ``gnt-cluster verify`` has been modified to check groups separately,
1171
  thereby improving performance.
1172
- Node group support has been added to ``gnt-cluster verify-disks``,
1173
  which now operates per node group.
1174
- Watcher has been changed to work better with node groups.
1175

    
1176
  - One process and state file per node group.
1177
  - Slow watcher in one group doesn't block other group's watcher.
1178

    
1179
- Added new command, ``gnt-group evacuate``, to move all instances in a
1180
  node group to other groups.
1181
- Added ``gnt-instance change-group`` to move an instance to another
1182
  node group.
1183
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1184
  per-group operations.
1185
- Node groups can be tagged.
1186
- Some operations switch from an exclusive to a shared lock as soon as
1187
  possible.
1188
- Instance's primary and secondary nodes' groups are now available as
1189
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1190
  and ``snodes.group.uuid``).
1191

    
1192
Misc
1193
~~~~
1194

    
1195
- Numerous updates to documentation and manpages.
1196

    
1197
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1198
    descriptions.
1199
  - Some opcode/job results are now also documented, see :doc:`RAPI
1200
    <rapi>`.
1201

    
1202
- A lockset's internal lock is now also visible in lock monitor.
1203
- Log messages from job queue workers now contain information about the
1204
  opcode they're processing.
1205
- ``gnt-instance console`` no longer requires the instance lock.
1206
- A short delay when waiting for job changes reduces the number of LUXI
1207
  requests significantly.
1208
- DRBD metadata volumes are overwritten with zeros during disk creation.
1209
- Out-of-band commands no longer acquire the cluster lock in exclusive
1210
  mode.
1211
- ``devel/upload`` now uses correct permissions for directories.
1212

    
1213

    
1214
Version 2.5.0 rc6
1215
-----------------
1216

    
1217
*(Released Fri, 23 Mar 2012)*
1218

    
1219
This was the sixth release candidate of the 2.5 series.
1220

    
1221

    
1222
Version 2.5.0 rc5
1223
-----------------
1224

    
1225
*(Released Mon, 9 Jan 2012)*
1226

    
1227
This was the fifth release candidate of the 2.5 series.
1228

    
1229

    
1230
Version 2.5.0 rc4
1231
-----------------
1232

    
1233
*(Released Thu, 27 Oct 2011)*
1234

    
1235
This was the fourth release candidate of the 2.5 series.
1236

    
1237

    
1238
Version 2.5.0 rc3
1239
-----------------
1240

    
1241
*(Released Wed, 26 Oct 2011)*
1242

    
1243
This was the third release candidate of the 2.5 series.
1244

    
1245

    
1246
Version 2.5.0 rc2
1247
-----------------
1248

    
1249
*(Released Tue, 18 Oct 2011)*
1250

    
1251
This was the second release candidate of the 2.5 series.
1252

    
1253

    
1254
Version 2.5.0 rc1
1255
-----------------
1256

    
1257
*(Released Tue, 4 Oct 2011)*
1258

    
1259
This was the first release candidate of the 2.5 series.
1260

    
1261

    
1262
Version 2.5.0 beta3
1263
-------------------
1264

    
1265
*(Released Wed, 31 Aug 2011)*
1266

    
1267
This was the third beta release of the 2.5 series.
1268

    
1269

    
1270
Version 2.5.0 beta2
1271
-------------------
1272

    
1273
*(Released Mon, 22 Aug 2011)*
1274

    
1275
This was the second beta release of the 2.5 series.
1276

    
1277

    
1278
Version 2.5.0 beta1
1279
-------------------
1280

    
1281
*(Released Fri, 12 Aug 2011)*
1282

    
1283
This was the first beta release of the 2.5 series.
1284

    
1285

    
1286
Version 2.4.5
1287
-------------
1288

    
1289
*(Released Thu, 27 Oct 2011)*
1290

    
1291
- Fixed bug when parsing command line parameter values ending in
1292
  backslash
1293
- Fixed assertion error after unclean master shutdown
1294
- Disable HTTP client pool for RPC, significantly reducing memory usage
1295
  of master daemon
1296
- Fixed queue archive creation with wrong permissions
1297

    
1298

    
1299
Version 2.4.4
1300
-------------
1301

    
1302
*(Released Tue, 23 Aug 2011)*
1303

    
1304
Small bug-fixes:
1305

    
1306
- Fixed documentation for importing with ``--src-dir`` option
1307
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1308
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1309

    
1310

    
1311
Version 2.4.3
1312
-------------
1313

    
1314
*(Released Fri, 5 Aug 2011)*
1315

    
1316
Many bug-fixes and a few small features:
1317

    
1318
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1319
  issues when you tried to add an instance with two MAC addresses in one
1320
  request
1321
- KVM: fixed per-instance stored UID value
1322
- KVM: configure bridged NICs at migration start
1323
- KVM: Fix a bug where instance will not start with never KVM versions
1324
  (>= 0.14)
1325
- Added OS search path to ``gnt-cluster info``
1326
- Fixed an issue with ``file_storage_dir`` where you were forced to
1327
  provide an absolute path, but the documentation states it is a
1328
  relative path, the documentation was right
1329
- Added a new parameter to instance stop/start called ``--no-remember``
1330
  that will make the state change to not be remembered
1331
- Implemented ``no_remember`` at RAPI level
1332
- Improved the documentation
1333
- Node evacuation: don't call IAllocator if node is already empty
1334
- Fixed bug in DRBD8 replace disks on current nodes
1335
- Fixed bug in recreate-disks for DRBD instances
1336
- Moved assertion checking locks in ``gnt-instance replace-disks``
1337
  causing it to abort with not owning the right locks for some situation
1338
- Job queue: Fixed potential race condition when cancelling queued jobs
1339
- Fixed off-by-one bug in job serial generation
1340
- ``gnt-node volumes``: Fix instance names
1341
- Fixed aliases in bash completion
1342
- Fixed a bug in reopening log files after being sent a SIGHUP
1343
- Added a flag to burnin to allow specifying VCPU count
1344
- Bugfixes to non-root Ganeti configuration
1345

    
1346

    
1347
Version 2.4.2
1348
-------------
1349

    
1350
*(Released Thu, 12 May 2011)*
1351

    
1352
Many bug-fixes and a few new small features:
1353

    
1354
- Fixed a bug related to log opening failures
1355
- Fixed a bug in instance listing with orphan instances
1356
- Fixed a bug which prevented resetting the cluster-level node parameter
1357
  ``oob_program`` to the default
1358
- Many fixes related to the ``cluster-merge`` tool
1359
- Fixed a race condition in the lock monitor, which caused failures
1360
  during (at least) creation of many instances in parallel
1361
- Improved output for gnt-job info
1362
- Removed the quiet flag on some ssh calls which prevented debugging
1363
  failures
1364
- Improved the N+1 failure messages in cluster verify by actually
1365
  showing the memory values (needed and available)
1366
- Increased lock attempt timeouts so that when executing long operations
1367
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1368
  too early and thus prevent the use of the 'fair' mechanism
1369
- Changed instance query data (``gnt-instance info``) to not acquire
1370
  locks unless needed, thus allowing its use on locked instance if only
1371
  static information is asked for
1372
- Improved behaviour with filesystems that do not support rename on an
1373
  opened file
1374
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1375
  kept locks on all nodes during the wipe, which is unneeded
1376
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1377
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1378
  10GiB) which caused the wipe to fail right at the end in some cases
1379
- Fixed master IP activation when doing master failover with no-voting
1380
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1381
  the master node itself
1382
- Fixed potential data-loss in under disk full conditions, where Ganeti
1383
  wouldn't check correctly the return code and would consider
1384
  partially-written files 'correct'
1385
- Fixed bug related to multiple VGs and DRBD disk replacing
1386
- Added new disk parameter ``metavg`` that allows placement of the meta
1387
  device for DRBD in a different volume group
1388
- Fixed error handling in the node daemon when the system libc doesn't
1389
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1390
- Fixed lock release during replace-disks, which kept cluster-wide locks
1391
  when doing disk replaces with an iallocator script
1392
- Added check for missing bridges in cluster verify
1393
- Handle EPIPE errors while writing to the terminal better, so that
1394
  piping the output to e.g. ``less`` doesn't cause a backtrace
1395
- Fixed rare case where a ^C during Luxi calls could have been
1396
  interpreted as server errors, instead of simply terminating
1397
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1398
  assign-nodes``)
1399
- Added a few more parameters to the KVM hypervisor, allowing a second
1400
  CDROM, custom disk type for CDROMs and a floppy image
1401
- Removed redundant message in instance rename when the name is given
1402
  already as a FQDN
1403
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1404
  disks on new nodes, allowing recreation when the original instance
1405
  nodes are completely gone
1406
- Added option when converting disk templates to DRBD to skip waiting
1407
  for the resync, in order to make the instance available sooner
1408
- Added two new variables to the OS scripts environment (containing the
1409
  instance's nodes)
1410
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1411
  to allow use of ``pvgrub`` as bootloader
1412
- Changed the instance memory modifications to only check out-of-memory
1413
  conditions on memory increases, and turned the secondary node warnings
1414
  into errors (they can still be overridden via ``--force``)
1415
- Fixed the handling of a corner case when the Python installation gets
1416
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1417
  to execute a command that doesn't exist
1418
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1419
  node of the instance returned failures during instance shutdown; this
1420
  adds the option ``--ignore-consistency`` to gnt-instance move
1421

    
1422
And as usual, various improvements to the error messages, documentation
1423
and man pages.
1424

    
1425

    
1426
Version 2.4.1
1427
-------------
1428

    
1429
*(Released Wed, 09 Mar 2011)*
1430

    
1431
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1432
the RAPI users file if run twice (even with ``--dry-run``).
1433

    
1434
The release fixes that bug (nothing else changed).
1435

    
1436

    
1437
Version 2.4.0
1438
-------------
1439

    
1440
*(Released Mon, 07 Mar 2011)*
1441

    
1442
Final 2.4.0 release. Just a few small fixes:
1443

    
1444
- Fixed RAPI node evacuate
1445
- Fixed the kvm-ifup script
1446
- Fixed internal error handling for special job cases
1447
- Updated man page to specify the escaping feature for options
1448

    
1449

    
1450
Version 2.4.0 rc3
1451
-----------------
1452

    
1453
*(Released Mon, 28 Feb 2011)*
1454

    
1455
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1456
that this feature wiped the disks of the wrong instance, leading to loss
1457
of data.
1458

    
1459
Other changes:
1460

    
1461
- Fixed title of query field containing instance name
1462
- Expanded the glossary in the documentation
1463
- Fixed one unittest (internal issue)
1464

    
1465

    
1466
Version 2.4.0 rc2
1467
-----------------
1468

    
1469
*(Released Mon, 21 Feb 2011)*
1470

    
1471
A number of bug fixes plus just a couple functionality changes.
1472

    
1473
On the user-visible side, the ``gnt-* list`` command output has changed
1474
with respect to "special" field states. The current rc1 style of display
1475
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1476
the default output mode special fields states are displayed as follows:
1477

    
1478
- Offline resource: ``*``
1479
- Unavailable/not applicable: ``-``
1480
- Data missing (RPC failure): ``?``
1481
- Unknown field: ``??``
1482

    
1483
Another user-visible change is the addition of ``--force-join`` to
1484
``gnt-node add``.
1485

    
1486
As for bug fixes:
1487

    
1488
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1489
- Fixed regression in RAPI/instance reinstall where all parameters were
1490
  required (instead of optional)
1491
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1492
- Fixed iallocator usage (offline nodes were not considered offline)
1493
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1494
- Fixed hypervisor and OS parameter validation with respect to
1495
  non-vm_capable nodes
1496
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1497
  cosmetic)
1498
- Fixed ``tools/listrunner`` with respect to agent-based usage
1499

    
1500

    
1501
Version 2.4.0 rc1
1502
-----------------
1503

    
1504
*(Released Fri,  4 Feb 2011)*
1505

    
1506
Many changes and fixes since the beta1 release. While there were some
1507
internal changes, the code has been mostly stabilised for the RC
1508
release.
1509

    
1510
Note: the dumb allocator was removed in this release, as it was not kept
1511
up-to-date with the IAllocator protocol changes. It is recommended to
1512
use the ``hail`` command from the ganeti-htools package.
1513

    
1514
Note: the 2.4 and up versions of Ganeti are not compatible with the
1515
0.2.x branch of ganeti-htools. You need to upgrade to
1516
ganeti-htools-0.3.0 (or later).
1517

    
1518
Regressions fixed from 2.3
1519
~~~~~~~~~~~~~~~~~~~~~~~~~~
1520

    
1521
- Fixed the ``gnt-cluster verify-disks`` command
1522
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1523
  serially on nodes)
1524
- Fixed disk adoption breakage
1525
- Fixed wrong headers in instance listing for field aliases
1526

    
1527
Other bugs fixed
1528
~~~~~~~~~~~~~~~~
1529

    
1530
- Fixed corner case in KVM handling of NICs
1531
- Fixed many cases of wrong handling of non-vm_capable nodes
1532
- Fixed a bug where a missing instance symlink was not possible to
1533
  recreate with any ``gnt-*`` command (now ``gnt-instance
1534
  activate-disks`` does it)
1535
- Fixed the volume group name as reported by ``gnt-cluster
1536
  verify-disks``
1537
- Increased timeouts for the import-export code, hopefully leading to
1538
  fewer aborts due network or instance timeouts
1539
- Fixed bug in ``gnt-node list-storage``
1540
- Fixed bug where not all daemons were started on cluster
1541
  initialisation, but only at the first watcher run
1542
- Fixed many bugs in the OOB implementation
1543
- Fixed watcher behaviour in presence of instances with offline
1544
  secondaries
1545
- Fixed instance list output for instances running on the wrong node
1546
- a few fixes to the cluster-merge tool, but it still cannot merge
1547
  multi-node groups (currently it is not recommended to use this tool)
1548

    
1549

    
1550
Improvements
1551
~~~~~~~~~~~~
1552

    
1553
- Improved network configuration for the KVM hypervisor
1554
- Added e1000 as a supported NIC for Xen-HVM
1555
- Improved the lvmstrap tool to also be able to use partitions, as
1556
  opposed to full disks
1557
- Improved speed of disk wiping (the cluster parameter
1558
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1559
  of instance creations
1560
- Added documentation for the OS parameters
1561
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1562
  hypervisor is not responding
1563
- Added display of blacklisted and hidden OS information in
1564
  ``gnt-cluster info``
1565
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1566
  NIC and node parameters, which might create problems with currently
1567
  invalid (but undetected) configuration files, but prevents validation
1568
  failures when unrelated parameters are modified
1569
- Changed cluster initialisation to wait for the master daemon to become
1570
  available
1571
- Expanded the RAPI interface:
1572

    
1573
  - Added config redistribution resource
1574
  - Added activation/deactivation of instance disks
1575
  - Added export of console information
1576

    
1577
- Implemented log file reopening on SIGHUP, which allows using
1578
  logrotate(8) for the Ganeti log files
1579
- Added a basic OOB helper script as an example
1580

    
1581

    
1582
Version 2.4.0 beta1
1583
-------------------
1584

    
1585
*(Released Fri, 14 Jan 2011)*
1586

    
1587
User-visible
1588
~~~~~~~~~~~~
1589

    
1590
- Fixed timezone issues when formatting timestamps
1591
- Added support for node groups, available via ``gnt-group`` and other
1592
  commands
1593
- Added out-of-band framework and management, see :doc:`design
1594
  document <design-oob>`
1595
- Removed support for roman numbers from ``gnt-node list`` and
1596
  ``gnt-instance list``.
1597
- Allowed modification of master network interface via ``gnt-cluster
1598
  modify --master-netdev``
1599
- Accept offline secondaries while shutting down instance disks
1600
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1601
- Added support for multiple LVM volume groups
1602
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1603
  requested
1604
- Added commands to list available fields:
1605

    
1606
  - ``gnt-node list-fields``
1607
  - ``gnt-group list-fields``
1608
  - ``gnt-instance list-fields``
1609

    
1610
- Updated documentation and man pages
1611

    
1612
Integration
1613
~~~~~~~~~~~
1614

    
1615
- Moved ``rapi_users`` file into separate directory, now named
1616
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1617
  symlink
1618
- Added new tool for running commands on many machines,
1619
  ``tools/ganeti-listrunner``
1620
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1621
  improving the ``gnt-instance console`` output
1622
- Allowed customisation of disk index separator at ``configure`` time
1623
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1624
- Added support for non-partitioned md disks in ``lvmstrap``
1625
- Added script to gracefully power off KVM instances
1626
- Split ``utils`` module into smaller parts
1627
- Changed query operations to return more detailed information, e.g.
1628
  whether an information is unavailable due to an offline node. To use
1629
  this new functionality, the LUXI call ``Query`` must be used. Field
1630
  information is now stored by the master daemon and can be retrieved
1631
  using ``QueryFields``. Instances, nodes and groups can also be queried
1632
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1633
  exposed via RAPI). The following commands make use of this
1634
  infrastructure change:
1635

    
1636
  - ``gnt-group list``
1637
  - ``gnt-group list-fields``
1638
  - ``gnt-node list``
1639
  - ``gnt-node list-fields``
1640
  - ``gnt-instance list``
1641
  - ``gnt-instance list-fields``
1642
  - ``gnt-debug locks``
1643

    
1644
Remote API
1645
~~~~~~~~~~
1646

    
1647
- New RAPI resources (see :doc:`rapi`):
1648

    
1649
  - ``/2/modify``
1650
  - ``/2/groups``
1651
  - ``/2/groups/[group_name]``
1652
  - ``/2/groups/[group_name]/assign-nodes``
1653
  - ``/2/groups/[group_name]/modify``
1654
  - ``/2/groups/[group_name]/rename``
1655
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1656

    
1657
- RAPI changes:
1658

    
1659
  - Implemented ``no_install`` for instance creation
1660
  - Implemented OS parameters for instance reinstallation, allowing
1661
    use of special settings on reinstallation (e.g. for preserving data)
1662

    
1663
Misc
1664
~~~~
1665

    
1666
- Added IPv6 support in import/export
1667
- Pause DRBD synchronization while wiping disks on instance creation
1668
- Updated unittests and QA scripts
1669
- Improved network parameters passed to KVM
1670
- Converted man pages from docbook to reStructuredText
1671

    
1672

    
1673
Version 2.3.1
1674
-------------
1675

    
1676
*(Released Mon, 20 Dec 2010)*
1677

    
1678
Released version 2.3.1~rc1 without any changes.
1679

    
1680

    
1681
Version 2.3.1 rc1
1682
-----------------
1683

    
1684
*(Released Wed, 1 Dec 2010)*
1685

    
1686
- impexpd: Disable OpenSSL compression in socat if possible (backport
1687
  from master, commit e90739d625b, see :doc:`installation guide
1688
  <install-quick>` for details)
1689
- Changed unittest coverage report to exclude test scripts
1690
- Added script to check version format
1691

    
1692

    
1693
Version 2.3.0
1694
-------------
1695

    
1696
*(Released Wed, 1 Dec 2010)*
1697

    
1698
Released version 2.3.0~rc1 without any changes.
1699

    
1700

    
1701
Version 2.3.0 rc1
1702
-----------------
1703

    
1704
*(Released Fri, 19 Nov 2010)*
1705

    
1706
A number of bugfixes and documentation updates:
1707

    
1708
- Update ganeti-os-interface documentation
1709
- Fixed a bug related to duplicate MACs or similar items which should be
1710
  unique
1711
- Fix breakage in OS state modify
1712
- Reinstall instance: disallow offline secondaries (fixes bug related to
1713
  OS changing but reinstall failing)
1714
- plus all the other fixes between 2.2.1 and 2.2.2
1715

    
1716

    
1717
Version 2.3.0 rc0
1718
-----------------
1719

    
1720
*(Released Tue, 2 Nov 2010)*
1721

    
1722
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1723
- Fixed master failover race with watcher
1724
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1725
  configuration
1726
- Accept previously stopped instance for export with instance removal
1727
- Simplify and extend the environment variables for instance OS scripts
1728
- Added new node flags, ``master_capable`` and ``vm_capable``
1729
- Added optional instance disk wiping prior during allocation. This is a
1730
  cluster-wide option and can be set/modified using
1731
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1732
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1733
  :doc:`install-quick`
1734
- Added a new watcher option (``--ignore-pause``)
1735
- Added option to ignore offline node on instance start/stop
1736
  (``--ignore-offline``)
1737
- Allow overriding OS parameters with ``gnt-instance reinstall``
1738
- Added ability to change node's secondary IP address using ``gnt-node
1739
  modify``
1740
- Implemented privilege separation for all daemons except
1741
  ``ganeti-noded``, see ``configure`` options
1742
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1743
  verify``
1744
- Implemented job priorities (see ``ganeti(7)`` manpage)
1745
- Ignore failures while shutting down instances during failover from
1746
  offline node
1747
- Exit daemon's bootstrap process only once daemon is ready
1748
- Export more information via ``LUInstanceQuery``/remote API
1749
- Improved documentation, QA and unittests
1750
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1751
  restart if the file was created or changed
1752
- Added LUXI protocol version sent with each request and response,
1753
  allowing detection of server/client mismatches
1754
- Moved the Python scripts among gnt-* and ganeti-* into modules
1755
- Moved all code related to setting up SSH to an external script,
1756
  ``setup-ssh``
1757
- Infrastructure changes for node group support in future versions
1758

    
1759

    
1760
Version 2.2.2
1761
-------------
1762

    
1763
*(Released Fri, 19 Nov 2010)*
1764

    
1765
A few small bugs fixed, and some improvements to the build system:
1766

    
1767
- Fix documentation regarding conversion to drbd
1768
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1769
  -B``)
1770
- Fix error handling in node modify with multiple changes
1771
- Allow remote imports without checked names
1772

    
1773

    
1774
Version 2.2.1
1775
-------------
1776

    
1777
*(Released Tue, 19 Oct 2010)*
1778

    
1779
- Disable SSL session ID cache in RPC client
1780

    
1781

    
1782
Version 2.2.1 rc1
1783
-----------------
1784

    
1785
*(Released Thu, 14 Oct 2010)*
1786

    
1787
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1788
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1789
  with GnuTLS
1790
- Fix problems with interaction between Curl and Python's HTTP server,
1791
  resulting in increased speed in many RPC calls
1792
- Improve our release script to prevent breakage with older aclocal and
1793
  Python 2.6
1794

    
1795

    
1796
Version 2.2.1 rc0
1797
-----------------
1798

    
1799
*(Released Thu, 7 Oct 2010)*
1800

    
1801
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1802
  value retrieved from master
1803
- Added support for blacklisted or hidden OS definitions
1804
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1805
- Added support for -mem-path in KVM hypervisor abstraction layer
1806
- Allow overriding instance parameters in tool for inter-cluster
1807
  instance moves (``tools/move-instance``)
1808
- Improved opcode summaries (e.g. in ``gnt-job list``)
1809
- Improve consistency of OS listing by sorting it
1810
- Documentation updates
1811

    
1812

    
1813
Version 2.2.0.1
1814
---------------
1815

    
1816
*(Released Fri, 8 Oct 2010)*
1817

    
1818
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1819

    
1820

    
1821
Version 2.2.0
1822
-------------
1823

    
1824
*(Released Mon, 4 Oct 2010)*
1825

    
1826
- Fixed regression in ``gnt-instance rename``
1827

    
1828

    
1829
Version 2.2.0 rc2
1830
-----------------
1831

    
1832
*(Released Wed, 22 Sep 2010)*
1833

    
1834
- Fixed OS_VARIANT variable for OS scripts
1835
- Fixed cluster tag operations via RAPI
1836
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1837
- Disabled RAPI CA checks in watcher
1838

    
1839

    
1840
Version 2.2.0 rc1
1841
-----------------
1842

    
1843
*(Released Mon, 23 Aug 2010)*
1844

    
1845
- Support DRBD versions of the format "a.b.c.d"
1846
- Updated manpages
1847
- Re-introduce support for usage from multiple threads in RAPI client
1848
- Instance renames and modify via RAPI
1849
- Work around race condition between processing and archival in job
1850
  queue
1851
- Mark opcodes following failed one as failed, too
1852
- Job field ``lock_status`` was removed due to difficulties making it
1853
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1854
  locks is expected for a later 2.2.x release
1855
- Fixed dry-run behaviour with many commands
1856
- Support ``ssh-agent`` again when adding nodes
1857
- Many additional bugfixes
1858

    
1859

    
1860
Version 2.2.0 rc0
1861
-----------------
1862

    
1863
*(Released Fri, 30 Jul 2010)*
1864

    
1865
Important change: the internal RPC mechanism between Ganeti nodes has
1866
changed from using a home-grown http library (based on the Python base
1867
libraries) to use the PycURL library. This requires that PycURL is
1868
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1869
against GnuTLS by default. cURL's support for GnuTLS had known issues
1870
before cURL 7.21.0 and we recommend using the latest cURL release or
1871
linking against OpenSSL. Most other distributions already link PycURL
1872
and cURL against OpenSSL. The command::
1873

    
1874
  python -c 'import pycurl; print pycurl.version'
1875

    
1876
can be used to determine the libraries PycURL and cURL are linked
1877
against.
1878

    
1879
Other significant changes:
1880

    
1881
- Rewrote much of the internals of the job queue, in order to achieve
1882
  better parallelism; this decouples job query operations from the job
1883
  processing, and it should allow much nicer behaviour of the master
1884
  daemon under load, and it also has uncovered some long-standing bugs
1885
  related to the job serialisation (now fixed)
1886
- Added a default iallocator setting to the cluster parameters,
1887
  eliminating the need to always pass nodes or an iallocator for
1888
  operations that require selection of new node(s)
1889
- Added experimental support for the LXC virtualization method
1890
- Added support for OS parameters, which allows the installation of
1891
  instances to pass parameter to OS scripts in order to customise the
1892
  instance
1893
- Added a hypervisor parameter controlling the migration type (live or
1894
  non-live), since hypervisors have various levels of reliability; this
1895
  has renamed the 'live' parameter to 'mode'
1896
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
1897
  logical volumes, meaning that cluster verify will ignore them and not
1898
  flag their presence as errors
1899
- The watcher will now reset the error count for failed instances after
1900
  8 hours, thus allowing self-healing if the problem that caused the
1901
  instances to be down/fail to start has cleared in the meantime
1902
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1903
  check for, and warn, if the drbd module parameter ``usermode_helper``
1904
  is not consistent with the cluster-wide setting; this is needed to
1905
  make diagnose easier of failed drbd creations
1906
- Started adding base IPv6 support, but this is not yet
1907
  enabled/available for use
1908
- Rename operations (cluster, instance) will now return the new name,
1909
  which is especially useful if a short name was passed in
1910
- Added support for instance migration in RAPI
1911
- Added a tool to pre-configure nodes for the SSH setup, before joining
1912
  them to the cluster; this will allow in the future a simplified model
1913
  for node joining (but not yet fully enabled in 2.2); this needs the
1914
  paramiko python library
1915
- Fixed handling of name-resolving errors
1916
- Fixed consistency of job results on the error path
1917
- Fixed master-failover race condition when executed multiple times in
1918
  sequence
1919
- Fixed many bugs related to the job queue (mostly introduced during the
1920
  2.2 development cycle, so not all are impacting 2.1)
1921
- Fixed instance migration with missing disk symlinks
1922
- Fixed handling of unknown jobs in ``gnt-job archive``
1923
- And many other small fixes/improvements
1924

    
1925
Internal changes:
1926

    
1927
- Enhanced both the unittest and the QA coverage
1928
- Switched the opcode validation to a generic model, and extended the
1929
  validation to all opcode parameters
1930
- Changed more parts of the code that write shell scripts to use the
1931
  same class for this
1932
- Switched the master daemon to use the asyncore library for the Luxi
1933
  server endpoint
1934

    
1935

    
1936
Version 2.2.0 beta0
1937
-------------------
1938

    
1939
*(Released Thu, 17 Jun 2010)*
1940

    
1941
- Added tool (``move-instance``) and infrastructure to move instances
1942
  between separate clusters (see :doc:`separate documentation
1943
  <move-instance>` and :doc:`design document <design-2.2>`)
1944
- Added per-request RPC timeout
1945
- RAPI now requires a Content-Type header for requests with a body (e.g.
1946
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1947
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
1948
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1949
  reachable
1950
- Implemented initial support for running Ganeti daemons as separate
1951
  users, see configure-time flags ``--with-user-prefix`` and
1952
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1953
  time)
1954
- Instances can be removed after export (``gnt-backup export
1955
  --remove-instance``)
1956
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1957
  key (instead of 1024 bit)
1958
- Added new cluster configuration file for cluster domain secret
1959
- Import/export now use SSL instead of SSH
1960
- Added support for showing estimated time when exporting an instance,
1961
  see the ``ganeti-os-interface(7)`` manpage and look for
1962
  ``EXP_SIZE_FD``
1963

    
1964

    
1965
Version 2.1.8
1966
-------------
1967

    
1968
*(Released Tue, 16 Nov 2010)*
1969

    
1970
Some more bugfixes. Unless critical bugs occur, this will be the last
1971
2.1 release:
1972

    
1973
- Fix case of MAC special-values
1974
- Fix mac checker regex
1975
- backend: Fix typo causing "out of range" error
1976
- Add missing --units in gnt-instance list man page
1977

    
1978

    
1979
Version 2.1.7
1980
-------------
1981

    
1982
*(Released Tue, 24 Aug 2010)*
1983

    
1984
Bugfixes only:
1985
  - Don't ignore secondary node silently on non-mirrored disk templates
1986
    (issue 113)
1987
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1988
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1989
  - Properly document the usb_mouse parameter
1990
  - Fix path in ganeti-rapi(8) (issue 116)
1991
  - Adjust error message when the ganeti user's .ssh directory is
1992
    missing
1993
  - Add same-node-check when changing the disk template to drbd
1994

    
1995

    
1996
Version 2.1.6
1997
-------------
1998

    
1999
*(Released Fri, 16 Jul 2010)*
2000

    
2001
Bugfixes only:
2002
  - Add an option to only select some reboot types during qa/burnin.
2003
    (on some hypervisors consequent reboots are not supported)
2004
  - Fix infrequent race condition in master failover. Sometimes the old
2005
    master ip address would be still detected as up for a short time
2006
    after it was removed, causing failover to fail.
2007
  - Decrease mlockall warnings when the ctypes module is missing. On
2008
    Python 2.4 we support running even if no ctypes module is installed,
2009
    but we were too verbose about this issue.
2010
  - Fix building on old distributions, on which man doesn't have a
2011
    --warnings option.
2012
  - Fix RAPI not to ignore the MAC address on instance creation
2013
  - Implement the old instance creation format in the RAPI client.
2014

    
2015

    
2016
Version 2.1.5
2017
-------------
2018

    
2019
*(Released Thu, 01 Jul 2010)*
2020

    
2021
A small bugfix release:
2022
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2023
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2024
    a non-existing option
2025
  - Fix instance create: the --force-variant option was ignored
2026
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2027
  - Fix modify node storage with non-FQDN arguments
2028
  - Fix RAPI client to authenticate under Python 2.6 when used
2029
    for more than 5 requests needing authentication
2030
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2031
    when converting a non-shutdown drbd instance to plain
2032

    
2033

    
2034
Version 2.1.4
2035
-------------
2036

    
2037
*(Released Fri, 18 Jun 2010)*
2038

    
2039
A small bugfix release:
2040

    
2041
  - Fix live migration of KVM instances started with older Ganeti
2042
    versions which had fewer hypervisor parameters
2043
  - Fix gnt-instance grow-disk on down instances
2044
  - Fix an error-reporting bug during instance migration
2045
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2046
    silently ignoring broken ones
2047
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2048
    users
2049
  - Fix bug triggered by different API version os-es on different nodes
2050
  - Fix a bug in instance startup with custom hvparams: OS level
2051
    parameters would fail to be applied.
2052
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2053
    make it work completely well with OpenSSL)
2054
  - Fix handling of errors when resolving names from DNS
2055

    
2056

    
2057
Version 2.1.3
2058
-------------
2059

    
2060
*(Released Thu, 3 Jun 2010)*
2061

    
2062
A medium sized development cycle. Some new features, and some
2063
fixes/small improvements/cleanups.
2064

    
2065
Significant features
2066
~~~~~~~~~~~~~~~~~~~~
2067

    
2068
The node deamon now tries to mlock itself into memory, unless the
2069
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2070
its logs, and falls back to console logging. This allows emergency
2071
features such as ``gnt-node powercycle`` to work even in the event of a
2072
broken node disk (tested offlining the disk hosting the node's
2073
filesystem and dropping its memory caches; don't try this at home)
2074

    
2075
KVM: add vhost-net acceleration support. It can be tested with a new
2076
enough version of the kernel and of qemu-kvm.
2077

    
2078
KVM: Add instance chrooting feature. If you use privilege dropping for
2079
your VMs you can also now force them to chroot to an empty directory,
2080
before starting the emulated guest.
2081

    
2082
KVM: Add maximum migration bandwith and maximum downtime tweaking
2083
support (requires a new-enough version of qemu-kvm).
2084

    
2085
Cluster verify will now warn if the master node doesn't have the master
2086
ip configured on it.
2087

    
2088
Add a new (incompatible) instance creation request format to RAPI which
2089
supports all parameters (previously only a subset was supported, and it
2090
wasn't possible to extend the old format to accomodate all the new
2091
features. The old format is still supported, and a client can check for
2092
this feature, before using it, by checking for its presence in the
2093
``features`` RAPI resource.
2094

    
2095
Now with ancient latin support. Try it passing the ``--roman`` option to
2096
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2097
(requires the python-roman module to be installed, in order to work).
2098

    
2099
Other changes
2100
~~~~~~~~~~~~~
2101

    
2102
As usual many internal code refactorings, documentation updates, and
2103
such. Among others:
2104

    
2105
  - Lots of improvements and cleanups to the experimental Remote API
2106
    (RAPI) client library.
2107
  - A new unit test suite for the core daemon libraries.
2108
  - A fix to creating missing directories makes sure the umask is not
2109
    applied anymore. This enforces the same directory permissions
2110
    everywhere.
2111
  - Better handling terminating daemons with ctrl+c (used when running
2112
    them in debugging mode).
2113
  - Fix a race condition in live migrating a KVM instance, when stat()
2114
    on the old proc status file returned EINVAL, which is an unexpected
2115
    value.
2116
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2117
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2118

    
2119

    
2120
Version 2.1.2.1
2121
---------------
2122

    
2123
*(Released Fri, 7 May 2010)*
2124

    
2125
Fix a bug which prevented untagged KVM instances from starting.
2126

    
2127

    
2128
Version 2.1.2
2129
-------------
2130

    
2131
*(Released Fri, 7 May 2010)*
2132

    
2133
Another release with a long development cycle, during which many
2134
different features were added.
2135

    
2136
Significant features
2137
~~~~~~~~~~~~~~~~~~~~
2138

    
2139
The KVM hypervisor now can run the individual instances as non-root, to
2140
reduce the impact of a VM being hijacked due to bugs in the
2141
hypervisor. It is possible to run all instances as a single (non-root)
2142
user, to manually specify a user for each instance, or to dynamically
2143
allocate a user out of a cluster-wide pool to each instance, with the
2144
guarantee that no two instances will run under the same user ID on any
2145
given node.
2146

    
2147
An experimental RAPI client library, that can be used standalone
2148
(without the other Ganeti libraries), is provided in the source tree as
2149
``lib/rapi/client.py``. Note this client might change its interface in
2150
the future, as we iterate on its capabilities.
2151

    
2152
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2153
replace the cluster's certificates and crypto keys. This might help in
2154
case they have been compromised, or have simply expired.
2155

    
2156
A new disk option for instance creation has been added that allows one
2157
to "adopt" currently existing logical volumes, with data
2158
preservation. This should allow easier migration to Ganeti from
2159
unmanaged (or managed via other software) instances.
2160

    
2161
Another disk improvement is the possibility to convert between redundant
2162
(DRBD) and plain (LVM) disk configuration for an instance. This should
2163
allow better scalability (starting with one node and growing the
2164
cluster, or shrinking a two-node cluster to one node).
2165

    
2166
A new feature that could help with automated node failovers has been
2167
implemented: if a node sees itself as offline (by querying the master
2168
candidates), it will try to shutdown (hard) all instances and any active
2169
DRBD devices. This reduces the risk of duplicate instances if an
2170
external script automatically failovers the instances on such nodes. To
2171
enable this, the cluster parameter ``maintain_node_health`` should be
2172
enabled; in the future this option (per the name) will enable other
2173
automatic maintenance features.
2174

    
2175
Instance export/import now will reuse the original instance
2176
specifications for all parameters; that means exporting an instance,
2177
deleting it and the importing it back should give an almost identical
2178
instance. Note that the default import behaviour has changed from
2179
before, where it created only one NIC; now it recreates the original
2180
number of NICs.
2181

    
2182
Cluster verify has added a few new checks: SSL certificates validity,
2183
/etc/hosts consistency across the cluster, etc.
2184

    
2185
Other changes
2186
~~~~~~~~~~~~~
2187

    
2188
As usual, many internal changes were done, documentation fixes,
2189
etc. Among others:
2190

    
2191
- Fixed cluster initialization with disabled cluster storage (regression
2192
  introduced in 2.1.1)
2193
- File-based storage supports growing the disks
2194
- Fixed behaviour of node role changes
2195
- Fixed cluster verify for some corner cases, plus a general rewrite of
2196
  cluster verify to allow future extension with more checks
2197
- Fixed log spamming by watcher and node daemon (regression introduced
2198
  in 2.1.1)
2199
- Fixed possible validation issues when changing the list of enabled
2200
  hypervisors
2201
- Fixed cleanup of /etc/hosts during node removal
2202
- Fixed RAPI response for invalid methods
2203
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2204
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2205
  from ide disks, etc.)
2206
- Allow OS changes without re-installation (to record a changed OS
2207
  outside of Ganeti, or to allow OS renames)
2208
- Allow instance creation without OS installation (useful for example if
2209
  the OS will be installed manually, or restored from a backup not in
2210
  Ganeti format)
2211
- Implemented option to make cluster ``copyfile`` use the replication
2212
  network
2213
- Added list of enabled hypervisors to ssconf (possibly useful for
2214
  external scripts)
2215
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2216
  1.2 clusters
2217
- A partial form of node re-IP is possible via node readd, which now
2218
  allows changed node primary IP
2219
- Command line utilities now show an informational message if the job is
2220
  waiting for a lock
2221
- The logs of the master daemon now show the PID/UID/GID of the
2222
  connected client
2223

    
2224

    
2225
Version 2.1.1
2226
-------------
2227

    
2228
*(Released Fri, 12 Mar 2010)*
2229

    
2230
During the 2.1.0 long release candidate cycle, a lot of improvements and
2231
changes have accumulated with were released later as 2.1.1.
2232

    
2233
Major changes
2234
~~~~~~~~~~~~~
2235

    
2236
The node evacuate command (``gnt-node evacuate``) was significantly
2237
rewritten, and as such the IAllocator protocol was changed - a new
2238
request type has been added. This unfortunate change during a stable
2239
series is designed to improve performance of node evacuations; on
2240
clusters with more than about five nodes and which are well-balanced,
2241
evacuation should proceed in parallel for all instances of the node
2242
being evacuated. As such, any existing IAllocator scripts need to be
2243
updated, otherwise the above command will fail due to the unknown
2244
request. The provided "dumb" allocator has not been updated; but the
2245
ganeti-htools package supports the new protocol since version 0.2.4.
2246

    
2247
Another important change is increased validation of node and instance
2248
names. This might create problems in special cases, if invalid host
2249
names are being used.
2250

    
2251
Also, a new layer of hypervisor parameters has been added, that sits at
2252
OS level between the cluster defaults and the instance ones. This allows
2253
customisation of virtualization parameters depending on the installed
2254
OS. For example instances with OS 'X' may have a different KVM kernel
2255
(or any other parameter) than the cluster defaults. This is intended to
2256
help managing a multiple OSes on the same cluster, without manual
2257
modification of each instance's parameters.
2258

    
2259
A tool for merging clusters, ``cluster-merge``, has been added in the
2260
tools sub-directory.
2261

    
2262
Bug fixes
2263
~~~~~~~~~
2264

    
2265
- Improved the int/float conversions that should make the code more
2266
  robust in face of errors from the node daemons
2267
- Fixed the remove node code in case of internal configuration errors
2268
- Fixed the node daemon behaviour in face of inconsistent queue
2269
  directory (e.g. read-only file-system where we can't open the files
2270
  read-write, etc.)
2271
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2272
  now it either aborts cleanly or, if given the new "auto_promote"
2273
  parameter, will automatically promote other nodes as needed
2274
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2275
  completely prevent Ganeti from working
2276
- Fixed bug for instance export when not all disks were successfully
2277
  exported
2278
- Fixed behaviour of node add when the new node is slow in starting up
2279
  the node daemon
2280
- Fixed handling of signals in the LUXI client, which should improve
2281
  behaviour of command-line scripts
2282
- Added checks for invalid node/instance names in the configuration (now
2283
  flagged during cluster verify)
2284
- Fixed watcher behaviour for disk activation errors
2285
- Fixed two potentially endless loops in http library, which led to the
2286
  RAPI daemon hanging and consuming 100% CPU in some cases
2287
- Fixed bug in RAPI daemon related to hashed passwords
2288
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2289
  instances
2290
- Enhanced compatibility with non-Debian OSes, but not using absolute
2291
  path in some commands and allowing customisation of the ssh
2292
  configuration directory
2293
- Fixed possible future issue with new Python versions by abiding to the
2294
  proper use of ``__slots__`` attribute on classes
2295
- Added checks that should prevent directory traversal attacks
2296
- Many documentation fixes based on feedback from users
2297

    
2298
New features
2299
~~~~~~~~~~~~
2300

    
2301
- Added an "early_release" more for instance replace disks and node
2302
  evacuate, where we release locks earlier and thus allow higher
2303
  parallelism within the cluster
2304
- Added watcher hooks, intended to allow the watcher to restart other
2305
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2306
  course for any other purpose
2307
- Added a compile-time disable for DRBD barriers, to increase
2308
  performance if the administrator trusts the power supply or the
2309
  storage system to not lose writes
2310
- Added the option of using syslog for logging instead of, or in
2311
  addition to, Ganeti's own log files
2312
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2313
  can indeed boot from a paravirtual NIC
2314
- Added a generic debug level for many operations; while this is not
2315
  used widely yet, it allows one to pass the debug value all the way to
2316
  the OS scripts
2317
- Enhanced the hooks environment for instance moves (failovers,
2318
  migrations) where the primary/secondary nodes changed during the
2319
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2320
- Enhanced data validations for many user-supplied values; one important
2321
  item is the restrictions imposed on instance and node names, which
2322
  might reject some (invalid) host names
2323
- Add a configure-time option to disable file-based storage, if it's not
2324
  needed; this allows greater security separation between the master
2325
  node and the other nodes from the point of view of the inter-node RPC
2326
  protocol
2327
- Added user notification in interactive tools if job is waiting in the
2328
  job queue or trying to acquire locks
2329
- Added log messages when a job is waiting for locks
2330
- Added filtering by node tags in instance operations which admit
2331
  multiple instances (start, stop, reboot, reinstall)
2332
- Added a new tool for cluster mergers, ``cluster-merge``
2333
- Parameters from command line which are of the form ``a=b,c=d`` can now
2334
  use backslash escapes to pass in values which contain commas,
2335
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2336
  ``b,c``
2337
- For KVM, the instance name is the first parameter passed to KVM, so
2338
  that it's more visible in the process list
2339

    
2340

    
2341
Version 2.1.0
2342
-------------
2343

    
2344
*(Released Tue, 2 Mar 2010)*
2345

    
2346
Ganeti 2.1 brings many improvements with it. Major changes:
2347

    
2348
- Added infrastructure to ease automated disk repairs
2349
- Added new daemon to export configuration data in a cheaper way than
2350
  using the remote API
2351
- Instance NICs can now be routed instead of being associated with a
2352
  networking bridge
2353
- Improved job locking logic to reduce impact of jobs acquiring multiple
2354
  locks waiting for other long-running jobs
2355

    
2356
In-depth implementation details can be found in the Ganeti 2.1 design
2357
document.
2358

    
2359
Details
2360
~~~~~~~
2361

    
2362
- Added chroot hypervisor
2363
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2364
  ``device_model``)
2365
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2366
  ``bootloader_path`` and ``bootloader_args``)
2367
- Added the ``use_localtime`` option for the xen-hvm and kvm
2368
  hypervisors, and the default value for this has changed to false (in
2369
  2.0 xen-hvm always enabled it)
2370
- Added luxi call to submit multiple jobs in one go
2371
- Added cluster initialization option to not modify ``/etc/hosts``
2372
  file on nodes
2373
- Added network interface parameters
2374
- Added dry run mode to some LUs
2375
- Added RAPI resources:
2376

    
2377
  - ``/2/instances/[instance_name]/info``
2378
  - ``/2/instances/[instance_name]/replace-disks``
2379
  - ``/2/nodes/[node_name]/evacuate``
2380
  - ``/2/nodes/[node_name]/migrate``
2381
  - ``/2/nodes/[node_name]/role``
2382
  - ``/2/nodes/[node_name]/storage``
2383
  - ``/2/nodes/[node_name]/storage/modify``
2384
  - ``/2/nodes/[node_name]/storage/repair``
2385

    
2386
- Added OpCodes to evacuate or migrate all instances on a node
2387
- Added new command to list storage elements on nodes (``gnt-node
2388
  list-storage``) and modify them (``gnt-node modify-storage``)
2389
- Added new ssconf files with master candidate IP address
2390
  (``ssconf_master_candidates_ips``), node primary IP address
2391
  (``ssconf_node_primary_ips``) and node secondary IP address
2392
  (``ssconf_node_secondary_ips``)
2393
- Added ``ganeti-confd`` and a client library to query the Ganeti
2394
  configuration via UDP
2395
- Added ability to run hooks after cluster initialization and before
2396
  cluster destruction
2397
- Added automatic mode for disk replace (``gnt-instance replace-disks
2398
  --auto``)
2399
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2400
  after catastrophic data-loss
2401
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2402
  groups
2403
- Added ``gnt-instance move`` command to move instances
2404
- Added ``gnt-cluster watcher`` command to control watcher
2405
- Added ``gnt-node powercycle`` command to powercycle nodes
2406
- Added new job status field ``lock_status``
2407
- Added parseable error codes to cluster verification (``gnt-cluster
2408
  verify --error-codes``) and made output less verbose (use
2409
  ``--verbose`` to restore previous behaviour)
2410
- Added UUIDs to the main config entities (cluster, nodes, instances)
2411
- Added support for OS variants
2412
- Added support for hashed passwords in the Ganeti remote API users file
2413
  (``rapi_users``)
2414
- Added option to specify maximum timeout on instance shutdown
2415
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2416
- Added new helper script to start and stop Ganeti daemons
2417
  (``daemon-util``), with the intent to reduce the work necessary to
2418
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2419
  from one place
2420
- Added more unittests
2421
- Fixed critical bug in ganeti-masterd startup
2422
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2423
  now customisable at the cluster level for both the KVM and Xen
2424
  hypervisors using the new ``migration_port`` parameter
2425
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2426
  reinstalling an instance
2427
- Allowed ``@`` in tag names
2428
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2429
- Many documentation updates
2430
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2431
- ``gnt-instance reinstall`` can now reinstall multiple instances
2432
- Updated many command line parameters
2433
- Introduced new OS API version 15
2434
- No longer support a default hypervisor
2435
- Treat virtual LVs as inexistent
2436
- Improved job locking logic to reduce lock contention
2437
- Match instance and node names case insensitively
2438
- Reimplemented bash completion script to be more complete
2439
- Improved burnin
2440

    
2441

    
2442
Version 2.0.6
2443
-------------
2444

    
2445
*(Released Thu, 4 Feb 2010)*
2446

    
2447
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2448
- Fix a string formatting bug
2449
- Improve safety of the code in some error paths
2450
- Improve data validation in the master of values returned from nodes
2451

    
2452

    
2453
Version 2.0.5
2454
-------------
2455

    
2456
*(Released Thu, 17 Dec 2009)*
2457

    
2458
- Fix security issue due to missing validation of iallocator names; this
2459
  allows local and remote execution of arbitrary executables
2460
- Fix failure of gnt-node list during instance removal
2461
- Ship the RAPI documentation in the archive
2462

    
2463

    
2464
Version 2.0.4
2465
-------------
2466

    
2467
*(Released Wed, 30 Sep 2009)*
2468

    
2469
- Fixed many wrong messages
2470
- Fixed a few bugs related to the locking library
2471
- Fixed MAC checking at instance creation time
2472
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2473
- Fixed a few issues related to signal handling in both daemons and
2474
  scripts
2475
- Fixed the example startup script provided
2476
- Fixed insserv dependencies in the example startup script (patch from
2477
  Debian)
2478
- Fixed handling of drained nodes in the iallocator framework
2479
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2480
  #528618)
2481
- Fixed error related to invalid job IDs in job polling
2482
- Fixed job/opcode persistence on unclean master shutdown
2483
- Fixed handling of partial job processing after unclean master
2484
  shutdown
2485
- Fixed error reporting from LUs, previously all errors were converted
2486
  into execution errors
2487
- Fixed error reporting from burnin
2488
- Decreased significantly the memory usage of the job queue
2489
- Optimised slightly multi-job submission
2490
- Optimised slightly opcode loading
2491
- Backported the multi-job submit framework from the development
2492
  branch; multi-instance start and stop should be faster
2493
- Added script to clean archived jobs after 21 days; this will reduce
2494
  the size of the queue directory
2495
- Added some extra checks in disk size tracking
2496
- Added an example ethers hook script
2497
- Added a cluster parameter that prevents Ganeti from modifying of
2498
  /etc/hosts
2499
- Added more node information to RAPI responses
2500
- Added a ``gnt-job watch`` command that allows following the ouput of a
2501
  job
2502
- Added a bind-address option to ganeti-rapi
2503
- Added more checks to the configuration verify
2504
- Enhanced the burnin script such that some operations can be retried
2505
  automatically
2506
- Converted instance reinstall to multi-instance model
2507

    
2508

    
2509
Version 2.0.3
2510
-------------
2511

    
2512
*(Released Fri, 7 Aug 2009)*
2513

    
2514
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2515
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2516
  instances have mismatched disk sizes in the configuration
2517
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2518
  any configuration mismatches for disk sizes
2519
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2520
  failover to work on two-node clusters
2521
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2522
  unusable
2523
- Fixed detection of OS script errors in ``gnt-backup export``
2524
- Fixed exit code of ``gnt-backup export``
2525

    
2526

    
2527
Version 2.0.2
2528
-------------
2529

    
2530
*(Released Fri, 17 Jul 2009)*
2531

    
2532
- Added experimental support for stripped logical volumes; this should
2533
  enhance performance but comes with a higher complexity in the block
2534
  device handling; stripping is only enabled when passing
2535
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2536
  affected even in the non-stripped mode
2537
- Improved resiliency against transient failures at the end of DRBD
2538
  resyncs, and in general of DRBD resync checks
2539
- Fixed a couple of issues with exports and snapshot errors
2540
- Fixed a couple of issues in instance listing
2541
- Added display of the disk size in ``gnt-instance info``
2542
- Fixed checking for valid OSes in instance creation
2543
- Fixed handling of the "vcpus" parameter in instance listing and in
2544
  general of invalid parameters
2545
- Fixed http server library, and thus RAPI, to handle invalid
2546
  username/password combinations correctly; this means that now they
2547
  report unauthorized for queries too, not only for modifications,
2548
  allowing earlier detect of configuration problems
2549
- Added a new "role" node list field, equivalent to the master/master
2550
  candidate/drained/offline flags combinations
2551
- Fixed cluster modify and changes of candidate pool size
2552
- Fixed cluster verify error messages for wrong files on regular nodes
2553
- Fixed a couple of issues with node demotion from master candidate role
2554
- Fixed node readd issues
2555
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2556
- Added a new ``--no-voting`` option for masterfailover to fix failover
2557
  on two-nodes clusters when the former master node is unreachable
2558
- Added instance reinstall over RAPI
2559

    
2560

    
2561
Version 2.0.1
2562
-------------
2563

    
2564
*(Released Tue, 16 Jun 2009)*
2565

    
2566
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2567
  allow re-adding the start in single-user option (regression from 1.2)
2568
- the watcher writes the instance status to a file, to allow monitoring
2569
  to report the instance status (from the master) based on cached
2570
  results of the watcher's queries; while this can get stale if the
2571
  watcher is being locked due to other work on the cluster, this is
2572
  still an improvement
2573
- the watcher now also restarts the node daemon and the rapi daemon if
2574
  they died
2575
- fixed the watcher to handle full and drained queue cases
2576
- hooks export more instance data in the environment, which helps if
2577
  hook scripts need to take action based on the instance's properties
2578
  (no longer need to query back into ganeti)
2579
- instance failovers when the instance is stopped do not check for free
2580
  RAM, so that failing over a stopped instance is possible in low memory
2581
  situations
2582
- rapi uses queries for tags instead of jobs (for less job traffic), and
2583
  for cluster tags it won't talk to masterd at all but read them from
2584
  ssconf
2585
- a couple of error handling fixes in RAPI
2586
- drbd handling: improved the error handling of inconsistent disks after
2587
  resync to reduce the frequency of "there are some degraded disks for
2588
  this instance" messages
2589
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2590
  error handling path called a wrong function name)
2591

    
2592

    
2593
Version 2.0.0
2594
-------------
2595

    
2596
*(Released Wed, 27 May 2009)*
2597

    
2598
- no changes from rc5
2599

    
2600

    
2601
Version 2.0 rc5
2602
---------------
2603

    
2604
*(Released Wed, 20 May 2009)*
2605

    
2606
- fix a couple of bugs (validation, argument checks)
2607
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2608
- some small improvements to RAPI and IAllocator
2609
- make watcher automatically start the master daemon if down
2610

    
2611

    
2612
Version 2.0 rc4
2613
---------------
2614

    
2615
*(Released Mon, 27 Apr 2009)*
2616

    
2617
- change the OS list to not require locks; this helps with big clusters
2618
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2619
  volume group is broken
2620
- ``gnt-instance info``, without any arguments, doesn't run for all
2621
  instances anymore; either pass ``--all`` or pass the desired
2622
  instances; this helps against mistakes on big clusters where listing
2623
  the information for all instances takes a long time
2624
- miscellaneous doc and man pages fixes
2625

    
2626

    
2627
Version 2.0 rc3
2628
---------------
2629

    
2630
*(Released Wed, 8 Apr 2009)*
2631

    
2632
- Change the internal locking model of some ``gnt-node`` commands, in
2633
  order to reduce contention (and blocking of master daemon) when
2634
  batching many creation/reinstall jobs
2635
- Fixes to Xen soft reboot
2636
- No longer build documentation at build time, instead distribute it in
2637
  the archive, in order to reduce the need for the whole docbook/rst
2638
  toolchains
2639

    
2640

    
2641
Version 2.0 rc2
2642
---------------
2643

    
2644
*(Released Fri, 27 Mar 2009)*
2645

    
2646
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2647
- Fix watcher startup sequence, improves the behaviour of busy clusters
2648
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2649
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2650
  bind address and other places
2651
- Some documentation fixes and updates
2652

    
2653

    
2654
Version 2.0 rc1
2655
---------------
2656

    
2657
*(Released Mon, 2 Mar 2009)*
2658

    
2659
- More documentation updates, now all docs should be more-or-less
2660
  up-to-date
2661
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2662
  etc.)
2663
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2664
  and KVM)
2665
- Fix an issue related to $libdir/run/ganeti and cluster creation
2666

    
2667

    
2668
Version 2.0 beta2
2669
-----------------
2670

    
2671
*(Released Thu, 19 Feb 2009)*
2672

    
2673
- Xen PVM and KVM have switched the default value for the instance root
2674
  disk to the first partition on the first drive, instead of the whole
2675
  drive; this means that the OS installation scripts must be changed
2676
  accordingly
2677
- Man pages have been updated
2678
- RAPI has been switched by default to HTTPS, and the exported functions
2679
  should all work correctly
2680
- RAPI v1 has been removed
2681
- Many improvements to the KVM hypervisor
2682
- Block device errors are now better reported
2683
- Many other bugfixes and small improvements
2684

    
2685

    
2686
Version 2.0 beta1
2687
-----------------
2688

    
2689
*(Released Mon, 26 Jan 2009)*
2690

    
2691
- Version 2 is a general rewrite of the code and therefore the
2692
  differences are too many to list, see the design document for 2.0 in
2693
  the ``doc/`` subdirectory for more details
2694
- In this beta version there is not yet a migration path from 1.2 (there
2695
  will be one in the final 2.0 release)
2696
- A few significant changes are:
2697

    
2698
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2699
    various ``gnt-*`` commands are just front-ends to it
2700
  - all the commands are entered into, and executed from a job queue,
2701
    see the ``gnt-job(8)`` manpage
2702
  - the RAPI daemon supports read-write operations, secured by basic
2703
    HTTP authentication on top of HTTPS
2704
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2705
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2706
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2707
  - DRBD devices are using statically allocated minor numbers, which
2708
    will be assigned to existing instances during the migration process
2709
  - there is support for both Xen PVM and Xen HVM instances running on
2710
    the same cluster
2711
  - KVM virtualization is supported too
2712
  - file-based storage has been implemented, which means that it is
2713
    possible to run the cluster without LVM and DRBD storage, for
2714
    example using a shared filesystem exported from shared storage (and
2715
    still have live migration)
2716

    
2717

    
2718
Version 1.2.7
2719
-------------
2720

    
2721
*(Released Tue, 13 Jan 2009)*
2722

    
2723
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2724
- Reuse the old instance mac address by default on instance import, if
2725
  the instance name is the same.
2726
- Handle situations in which the node info rpc returns incomplete
2727
  results (issue 46)
2728
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2729
- Improved version of batcher:
2730

    
2731
  - state file support
2732
  - instance mac address support
2733
  - support for HVM clusters/instances
2734

    
2735
- Add an option to show the number of cpu sockets and nodes in
2736
  ``gnt-node list``
2737
- Support OSes that handle more than one version of the OS api (but do
2738
  not change the current API in any other way)
2739
- Fix ``gnt-node migrate``
2740
- ``gnt-debug`` man page
2741
- Fixes various more typos and small issues
2742
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2743

    
2744

    
2745
Version 1.2.6
2746
-------------
2747

    
2748
*(Released Wed, 24 Sep 2008)*
2749

    
2750
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2751
  type of disk exported to fully virtualized instances.
2752
- provide access to the serial console of HVM instances
2753
- instance auto_balance flag, set by default. If turned off it will
2754
  avoid warnings on cluster verify if there is not enough memory to fail
2755
  over an instance. in the future it will prevent automatically failing
2756
  it over when we will support that.
2757
- batcher tool for instance creation, see ``tools/README.batcher``
2758
- ``gnt-instance reinstall --select-os`` to interactively select a new
2759
  operating system when reinstalling an instance.
2760
- when changing the memory amount on instance modify a check has been
2761
  added that the instance will be able to start. also warnings are
2762
  emitted if the instance will not be able to fail over, if auto_balance
2763
  is true.
2764
- documentation fixes
2765
- sync fields between ``gnt-instance list/modify/add/import``
2766
- fix a race condition in drbd when the sync speed was set after giving
2767
  the device a remote peer.
2768

    
2769

    
2770
Version 1.2.5
2771
-------------
2772

    
2773
*(Released Tue, 22 Jul 2008)*
2774

    
2775
- note: the allowed size and number of tags per object were reduced
2776
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2777
- fixed twisted 8.x compatibility
2778
- fixed ``gnt-instance replace-disks`` with iallocator
2779
- add TCP keepalives on twisted connections to detect restarted nodes
2780
- disk increase support, see ``gnt-instance grow-disk``
2781
- implement bulk node/instance query for RAPI
2782
- add tags in node/instance listing (optional)
2783
- experimental migration (and live migration) support, read the man page
2784
  for ``gnt-instance migrate``
2785
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2786
  has some small improvements in handling its state file
2787

    
2788

    
2789
Version 1.2.4
2790
-------------
2791

    
2792
*(Released Fri, 13 Jun 2008)*
2793

    
2794
- Experimental readonly, REST-based remote API implementation;
2795
  automatically started on master node, TCP port 5080, if enabled by
2796
  ``--enable-rapi`` parameter to configure script.
2797
- Instance allocator support. Add and import instance accept a
2798
  ``--iallocator`` parameter, and call that instance allocator to decide
2799
  which node to use for the instance. The iallocator document describes
2800
  what's expected from an allocator script.
2801
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2802
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2803
  node is lost there is still enough memory to fail over the instances
2804
  that reside on it.
2805
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2806
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2807
  hooks will run, and their output, if any, will be displayed. Any
2808
  failing hook will make the verification return an error value.
2809
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2810
  primary and secondary interfaces
2811
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2812
  that is still declared as part of the cluster and has failed.
2813
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2814
  specifying output fields, that just adds the chosen fields to the
2815
  default ones.
2816
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2817
  export from the filesystem.
2818
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2819
  have been added. Using them you can enable/disable acpi and pae
2820
  support, and specify a path for a cd image to be exported to the
2821
  instance. These parameters as the name suggest only work on HVM
2822
  clusters.
2823
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2824
  PAE support will be set to the previously hardcoded values, but the
2825
  (previously hardcoded) path to the CDROM ISO image will be unset and
2826
  if required, needs to be set manually with ``gnt-instance modify``
2827
  after the upgrade.
2828
- The address to which an instance's VNC console is bound is now
2829
  selectable per-instance, rather than being cluster wide. Of course
2830
  this only applies to instances controlled via VNC, so currently just
2831
  applies to HVM clusters.
2832

    
2833

    
2834
Version 1.2.3
2835
-------------
2836

    
2837
*(Released Mon, 18 Feb 2008)*
2838

    
2839
- more tweaks to the disk activation code (especially helpful for DRBD)
2840
- change the default ``gnt-instance list`` output format, now there is
2841
  one combined status field (see the manpage for the exact values this
2842
  field will have)
2843
- some more fixes for the mac export to hooks change
2844
- make Ganeti not break with DRBD 8.2.x (which changed the version
2845
  format in ``/proc/drbd``) (issue 24)
2846
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2847
  template, allowing migration from DRBD0.7+MD to DRBD8
2848

    
2849

    
2850
Version 1.2.2
2851
-------------
2852

    
2853
*(Released Wed, 30 Jan 2008)*
2854

    
2855
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2856
  support (issue 23)
2857
- add command aliases infrastructure and a few aliases
2858
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2859
  man pages and the ``--help`` option of ``gnt-node
2860
  list``/``gnt-instance list``
2861
- fix ``gnt-backup list`` with down nodes (issue 21)
2862
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2863
- fix the dist archive and add a check for including svn/git files in
2864
  the future
2865
- some developer-related changes: improve the burnin and the QA suite,
2866
  add an upload script for testing during development
2867

    
2868

    
2869
Version 1.2.1
2870
-------------
2871

    
2872
*(Released Wed, 16 Jan 2008)*
2873

    
2874
- experimental HVM support, read the install document, section
2875
  "Initializing the cluster"
2876
- allow for the PVM hypervisor per-instance kernel and initrd paths
2877
- add a new command ``gnt-cluster verify-disks`` which uses a new
2878
  algorithm to improve the reconnection of the DRBD pairs if the device
2879
  on the secondary node has gone away
2880
- make logical volume code auto-activate LVs at disk activation time
2881
- slightly improve the speed of activating disks
2882
- allow specification of the MAC address at instance creation time, and
2883
  changing it later via ``gnt-instance modify``
2884
- fix handling of external commands that generate lots of output on
2885
  stderr
2886
- update documentation with regard to minimum version of DRBD8 supported
2887

    
2888

    
2889
Version 1.2.0
2890
-------------
2891

    
2892
*(Released Tue, 4 Dec 2007)*
2893

    
2894
- Log the ``xm create`` output to the node daemon log on failure (to
2895
  help diagnosing the error)
2896
- In debug mode, log all external commands output if failed to the logs
2897
- Change parsing of lvm commands to ignore stderr
2898

    
2899

    
2900
Version 1.2 beta3
2901
-----------------
2902

    
2903
*(Released Wed, 28 Nov 2007)*
2904

    
2905
- Another round of updates to the DRBD 8 code to deal with more failures
2906
  in the replace secondary node operation
2907
- Some more logging of failures in disk operations (lvm, drbd)
2908
- A few documentation updates
2909
- QA updates
2910

    
2911

    
2912
Version 1.2 beta2
2913
-----------------
2914

    
2915
*(Released Tue, 13 Nov 2007)*
2916

    
2917
- Change configuration file format from Python's Pickle to JSON.
2918
  Upgrading is possible using the cfgupgrade utility.
2919
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2920
  faster replace disks and is more stable (DRBD 8 has many improvements
2921
  compared to DRBD 0.7)
2922
- Added command line tags support (see man pages for ``gnt-instance``,
2923
  ``gnt-node``, ``gnt-cluster``)
2924
- Added instance rename support
2925
- Added multi-instance startup/shutdown
2926
- Added cluster rename support
2927
- Added ``gnt-node evacuate`` to simplify some node operations
2928
- Added instance reboot operation that can speedup reboot as compared to
2929
  stop and start
2930
- Soften the requirement that hostnames are in FQDN format
2931
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
2932
  reboots
2933
- Removed dependency on debian's patched fping that uses the
2934
  non-standard ``-S`` option
2935
- Now the OS definitions are searched for in multiple, configurable
2936
  paths (easier for distros to package)
2937
- Some changes to the hooks infrastructure (especially the new
2938
  post-configuration update hook)
2939
- Other small bugfixes
2940

    
2941
.. vim: set textwidth=72 syntax=rst :
2942
.. Local Variables:
2943
.. mode: rst
2944
.. fill-column: 72
2945
.. End: