Statistics
| Branch: | Tag: | Revision:

root / NEWS @ e6e4ff4c

History | View | Annotate | Download (111.7 kB)

1
News
2
====
3

    
4

    
5
Version 2.8.4
6
-------------
7

    
8
*(unreleased)*
9

    
10
- Change the list separator for the usb_devices parameter from comma to space.
11
  Commas could not work because they are already the hypervisor option
12
  separator (Issue 649)
13

    
14

    
15
Version 2.8.3
16
-------------
17

    
18
*(Released Thu, 12 Dec 2013)*
19

    
20
- Fixed Luxi daemon socket permissions after master-failover
21
- Improve IP version detection code directly checking for colons rather than
22
  passing the family from the cluster object
23
- Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
24
  opportunistically anymore (Issue 622)
25
- Allow link local IPv6 gateways (Issue 624)
26
- Fix error printing (Issue 616)
27
- Fix a bug in InstanceSetParams concerning names: in case no name is passed in
28
  disk modifications, keep the old one. If name=none then set disk name to
29
  None.
30
- Update build_chroot script to work with the latest hackage packages
31
- Add a packet number limit to "fping" in master-ip-setup (Issue 630)
32
- Fix evacuation out of drained node (Issue 615)
33
- Add default file_driver if missing (Issue 571)
34
- Fix job error message after unclean master shutdown (Issue 618)
35
- Lock group(s) when creating instances (Issue 621)
36
- SetDiskID() before accepting an instance (Issue 633)
37
- Allow the ext template disks to receive arbitrary parameters, both at creation
38
  time and while being modified
39
- Xen handle domain shutdown (future proofing cherry-pick)
40
- Refactor reading live data in htools (future proofing cherry-pick)
41

    
42

    
43
Version 2.8.2
44
-------------
45

    
46
*(Released Thu, 07 Nov 2013)*
47

    
48
- DRBD: ensure peers are UpToDate for dual-primary
49
- Improve error message for replace-disks
50
- More dependency checks at configure time
51
- Placate warnings on ganeti.outils_unittest.py
52

    
53

    
54
Version 2.8.1
55
-------------
56

    
57
*(Released Thu, 17 Oct 2013)*
58

    
59
- Correctly start/stop luxid during gnt-cluster master-failover
60
- Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
61
- Fix path for the job queue serial file
62
- Improved harep man page
63
- Minor documentation improvements
64

    
65

    
66
Version 2.8.0
67
-------------
68

    
69
*(Released Mon, 30 Sep 2013)*
70

    
71
Incompatible/important changes
72
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
73

    
74
- Instance policy can contain multiple instance specs, as described in
75
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
76
  <design-partitioned>`. As a consequence, it's not possible to partially change
77
  or override instance specs. Bounding specs (min and max) can be specified as a
78
  whole using the new option ``--ipolicy-bounds-specs``, while standard
79
  specs use the new option ``--ipolicy-std-specs``.
80
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
81
  gnt-instance is a valid YAML object.
82
- hail now honors network restrictions when allocating nodes. This led to an
83
  update of the IAllocator protocol. See the IAllocator documentation for
84
  details.
85
- confd now only answers static configuration request over the network. luxid
86
  was extracted, listens on the local LUXI socket and responds to live queries.
87
  This allows finer grained permissions if using separate users.
88

    
89
New features
90
~~~~~~~~~~~~
91

    
92
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
93
  to always require authentication, ``--require-authentication``. It can
94
  be specified in ``$sysconfdir/default/ganeti``.
95
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
96
  be used to manage the disk templates to be used by instances in the cluster.
97
  Initially, it will be set to a list that includes plain, drbd, if they were
98
  enabled by specifying a volume group name, and file and sharedfile, if those
99
  were enabled at configure time. Additionally, it will include all disk
100
  templates that are currently used by instances. The order of disk templates
101
  will be based on Ganeti's history of supporting them. In the future, the
102
  first entry of the list will be used as a default disk template on instance
103
  creation.
104
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
105
  configuration back to the previous stable version.
106
- Disk templates in group ipolicy can be restored to the default value.
107
- Initial support for diskless instances and virtual clusters in QA.
108
- More QA and unit tests for instance policies.
109
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
110
  describing why the opcode itself was executed.
111
- The monitoring daemon is now available. It allows users to query the cluster
112
  for obtaining information about the status of the system. The daemon is only
113
  responsible for providing the information over the network: the actual data
114
  gathering is performed by data collectors (currently, only the DRBD status
115
  collector is available).
116
- In order to help developers work on Ganeti, a new script
117
  (``devel/build_chroot``) is provided, for building a chroot that contains all
118
  the required development libraries and tools for compiling Ganeti on a Debian
119
  Squeeze system.
120
- A new tool, ``harep``, for performing self-repair and recreation of instances
121
  in Ganeti has been added.
122
- Split queries are enabled for tags, network, exports, cluster info, groups,
123
  jobs, nodes.
124
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
125
  It prints the command line to set the current policies, to ease
126
  changing them.
127
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
128
  devices for KVM virtio-net interfaces will get created with VNET_HDR
129
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
130
  virtio-net interfaces, which prevents host kernel tainting and log
131
  flooding, when dealing with broken or malicious virtio-net drivers.
132
  It's set to true by default.
133
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
134
  failures.
135
- Support 'viridian' parameter in Xen HVM
136
- Support DSA SSH keys in bootstrap
137
- To simplify the work of packaging frameworks that want to add the needed users
138
  and groups in a split-user setup themselves, at build time three files in
139
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
140
  the groups to be generated, the ``users`` file contains, one per line, the
141
  users to be generated, optionally followed by their primary group, where
142
  important. The ``groupmemberships`` file contains, one per line, additional
143
  user-group membership relations that need to be established. The syntax of
144
  these files will remain stable in all future versions.
145

    
146

    
147
New dependencies
148
~~~~~~~~~~~~~~~~
149
The following new dependencies have been added:
150

    
151
For Haskell:
152
- The ``curl`` library is not optional anymore for compiling the Haskell code.
153
- ``snap-server`` library (if monitoring is enabled).
154

    
155
For Python:
156
- The minimum Python version needed to run Ganeti is now 2.6.
157
- ``yaml`` library (only for running the QA).
158

    
159
Since 2.8.0 rc3
160
~~~~~~~~~~~~~~~
161
- Perform proper cleanup on termination of Haskell daemons
162
- Fix corner-case in handling of remaining retry time
163

    
164

    
165
Version 2.8.0 rc3
166
-----------------
167

    
168
*(Released Tue, 17 Sep 2013)*
169

    
170
- To simplify the work of packaging frameworks that want to add the needed users
171
  and groups in a split-user setup themselves, at build time three files in
172
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
173
  the groups to be generated, the ``users`` file contains, one per line, the
174
  users to be generated, optionally followed by their primary group, where
175
  important. The ``groupmemberships`` file contains, one per line, additional
176
  user-group membership relations that need to be established. The syntax of
177
  these files will remain stable in all future versions.
178
- Add a default to file-driver when unspecified over RAPI (Issue 571)
179
- Mark the DSA host pubkey as optional, and remove it during config downgrade
180
  (Issue 560)
181
- Some documentation fixes
182

    
183

    
184
Version 2.8.0 rc2
185
-----------------
186

    
187
*(Released Tue, 27 Aug 2013)*
188

    
189
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
190

    
191
- Support 'viridian' parameter in Xen HVM (Issue 233)
192
- Include VCS version in ``gnt-cluster version``
193
- Support DSA SSH keys in bootstrap (Issue 338)
194
- Fix batch creation of instances
195
- Use FQDN to check master node status (Issue 551)
196
- Make the DRBD collector more failure-resilient
197

    
198

    
199
Version 2.8.0 rc1
200
-----------------
201

    
202
*(Released Fri, 2 Aug 2013)*
203

    
204
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
205

    
206
- Fix upgrading/downgrading from 2.7
207
- Increase maximum RAPI message size
208
- Documentation updates
209
- Split ``confd`` between ``luxid`` and ``confd``
210
- Merge 2.7 series up to the 2.7.1 release
211
- Allow the ``modify_etc_hosts`` option to be changed
212
- Add better debugging for ``luxid`` queries
213
- Expose bulk parameter for GetJobs in RAPI client
214
- Expose missing ``network`` fields in RAPI
215
- Add some ``cluster verify`` tests
216
- Some unittest fixes
217
- Fix a malfunction in ``hspace``'s tiered allocation
218
- Fix query compatibility between haskell and python implementations
219
- Add the ``vnet_hdr`` HV parameter for KVM
220
- Add ``--cleanup`` to instance failover
221
- Change the connected groups format in ``gnt-network info`` output; it
222
  was previously displayed as a raw list by mistake. (Merged from 2.7)
223

    
224

    
225
Version 2.8.0 beta1
226
-------------------
227

    
228
*(Released Mon, 24 Jun 2013)*
229

    
230
This was the first beta release of the 2.8 series. All important changes
231
are listed in the latest 2.8 entry.
232

    
233

    
234
Version 2.7.2
235
-------------
236

    
237
*(Released Thu, 26 Sep 2013)*
238

    
239
- Change the connected groups format in ``gnt-network info`` output; it
240
  was previously displayed as a raw list by mistake
241
- Check disk template in right dict when copying
242
- Support multi-instance allocs without iallocator
243
- Fix some errors in the documentation
244
- Fix formatting of tuple in an error message
245

    
246

    
247
Version 2.7.1
248
-------------
249

    
250
*(Released Thu, 25 Jul 2013)*
251

    
252
- Add logrotate functionality in daemon-util
253
- Add logrotate example file
254
- Add missing fields to network queries over rapi
255
- Fix network object timestamps
256
- Add support for querying network timestamps
257
- Fix a typo in the example crontab
258
- Fix a documentation typo
259

    
260

    
261
Version 2.7.0
262
-------------
263

    
264
*(Released Thu, 04 Jul 2013)*
265

    
266
Incompatible/important changes
267
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
268

    
269
- Instance policies for disk size were documented to be on a per-disk
270
  basis, but hail applied them to the sum of all disks. This has been
271
  fixed.
272
- ``hbal`` will now exit with status 0 if, during job execution over
273
  LUXI, early exit has been requested and all jobs are successful;
274
  before, exit status 1 was used, which cannot be differentiated from
275
  "job error" case
276
- Compatibility with newer versions of rbd has been fixed
277
- ``gnt-instance batch-create`` has been changed to use the bulk create
278
  opcode from Ganeti. This lead to incompatible changes in the format of
279
  the JSON file. It's now not a custom dict anymore but a dict
280
  compatible with the ``OpInstanceCreate`` opcode.
281
- Parent directories for file storage need to be listed in
282
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
283
  write the file automatically based on old configuration values, but it
284
  can not distribute it across all nodes and the file contents should be
285
  verified. Use ``gnt-cluster copyfile
286
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
287
  upgraded. The reason for requiring this list of paths now is that
288
  before it would have been possible to inject new paths via RPC,
289
  allowing files to be created in arbitrary locations. The RPC protocol
290
  is protected using SSL/X.509 certificates, but as a design principle
291
  Ganeti does not permit arbitrary paths to be passed.
292
- The parsing of the variants file for OSes (see
293
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
294
  lines and comment lines (starting with ``#``) are ignored for better
295
  readability.
296
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
297
  longer available. ``gnt-node add`` now invokes a new tool on the
298
  destination node, named ``prepare-node-join``, to configure the SSH
299
  daemon. Paramiko is no longer necessary to configure nodes' SSH
300
  daemons via ``gnt-node add``.
301
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
302
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
303
  the flag is not reset after a master failover.
304
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
305
  is recommended. 2.6 will be mandatory from the 2.8 series.
306

    
307

    
308
New features
309
~~~~~~~~~~~~
310

    
311
- New network management functionality to support automatic allocation
312
  of IP addresses and managing of network parameters. See
313
  :manpage:`gnt-network(8)` for more details.
314
- New external storage backend, to allow managing arbitrary storage
315
  systems external to the cluster. See
316
  :manpage:`ganeti-extstorage-interface(7)`.
317
- New ``exclusive-storage`` node parameter added, restricted to
318
  nodegroup level. When it's set to true, physical disks are assigned in
319
  an exclusive fashion to instances, as documented in :doc:`Partitioned
320
  Ganeti <design-partitioned>`.  Currently, only instances using the
321
  ``plain`` disk template are supported.
322
- The KVM hypervisor has been updated with many new hypervisor
323
  parameters, including a generic one for passing arbitrary command line
324
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
325
  compatible up to qemu 1.4.
326
- A new tool, called ``mon-collector``, is the stand-alone executor of
327
  the data collectors for a monitoring system. As of this version, it
328
  just includes the DRBD data collector, that can be executed by calling
329
  ``mon-collector`` using the ``drbd`` parameter. See
330
  :manpage:`mon-collector(7)`.
331
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
332
  for RAPI users. It allows granting permissions to query for
333
  information to a specific user without giving
334
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
335
- A new tool named ``node-cleanup`` has been added. It cleans remains of
336
  a cluster from a machine by stopping all daemons, removing
337
  certificates and ssconf files. Unless the ``--no-backup`` option is
338
  given, copies of the certificates are made.
339
- Instance creations now support the use of opportunistic locking,
340
  potentially speeding up the (parallel) creation of multiple instances.
341
  This feature is currently only available via the :doc:`RAPI
342
  <rapi>` interface and when an instance allocator is used. If the
343
  ``opportunistic_locking`` parameter is set the opcode will try to
344
  acquire as many locks as possible, but will not wait for any locks
345
  held by other opcodes. If not enough resources can be found to
346
  allocate the instance, the temporary error code
347
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
348
  retried thereafter, with or without opportunistic locking.
349
- New experimental linux-ha resource scripts.
350
- Restricted-commands support: ganeti can now be asked (via command line
351
  or rapi) to perform commands on a node. These are passed via ganeti
352
  RPC rather than ssh. This functionality is restricted to commands
353
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
354
  reasons. The file is not copied automatically.
355

    
356

    
357
Misc changes
358
~~~~~~~~~~~~
359

    
360
- Diskless instances are now externally mirrored (Issue 237). This for
361
  now has only been tested in conjunction with explicit target nodes for
362
  migration/failover.
363
- Queries not needing locks or RPC access to the node can now be
364
  performed by the confd daemon, making them independent from jobs, and
365
  thus faster to execute. This is selectable at configure time.
366
- The functionality for allocating multiple instances at once has been
367
  overhauled and is now also available through :doc:`RAPI <rapi>`.
368

    
369
There are no significant changes from version 2.7.0~rc3.
370

    
371

    
372
Version 2.7.0 rc3
373
-----------------
374

    
375
*(Released Tue, 25 Jun 2013)*
376

    
377
- Fix permissions on the confd query socket (Issue 477)
378
- Fix permissions on the job archive dir (Issue 498)
379
- Fix handling of an internal exception in replace-disks (Issue 472)
380
- Fix gnt-node info handling of shortened names (Issue 497)
381
- Fix gnt-instance grow-disk when wiping is enabled
382
- Documentation improvements, and support for newer pandoc
383
- Fix hspace honoring ipolicy for disks (Issue 484)
384
- Improve handling of the ``kvm_extra`` HV parameter
385

    
386

    
387
Version 2.7.0 rc2
388
-----------------
389

    
390
*(Released Fri, 24 May 2013)*
391

    
392
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
393
  symlink.
394
- Disks added through ``gnt-instance modify`` or created through
395
  ``gnt-instance recreate-disks`` are wiped, if the
396
  ``prealloc_wipe_disks`` flag is set.
397
- If wiping newly created disks fails, the disks are removed. Also,
398
  partial failures in creating disks through ``gnt-instance modify``
399
  triggers a cleanup of the partially-created disks.
400
- Removing the master IP address doesn't fail if the address has been
401
  already removed.
402
- Fix ownership of the OS log dir
403
- Workaround missing SO_PEERCRED constant (Issue 191)
404

    
405

    
406
Version 2.7.0 rc1
407
-----------------
408

    
409
*(Released Fri, 3 May 2013)*
410

    
411
This was the first release candidate of the 2.7 series. Since beta3:
412

    
413
- Fix kvm compatibility with qemu 1.4 (Issue 389)
414
- Documentation updates (admin guide, upgrade notes, install
415
  instructions) (Issue 372)
416
- Fix gnt-group list nodes and instances count (Issue 436)
417
- Fix compilation without non-mandatory libraries (Issue 441)
418
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
419
- Make confd logging more verbose at INFO level (Issue 435)
420
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
421
- Fix failure path for instance storage type conversion (Issue 229)
422
- Update htools text backend documentation
423
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
424
- Disable inter-cluster instance move for file-based instances, because
425
  it is dependant on instance export, which is not supported for
426
  file-based instances. (Issue 414)
427
- Fix gnt-job crashes on non-ascii characters (Issue 427)
428
- Fix volume group checks on non-vm-capable nodes (Issue 432)
429

    
430

    
431
Version 2.7.0 beta3
432
-------------------
433

    
434
*(Released Mon, 22 Apr 2013)*
435

    
436
This was the third beta release of the 2.7 series. Since beta2:
437

    
438
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
439
- Fix data loss on wrong usage of ``gnt-instance move``
440
- Properly export errors in confd-based job queries
441
- Add ``users-setup`` tool
442
- Fix iallocator protocol to report 0 as a disk size for diskless
443
  instances. This avoids hail breaking when a diskless instance is
444
  present.
445
- Fix job queue directory permission problem that made confd job queries
446
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
447
  for access to archived jobs (Issue 406).
448
- Limit the sizes of networks supported by ``gnt-network`` to something
449
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
450
- Fix bugs in instance disk template conversion
451
- Fix GHC 7 compatibility
452
- Fix ``burnin`` install path (Issue 426).
453
- Allow very small disk grows (Issue 347).
454
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
455
  that noded doesn't import masterd code (Issue 419).
456
- Make sure the default metavg at cluster init is the same as the vg, if
457
  unspecified (Issue 358).
458
- Fix cleanup of partially created disks (part of Issue 416)
459

    
460

    
461
Version 2.7.0 beta2
462
-------------------
463

    
464
*(Released Tue, 2 Apr 2013)*
465

    
466
This was the second beta release of the 2.7 series. Since beta1:
467

    
468
- Networks no longer have a "type" slot, since this information was
469
  unused in Ganeti: instead of it tags should be used.
470
- The rapi client now has a ``target_node`` option to MigrateInstance.
471
- Fix early exit return code for hbal (Issue 386).
472
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
473
- Fix ``rbd showmapped`` output parsing (Issue 312).
474
- Networks are now referenced indexed by UUID, rather than name. This
475
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
476
  use.
477
- The OS environment now includes network information.
478
- Deleting of a network is now disallowed if any instance nic is using
479
  it, to prevent dangling references.
480
- External storage is now documented in man pages.
481
- The exclusive_storage flag can now only be set at nodegroup level.
482
- Hbal can now submit an explicit priority with its jobs.
483
- Many network related locking fixes.
484
- Bump up the required pylint version to 0.25.1.
485
- Fix the ``no_remember`` option in RAPI client.
486
- Many ipolicy related tests, qa, and fixes.
487
- Many documentation improvements and fixes.
488
- Fix building with ``--disable-file-storage``.
489
- Fix ``-q`` option in htools, which was broken if passed more than
490
  once.
491
- Some haskell/python interaction improvements and fixes.
492
- Fix iallocator in case of missing LVM storage.
493
- Fix confd config load in case of ``--no-lvm-storage``.
494
- The confd/query functionality is now mentioned in the security
495
  documentation.
496

    
497

    
498
Version 2.7.0 beta1
499
-------------------
500

    
501
*(Released Wed, 6 Feb 2013)*
502

    
503
This was the first beta release of the 2.7 series. All important changes
504
are listed in the latest 2.7 entry.
505

    
506

    
507
Version 2.6.2
508
-------------
509

    
510
*(Released Fri, 21 Dec 2012)*
511

    
512
Important behaviour change: hbal won't rebalance anymore instances which
513
have the ``auto_balance`` attribute set to false. This was the intention
514
all along, but until now it only skipped those from the N+1 memory
515
reservation (DRBD-specific).
516

    
517
A significant number of bug fixes in this release:
518

    
519
- Fixed disk adoption interaction with ipolicy checks.
520
- Fixed networking issues when instances are started, stopped or
521
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
522
- Fixed the warning in cluster verify for shared storage instances not
523
  being redundant.
524
- Fixed removal of storage directory on shared file storage (issue 262).
525
- Fixed validation of LVM volume group name in OpClusterSetParams
526
  (``gnt-cluster modify``) (issue 285).
527
- Fixed runtime memory increases (``gnt-instance modify -m``).
528
- Fixed live migration under Xen's ``xl`` mode.
529
- Fixed ``gnt-instance console`` with ``xl``.
530
- Fixed building with newer Haskell compiler/libraries.
531
- Fixed PID file writing in Haskell daemons (confd); this prevents
532
  restart issues if confd was launched manually (outside of
533
  ``daemon-util``) while another copy of it was running
534
- Fixed a type error when doing live migrations with KVM (issue 297) and
535
  the error messages for failing migrations have been improved.
536
- Fixed opcode validation for the out-of-band commands (``gnt-node
537
  power``).
538
- Fixed a type error when unsetting OS hypervisor parameters (issue
539
  311); now it's possible to unset all OS-specific hypervisor
540
  parameters.
541
- Fixed the ``dry-run`` mode for many operations: verification of
542
  results was over-zealous but didn't take into account the ``dry-run``
543
  operation, resulting in "wrong" failures.
544
- Fixed bash completion in ``gnt-job list`` when the job queue has
545
  hundreds of entries; especially with older ``bash`` versions, this
546
  results in significant CPU usage.
547

    
548
And lastly, a few other improvements have been made:
549

    
550
- Added option to force master-failover without voting (issue 282).
551
- Clarified error message on lock conflict (issue 287).
552
- Logging of newly submitted jobs has been improved (issue 290).
553
- Hostname checks have been made uniform between instance rename and
554
  create (issue 291).
555
- The ``--submit`` option is now supported by ``gnt-debug delay``.
556
- Shutting down the master daemon by sending SIGTERM now stops it from
557
  processing jobs waiting for locks; instead, those jobs will be started
558
  once again after the master daemon is started the next time (issue
559
  296).
560
- Support for Xen's ``xl`` program has been improved (besides the fixes
561
  above).
562
- Reduced logging noise in the Haskell confd daemon (only show one log
563
  entry for each config reload, instead of two).
564
- Several man page updates and typo fixes.
565

    
566

    
567
Version 2.6.1
568
-------------
569

    
570
*(Released Fri, 12 Oct 2012)*
571

    
572
A small bugfix release. Among the bugs fixed:
573

    
574
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
575
  made the command unusable.
576
- Commands that issue many jobs don't fail anymore just because some jobs
577
  take so long that other jobs are archived.
578
- Failures during ``gnt-instance reinstall`` are reflected by the exit
579
  status.
580
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
581
  DRBD is enabled.
582
- When ``always_failover`` is set, ``--allow-failover`` is not required
583
  in migrate commands anymore.
584
- ``bash_completion`` works even if extglob is disabled.
585
- Fixed bug with locks that made failover for RDB-based instances fail.
586
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
587
  a random node instead of one based on the allocator metric.
588
- Support for newer versions of pylint and pep8.
589
- Hail doesn't fail anymore when trying to add an instance of type
590
  ``file``, ``sharedfile`` or ``rbd``.
591
- Added new Makefile target to rebuild the whole distribution, so that
592
  all files are included.
593

    
594

    
595
Version 2.6.0
596
-------------
597

    
598
*(Released Fri, 27 Jul 2012)*
599

    
600

    
601
.. attention:: The ``LUXI`` protocol has been made more consistent
602
   regarding its handling of command arguments. This, however, leads to
603
   incompatibility issues with previous versions. Please ensure that you
604
   restart Ganeti daemons soon after the upgrade, otherwise most
605
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
606
   pausing/resuming the watcher, cancelling and archiving jobs, querying
607
   the cluster configuration) will fail.
608

    
609

    
610
New features
611
~~~~~~~~~~~~
612

    
613
Instance run status
614
+++++++++++++++++++
615

    
616
The current ``admin_up`` field, which used to denote whether an instance
617
should be running or not, has been removed. Instead, ``admin_state`` is
618
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
619

    
620
The rational behind this is that an instance being “down” can have
621
different meanings:
622

    
623
- it could be down during a reboot
624
- it could be temporarily be down for a reinstall
625
- or it could be down because it is deprecated and kept just for its
626
  disk
627

    
628
The previous Boolean state was making it difficult to do capacity
629
calculations: should Ganeti reserve memory for a down instance? Now, the
630
tri-state field makes it clear:
631

    
632
- in ``up`` and ``down`` state, all resources are reserved for the
633
  instance, and it can be at any time brought up if it is down
634
- in ``offline`` state, only disk space is reserved for it, but not
635
  memory or CPUs
636

    
637
The field can have an extra use: since the transition between ``up`` and
638
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
639
transition between ``offline`` and ``down`` is done via ``gnt-instance
640
modify``, it is possible to given different rights to users. For
641
example, owners of an instance could be allowed to start/stop it, but
642
not transition it out of the offline state.
643

    
644
Instance policies and specs
645
+++++++++++++++++++++++++++
646

    
647
In previous Ganeti versions, an instance creation request was not
648
limited on the minimum size and on the maximum size just by the cluster
649
resources. As such, any policy could be implemented only in third-party
650
clients (RAPI clients, or shell wrappers over ``gnt-*``
651
tools). Furthermore, calculating cluster capacity via ``hspace`` again
652
required external input with regards to instance sizes.
653

    
654
In order to improve these workflows and to allow for example better
655
per-node group differentiation, we introduced instance specs, which
656
allow declaring:
657

    
658
- minimum instance disk size, disk count, memory size, cpu count
659
- maximum values for the above metrics
660
- and “standard” values (used in ``hspace`` to calculate the standard
661
  sized instances)
662

    
663
The minimum/maximum values can be also customised at node-group level,
664
for example allowing more powerful hardware to support bigger instance
665
memory sizes.
666

    
667
Beside the instance specs, there are a few other settings belonging to
668
the instance policy framework. It is possible now to customise, per
669
cluster and node-group:
670

    
671
- the list of allowed disk templates
672
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
673
- the maximum ratio of instance to spindles (see below for more
674
  information) for local storage
675

    
676
All these together should allow all tools that talk to Ganeti to know
677
what are the ranges of allowed values for instances and the
678
over-subscription that is allowed.
679

    
680
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
681
instance configuration, and the physical CPU configuration from the
682
node. For the spindle ratios however, we didn't track before these
683
values, so new parameters have been added:
684

    
685
- a new node parameter ``spindle_count``, defaults to 1, customisable at
686
  node group or node level
687
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
688

    
689
Note that spindles in this context doesn't need to mean actual
690
mechanical hard-drives; it's just a relative number for both the node
691
I/O capacity and instance I/O consumption.
692

    
693
Instance migration behaviour
694
++++++++++++++++++++++++++++
695

    
696
While live-migration is in general desirable over failover, it is
697
possible that for some workloads it is actually worse, due to the
698
variable time of the “suspend” phase during live migration.
699

    
700
To allow the tools to work consistently over such instances (without
701
having to hard-code instance names), a new backend parameter
702
``always_failover`` has been added to control the migration/failover
703
behaviour. When set to True, all migration requests for an instance will
704
instead fall-back to failover.
705

    
706
Instance memory ballooning
707
++++++++++++++++++++++++++
708

    
709
Initial support for memory ballooning has been added. The memory for an
710
instance is no longer fixed (backend parameter ``memory``), but instead
711
can vary between minimum and maximum values (backend parameters
712
``minmem`` and ``maxmem``). Currently we only change an instance's
713
memory when:
714

    
715
- live migrating or failing over and instance and the target node
716
  doesn't have enough memory
717
- user requests changing the memory via ``gnt-instance modify
718
  --runtime-memory``
719

    
720
Instance CPU pinning
721
++++++++++++++++++++
722

    
723
In order to control the use of specific CPUs by instance, support for
724
controlling CPU pinning has been added for the Xen, HVM and LXC
725
hypervisors. This is controlled by a new hypervisor parameter
726
``cpu_mask``; details about possible values for this are in the
727
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
728
VCPU-to-CPU mapping) form will work well only when all nodes in your
729
cluster have the same amount of CPUs.
730

    
731
Disk parameters
732
+++++++++++++++
733

    
734
Another area in which Ganeti was not customisable were the parameters
735
used for storage configuration, e.g. how many stripes to use for LVM,
736
DRBD resync configuration, etc.
737

    
738
To improve this area, we've added disks parameters, which are
739
customisable at cluster and node group level, and which allow to
740
specify various parameters for disks (DRBD has the most parameters
741
currently), for example:
742

    
743
- DRBD resync algorithm and parameters (e.g. speed)
744
- the default VG for meta-data volumes for DRBD
745
- number of stripes for LVM (plain disk template)
746
- the RBD pool
747

    
748
These parameters can be modified via ``gnt-cluster modify -D …`` and
749
``gnt-group modify -D …``, and are used at either instance creation (in
750
case of LVM stripes, for example) or at disk “activation” time
751
(e.g. resync speed).
752

    
753
Rados block device support
754
++++++++++++++++++++++++++
755

    
756
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
757
denoted by the ``rbd`` disk template type. This is considered
758
experimental, feedback is welcome. For details on configuring it, see
759
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
760

    
761
Master IP setup
762
+++++++++++++++
763

    
764
The existing master IP functionality works well only in simple setups (a
765
single network shared by all nodes); however, if nodes belong to
766
different networks, then the ``/32`` setup and lack of routing
767
information is not enough.
768

    
769
To allow the master IP to function well in more complex cases, the
770
system was reworked as follows:
771

    
772
- a master IP netmask setting has been added
773
- the master IP activation/turn-down code was moved from the node daemon
774
  to a separate script
775
- whether to run the Ganeti-supplied master IP script or a user-supplied
776
  on is a ``gnt-cluster init`` setting
777

    
778
Details about the location of the standard and custom setup scripts are
779
in the man page :manpage:`gnt-cluster(8)`; for information about the
780
setup script protocol, look at the Ganeti-supplied script.
781

    
782
SPICE support
783
+++++++++++++
784

    
785
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
786
improved.
787

    
788
It is now possible to use TLS-protected connections, and when renewing
789
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
790
it is now possible to specify spice or spice CA certificates. Also, it
791
is possible to configure a password for SPICE sessions via the
792
hypervisor parameter ``spice_password_file``.
793

    
794
There are also new parameters to control the compression and streaming
795
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
796
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
797
for the spice parameters.
798

    
799
Lastly, it is now possible to see the SPICE connection information via
800
``gnt-instance console``.
801

    
802
OVF converter
803
+++++++++++++
804

    
805
A new tool (``tools/ovfconverter``) has been added that supports
806
conversion between Ganeti and the `Open Virtualization Format
807
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
808
from).
809

    
810
This relies on the ``qemu-img`` tool to convert the disk formats, so the
811
actual compatibility with other virtualization solutions depends on it.
812

    
813
Confd daemon changes
814
++++++++++++++++++++
815

    
816
The configuration query daemon (``ganeti-confd``) is now optional, and
817
has been rewritten in Haskell; whether to use the daemon at all, use the
818
Python (default) or the Haskell version is selectable at configure time
819
via the ``--enable-confd`` parameter, which can take one of the
820
``haskell``, ``python`` or ``no`` values. If not used, disabling the
821
daemon will result in a smaller footprint; for larger systems, we
822
welcome feedback on the Haskell version which might become the default
823
in future versions.
824

    
825
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
826
daemon running. The Python version doesn't implement the new call.
827

    
828

    
829
User interface changes
830
~~~~~~~~~~~~~~~~~~~~~~
831

    
832
We have replaced the ``--disks`` option of ``gnt-instance
833
replace-disks`` with a more flexible ``--disk`` option, which allows
834
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
835
disk size and mode can be changed upon recreation (via ``gnt-instance
836
recreate-disks``, which accepts the same ``--disk`` option).
837

    
838
As many people are used to a ``show`` command, we have added that as an
839
alias to ``info`` on all ``gnt-*`` commands.
840

    
841
The ``gnt-instance grow-disk`` command has a new mode in which it can
842
accept the target size of the disk, instead of the delta; this can be
843
more safe since two runs in absolute mode will be idempotent, and
844
sometimes it's also easier to specify the desired size directly.
845

    
846
Also the handling of instances with regard to offline secondaries has
847
been improved. Instance operations should not fail because one of it's
848
secondary nodes is offline, even though it's safe to proceed.
849

    
850
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
851
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
852
minors to instance name.
853

    
854
API changes
855
~~~~~~~~~~~
856

    
857
RAPI coverage has improved, with (for example) new resources for
858
recreate-disks, node power-cycle, etc.
859

    
860
Compatibility
861
~~~~~~~~~~~~~
862

    
863
There is partial support for ``xl`` in the Xen hypervisor; feedback is
864
welcome.
865

    
866
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
867
whether to still support Python 2.4 or move to Python 2.6 as minimum
868
required version.
869

    
870
Support for Fedora has been slightly improved; the provided example
871
init.d script should work better on it and the INSTALL file should
872
document the needed dependencies.
873

    
874
Internal changes
875
~~~~~~~~~~~~~~~~
876

    
877
The deprecated ``QueryLocks`` LUXI request has been removed. Use
878
``Query(what=QR_LOCK, ...)`` instead.
879

    
880
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
881
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
882
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
883
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
884
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
885

    
886
RAPI client: ``CertificateError`` now derives from
887
``GanetiApiError``. This should make it more easy to handle Ganeti
888
errors.
889

    
890
Deprecation warnings due to PyCrypto/paramiko import in
891
``tools/setup-ssh`` have been silenced, as usually they are safe; please
892
make sure to run an up-to-date paramiko version, if you use this tool.
893

    
894
The QA scripts now depend on Python 2.5 or above (the main code base
895
still works with Python 2.4).
896

    
897
The configuration file (``config.data``) is now written without
898
indentation for performance reasons; if you want to edit it, it can be
899
re-formatted via ``tools/fmtjson``.
900

    
901
A number of bugs has been fixed in the cluster merge tool.
902

    
903
``x509`` certification verification (used in import-export) has been
904
changed to allow the same clock skew as permitted by the cluster
905
verification. This will remove some rare but hard to diagnose errors in
906
import-export.
907

    
908

    
909
Version 2.6.0 rc4
910
-----------------
911

    
912
*(Released Thu, 19 Jul 2012)*
913

    
914
Very few changes from rc4 to the final release, only bugfixes:
915

    
916
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
917
  instance, fix CDROM booting for KVM instances)
918
- fixed node group modification of node parameters
919
- fixed issue in LUClusterVerifyGroup with multi-group clusters
920
- fixed generation of bash completion to ensure a stable ordering
921
- fixed a few typos
922

    
923

    
924
Version 2.6.0 rc3
925
-----------------
926

    
927
*(Released Fri, 13 Jul 2012)*
928

    
929
Third release candidate for 2.6. The following changes were done from
930
rc3 to rc4:
931

    
932
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
933
- Fixed an inconsistency in the LUXI protocol with the provided
934
  arguments (NOT backwards compatible)
935
- Fixed a bug with node groups ipolicy where ``min`` was greater than
936
  the cluster ``std`` value
937
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
938
  easier instance debugging on nodes (requires ``hconfd`` to work)
939

    
940

    
941
Version 2.6.0 rc2
942
-----------------
943

    
944
*(Released Tue, 03 Jul 2012)*
945

    
946
Second release candidate for 2.6. The following changes were done from
947
rc2 to rc3:
948

    
949
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
950
  master candidates
951
- Fixed a RAPI regression on missing beparams/memory
952
- Fixed redistribution of files on offline nodes
953
- Added possibility to run activate-disks even though secondaries are
954
  offline. With this change it relaxes also the strictness on some other
955
  commands which use activate disks internally:
956
  * ``gnt-instance start|reboot|rename|backup|export``
957
- Made it possible to remove safely an instance if its secondaries are
958
  offline
959
- Made it possible to reinstall even though secondaries are offline
960

    
961

    
962
Version 2.6.0 rc1
963
-----------------
964

    
965
*(Released Mon, 25 Jun 2012)*
966

    
967
First release candidate for 2.6. The following changes were done from
968
rc1 to rc2:
969

    
970
- Fixed bugs with disk parameters and ``rbd`` templates as well as
971
  ``instance_os_add``
972
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
973
  behaviour. It supports now the modify operation
974
- ``hcheck`` implemented to analyze cluster health and possibility of
975
  improving health by rebalance
976
- ``hbal`` has been improved in dealing with split instances
977

    
978

    
979
Version 2.6.0 beta2
980
-------------------
981

    
982
*(Released Mon, 11 Jun 2012)*
983

    
984
Second beta release of 2.6. The following changes were done from beta2
985
to rc1:
986

    
987
- Fixed ``daemon-util`` with non-root user models
988
- Fixed creation of plain instances with ``--no-wait-for-sync``
989
- Fix wrong iv_names when running ``cfgupgrade``
990
- Export more information in RAPI group queries
991
- Fixed bug when changing instance network interfaces
992
- Extended burnin to do NIC changes
993
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
994
- Changed default for DRBD barriers
995
- Fixed DRBD error reporting for syncer rate
996
- Verify the options on disk parameters
997

    
998
And of course various fixes to documentation and improved unittests and
999
QA.
1000

    
1001

    
1002
Version 2.6.0 beta1
1003
-------------------
1004

    
1005
*(Released Wed, 23 May 2012)*
1006

    
1007
First beta release of 2.6. The following changes were done from beta1 to
1008
beta2:
1009

    
1010
- integrated patch for distributions without ``start-stop-daemon``
1011
- adapted example init.d script to work on Fedora
1012
- fixed log handling in Haskell daemons
1013
- adapted checks in the watcher for pycurl linked against libnss
1014
- add partial support for ``xl`` instead of ``xm`` for Xen
1015
- fixed a type issue in cluster verification
1016
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1017
  clusters)
1018

    
1019
Plus integrated fixes from the 2.5 branch:
1020

    
1021
- fixed ``kvm-ifup`` to use ``/bin/bash``
1022
- fixed parallel build failures
1023
- KVM live migration when using a custom keymap
1024

    
1025

    
1026
Version 2.5.2
1027
-------------
1028

    
1029
*(Released Tue, 24 Jul 2012)*
1030

    
1031
A small bugfix release, with no new features:
1032

    
1033
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1034
  different default shell (e.g. Debian, Ubuntu)
1035
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1036
  243 and Debian bug #650664)
1037
- fixed compatibility with KVM versions that don't support multiple boot
1038
  devices (fixes Issue 230 and Debian bug #624256)
1039

    
1040
Additionally, a few fixes were done to the build system (fixed parallel
1041
build failures) and to the unittests (fixed race condition in test for
1042
FileID functions, and the default enable/disable mode for QA test is now
1043
customisable).
1044

    
1045

    
1046
Version 2.5.1
1047
-------------
1048

    
1049
*(Released Fri, 11 May 2012)*
1050

    
1051
A small bugfix release.
1052

    
1053
The main issues solved are on the topic of compatibility with newer LVM
1054
releases:
1055

    
1056
- fixed parsing of ``lv_attr`` field
1057
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1058
  the ``--force`` flag is needed
1059

    
1060
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1061
to accept kernel 3.x too (was hardcoded to 2.6.*).
1062

    
1063
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1064
of hook results and that also made display of other errors suboptimal
1065
was fixed; the code behaves now like 2.4 and earlier.
1066

    
1067
Another change in 2.5, the cleanup of the OS scripts environment, is too
1068
aggressive: it removed even the ``PATH`` variable, which requires the OS
1069
scripts to *always* need to export it. Since this is a bit too strict,
1070
we now export a minimal PATH, the same that we export for hooks.
1071

    
1072
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1073
integrated into this release.
1074

    
1075
Finally, a few other miscellaneous changes were done (no new features,
1076
just small improvements):
1077

    
1078
- Fix ``gnt-group --help`` display
1079
- Fix hardcoded Xen kernel path
1080
- Fix grow-disk handling of invalid units
1081
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1082
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1083

    
1084

    
1085
Version 2.5.0
1086
-------------
1087

    
1088
*(Released Thu, 12 Apr 2012)*
1089

    
1090
Incompatible/important changes and bugfixes
1091
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1092

    
1093
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1094
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1095
  to match the underlying LUXI interface.
1096
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1097
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1098
  not maintain backwards-compatibility as the underlying operation
1099
  changed in an incompatible way. The RAPI client can talk to old
1100
  servers, but it needs to be told so as the return value changed.
1101
- When creating file-based instances via RAPI, the ``file_driver``
1102
  parameter no longer defaults to ``loop`` and must be specified.
1103
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1104
  ``link`` instead.
1105
- Support for the undocumented and deprecated RAPI instance creation
1106
  request format version 0 has been dropped. Use version 1, supported
1107
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1108
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1109
  documentation <install>`.
1110
- The "cluster-verify" hooks are now executed per group by the
1111
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1112
  if you just run ``gnt-cluster verify``, which generates one opcode per
1113
  group.
1114
- The environment as passed to the OS scripts is cleared, and thus no
1115
  environment variables defined in the node daemon's environment will be
1116
  inherited by the scripts.
1117
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1118
  deprecated.
1119
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1120
  support operations involving multiple node groups.
1121
- Offline nodes are ignored when failing over an instance.
1122
- Support for KVM version 1.0, which changed the version reporting format
1123
  from 3 to 2 digits.
1124
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1125
  removal.
1126
- ``Makefile`` is now compatible with Automake 1.11.2
1127
- Includes all bugfixes made in the 2.4 series
1128

    
1129
New features
1130
~~~~~~~~~~~~
1131

    
1132
- The ganeti-htools project has been merged into the ganeti-core source
1133
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1134
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1135
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1136
  GPT-style partition tables (requires `parted
1137
  <http://www.gnu.org/s/parted/>`_).
1138
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1139
- Allowed adding tags on instance creation.
1140
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1141
- Allow instances to be started in a paused state, enabling the user to
1142
  see the complete console output on boot using the console.
1143
- Added new hypervisor flag to control default reboot behaviour
1144
  (``reboot_behavior``).
1145
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1146
- Improved out-of-band management support:
1147

    
1148
  - Added ``gnt-node health`` command reporting the health status of
1149
    nodes.
1150
  - Added ``gnt-node power`` command to manage power status of nodes.
1151
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1152

    
1153
- Instance migration can fall back to failover if instance is not
1154
  running.
1155
- Filters can be used when listing nodes, instances, groups and locks;
1156
  see :manpage:`ganeti(7)` manpage.
1157
- Added post-execution status as variables to :doc:`hooks <hooks>`
1158
  environment.
1159
- Instance tags are exported/imported together with the instance.
1160
- When given an explicit job ID, ``gnt-job info`` will work for archived
1161
  jobs.
1162
- Jobs can define dependencies on other jobs (not yet supported via
1163
  RAPI or command line, but used by internal commands and usable via
1164
  LUXI).
1165

    
1166
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1167
    dependencies.
1168

    
1169
- Instance failover is now available as a RAPI resource
1170
  (``/2/instances/[instance_name]/failover``).
1171
- ``gnt-instance info`` defaults to static information if primary node
1172
  is offline.
1173
- Opcodes have a new ``comment`` attribute.
1174
- Added basic SPICE support to KVM hypervisor.
1175
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1176

    
1177
Node group improvements
1178
~~~~~~~~~~~~~~~~~~~~~~~
1179

    
1180
- ``gnt-cluster verify`` has been modified to check groups separately,
1181
  thereby improving performance.
1182
- Node group support has been added to ``gnt-cluster verify-disks``,
1183
  which now operates per node group.
1184
- Watcher has been changed to work better with node groups.
1185

    
1186
  - One process and state file per node group.
1187
  - Slow watcher in one group doesn't block other group's watcher.
1188

    
1189
- Added new command, ``gnt-group evacuate``, to move all instances in a
1190
  node group to other groups.
1191
- Added ``gnt-instance change-group`` to move an instance to another
1192
  node group.
1193
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1194
  per-group operations.
1195
- Node groups can be tagged.
1196
- Some operations switch from an exclusive to a shared lock as soon as
1197
  possible.
1198
- Instance's primary and secondary nodes' groups are now available as
1199
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1200
  and ``snodes.group.uuid``).
1201

    
1202
Misc
1203
~~~~
1204

    
1205
- Numerous updates to documentation and manpages.
1206

    
1207
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1208
    descriptions.
1209
  - Some opcode/job results are now also documented, see :doc:`RAPI
1210
    <rapi>`.
1211

    
1212
- A lockset's internal lock is now also visible in lock monitor.
1213
- Log messages from job queue workers now contain information about the
1214
  opcode they're processing.
1215
- ``gnt-instance console`` no longer requires the instance lock.
1216
- A short delay when waiting for job changes reduces the number of LUXI
1217
  requests significantly.
1218
- DRBD metadata volumes are overwritten with zeros during disk creation.
1219
- Out-of-band commands no longer acquire the cluster lock in exclusive
1220
  mode.
1221
- ``devel/upload`` now uses correct permissions for directories.
1222

    
1223

    
1224
Version 2.5.0 rc6
1225
-----------------
1226

    
1227
*(Released Fri, 23 Mar 2012)*
1228

    
1229
This was the sixth release candidate of the 2.5 series.
1230

    
1231

    
1232
Version 2.5.0 rc5
1233
-----------------
1234

    
1235
*(Released Mon, 9 Jan 2012)*
1236

    
1237
This was the fifth release candidate of the 2.5 series.
1238

    
1239

    
1240
Version 2.5.0 rc4
1241
-----------------
1242

    
1243
*(Released Thu, 27 Oct 2011)*
1244

    
1245
This was the fourth release candidate of the 2.5 series.
1246

    
1247

    
1248
Version 2.5.0 rc3
1249
-----------------
1250

    
1251
*(Released Wed, 26 Oct 2011)*
1252

    
1253
This was the third release candidate of the 2.5 series.
1254

    
1255

    
1256
Version 2.5.0 rc2
1257
-----------------
1258

    
1259
*(Released Tue, 18 Oct 2011)*
1260

    
1261
This was the second release candidate of the 2.5 series.
1262

    
1263

    
1264
Version 2.5.0 rc1
1265
-----------------
1266

    
1267
*(Released Tue, 4 Oct 2011)*
1268

    
1269
This was the first release candidate of the 2.5 series.
1270

    
1271

    
1272
Version 2.5.0 beta3
1273
-------------------
1274

    
1275
*(Released Wed, 31 Aug 2011)*
1276

    
1277
This was the third beta release of the 2.5 series.
1278

    
1279

    
1280
Version 2.5.0 beta2
1281
-------------------
1282

    
1283
*(Released Mon, 22 Aug 2011)*
1284

    
1285
This was the second beta release of the 2.5 series.
1286

    
1287

    
1288
Version 2.5.0 beta1
1289
-------------------
1290

    
1291
*(Released Fri, 12 Aug 2011)*
1292

    
1293
This was the first beta release of the 2.5 series.
1294

    
1295

    
1296
Version 2.4.5
1297
-------------
1298

    
1299
*(Released Thu, 27 Oct 2011)*
1300

    
1301
- Fixed bug when parsing command line parameter values ending in
1302
  backslash
1303
- Fixed assertion error after unclean master shutdown
1304
- Disable HTTP client pool for RPC, significantly reducing memory usage
1305
  of master daemon
1306
- Fixed queue archive creation with wrong permissions
1307

    
1308

    
1309
Version 2.4.4
1310
-------------
1311

    
1312
*(Released Tue, 23 Aug 2011)*
1313

    
1314
Small bug-fixes:
1315

    
1316
- Fixed documentation for importing with ``--src-dir`` option
1317
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1318
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1319

    
1320

    
1321
Version 2.4.3
1322
-------------
1323

    
1324
*(Released Fri, 5 Aug 2011)*
1325

    
1326
Many bug-fixes and a few small features:
1327

    
1328
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1329
  issues when you tried to add an instance with two MAC addresses in one
1330
  request
1331
- KVM: fixed per-instance stored UID value
1332
- KVM: configure bridged NICs at migration start
1333
- KVM: Fix a bug where instance will not start with never KVM versions
1334
  (>= 0.14)
1335
- Added OS search path to ``gnt-cluster info``
1336
- Fixed an issue with ``file_storage_dir`` where you were forced to
1337
  provide an absolute path, but the documentation states it is a
1338
  relative path, the documentation was right
1339
- Added a new parameter to instance stop/start called ``--no-remember``
1340
  that will make the state change to not be remembered
1341
- Implemented ``no_remember`` at RAPI level
1342
- Improved the documentation
1343
- Node evacuation: don't call IAllocator if node is already empty
1344
- Fixed bug in DRBD8 replace disks on current nodes
1345
- Fixed bug in recreate-disks for DRBD instances
1346
- Moved assertion checking locks in ``gnt-instance replace-disks``
1347
  causing it to abort with not owning the right locks for some situation
1348
- Job queue: Fixed potential race condition when cancelling queued jobs
1349
- Fixed off-by-one bug in job serial generation
1350
- ``gnt-node volumes``: Fix instance names
1351
- Fixed aliases in bash completion
1352
- Fixed a bug in reopening log files after being sent a SIGHUP
1353
- Added a flag to burnin to allow specifying VCPU count
1354
- Bugfixes to non-root Ganeti configuration
1355

    
1356

    
1357
Version 2.4.2
1358
-------------
1359

    
1360
*(Released Thu, 12 May 2011)*
1361

    
1362
Many bug-fixes and a few new small features:
1363

    
1364
- Fixed a bug related to log opening failures
1365
- Fixed a bug in instance listing with orphan instances
1366
- Fixed a bug which prevented resetting the cluster-level node parameter
1367
  ``oob_program`` to the default
1368
- Many fixes related to the ``cluster-merge`` tool
1369
- Fixed a race condition in the lock monitor, which caused failures
1370
  during (at least) creation of many instances in parallel
1371
- Improved output for gnt-job info
1372
- Removed the quiet flag on some ssh calls which prevented debugging
1373
  failures
1374
- Improved the N+1 failure messages in cluster verify by actually
1375
  showing the memory values (needed and available)
1376
- Increased lock attempt timeouts so that when executing long operations
1377
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1378
  too early and thus prevent the use of the 'fair' mechanism
1379
- Changed instance query data (``gnt-instance info``) to not acquire
1380
  locks unless needed, thus allowing its use on locked instance if only
1381
  static information is asked for
1382
- Improved behaviour with filesystems that do not support rename on an
1383
  opened file
1384
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1385
  kept locks on all nodes during the wipe, which is unneeded
1386
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1387
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1388
  10GiB) which caused the wipe to fail right at the end in some cases
1389
- Fixed master IP activation when doing master failover with no-voting
1390
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1391
  the master node itself
1392
- Fixed potential data-loss in under disk full conditions, where Ganeti
1393
  wouldn't check correctly the return code and would consider
1394
  partially-written files 'correct'
1395
- Fixed bug related to multiple VGs and DRBD disk replacing
1396
- Added new disk parameter ``metavg`` that allows placement of the meta
1397
  device for DRBD in a different volume group
1398
- Fixed error handling in the node daemon when the system libc doesn't
1399
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1400
- Fixed lock release during replace-disks, which kept cluster-wide locks
1401
  when doing disk replaces with an iallocator script
1402
- Added check for missing bridges in cluster verify
1403
- Handle EPIPE errors while writing to the terminal better, so that
1404
  piping the output to e.g. ``less`` doesn't cause a backtrace
1405
- Fixed rare case where a ^C during Luxi calls could have been
1406
  interpreted as server errors, instead of simply terminating
1407
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1408
  assign-nodes``)
1409
- Added a few more parameters to the KVM hypervisor, allowing a second
1410
  CDROM, custom disk type for CDROMs and a floppy image
1411
- Removed redundant message in instance rename when the name is given
1412
  already as a FQDN
1413
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1414
  disks on new nodes, allowing recreation when the original instance
1415
  nodes are completely gone
1416
- Added option when converting disk templates to DRBD to skip waiting
1417
  for the resync, in order to make the instance available sooner
1418
- Added two new variables to the OS scripts environment (containing the
1419
  instance's nodes)
1420
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1421
  to allow use of ``pvgrub`` as bootloader
1422
- Changed the instance memory modifications to only check out-of-memory
1423
  conditions on memory increases, and turned the secondary node warnings
1424
  into errors (they can still be overridden via ``--force``)
1425
- Fixed the handling of a corner case when the Python installation gets
1426
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1427
  to execute a command that doesn't exist
1428
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1429
  node of the instance returned failures during instance shutdown; this
1430
  adds the option ``--ignore-consistency`` to gnt-instance move
1431

    
1432
And as usual, various improvements to the error messages, documentation
1433
and man pages.
1434

    
1435

    
1436
Version 2.4.1
1437
-------------
1438

    
1439
*(Released Wed, 09 Mar 2011)*
1440

    
1441
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1442
the RAPI users file if run twice (even with ``--dry-run``).
1443

    
1444
The release fixes that bug (nothing else changed).
1445

    
1446

    
1447
Version 2.4.0
1448
-------------
1449

    
1450
*(Released Mon, 07 Mar 2011)*
1451

    
1452
Final 2.4.0 release. Just a few small fixes:
1453

    
1454
- Fixed RAPI node evacuate
1455
- Fixed the kvm-ifup script
1456
- Fixed internal error handling for special job cases
1457
- Updated man page to specify the escaping feature for options
1458

    
1459

    
1460
Version 2.4.0 rc3
1461
-----------------
1462

    
1463
*(Released Mon, 28 Feb 2011)*
1464

    
1465
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1466
that this feature wiped the disks of the wrong instance, leading to loss
1467
of data.
1468

    
1469
Other changes:
1470

    
1471
- Fixed title of query field containing instance name
1472
- Expanded the glossary in the documentation
1473
- Fixed one unittest (internal issue)
1474

    
1475

    
1476
Version 2.4.0 rc2
1477
-----------------
1478

    
1479
*(Released Mon, 21 Feb 2011)*
1480

    
1481
A number of bug fixes plus just a couple functionality changes.
1482

    
1483
On the user-visible side, the ``gnt-* list`` command output has changed
1484
with respect to "special" field states. The current rc1 style of display
1485
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1486
the default output mode special fields states are displayed as follows:
1487

    
1488
- Offline resource: ``*``
1489
- Unavailable/not applicable: ``-``
1490
- Data missing (RPC failure): ``?``
1491
- Unknown field: ``??``
1492

    
1493
Another user-visible change is the addition of ``--force-join`` to
1494
``gnt-node add``.
1495

    
1496
As for bug fixes:
1497

    
1498
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1499
- Fixed regression in RAPI/instance reinstall where all parameters were
1500
  required (instead of optional)
1501
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1502
- Fixed iallocator usage (offline nodes were not considered offline)
1503
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1504
- Fixed hypervisor and OS parameter validation with respect to
1505
  non-vm_capable nodes
1506
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1507
  cosmetic)
1508
- Fixed ``tools/listrunner`` with respect to agent-based usage
1509

    
1510

    
1511
Version 2.4.0 rc1
1512
-----------------
1513

    
1514
*(Released Fri,  4 Feb 2011)*
1515

    
1516
Many changes and fixes since the beta1 release. While there were some
1517
internal changes, the code has been mostly stabilised for the RC
1518
release.
1519

    
1520
Note: the dumb allocator was removed in this release, as it was not kept
1521
up-to-date with the IAllocator protocol changes. It is recommended to
1522
use the ``hail`` command from the ganeti-htools package.
1523

    
1524
Note: the 2.4 and up versions of Ganeti are not compatible with the
1525
0.2.x branch of ganeti-htools. You need to upgrade to
1526
ganeti-htools-0.3.0 (or later).
1527

    
1528
Regressions fixed from 2.3
1529
~~~~~~~~~~~~~~~~~~~~~~~~~~
1530

    
1531
- Fixed the ``gnt-cluster verify-disks`` command
1532
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1533
  serially on nodes)
1534
- Fixed disk adoption breakage
1535
- Fixed wrong headers in instance listing for field aliases
1536

    
1537
Other bugs fixed
1538
~~~~~~~~~~~~~~~~
1539

    
1540
- Fixed corner case in KVM handling of NICs
1541
- Fixed many cases of wrong handling of non-vm_capable nodes
1542
- Fixed a bug where a missing instance symlink was not possible to
1543
  recreate with any ``gnt-*`` command (now ``gnt-instance
1544
  activate-disks`` does it)
1545
- Fixed the volume group name as reported by ``gnt-cluster
1546
  verify-disks``
1547
- Increased timeouts for the import-export code, hopefully leading to
1548
  fewer aborts due network or instance timeouts
1549
- Fixed bug in ``gnt-node list-storage``
1550
- Fixed bug where not all daemons were started on cluster
1551
  initialisation, but only at the first watcher run
1552
- Fixed many bugs in the OOB implementation
1553
- Fixed watcher behaviour in presence of instances with offline
1554
  secondaries
1555
- Fixed instance list output for instances running on the wrong node
1556
- a few fixes to the cluster-merge tool, but it still cannot merge
1557
  multi-node groups (currently it is not recommended to use this tool)
1558

    
1559

    
1560
Improvements
1561
~~~~~~~~~~~~
1562

    
1563
- Improved network configuration for the KVM hypervisor
1564
- Added e1000 as a supported NIC for Xen-HVM
1565
- Improved the lvmstrap tool to also be able to use partitions, as
1566
  opposed to full disks
1567
- Improved speed of disk wiping (the cluster parameter
1568
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1569
  of instance creations
1570
- Added documentation for the OS parameters
1571
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1572
  hypervisor is not responding
1573
- Added display of blacklisted and hidden OS information in
1574
  ``gnt-cluster info``
1575
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1576
  NIC and node parameters, which might create problems with currently
1577
  invalid (but undetected) configuration files, but prevents validation
1578
  failures when unrelated parameters are modified
1579
- Changed cluster initialisation to wait for the master daemon to become
1580
  available
1581
- Expanded the RAPI interface:
1582

    
1583
  - Added config redistribution resource
1584
  - Added activation/deactivation of instance disks
1585
  - Added export of console information
1586

    
1587
- Implemented log file reopening on SIGHUP, which allows using
1588
  logrotate(8) for the Ganeti log files
1589
- Added a basic OOB helper script as an example
1590

    
1591

    
1592
Version 2.4.0 beta1
1593
-------------------
1594

    
1595
*(Released Fri, 14 Jan 2011)*
1596

    
1597
User-visible
1598
~~~~~~~~~~~~
1599

    
1600
- Fixed timezone issues when formatting timestamps
1601
- Added support for node groups, available via ``gnt-group`` and other
1602
  commands
1603
- Added out-of-band framework and management, see :doc:`design
1604
  document <design-oob>`
1605
- Removed support for roman numbers from ``gnt-node list`` and
1606
  ``gnt-instance list``.
1607
- Allowed modification of master network interface via ``gnt-cluster
1608
  modify --master-netdev``
1609
- Accept offline secondaries while shutting down instance disks
1610
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1611
- Added support for multiple LVM volume groups
1612
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1613
  requested
1614
- Added commands to list available fields:
1615

    
1616
  - ``gnt-node list-fields``
1617
  - ``gnt-group list-fields``
1618
  - ``gnt-instance list-fields``
1619

    
1620
- Updated documentation and man pages
1621

    
1622
Integration
1623
~~~~~~~~~~~
1624

    
1625
- Moved ``rapi_users`` file into separate directory, now named
1626
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1627
  symlink
1628
- Added new tool for running commands on many machines,
1629
  ``tools/ganeti-listrunner``
1630
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1631
  improving the ``gnt-instance console`` output
1632
- Allowed customisation of disk index separator at ``configure`` time
1633
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1634
- Added support for non-partitioned md disks in ``lvmstrap``
1635
- Added script to gracefully power off KVM instances
1636
- Split ``utils`` module into smaller parts
1637
- Changed query operations to return more detailed information, e.g.
1638
  whether an information is unavailable due to an offline node. To use
1639
  this new functionality, the LUXI call ``Query`` must be used. Field
1640
  information is now stored by the master daemon and can be retrieved
1641
  using ``QueryFields``. Instances, nodes and groups can also be queried
1642
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1643
  exposed via RAPI). The following commands make use of this
1644
  infrastructure change:
1645

    
1646
  - ``gnt-group list``
1647
  - ``gnt-group list-fields``
1648
  - ``gnt-node list``
1649
  - ``gnt-node list-fields``
1650
  - ``gnt-instance list``
1651
  - ``gnt-instance list-fields``
1652
  - ``gnt-debug locks``
1653

    
1654
Remote API
1655
~~~~~~~~~~
1656

    
1657
- New RAPI resources (see :doc:`rapi`):
1658

    
1659
  - ``/2/modify``
1660
  - ``/2/groups``
1661
  - ``/2/groups/[group_name]``
1662
  - ``/2/groups/[group_name]/assign-nodes``
1663
  - ``/2/groups/[group_name]/modify``
1664
  - ``/2/groups/[group_name]/rename``
1665
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1666

    
1667
- RAPI changes:
1668

    
1669
  - Implemented ``no_install`` for instance creation
1670
  - Implemented OS parameters for instance reinstallation, allowing
1671
    use of special settings on reinstallation (e.g. for preserving data)
1672

    
1673
Misc
1674
~~~~
1675

    
1676
- Added IPv6 support in import/export
1677
- Pause DRBD synchronization while wiping disks on instance creation
1678
- Updated unittests and QA scripts
1679
- Improved network parameters passed to KVM
1680
- Converted man pages from docbook to reStructuredText
1681

    
1682

    
1683
Version 2.3.1
1684
-------------
1685

    
1686
*(Released Mon, 20 Dec 2010)*
1687

    
1688
Released version 2.3.1~rc1 without any changes.
1689

    
1690

    
1691
Version 2.3.1 rc1
1692
-----------------
1693

    
1694
*(Released Wed, 1 Dec 2010)*
1695

    
1696
- impexpd: Disable OpenSSL compression in socat if possible (backport
1697
  from master, commit e90739d625b, see :doc:`installation guide
1698
  <install-quick>` for details)
1699
- Changed unittest coverage report to exclude test scripts
1700
- Added script to check version format
1701

    
1702

    
1703
Version 2.3.0
1704
-------------
1705

    
1706
*(Released Wed, 1 Dec 2010)*
1707

    
1708
Released version 2.3.0~rc1 without any changes.
1709

    
1710

    
1711
Version 2.3.0 rc1
1712
-----------------
1713

    
1714
*(Released Fri, 19 Nov 2010)*
1715

    
1716
A number of bugfixes and documentation updates:
1717

    
1718
- Update ganeti-os-interface documentation
1719
- Fixed a bug related to duplicate MACs or similar items which should be
1720
  unique
1721
- Fix breakage in OS state modify
1722
- Reinstall instance: disallow offline secondaries (fixes bug related to
1723
  OS changing but reinstall failing)
1724
- plus all the other fixes between 2.2.1 and 2.2.2
1725

    
1726

    
1727
Version 2.3.0 rc0
1728
-----------------
1729

    
1730
*(Released Tue, 2 Nov 2010)*
1731

    
1732
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1733
- Fixed master failover race with watcher
1734
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1735
  configuration
1736
- Accept previously stopped instance for export with instance removal
1737
- Simplify and extend the environment variables for instance OS scripts
1738
- Added new node flags, ``master_capable`` and ``vm_capable``
1739
- Added optional instance disk wiping prior during allocation. This is a
1740
  cluster-wide option and can be set/modified using
1741
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1742
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1743
  :doc:`install-quick`
1744
- Added a new watcher option (``--ignore-pause``)
1745
- Added option to ignore offline node on instance start/stop
1746
  (``--ignore-offline``)
1747
- Allow overriding OS parameters with ``gnt-instance reinstall``
1748
- Added ability to change node's secondary IP address using ``gnt-node
1749
  modify``
1750
- Implemented privilege separation for all daemons except
1751
  ``ganeti-noded``, see ``configure`` options
1752
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1753
  verify``
1754
- Implemented job priorities (see ``ganeti(7)`` manpage)
1755
- Ignore failures while shutting down instances during failover from
1756
  offline node
1757
- Exit daemon's bootstrap process only once daemon is ready
1758
- Export more information via ``LUInstanceQuery``/remote API
1759
- Improved documentation, QA and unittests
1760
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1761
  restart if the file was created or changed
1762
- Added LUXI protocol version sent with each request and response,
1763
  allowing detection of server/client mismatches
1764
- Moved the Python scripts among gnt-* and ganeti-* into modules
1765
- Moved all code related to setting up SSH to an external script,
1766
  ``setup-ssh``
1767
- Infrastructure changes for node group support in future versions
1768

    
1769

    
1770
Version 2.2.2
1771
-------------
1772

    
1773
*(Released Fri, 19 Nov 2010)*
1774

    
1775
A few small bugs fixed, and some improvements to the build system:
1776

    
1777
- Fix documentation regarding conversion to drbd
1778
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1779
  -B``)
1780
- Fix error handling in node modify with multiple changes
1781
- Allow remote imports without checked names
1782

    
1783

    
1784
Version 2.2.1
1785
-------------
1786

    
1787
*(Released Tue, 19 Oct 2010)*
1788

    
1789
- Disable SSL session ID cache in RPC client
1790

    
1791

    
1792
Version 2.2.1 rc1
1793
-----------------
1794

    
1795
*(Released Thu, 14 Oct 2010)*
1796

    
1797
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1798
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1799
  with GnuTLS
1800
- Fix problems with interaction between Curl and Python's HTTP server,
1801
  resulting in increased speed in many RPC calls
1802
- Improve our release script to prevent breakage with older aclocal and
1803
  Python 2.6
1804

    
1805

    
1806
Version 2.2.1 rc0
1807
-----------------
1808

    
1809
*(Released Thu, 7 Oct 2010)*
1810

    
1811
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1812
  value retrieved from master
1813
- Added support for blacklisted or hidden OS definitions
1814
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1815
- Added support for -mem-path in KVM hypervisor abstraction layer
1816
- Allow overriding instance parameters in tool for inter-cluster
1817
  instance moves (``tools/move-instance``)
1818
- Improved opcode summaries (e.g. in ``gnt-job list``)
1819
- Improve consistency of OS listing by sorting it
1820
- Documentation updates
1821

    
1822

    
1823
Version 2.2.0.1
1824
---------------
1825

    
1826
*(Released Fri, 8 Oct 2010)*
1827

    
1828
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1829

    
1830

    
1831
Version 2.2.0
1832
-------------
1833

    
1834
*(Released Mon, 4 Oct 2010)*
1835

    
1836
- Fixed regression in ``gnt-instance rename``
1837

    
1838

    
1839
Version 2.2.0 rc2
1840
-----------------
1841

    
1842
*(Released Wed, 22 Sep 2010)*
1843

    
1844
- Fixed OS_VARIANT variable for OS scripts
1845
- Fixed cluster tag operations via RAPI
1846
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1847
- Disabled RAPI CA checks in watcher
1848

    
1849

    
1850
Version 2.2.0 rc1
1851
-----------------
1852

    
1853
*(Released Mon, 23 Aug 2010)*
1854

    
1855
- Support DRBD versions of the format "a.b.c.d"
1856
- Updated manpages
1857
- Re-introduce support for usage from multiple threads in RAPI client
1858
- Instance renames and modify via RAPI
1859
- Work around race condition between processing and archival in job
1860
  queue
1861
- Mark opcodes following failed one as failed, too
1862
- Job field ``lock_status`` was removed due to difficulties making it
1863
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1864
  locks is expected for a later 2.2.x release
1865
- Fixed dry-run behaviour with many commands
1866
- Support ``ssh-agent`` again when adding nodes
1867
- Many additional bugfixes
1868

    
1869

    
1870
Version 2.2.0 rc0
1871
-----------------
1872

    
1873
*(Released Fri, 30 Jul 2010)*
1874

    
1875
Important change: the internal RPC mechanism between Ganeti nodes has
1876
changed from using a home-grown http library (based on the Python base
1877
libraries) to use the PycURL library. This requires that PycURL is
1878
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1879
against GnuTLS by default. cURL's support for GnuTLS had known issues
1880
before cURL 7.21.0 and we recommend using the latest cURL release or
1881
linking against OpenSSL. Most other distributions already link PycURL
1882
and cURL against OpenSSL. The command::
1883

    
1884
  python -c 'import pycurl; print pycurl.version'
1885

    
1886
can be used to determine the libraries PycURL and cURL are linked
1887
against.
1888

    
1889
Other significant changes:
1890

    
1891
- Rewrote much of the internals of the job queue, in order to achieve
1892
  better parallelism; this decouples job query operations from the job
1893
  processing, and it should allow much nicer behaviour of the master
1894
  daemon under load, and it also has uncovered some long-standing bugs
1895
  related to the job serialisation (now fixed)
1896
- Added a default iallocator setting to the cluster parameters,
1897
  eliminating the need to always pass nodes or an iallocator for
1898
  operations that require selection of new node(s)
1899
- Added experimental support for the LXC virtualization method
1900
- Added support for OS parameters, which allows the installation of
1901
  instances to pass parameter to OS scripts in order to customise the
1902
  instance
1903
- Added a hypervisor parameter controlling the migration type (live or
1904
  non-live), since hypervisors have various levels of reliability; this
1905
  has renamed the 'live' parameter to 'mode'
1906
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
1907
  logical volumes, meaning that cluster verify will ignore them and not
1908
  flag their presence as errors
1909
- The watcher will now reset the error count for failed instances after
1910
  8 hours, thus allowing self-healing if the problem that caused the
1911
  instances to be down/fail to start has cleared in the meantime
1912
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1913
  check for, and warn, if the drbd module parameter ``usermode_helper``
1914
  is not consistent with the cluster-wide setting; this is needed to
1915
  make diagnose easier of failed drbd creations
1916
- Started adding base IPv6 support, but this is not yet
1917
  enabled/available for use
1918
- Rename operations (cluster, instance) will now return the new name,
1919
  which is especially useful if a short name was passed in
1920
- Added support for instance migration in RAPI
1921
- Added a tool to pre-configure nodes for the SSH setup, before joining
1922
  them to the cluster; this will allow in the future a simplified model
1923
  for node joining (but not yet fully enabled in 2.2); this needs the
1924
  paramiko python library
1925
- Fixed handling of name-resolving errors
1926
- Fixed consistency of job results on the error path
1927
- Fixed master-failover race condition when executed multiple times in
1928
  sequence
1929
- Fixed many bugs related to the job queue (mostly introduced during the
1930
  2.2 development cycle, so not all are impacting 2.1)
1931
- Fixed instance migration with missing disk symlinks
1932
- Fixed handling of unknown jobs in ``gnt-job archive``
1933
- And many other small fixes/improvements
1934

    
1935
Internal changes:
1936

    
1937
- Enhanced both the unittest and the QA coverage
1938
- Switched the opcode validation to a generic model, and extended the
1939
  validation to all opcode parameters
1940
- Changed more parts of the code that write shell scripts to use the
1941
  same class for this
1942
- Switched the master daemon to use the asyncore library for the Luxi
1943
  server endpoint
1944

    
1945

    
1946
Version 2.2.0 beta0
1947
-------------------
1948

    
1949
*(Released Thu, 17 Jun 2010)*
1950

    
1951
- Added tool (``move-instance``) and infrastructure to move instances
1952
  between separate clusters (see :doc:`separate documentation
1953
  <move-instance>` and :doc:`design document <design-2.2>`)
1954
- Added per-request RPC timeout
1955
- RAPI now requires a Content-Type header for requests with a body (e.g.
1956
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1957
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
1958
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1959
  reachable
1960
- Implemented initial support for running Ganeti daemons as separate
1961
  users, see configure-time flags ``--with-user-prefix`` and
1962
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1963
  time)
1964
- Instances can be removed after export (``gnt-backup export
1965
  --remove-instance``)
1966
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1967
  key (instead of 1024 bit)
1968
- Added new cluster configuration file for cluster domain secret
1969
- Import/export now use SSL instead of SSH
1970
- Added support for showing estimated time when exporting an instance,
1971
  see the ``ganeti-os-interface(7)`` manpage and look for
1972
  ``EXP_SIZE_FD``
1973

    
1974

    
1975
Version 2.1.8
1976
-------------
1977

    
1978
*(Released Tue, 16 Nov 2010)*
1979

    
1980
Some more bugfixes. Unless critical bugs occur, this will be the last
1981
2.1 release:
1982

    
1983
- Fix case of MAC special-values
1984
- Fix mac checker regex
1985
- backend: Fix typo causing "out of range" error
1986
- Add missing --units in gnt-instance list man page
1987

    
1988

    
1989
Version 2.1.7
1990
-------------
1991

    
1992
*(Released Tue, 24 Aug 2010)*
1993

    
1994
Bugfixes only:
1995
  - Don't ignore secondary node silently on non-mirrored disk templates
1996
    (issue 113)
1997
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1998
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1999
  - Properly document the usb_mouse parameter
2000
  - Fix path in ganeti-rapi(8) (issue 116)
2001
  - Adjust error message when the ganeti user's .ssh directory is
2002
    missing
2003
  - Add same-node-check when changing the disk template to drbd
2004

    
2005

    
2006
Version 2.1.6
2007
-------------
2008

    
2009
*(Released Fri, 16 Jul 2010)*
2010

    
2011
Bugfixes only:
2012
  - Add an option to only select some reboot types during qa/burnin.
2013
    (on some hypervisors consequent reboots are not supported)
2014
  - Fix infrequent race condition in master failover. Sometimes the old
2015
    master ip address would be still detected as up for a short time
2016
    after it was removed, causing failover to fail.
2017
  - Decrease mlockall warnings when the ctypes module is missing. On
2018
    Python 2.4 we support running even if no ctypes module is installed,
2019
    but we were too verbose about this issue.
2020
  - Fix building on old distributions, on which man doesn't have a
2021
    --warnings option.
2022
  - Fix RAPI not to ignore the MAC address on instance creation
2023
  - Implement the old instance creation format in the RAPI client.
2024

    
2025

    
2026
Version 2.1.5
2027
-------------
2028

    
2029
*(Released Thu, 01 Jul 2010)*
2030

    
2031
A small bugfix release:
2032
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2033
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2034
    a non-existing option
2035
  - Fix instance create: the --force-variant option was ignored
2036
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2037
  - Fix modify node storage with non-FQDN arguments
2038
  - Fix RAPI client to authenticate under Python 2.6 when used
2039
    for more than 5 requests needing authentication
2040
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2041
    when converting a non-shutdown drbd instance to plain
2042

    
2043

    
2044
Version 2.1.4
2045
-------------
2046

    
2047
*(Released Fri, 18 Jun 2010)*
2048

    
2049
A small bugfix release:
2050

    
2051
  - Fix live migration of KVM instances started with older Ganeti
2052
    versions which had fewer hypervisor parameters
2053
  - Fix gnt-instance grow-disk on down instances
2054
  - Fix an error-reporting bug during instance migration
2055
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2056
    silently ignoring broken ones
2057
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2058
    users
2059
  - Fix bug triggered by different API version os-es on different nodes
2060
  - Fix a bug in instance startup with custom hvparams: OS level
2061
    parameters would fail to be applied.
2062
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2063
    make it work completely well with OpenSSL)
2064
  - Fix handling of errors when resolving names from DNS
2065

    
2066

    
2067
Version 2.1.3
2068
-------------
2069

    
2070
*(Released Thu, 3 Jun 2010)*
2071

    
2072
A medium sized development cycle. Some new features, and some
2073
fixes/small improvements/cleanups.
2074

    
2075
Significant features
2076
~~~~~~~~~~~~~~~~~~~~
2077

    
2078
The node deamon now tries to mlock itself into memory, unless the
2079
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2080
its logs, and falls back to console logging. This allows emergency
2081
features such as ``gnt-node powercycle`` to work even in the event of a
2082
broken node disk (tested offlining the disk hosting the node's
2083
filesystem and dropping its memory caches; don't try this at home)
2084

    
2085
KVM: add vhost-net acceleration support. It can be tested with a new
2086
enough version of the kernel and of qemu-kvm.
2087

    
2088
KVM: Add instance chrooting feature. If you use privilege dropping for
2089
your VMs you can also now force them to chroot to an empty directory,
2090
before starting the emulated guest.
2091

    
2092
KVM: Add maximum migration bandwith and maximum downtime tweaking
2093
support (requires a new-enough version of qemu-kvm).
2094

    
2095
Cluster verify will now warn if the master node doesn't have the master
2096
ip configured on it.
2097

    
2098
Add a new (incompatible) instance creation request format to RAPI which
2099
supports all parameters (previously only a subset was supported, and it
2100
wasn't possible to extend the old format to accomodate all the new
2101
features. The old format is still supported, and a client can check for
2102
this feature, before using it, by checking for its presence in the
2103
``features`` RAPI resource.
2104

    
2105
Now with ancient latin support. Try it passing the ``--roman`` option to
2106
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2107
(requires the python-roman module to be installed, in order to work).
2108

    
2109
Other changes
2110
~~~~~~~~~~~~~
2111

    
2112
As usual many internal code refactorings, documentation updates, and
2113
such. Among others:
2114

    
2115
  - Lots of improvements and cleanups to the experimental Remote API
2116
    (RAPI) client library.
2117
  - A new unit test suite for the core daemon libraries.
2118
  - A fix to creating missing directories makes sure the umask is not
2119
    applied anymore. This enforces the same directory permissions
2120
    everywhere.
2121
  - Better handling terminating daemons with ctrl+c (used when running
2122
    them in debugging mode).
2123
  - Fix a race condition in live migrating a KVM instance, when stat()
2124
    on the old proc status file returned EINVAL, which is an unexpected
2125
    value.
2126
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2127
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2128

    
2129

    
2130
Version 2.1.2.1
2131
---------------
2132

    
2133
*(Released Fri, 7 May 2010)*
2134

    
2135
Fix a bug which prevented untagged KVM instances from starting.
2136

    
2137

    
2138
Version 2.1.2
2139
-------------
2140

    
2141
*(Released Fri, 7 May 2010)*
2142

    
2143
Another release with a long development cycle, during which many
2144
different features were added.
2145

    
2146
Significant features
2147
~~~~~~~~~~~~~~~~~~~~
2148

    
2149
The KVM hypervisor now can run the individual instances as non-root, to
2150
reduce the impact of a VM being hijacked due to bugs in the
2151
hypervisor. It is possible to run all instances as a single (non-root)
2152
user, to manually specify a user for each instance, or to dynamically
2153
allocate a user out of a cluster-wide pool to each instance, with the
2154
guarantee that no two instances will run under the same user ID on any
2155
given node.
2156

    
2157
An experimental RAPI client library, that can be used standalone
2158
(without the other Ganeti libraries), is provided in the source tree as
2159
``lib/rapi/client.py``. Note this client might change its interface in
2160
the future, as we iterate on its capabilities.
2161

    
2162
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2163
replace the cluster's certificates and crypto keys. This might help in
2164
case they have been compromised, or have simply expired.
2165

    
2166
A new disk option for instance creation has been added that allows one
2167
to "adopt" currently existing logical volumes, with data
2168
preservation. This should allow easier migration to Ganeti from
2169
unmanaged (or managed via other software) instances.
2170

    
2171
Another disk improvement is the possibility to convert between redundant
2172
(DRBD) and plain (LVM) disk configuration for an instance. This should
2173
allow better scalability (starting with one node and growing the
2174
cluster, or shrinking a two-node cluster to one node).
2175

    
2176
A new feature that could help with automated node failovers has been
2177
implemented: if a node sees itself as offline (by querying the master
2178
candidates), it will try to shutdown (hard) all instances and any active
2179
DRBD devices. This reduces the risk of duplicate instances if an
2180
external script automatically failovers the instances on such nodes. To
2181
enable this, the cluster parameter ``maintain_node_health`` should be
2182
enabled; in the future this option (per the name) will enable other
2183
automatic maintenance features.
2184

    
2185
Instance export/import now will reuse the original instance
2186
specifications for all parameters; that means exporting an instance,
2187
deleting it and the importing it back should give an almost identical
2188
instance. Note that the default import behaviour has changed from
2189
before, where it created only one NIC; now it recreates the original
2190
number of NICs.
2191

    
2192
Cluster verify has added a few new checks: SSL certificates validity,
2193
/etc/hosts consistency across the cluster, etc.
2194

    
2195
Other changes
2196
~~~~~~~~~~~~~
2197

    
2198
As usual, many internal changes were done, documentation fixes,
2199
etc. Among others:
2200

    
2201
- Fixed cluster initialization with disabled cluster storage (regression
2202
  introduced in 2.1.1)
2203
- File-based storage supports growing the disks
2204
- Fixed behaviour of node role changes
2205
- Fixed cluster verify for some corner cases, plus a general rewrite of
2206
  cluster verify to allow future extension with more checks
2207
- Fixed log spamming by watcher and node daemon (regression introduced
2208
  in 2.1.1)
2209
- Fixed possible validation issues when changing the list of enabled
2210
  hypervisors
2211
- Fixed cleanup of /etc/hosts during node removal
2212
- Fixed RAPI response for invalid methods
2213
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2214
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2215
  from ide disks, etc.)
2216
- Allow OS changes without re-installation (to record a changed OS
2217
  outside of Ganeti, or to allow OS renames)
2218
- Allow instance creation without OS installation (useful for example if
2219
  the OS will be installed manually, or restored from a backup not in
2220
  Ganeti format)
2221
- Implemented option to make cluster ``copyfile`` use the replication
2222
  network
2223
- Added list of enabled hypervisors to ssconf (possibly useful for
2224
  external scripts)
2225
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2226
  1.2 clusters
2227
- A partial form of node re-IP is possible via node readd, which now
2228
  allows changed node primary IP
2229
- Command line utilities now show an informational message if the job is
2230
  waiting for a lock
2231
- The logs of the master daemon now show the PID/UID/GID of the
2232
  connected client
2233

    
2234

    
2235
Version 2.1.1
2236
-------------
2237

    
2238
*(Released Fri, 12 Mar 2010)*
2239

    
2240
During the 2.1.0 long release candidate cycle, a lot of improvements and
2241
changes have accumulated with were released later as 2.1.1.
2242

    
2243
Major changes
2244
~~~~~~~~~~~~~
2245

    
2246
The node evacuate command (``gnt-node evacuate``) was significantly
2247
rewritten, and as such the IAllocator protocol was changed - a new
2248
request type has been added. This unfortunate change during a stable
2249
series is designed to improve performance of node evacuations; on
2250
clusters with more than about five nodes and which are well-balanced,
2251
evacuation should proceed in parallel for all instances of the node
2252
being evacuated. As such, any existing IAllocator scripts need to be
2253
updated, otherwise the above command will fail due to the unknown
2254
request. The provided "dumb" allocator has not been updated; but the
2255
ganeti-htools package supports the new protocol since version 0.2.4.
2256

    
2257
Another important change is increased validation of node and instance
2258
names. This might create problems in special cases, if invalid host
2259
names are being used.
2260

    
2261
Also, a new layer of hypervisor parameters has been added, that sits at
2262
OS level between the cluster defaults and the instance ones. This allows
2263
customisation of virtualization parameters depending on the installed
2264
OS. For example instances with OS 'X' may have a different KVM kernel
2265
(or any other parameter) than the cluster defaults. This is intended to
2266
help managing a multiple OSes on the same cluster, without manual
2267
modification of each instance's parameters.
2268

    
2269
A tool for merging clusters, ``cluster-merge``, has been added in the
2270
tools sub-directory.
2271

    
2272
Bug fixes
2273
~~~~~~~~~
2274

    
2275
- Improved the int/float conversions that should make the code more
2276
  robust in face of errors from the node daemons
2277
- Fixed the remove node code in case of internal configuration errors
2278
- Fixed the node daemon behaviour in face of inconsistent queue
2279
  directory (e.g. read-only file-system where we can't open the files
2280
  read-write, etc.)
2281
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2282
  now it either aborts cleanly or, if given the new "auto_promote"
2283
  parameter, will automatically promote other nodes as needed
2284
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2285
  completely prevent Ganeti from working
2286
- Fixed bug for instance export when not all disks were successfully
2287
  exported
2288
- Fixed behaviour of node add when the new node is slow in starting up
2289
  the node daemon
2290
- Fixed handling of signals in the LUXI client, which should improve
2291
  behaviour of command-line scripts
2292
- Added checks for invalid node/instance names in the configuration (now
2293
  flagged during cluster verify)
2294
- Fixed watcher behaviour for disk activation errors
2295
- Fixed two potentially endless loops in http library, which led to the
2296
  RAPI daemon hanging and consuming 100% CPU in some cases
2297
- Fixed bug in RAPI daemon related to hashed passwords
2298
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2299
  instances
2300
- Enhanced compatibility with non-Debian OSes, but not using absolute
2301
  path in some commands and allowing customisation of the ssh
2302
  configuration directory
2303
- Fixed possible future issue with new Python versions by abiding to the
2304
  proper use of ``__slots__`` attribute on classes
2305
- Added checks that should prevent directory traversal attacks
2306
- Many documentation fixes based on feedback from users
2307

    
2308
New features
2309
~~~~~~~~~~~~
2310

    
2311
- Added an "early_release" more for instance replace disks and node
2312
  evacuate, where we release locks earlier and thus allow higher
2313
  parallelism within the cluster
2314
- Added watcher hooks, intended to allow the watcher to restart other
2315
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2316
  course for any other purpose
2317
- Added a compile-time disable for DRBD barriers, to increase
2318
  performance if the administrator trusts the power supply or the
2319
  storage system to not lose writes
2320
- Added the option of using syslog for logging instead of, or in
2321
  addition to, Ganeti's own log files
2322
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2323
  can indeed boot from a paravirtual NIC
2324
- Added a generic debug level for many operations; while this is not
2325
  used widely yet, it allows one to pass the debug value all the way to
2326
  the OS scripts
2327
- Enhanced the hooks environment for instance moves (failovers,
2328
  migrations) where the primary/secondary nodes changed during the
2329
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2330
- Enhanced data validations for many user-supplied values; one important
2331
  item is the restrictions imposed on instance and node names, which
2332
  might reject some (invalid) host names
2333
- Add a configure-time option to disable file-based storage, if it's not
2334
  needed; this allows greater security separation between the master
2335
  node and the other nodes from the point of view of the inter-node RPC
2336
  protocol
2337
- Added user notification in interactive tools if job is waiting in the
2338
  job queue or trying to acquire locks
2339
- Added log messages when a job is waiting for locks
2340
- Added filtering by node tags in instance operations which admit
2341
  multiple instances (start, stop, reboot, reinstall)
2342
- Added a new tool for cluster mergers, ``cluster-merge``
2343
- Parameters from command line which are of the form ``a=b,c=d`` can now
2344
  use backslash escapes to pass in values which contain commas,
2345
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2346
  ``b,c``
2347
- For KVM, the instance name is the first parameter passed to KVM, so
2348
  that it's more visible in the process list
2349

    
2350

    
2351
Version 2.1.0
2352
-------------
2353

    
2354
*(Released Tue, 2 Mar 2010)*
2355

    
2356
Ganeti 2.1 brings many improvements with it. Major changes:
2357

    
2358
- Added infrastructure to ease automated disk repairs
2359
- Added new daemon to export configuration data in a cheaper way than
2360
  using the remote API
2361
- Instance NICs can now be routed instead of being associated with a
2362
  networking bridge
2363
- Improved job locking logic to reduce impact of jobs acquiring multiple
2364
  locks waiting for other long-running jobs
2365

    
2366
In-depth implementation details can be found in the Ganeti 2.1 design
2367
document.
2368

    
2369
Details
2370
~~~~~~~
2371

    
2372
- Added chroot hypervisor
2373
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2374
  ``device_model``)
2375
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2376
  ``bootloader_path`` and ``bootloader_args``)
2377
- Added the ``use_localtime`` option for the xen-hvm and kvm
2378
  hypervisors, and the default value for this has changed to false (in
2379
  2.0 xen-hvm always enabled it)
2380
- Added luxi call to submit multiple jobs in one go
2381
- Added cluster initialization option to not modify ``/etc/hosts``
2382
  file on nodes
2383
- Added network interface parameters
2384
- Added dry run mode to some LUs
2385
- Added RAPI resources:
2386

    
2387
  - ``/2/instances/[instance_name]/info``
2388
  - ``/2/instances/[instance_name]/replace-disks``
2389
  - ``/2/nodes/[node_name]/evacuate``
2390
  - ``/2/nodes/[node_name]/migrate``
2391
  - ``/2/nodes/[node_name]/role``
2392
  - ``/2/nodes/[node_name]/storage``
2393
  - ``/2/nodes/[node_name]/storage/modify``
2394
  - ``/2/nodes/[node_name]/storage/repair``
2395

    
2396
- Added OpCodes to evacuate or migrate all instances on a node
2397
- Added new command to list storage elements on nodes (``gnt-node
2398
  list-storage``) and modify them (``gnt-node modify-storage``)
2399
- Added new ssconf files with master candidate IP address
2400
  (``ssconf_master_candidates_ips``), node primary IP address
2401
  (``ssconf_node_primary_ips``) and node secondary IP address
2402
  (``ssconf_node_secondary_ips``)
2403
- Added ``ganeti-confd`` and a client library to query the Ganeti
2404
  configuration via UDP
2405
- Added ability to run hooks after cluster initialization and before
2406
  cluster destruction
2407
- Added automatic mode for disk replace (``gnt-instance replace-disks
2408
  --auto``)
2409
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2410
  after catastrophic data-loss
2411
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2412
  groups
2413
- Added ``gnt-instance move`` command to move instances
2414
- Added ``gnt-cluster watcher`` command to control watcher
2415
- Added ``gnt-node powercycle`` command to powercycle nodes
2416
- Added new job status field ``lock_status``
2417
- Added parseable error codes to cluster verification (``gnt-cluster
2418
  verify --error-codes``) and made output less verbose (use
2419
  ``--verbose`` to restore previous behaviour)
2420
- Added UUIDs to the main config entities (cluster, nodes, instances)
2421
- Added support for OS variants
2422
- Added support for hashed passwords in the Ganeti remote API users file
2423
  (``rapi_users``)
2424
- Added option to specify maximum timeout on instance shutdown
2425
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2426
- Added new helper script to start and stop Ganeti daemons
2427
  (``daemon-util``), with the intent to reduce the work necessary to
2428
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2429
  from one place
2430
- Added more unittests
2431
- Fixed critical bug in ganeti-masterd startup
2432
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2433
  now customisable at the cluster level for both the KVM and Xen
2434
  hypervisors using the new ``migration_port`` parameter
2435
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2436
  reinstalling an instance
2437
- Allowed ``@`` in tag names
2438
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2439
- Many documentation updates
2440
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2441
- ``gnt-instance reinstall`` can now reinstall multiple instances
2442
- Updated many command line parameters
2443
- Introduced new OS API version 15
2444
- No longer support a default hypervisor
2445
- Treat virtual LVs as inexistent
2446
- Improved job locking logic to reduce lock contention
2447
- Match instance and node names case insensitively
2448
- Reimplemented bash completion script to be more complete
2449
- Improved burnin
2450

    
2451

    
2452
Version 2.0.6
2453
-------------
2454

    
2455
*(Released Thu, 4 Feb 2010)*
2456

    
2457
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2458
- Fix a string formatting bug
2459
- Improve safety of the code in some error paths
2460
- Improve data validation in the master of values returned from nodes
2461

    
2462

    
2463
Version 2.0.5
2464
-------------
2465

    
2466
*(Released Thu, 17 Dec 2009)*
2467

    
2468
- Fix security issue due to missing validation of iallocator names; this
2469
  allows local and remote execution of arbitrary executables
2470
- Fix failure of gnt-node list during instance removal
2471
- Ship the RAPI documentation in the archive
2472

    
2473

    
2474
Version 2.0.4
2475
-------------
2476

    
2477
*(Released Wed, 30 Sep 2009)*
2478

    
2479
- Fixed many wrong messages
2480
- Fixed a few bugs related to the locking library
2481
- Fixed MAC checking at instance creation time
2482
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2483
- Fixed a few issues related to signal handling in both daemons and
2484
  scripts
2485
- Fixed the example startup script provided
2486
- Fixed insserv dependencies in the example startup script (patch from
2487
  Debian)
2488
- Fixed handling of drained nodes in the iallocator framework
2489
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2490
  #528618)
2491
- Fixed error related to invalid job IDs in job polling
2492
- Fixed job/opcode persistence on unclean master shutdown
2493
- Fixed handling of partial job processing after unclean master
2494
  shutdown
2495
- Fixed error reporting from LUs, previously all errors were converted
2496
  into execution errors
2497
- Fixed error reporting from burnin
2498
- Decreased significantly the memory usage of the job queue
2499
- Optimised slightly multi-job submission
2500
- Optimised slightly opcode loading
2501
- Backported the multi-job submit framework from the development
2502
  branch; multi-instance start and stop should be faster
2503
- Added script to clean archived jobs after 21 days; this will reduce
2504
  the size of the queue directory
2505
- Added some extra checks in disk size tracking
2506
- Added an example ethers hook script
2507
- Added a cluster parameter that prevents Ganeti from modifying of
2508
  /etc/hosts
2509
- Added more node information to RAPI responses
2510
- Added a ``gnt-job watch`` command that allows following the ouput of a
2511
  job
2512
- Added a bind-address option to ganeti-rapi
2513
- Added more checks to the configuration verify
2514
- Enhanced the burnin script such that some operations can be retried
2515
  automatically
2516
- Converted instance reinstall to multi-instance model
2517

    
2518

    
2519
Version 2.0.3
2520
-------------
2521

    
2522
*(Released Fri, 7 Aug 2009)*
2523

    
2524
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2525
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2526
  instances have mismatched disk sizes in the configuration
2527
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2528
  any configuration mismatches for disk sizes
2529
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2530
  failover to work on two-node clusters
2531
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2532
  unusable
2533
- Fixed detection of OS script errors in ``gnt-backup export``
2534
- Fixed exit code of ``gnt-backup export``
2535

    
2536

    
2537
Version 2.0.2
2538
-------------
2539

    
2540
*(Released Fri, 17 Jul 2009)*
2541

    
2542
- Added experimental support for stripped logical volumes; this should
2543
  enhance performance but comes with a higher complexity in the block
2544
  device handling; stripping is only enabled when passing
2545
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2546
  affected even in the non-stripped mode
2547
- Improved resiliency against transient failures at the end of DRBD
2548
  resyncs, and in general of DRBD resync checks
2549
- Fixed a couple of issues with exports and snapshot errors
2550
- Fixed a couple of issues in instance listing
2551
- Added display of the disk size in ``gnt-instance info``
2552
- Fixed checking for valid OSes in instance creation
2553
- Fixed handling of the "vcpus" parameter in instance listing and in
2554
  general of invalid parameters
2555
- Fixed http server library, and thus RAPI, to handle invalid
2556
  username/password combinations correctly; this means that now they
2557
  report unauthorized for queries too, not only for modifications,
2558
  allowing earlier detect of configuration problems
2559
- Added a new "role" node list field, equivalent to the master/master
2560
  candidate/drained/offline flags combinations
2561
- Fixed cluster modify and changes of candidate pool size
2562
- Fixed cluster verify error messages for wrong files on regular nodes
2563
- Fixed a couple of issues with node demotion from master candidate role
2564
- Fixed node readd issues
2565
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2566
- Added a new ``--no-voting`` option for masterfailover to fix failover
2567
  on two-nodes clusters when the former master node is unreachable
2568
- Added instance reinstall over RAPI
2569

    
2570

    
2571
Version 2.0.1
2572
-------------
2573

    
2574
*(Released Tue, 16 Jun 2009)*
2575

    
2576
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2577
  allow re-adding the start in single-user option (regression from 1.2)
2578
- the watcher writes the instance status to a file, to allow monitoring
2579
  to report the instance status (from the master) based on cached
2580
  results of the watcher's queries; while this can get stale if the
2581
  watcher is being locked due to other work on the cluster, this is
2582
  still an improvement
2583
- the watcher now also restarts the node daemon and the rapi daemon if
2584
  they died
2585
- fixed the watcher to handle full and drained queue cases
2586
- hooks export more instance data in the environment, which helps if
2587
  hook scripts need to take action based on the instance's properties
2588
  (no longer need to query back into ganeti)
2589
- instance failovers when the instance is stopped do not check for free
2590
  RAM, so that failing over a stopped instance is possible in low memory
2591
  situations
2592
- rapi uses queries for tags instead of jobs (for less job traffic), and
2593
  for cluster tags it won't talk to masterd at all but read them from
2594
  ssconf
2595
- a couple of error handling fixes in RAPI
2596
- drbd handling: improved the error handling of inconsistent disks after
2597
  resync to reduce the frequency of "there are some degraded disks for
2598
  this instance" messages
2599
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2600
  error handling path called a wrong function name)
2601

    
2602

    
2603
Version 2.0.0
2604
-------------
2605

    
2606
*(Released Wed, 27 May 2009)*
2607

    
2608
- no changes from rc5
2609

    
2610

    
2611
Version 2.0 rc5
2612
---------------
2613

    
2614
*(Released Wed, 20 May 2009)*
2615

    
2616
- fix a couple of bugs (validation, argument checks)
2617
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2618
- some small improvements to RAPI and IAllocator
2619
- make watcher automatically start the master daemon if down
2620

    
2621

    
2622
Version 2.0 rc4
2623
---------------
2624

    
2625
*(Released Mon, 27 Apr 2009)*
2626

    
2627
- change the OS list to not require locks; this helps with big clusters
2628
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2629
  volume group is broken
2630
- ``gnt-instance info``, without any arguments, doesn't run for all
2631
  instances anymore; either pass ``--all`` or pass the desired
2632
  instances; this helps against mistakes on big clusters where listing
2633
  the information for all instances takes a long time
2634
- miscellaneous doc and man pages fixes
2635

    
2636

    
2637
Version 2.0 rc3
2638
---------------
2639

    
2640
*(Released Wed, 8 Apr 2009)*
2641

    
2642
- Change the internal locking model of some ``gnt-node`` commands, in
2643
  order to reduce contention (and blocking of master daemon) when
2644
  batching many creation/reinstall jobs
2645
- Fixes to Xen soft reboot
2646
- No longer build documentation at build time, instead distribute it in
2647
  the archive, in order to reduce the need for the whole docbook/rst
2648
  toolchains
2649

    
2650

    
2651
Version 2.0 rc2
2652
---------------
2653

    
2654
*(Released Fri, 27 Mar 2009)*
2655

    
2656
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2657
- Fix watcher startup sequence, improves the behaviour of busy clusters
2658
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2659
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2660
  bind address and other places
2661
- Some documentation fixes and updates
2662

    
2663

    
2664
Version 2.0 rc1
2665
---------------
2666

    
2667
*(Released Mon, 2 Mar 2009)*
2668

    
2669
- More documentation updates, now all docs should be more-or-less
2670
  up-to-date
2671
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2672
  etc.)
2673
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2674
  and KVM)
2675
- Fix an issue related to $libdir/run/ganeti and cluster creation
2676

    
2677

    
2678
Version 2.0 beta2
2679
-----------------
2680

    
2681
*(Released Thu, 19 Feb 2009)*
2682

    
2683
- Xen PVM and KVM have switched the default value for the instance root
2684
  disk to the first partition on the first drive, instead of the whole
2685
  drive; this means that the OS installation scripts must be changed
2686
  accordingly
2687
- Man pages have been updated
2688
- RAPI has been switched by default to HTTPS, and the exported functions
2689
  should all work correctly
2690
- RAPI v1 has been removed
2691
- Many improvements to the KVM hypervisor
2692
- Block device errors are now better reported
2693
- Many other bugfixes and small improvements
2694

    
2695

    
2696
Version 2.0 beta1
2697
-----------------
2698

    
2699
*(Released Mon, 26 Jan 2009)*
2700

    
2701
- Version 2 is a general rewrite of the code and therefore the
2702
  differences are too many to list, see the design document for 2.0 in
2703
  the ``doc/`` subdirectory for more details
2704
- In this beta version there is not yet a migration path from 1.2 (there
2705
  will be one in the final 2.0 release)
2706
- A few significant changes are:
2707

    
2708
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2709
    various ``gnt-*`` commands are just front-ends to it
2710
  - all the commands are entered into, and executed from a job queue,
2711
    see the ``gnt-job(8)`` manpage
2712
  - the RAPI daemon supports read-write operations, secured by basic
2713
    HTTP authentication on top of HTTPS
2714
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2715
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2716
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2717
  - DRBD devices are using statically allocated minor numbers, which
2718
    will be assigned to existing instances during the migration process
2719
  - there is support for both Xen PVM and Xen HVM instances running on
2720
    the same cluster
2721
  - KVM virtualization is supported too
2722
  - file-based storage has been implemented, which means that it is
2723
    possible to run the cluster without LVM and DRBD storage, for
2724
    example using a shared filesystem exported from shared storage (and
2725
    still have live migration)
2726

    
2727

    
2728
Version 1.2.7
2729
-------------
2730

    
2731
*(Released Tue, 13 Jan 2009)*
2732

    
2733
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2734
- Reuse the old instance mac address by default on instance import, if
2735
  the instance name is the same.
2736
- Handle situations in which the node info rpc returns incomplete
2737
  results (issue 46)
2738
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2739
- Improved version of batcher:
2740

    
2741
  - state file support
2742
  - instance mac address support
2743
  - support for HVM clusters/instances
2744

    
2745
- Add an option to show the number of cpu sockets and nodes in
2746
  ``gnt-node list``
2747
- Support OSes that handle more than one version of the OS api (but do
2748
  not change the current API in any other way)
2749
- Fix ``gnt-node migrate``
2750
- ``gnt-debug`` man page
2751
- Fixes various more typos and small issues
2752
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2753

    
2754

    
2755
Version 1.2.6
2756
-------------
2757

    
2758
*(Released Wed, 24 Sep 2008)*
2759

    
2760
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2761
  type of disk exported to fully virtualized instances.
2762
- provide access to the serial console of HVM instances
2763
- instance auto_balance flag, set by default. If turned off it will
2764
  avoid warnings on cluster verify if there is not enough memory to fail
2765
  over an instance. in the future it will prevent automatically failing
2766
  it over when we will support that.
2767
- batcher tool for instance creation, see ``tools/README.batcher``
2768
- ``gnt-instance reinstall --select-os`` to interactively select a new
2769
  operating system when reinstalling an instance.
2770
- when changing the memory amount on instance modify a check has been
2771
  added that the instance will be able to start. also warnings are
2772
  emitted if the instance will not be able to fail over, if auto_balance
2773
  is true.
2774
- documentation fixes
2775
- sync fields between ``gnt-instance list/modify/add/import``
2776
- fix a race condition in drbd when the sync speed was set after giving
2777
  the device a remote peer.
2778

    
2779

    
2780
Version 1.2.5
2781
-------------
2782

    
2783
*(Released Tue, 22 Jul 2008)*
2784

    
2785
- note: the allowed size and number of tags per object were reduced
2786
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2787
- fixed twisted 8.x compatibility
2788
- fixed ``gnt-instance replace-disks`` with iallocator
2789
- add TCP keepalives on twisted connections to detect restarted nodes
2790
- disk increase support, see ``gnt-instance grow-disk``
2791
- implement bulk node/instance query for RAPI
2792
- add tags in node/instance listing (optional)
2793
- experimental migration (and live migration) support, read the man page
2794
  for ``gnt-instance migrate``
2795
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2796
  has some small improvements in handling its state file
2797

    
2798

    
2799
Version 1.2.4
2800
-------------
2801

    
2802
*(Released Fri, 13 Jun 2008)*
2803

    
2804
- Experimental readonly, REST-based remote API implementation;
2805
  automatically started on master node, TCP port 5080, if enabled by
2806
  ``--enable-rapi`` parameter to configure script.
2807
- Instance allocator support. Add and import instance accept a
2808
  ``--iallocator`` parameter, and call that instance allocator to decide
2809
  which node to use for the instance. The iallocator document describes
2810
  what's expected from an allocator script.
2811
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2812
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2813
  node is lost there is still enough memory to fail over the instances
2814
  that reside on it.
2815
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2816
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2817
  hooks will run, and their output, if any, will be displayed. Any
2818
  failing hook will make the verification return an error value.
2819
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2820
  primary and secondary interfaces
2821
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2822
  that is still declared as part of the cluster and has failed.
2823
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2824
  specifying output fields, that just adds the chosen fields to the
2825
  default ones.
2826
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2827
  export from the filesystem.
2828
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2829
  have been added. Using them you can enable/disable acpi and pae
2830
  support, and specify a path for a cd image to be exported to the
2831
  instance. These parameters as the name suggest only work on HVM
2832
  clusters.
2833
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2834
  PAE support will be set to the previously hardcoded values, but the
2835
  (previously hardcoded) path to the CDROM ISO image will be unset and
2836
  if required, needs to be set manually with ``gnt-instance modify``
2837
  after the upgrade.
2838
- The address to which an instance's VNC console is bound is now
2839
  selectable per-instance, rather than being cluster wide. Of course
2840
  this only applies to instances controlled via VNC, so currently just
2841
  applies to HVM clusters.
2842

    
2843

    
2844
Version 1.2.3
2845
-------------
2846

    
2847
*(Released Mon, 18 Feb 2008)*
2848

    
2849
- more tweaks to the disk activation code (especially helpful for DRBD)
2850
- change the default ``gnt-instance list`` output format, now there is
2851
  one combined status field (see the manpage for the exact values this
2852
  field will have)
2853
- some more fixes for the mac export to hooks change
2854
- make Ganeti not break with DRBD 8.2.x (which changed the version
2855
  format in ``/proc/drbd``) (issue 24)
2856
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2857
  template, allowing migration from DRBD0.7+MD to DRBD8
2858

    
2859

    
2860
Version 1.2.2
2861
-------------
2862

    
2863
*(Released Wed, 30 Jan 2008)*
2864

    
2865
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2866
  support (issue 23)
2867
- add command aliases infrastructure and a few aliases
2868
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2869
  man pages and the ``--help`` option of ``gnt-node
2870
  list``/``gnt-instance list``
2871
- fix ``gnt-backup list`` with down nodes (issue 21)
2872
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2873
- fix the dist archive and add a check for including svn/git files in
2874
  the future
2875
- some developer-related changes: improve the burnin and the QA suite,
2876
  add an upload script for testing during development
2877

    
2878

    
2879
Version 1.2.1
2880
-------------
2881

    
2882
*(Released Wed, 16 Jan 2008)*
2883

    
2884
- experimental HVM support, read the install document, section
2885
  "Initializing the cluster"
2886
- allow for the PVM hypervisor per-instance kernel and initrd paths
2887
- add a new command ``gnt-cluster verify-disks`` which uses a new
2888
  algorithm to improve the reconnection of the DRBD pairs if the device
2889
  on the secondary node has gone away
2890
- make logical volume code auto-activate LVs at disk activation time
2891
- slightly improve the speed of activating disks
2892
- allow specification of the MAC address at instance creation time, and
2893
  changing it later via ``gnt-instance modify``
2894
- fix handling of external commands that generate lots of output on
2895
  stderr
2896
- update documentation with regard to minimum version of DRBD8 supported
2897

    
2898

    
2899
Version 1.2.0
2900
-------------
2901

    
2902
*(Released Tue, 4 Dec 2007)*
2903

    
2904
- Log the ``xm create`` output to the node daemon log on failure (to
2905
  help diagnosing the error)
2906
- In debug mode, log all external commands output if failed to the logs
2907
- Change parsing of lvm commands to ignore stderr
2908

    
2909

    
2910
Version 1.2 beta3
2911
-----------------
2912

    
2913
*(Released Wed, 28 Nov 2007)*
2914

    
2915
- Another round of updates to the DRBD 8 code to deal with more failures
2916
  in the replace secondary node operation
2917
- Some more logging of failures in disk operations (lvm, drbd)
2918
- A few documentation updates
2919
- QA updates
2920

    
2921

    
2922
Version 1.2 beta2
2923
-----------------
2924

    
2925
*(Released Tue, 13 Nov 2007)*
2926

    
2927
- Change configuration file format from Python's Pickle to JSON.
2928
  Upgrading is possible using the cfgupgrade utility.
2929
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2930
  faster replace disks and is more stable (DRBD 8 has many improvements
2931
  compared to DRBD 0.7)
2932
- Added command line tags support (see man pages for ``gnt-instance``,
2933
  ``gnt-node``, ``gnt-cluster``)
2934
- Added instance rename support
2935
- Added multi-instance startup/shutdown
2936
- Added cluster rename support
2937
- Added ``gnt-node evacuate`` to simplify some node operations
2938
- Added instance reboot operation that can speedup reboot as compared to
2939
  stop and start
2940
- Soften the requirement that hostnames are in FQDN format
2941
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
2942
  reboots
2943
- Removed dependency on debian's patched fping that uses the
2944
  non-standard ``-S`` option
2945
- Now the OS definitions are searched for in multiple, configurable
2946
  paths (easier for distros to package)
2947
- Some changes to the hooks infrastructure (especially the new
2948
  post-configuration update hook)
2949
- Other small bugfixes
2950

    
2951
.. vim: set textwidth=72 syntax=rst :
2952
.. Local Variables:
2953
.. mode: rst
2954
.. fill-column: 72
2955
.. End: