Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 6a360606

History | View | Annotate | Download (115.6 kB)

1
News
2
====
3

    
4

    
5
Version 2.10.0 alpha1
6
---------------------
7

    
8
*(unreleased)*
9

    
10
Incompatible/important changes
11
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
12

    
13
- Adding disks with 'gnt-instance modify' now waits for the disks to sync per
14
  default. Specify --no-wait-for-sync to override this behavior.
15
- The Ganeti python code now adheres to a private-module layout. In particular,
16
  the module 'ganeti' is no longer in the python search path.
17
- On instance allocation, the iallocator now considers non-LVM storage
18
  properly. In particular, actual file storage space information is used
19
  when allocating space for a file/sharedfile instance.
20
- When disabling disk templates cluster-wide, the cluster now first
21
  checks whether there are instances still using those templates.
22
- 'gnt-node list-storage' now also reports storage information about
23
  file-based storage types.
24

    
25
New features
26
~~~~~~~~~~~~
27
- KVM hypervisors can now access RBD storage directly without having to
28
  go through a block device.
29
- A new command 'gnt-cluster upgrade' was added that automates the upgrade
30
  procedure between two Ganeti versions that are both 2.10 or higher.
31
- The move-instance command can now change disk templates when moving
32
  instances, and does not require any node placement options to be
33
  specified if the destination cluster has a default iallocator.
34

    
35
Misc changes
36
~~~~~~~~~~~~
37

    
38
- Opcodes are entirely generated from Haskell using the tool 'hs2py' and
39
  the module 'src/Ganeti/OpCodes.hs'.
40
- Constants are also generated from Haskell using the tool
41
  'hs2py-constants' and the module 'src/Ganeti/HsConstants.hs', with the
42
  exception of socket related constants, which require changing the
43
  cluster configuration file, and HVS related constants, because they
44
  are part of a port of instance queries to Haskell.  As a result, these
45
  changes will be part of the next release of Ganeti.
46

    
47

    
48
Version 2.9.0 rc3
49
-----------------
50

    
51
*(Released Tue, 15 Oct 2013)*
52

    
53
Incompatible/important changes
54
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
55

    
56
- hroller now also plans for capacity to move non-redundant instances off
57
  any node to be rebooted; the old behavior of completely ignoring any
58
  non-redundant instances can be restored by adding the --ignore-non-redundant
59
  option.
60
- The cluster option '--no-lvm-storage' was removed in favor of the new option
61
  '--enabled-disk-templates'.
62
- On instance creation, disk templates no longer need to be specified
63
  with '-t'. The default disk template will be taken from the list of
64
  enabled disk templates.
65
- The monitoring daemon is now running as root, in order to be able to collect
66
  information only available to root (such as the state of Xen instances).
67
- The ConfD client is now IPv6 compatible.
68
- File and shared file storage is no longer dis/enabled at configure time,
69
  but using the option '--enabled-disk-templates' at cluster initialization and
70
  modification.
71
- The default directories for file and shared file storage are not anymore
72
  specified at configure time, but taken from the cluster's configuration.
73
  They can be set at cluster initialization and modification with
74
  '--file-storage-dir' and '--shared-file-storage-dir'.
75
- Cluster verification now includes stricter checks regarding the
76
  default file and shared file storage directories. It now checks that
77
  the directories are explicitely allowed in the 'file-storage-paths' file and
78
  that the directories exist on all nodes.
79
- The list of allowed disk templates in the instance policy and the list
80
  of cluster-wide enabled disk templates is now checked for consistency
81
  on cluster or group modification. On cluster initialization, the ipolicy
82
  disk templates are ensured to be a subset of the cluster-wide enabled
83
  disk templates.
84

    
85
New features
86
~~~~~~~~~~~~
87

    
88
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
89
  the correct command syntax. It is possible to use different DRBD versions
90
  on different nodes as long as they are compatible to each other. This
91
  enables rolling upgrades of DRBD with no downtime. As permanent operation
92
  of different DRBD versions within a node group is discouraged,
93
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
94
- New "inst-status-xen" data collector for the monitoring daemon, providing
95
  information about the state of the xen instances on the nodes.
96
- New "lv" data collector for the monitoring daemon, collecting data about the
97
  logical volumes on the nodes, and pairing them with the name of the instances
98
  they belong to.
99
- New "diskstats" data collector, collecting the data from /proc/diskstats and
100
  presenting them over the monitoring daemon interface.
101
- The ConfD client is now IPv6 compatible.
102

    
103
New dependencies
104
~~~~~~~~~~~~~~~~
105
The following new dependencies have been added.
106

    
107
Python
108

    
109
- ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
110
  for the unit tests (and only used for testing).
111

    
112
Haskell
113

    
114
- ``hslogger`` (http://software.complete.org/hslogger) is now always
115
  required, even if confd is not enabled.
116

    
117
Since 2.9.0 rc2
118
~~~~~~~~~~~~~~~
119

    
120
- in implicit configuration upgrade, match ipolicy with enabled disk templates
121
- improved harep documentation (inherited from stable-2.8)
122

    
123

    
124
Version 2.9.0 rc2
125
-----------------
126

    
127
*(Released Wed, 9 Oct 2013)*
128

    
129
The second release candidate in the 2.9 series. Since 2.9.0 rc1:
130

    
131
- Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
132
  at least one DRBD instance.
133
- Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
134
  downgrading.
135

    
136

    
137
Version 2.9.0 rc1
138
-----------------
139

    
140
*(Released Tue, 1 Oct 2013)*
141

    
142
The first release candidate in the 2.9 series. Since 2.9.0 beta1:
143

    
144
- various bug fixes
145
- update of the documentation, in particular installation instructions
146
- merging of LD_* constants into DT_* constants
147
- python style changes to be compatible with newer versions of pylint
148

    
149

    
150
Version 2.9.0 beta1
151
-------------------
152

    
153
*(Released Thu, 29 Aug 2013)*
154

    
155
This was the first beta release of the 2.9 series. All important changes
156
are listed in the latest 2.9 entry.
157

    
158

    
159
Version 2.8.0
160
-------------
161

    
162
*(Released Mon, 30 Sep 2013)*
163

    
164
Incompatible/important changes
165
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
166

    
167
- Instance policy can contain multiple instance specs, as described in
168
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
169
  <design-partitioned>`. As a consequence, it's not possible to partially change
170
  or override instance specs. Bounding specs (min and max) can be specified as a
171
  whole using the new option ``--ipolicy-bounds-specs``, while standard
172
  specs use the new option ``--ipolicy-std-specs``.
173
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
174
  gnt-instance is a valid YAML object.
175
- hail now honors network restrictions when allocating nodes. This led to an
176
  update of the IAllocator protocol. See the IAllocator documentation for
177
  details.
178
- confd now only answers static configuration request over the network. luxid
179
  was extracted, listens on the local LUXI socket and responds to live queries.
180
  This allows finer grained permissions if using separate users.
181

    
182
New features
183
~~~~~~~~~~~~
184

    
185
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
186
  to always require authentication, ``--require-authentication``. It can
187
  be specified in ``$sysconfdir/default/ganeti``.
188
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
189
  be used to manage the disk templates to be used by instances in the cluster.
190
  Initially, it will be set to a list that includes plain, drbd, if they were
191
  enabled by specifying a volume group name, and file and sharedfile, if those
192
  were enabled at configure time. Additionally, it will include all disk
193
  templates that are currently used by instances. The order of disk templates
194
  will be based on Ganeti's history of supporting them. In the future, the
195
  first entry of the list will be used as a default disk template on instance
196
  creation.
197
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
198
  configuration back to the previous stable version.
199
- Disk templates in group ipolicy can be restored to the default value.
200
- Initial support for diskless instances and virtual clusters in QA.
201
- More QA and unit tests for instance policies.
202
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
203
  describing why the opcode itself was executed.
204
- The monitoring daemon is now available. It allows users to query the cluster
205
  for obtaining information about the status of the system. The daemon is only
206
  responsible for providing the information over the network: the actual data
207
  gathering is performed by data collectors (currently, only the DRBD status
208
  collector is available).
209
- In order to help developers work on Ganeti, a new script
210
  (``devel/build_chroot``) is provided, for building a chroot that contains all
211
  the required development libraries and tools for compiling Ganeti on a Debian
212
  Squeeze system.
213
- A new tool, ``harep``, for performing self-repair and recreation of instances
214
  in Ganeti has been added.
215
- Split queries are enabled for tags, network, exports, cluster info, groups,
216
  jobs, nodes.
217
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
218
  It prints the command line to set the current policies, to ease
219
  changing them.
220
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
221
  devices for KVM virtio-net interfaces will get created with VNET_HDR
222
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
223
  virtio-net interfaces, which prevents host kernel tainting and log
224
  flooding, when dealing with broken or malicious virtio-net drivers.
225
  It's set to true by default.
226
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
227
  failures.
228
- Support 'viridian' parameter in Xen HVM
229
- Support DSA SSH keys in bootstrap
230
- To simplify the work of packaging frameworks that want to add the needed users
231
  and groups in a split-user setup themselves, at build time three files in
232
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
233
  the groups to be generated, the ``users`` file contains, one per line, the
234
  users to be generated, optionally followed by their primary group, where
235
  important. The ``groupmemberships`` file contains, one per line, additional
236
  user-group membership relations that need to be established. The syntax of
237
  these files will remain stable in all future versions.
238

    
239

    
240
New dependencies
241
~~~~~~~~~~~~~~~~
242
The following new dependencies have been added:
243

    
244
For Haskell:
245
- The ``curl`` library is not optional anymore for compiling the Haskell code.
246
- ``snap-server`` library (if monitoring is enabled).
247

    
248
For Python:
249
- The minimum Python version needed to run Ganeti is now 2.6.
250
- ``yaml`` library (only for running the QA).
251

    
252
Since 2.8.0 rc3
253
~~~~~~~~~~~~~~~
254
- Perform proper cleanup on termination of Haskell daemons
255
- Fix corner-case in handling of remaining retry time
256

    
257

    
258
Version 2.8.0 rc3
259
-----------------
260

    
261
*(Released Tue, 17 Sep 2013)*
262

    
263
- To simplify the work of packaging frameworks that want to add the needed users
264
  and groups in a split-user setup themselves, at build time three files in
265
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
266
  the groups to be generated, the ``users`` file contains, one per line, the
267
  users to be generated, optionally followed by their primary group, where
268
  important. The ``groupmemberships`` file contains, one per line, additional
269
  user-group membership relations that need to be established. The syntax of
270
  these files will remain stable in all future versions.
271
- Add a default to file-driver when unspecified over RAPI (Issue 571)
272
- Mark the DSA host pubkey as optional, and remove it during config downgrade
273
  (Issue 560)
274
- Some documentation fixes
275

    
276

    
277
Version 2.8.0 rc2
278
-----------------
279

    
280
*(Released Tue, 27 Aug 2013)*
281

    
282
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
283

    
284
- Support 'viridian' parameter in Xen HVM (Issue 233)
285
- Include VCS version in ``gnt-cluster version``
286
- Support DSA SSH keys in bootstrap (Issue 338)
287
- Fix batch creation of instances
288
- Use FQDN to check master node status (Issue 551)
289
- Make the DRBD collector more failure-resilient
290

    
291

    
292
Version 2.8.0 rc1
293
-----------------
294

    
295
*(Released Fri, 2 Aug 2013)*
296

    
297
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
298

    
299
- Fix upgrading/downgrading from 2.7
300
- Increase maximum RAPI message size
301
- Documentation updates
302
- Split ``confd`` between ``luxid`` and ``confd``
303
- Merge 2.7 series up to the 2.7.1 release
304
- Allow the ``modify_etc_hosts`` option to be changed
305
- Add better debugging for ``luxid`` queries
306
- Expose bulk parameter for GetJobs in RAPI client
307
- Expose missing ``network`` fields in RAPI
308
- Add some ``cluster verify`` tests
309
- Some unittest fixes
310
- Fix a malfunction in ``hspace``'s tiered allocation
311
- Fix query compatibility between haskell and python implementations
312
- Add the ``vnet_hdr`` HV parameter for KVM
313
- Add ``--cleanup`` to instance failover
314
- Change the connected groups format in ``gnt-network info`` output; it
315
  was previously displayed as a raw list by mistake. (Merged from 2.7)
316

    
317

    
318
Version 2.8.0 beta1
319
-------------------
320

    
321
*(Released Mon, 24 Jun 2013)*
322

    
323
This was the first beta release of the 2.8 series. All important changes
324
are listed in the latest 2.8 entry.
325

    
326

    
327
Version 2.7.2
328
-------------
329

    
330
*(Released Thu, 26 Sep 2013)*
331

    
332
- Change the connected groups format in ``gnt-network info`` output; it
333
  was previously displayed as a raw list by mistake
334
- Check disk template in right dict when copying
335
- Support multi-instance allocs without iallocator
336
- Fix some errors in the documentation
337
- Fix formatting of tuple in an error message
338

    
339

    
340
Version 2.7.1
341
-------------
342

    
343
*(Released Thu, 25 Jul 2013)*
344

    
345
- Add logrotate functionality in daemon-util
346
- Add logrotate example file
347
- Add missing fields to network queries over rapi
348
- Fix network object timestamps
349
- Add support for querying network timestamps
350
- Fix a typo in the example crontab
351
- Fix a documentation typo
352

    
353

    
354
Version 2.7.0
355
-------------
356

    
357
*(Released Thu, 04 Jul 2013)*
358

    
359
Incompatible/important changes
360
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
361

    
362
- Instance policies for disk size were documented to be on a per-disk
363
  basis, but hail applied them to the sum of all disks. This has been
364
  fixed.
365
- ``hbal`` will now exit with status 0 if, during job execution over
366
  LUXI, early exit has been requested and all jobs are successful;
367
  before, exit status 1 was used, which cannot be differentiated from
368
  "job error" case
369
- Compatibility with newer versions of rbd has been fixed
370
- ``gnt-instance batch-create`` has been changed to use the bulk create
371
  opcode from Ganeti. This lead to incompatible changes in the format of
372
  the JSON file. It's now not a custom dict anymore but a dict
373
  compatible with the ``OpInstanceCreate`` opcode.
374
- Parent directories for file storage need to be listed in
375
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
376
  write the file automatically based on old configuration values, but it
377
  can not distribute it across all nodes and the file contents should be
378
  verified. Use ``gnt-cluster copyfile
379
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
380
  upgraded. The reason for requiring this list of paths now is that
381
  before it would have been possible to inject new paths via RPC,
382
  allowing files to be created in arbitrary locations. The RPC protocol
383
  is protected using SSL/X.509 certificates, but as a design principle
384
  Ganeti does not permit arbitrary paths to be passed.
385
- The parsing of the variants file for OSes (see
386
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
387
  lines and comment lines (starting with ``#``) are ignored for better
388
  readability.
389
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
390
  longer available. ``gnt-node add`` now invokes a new tool on the
391
  destination node, named ``prepare-node-join``, to configure the SSH
392
  daemon. Paramiko is no longer necessary to configure nodes' SSH
393
  daemons via ``gnt-node add``.
394
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
395
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
396
  the flag is not reset after a master failover.
397
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
398
  is recommended. 2.6 will be mandatory from the 2.8 series.
399

    
400

    
401
New features
402
~~~~~~~~~~~~
403

    
404
- New network management functionality to support automatic allocation
405
  of IP addresses and managing of network parameters. See
406
  :manpage:`gnt-network(8)` for more details.
407
- New external storage backend, to allow managing arbitrary storage
408
  systems external to the cluster. See
409
  :manpage:`ganeti-extstorage-interface(7)`.
410
- New ``exclusive-storage`` node parameter added, restricted to
411
  nodegroup level. When it's set to true, physical disks are assigned in
412
  an exclusive fashion to instances, as documented in :doc:`Partitioned
413
  Ganeti <design-partitioned>`.  Currently, only instances using the
414
  ``plain`` disk template are supported.
415
- The KVM hypervisor has been updated with many new hypervisor
416
  parameters, including a generic one for passing arbitrary command line
417
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
418
  compatible up to qemu 1.4.
419
- A new tool, called ``mon-collector``, is the stand-alone executor of
420
  the data collectors for a monitoring system. As of this version, it
421
  just includes the DRBD data collector, that can be executed by calling
422
  ``mon-collector`` using the ``drbd`` parameter. See
423
  :manpage:`mon-collector(7)`.
424
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
425
  for RAPI users. It allows granting permissions to query for
426
  information to a specific user without giving
427
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
428
- A new tool named ``node-cleanup`` has been added. It cleans remains of
429
  a cluster from a machine by stopping all daemons, removing
430
  certificates and ssconf files. Unless the ``--no-backup`` option is
431
  given, copies of the certificates are made.
432
- Instance creations now support the use of opportunistic locking,
433
  potentially speeding up the (parallel) creation of multiple instances.
434
  This feature is currently only available via the :doc:`RAPI
435
  <rapi>` interface and when an instance allocator is used. If the
436
  ``opportunistic_locking`` parameter is set the opcode will try to
437
  acquire as many locks as possible, but will not wait for any locks
438
  held by other opcodes. If not enough resources can be found to
439
  allocate the instance, the temporary error code
440
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
441
  retried thereafter, with or without opportunistic locking.
442
- New experimental linux-ha resource scripts.
443
- Restricted-commands support: ganeti can now be asked (via command line
444
  or rapi) to perform commands on a node. These are passed via ganeti
445
  RPC rather than ssh. This functionality is restricted to commands
446
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
447
  reasons. The file is not copied automatically.
448

    
449

    
450
Misc changes
451
~~~~~~~~~~~~
452

    
453
- Diskless instances are now externally mirrored (Issue 237). This for
454
  now has only been tested in conjunction with explicit target nodes for
455
  migration/failover.
456
- Queries not needing locks or RPC access to the node can now be
457
  performed by the confd daemon, making them independent from jobs, and
458
  thus faster to execute. This is selectable at configure time.
459
- The functionality for allocating multiple instances at once has been
460
  overhauled and is now also available through :doc:`RAPI <rapi>`.
461

    
462
There are no significant changes from version 2.7.0~rc3.
463

    
464

    
465
Version 2.7.0 rc3
466
-----------------
467

    
468
*(Released Tue, 25 Jun 2013)*
469

    
470
- Fix permissions on the confd query socket (Issue 477)
471
- Fix permissions on the job archive dir (Issue 498)
472
- Fix handling of an internal exception in replace-disks (Issue 472)
473
- Fix gnt-node info handling of shortened names (Issue 497)
474
- Fix gnt-instance grow-disk when wiping is enabled
475
- Documentation improvements, and support for newer pandoc
476
- Fix hspace honoring ipolicy for disks (Issue 484)
477
- Improve handling of the ``kvm_extra`` HV parameter
478

    
479

    
480
Version 2.7.0 rc2
481
-----------------
482

    
483
*(Released Fri, 24 May 2013)*
484

    
485
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
486
  symlink.
487
- Disks added through ``gnt-instance modify`` or created through
488
  ``gnt-instance recreate-disks`` are wiped, if the
489
  ``prealloc_wipe_disks`` flag is set.
490
- If wiping newly created disks fails, the disks are removed. Also,
491
  partial failures in creating disks through ``gnt-instance modify``
492
  triggers a cleanup of the partially-created disks.
493
- Removing the master IP address doesn't fail if the address has been
494
  already removed.
495
- Fix ownership of the OS log dir
496
- Workaround missing SO_PEERCRED constant (Issue 191)
497

    
498

    
499
Version 2.7.0 rc1
500
-----------------
501

    
502
*(Released Fri, 3 May 2013)*
503

    
504
This was the first release candidate of the 2.7 series. Since beta3:
505

    
506
- Fix kvm compatibility with qemu 1.4 (Issue 389)
507
- Documentation updates (admin guide, upgrade notes, install
508
  instructions) (Issue 372)
509
- Fix gnt-group list nodes and instances count (Issue 436)
510
- Fix compilation without non-mandatory libraries (Issue 441)
511
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
512
- Make confd logging more verbose at INFO level (Issue 435)
513
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
514
- Fix failure path for instance storage type conversion (Issue 229)
515
- Update htools text backend documentation
516
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
517
- Disable inter-cluster instance move for file-based instances, because
518
  it is dependant on instance export, which is not supported for
519
  file-based instances. (Issue 414)
520
- Fix gnt-job crashes on non-ascii characters (Issue 427)
521
- Fix volume group checks on non-vm-capable nodes (Issue 432)
522

    
523

    
524
Version 2.7.0 beta3
525
-------------------
526

    
527
*(Released Mon, 22 Apr 2013)*
528

    
529
This was the third beta release of the 2.7 series. Since beta2:
530

    
531
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
532
- Fix data loss on wrong usage of ``gnt-instance move``
533
- Properly export errors in confd-based job queries
534
- Add ``users-setup`` tool
535
- Fix iallocator protocol to report 0 as a disk size for diskless
536
  instances. This avoids hail breaking when a diskless instance is
537
  present.
538
- Fix job queue directory permission problem that made confd job queries
539
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
540
  for access to archived jobs (Issue 406).
541
- Limit the sizes of networks supported by ``gnt-network`` to something
542
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
543
- Fix bugs in instance disk template conversion
544
- Fix GHC 7 compatibility
545
- Fix ``burnin`` install path (Issue 426).
546
- Allow very small disk grows (Issue 347).
547
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
548
  that noded doesn't import masterd code (Issue 419).
549
- Make sure the default metavg at cluster init is the same as the vg, if
550
  unspecified (Issue 358).
551
- Fix cleanup of partially created disks (part of Issue 416)
552

    
553

    
554
Version 2.7.0 beta2
555
-------------------
556

    
557
*(Released Tue, 2 Apr 2013)*
558

    
559
This was the second beta release of the 2.7 series. Since beta1:
560

    
561
- Networks no longer have a "type" slot, since this information was
562
  unused in Ganeti: instead of it tags should be used.
563
- The rapi client now has a ``target_node`` option to MigrateInstance.
564
- Fix early exit return code for hbal (Issue 386).
565
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
566
- Fix ``rbd showmapped`` output parsing (Issue 312).
567
- Networks are now referenced indexed by UUID, rather than name. This
568
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
569
  use.
570
- The OS environment now includes network information.
571
- Deleting of a network is now disallowed if any instance nic is using
572
  it, to prevent dangling references.
573
- External storage is now documented in man pages.
574
- The exclusive_storage flag can now only be set at nodegroup level.
575
- Hbal can now submit an explicit priority with its jobs.
576
- Many network related locking fixes.
577
- Bump up the required pylint version to 0.25.1.
578
- Fix the ``no_remember`` option in RAPI client.
579
- Many ipolicy related tests, qa, and fixes.
580
- Many documentation improvements and fixes.
581
- Fix building with ``--disable-file-storage``.
582
- Fix ``-q`` option in htools, which was broken if passed more than
583
  once.
584
- Some haskell/python interaction improvements and fixes.
585
- Fix iallocator in case of missing LVM storage.
586
- Fix confd config load in case of ``--no-lvm-storage``.
587
- The confd/query functionality is now mentioned in the security
588
  documentation.
589

    
590

    
591
Version 2.7.0 beta1
592
-------------------
593

    
594
*(Released Wed, 6 Feb 2013)*
595

    
596
This was the first beta release of the 2.7 series. All important changes
597
are listed in the latest 2.7 entry.
598

    
599

    
600
Version 2.6.2
601
-------------
602

    
603
*(Released Fri, 21 Dec 2012)*
604

    
605
Important behaviour change: hbal won't rebalance anymore instances which
606
have the ``auto_balance`` attribute set to false. This was the intention
607
all along, but until now it only skipped those from the N+1 memory
608
reservation (DRBD-specific).
609

    
610
A significant number of bug fixes in this release:
611

    
612
- Fixed disk adoption interaction with ipolicy checks.
613
- Fixed networking issues when instances are started, stopped or
614
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
615
- Fixed the warning in cluster verify for shared storage instances not
616
  being redundant.
617
- Fixed removal of storage directory on shared file storage (issue 262).
618
- Fixed validation of LVM volume group name in OpClusterSetParams
619
  (``gnt-cluster modify``) (issue 285).
620
- Fixed runtime memory increases (``gnt-instance modify -m``).
621
- Fixed live migration under Xen's ``xl`` mode.
622
- Fixed ``gnt-instance console`` with ``xl``.
623
- Fixed building with newer Haskell compiler/libraries.
624
- Fixed PID file writing in Haskell daemons (confd); this prevents
625
  restart issues if confd was launched manually (outside of
626
  ``daemon-util``) while another copy of it was running
627
- Fixed a type error when doing live migrations with KVM (issue 297) and
628
  the error messages for failing migrations have been improved.
629
- Fixed opcode validation for the out-of-band commands (``gnt-node
630
  power``).
631
- Fixed a type error when unsetting OS hypervisor parameters (issue
632
  311); now it's possible to unset all OS-specific hypervisor
633
  parameters.
634
- Fixed the ``dry-run`` mode for many operations: verification of
635
  results was over-zealous but didn't take into account the ``dry-run``
636
  operation, resulting in "wrong" failures.
637
- Fixed bash completion in ``gnt-job list`` when the job queue has
638
  hundreds of entries; especially with older ``bash`` versions, this
639
  results in significant CPU usage.
640

    
641
And lastly, a few other improvements have been made:
642

    
643
- Added option to force master-failover without voting (issue 282).
644
- Clarified error message on lock conflict (issue 287).
645
- Logging of newly submitted jobs has been improved (issue 290).
646
- Hostname checks have been made uniform between instance rename and
647
  create (issue 291).
648
- The ``--submit`` option is now supported by ``gnt-debug delay``.
649
- Shutting down the master daemon by sending SIGTERM now stops it from
650
  processing jobs waiting for locks; instead, those jobs will be started
651
  once again after the master daemon is started the next time (issue
652
  296).
653
- Support for Xen's ``xl`` program has been improved (besides the fixes
654
  above).
655
- Reduced logging noise in the Haskell confd daemon (only show one log
656
  entry for each config reload, instead of two).
657
- Several man page updates and typo fixes.
658

    
659

    
660
Version 2.6.1
661
-------------
662

    
663
*(Released Fri, 12 Oct 2012)*
664

    
665
A small bugfix release. Among the bugs fixed:
666

    
667
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
668
  made the command unusable.
669
- Commands that issue many jobs don't fail anymore just because some jobs
670
  take so long that other jobs are archived.
671
- Failures during ``gnt-instance reinstall`` are reflected by the exit
672
  status.
673
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
674
  DRBD is enabled.
675
- When ``always_failover`` is set, ``--allow-failover`` is not required
676
  in migrate commands anymore.
677
- ``bash_completion`` works even if extglob is disabled.
678
- Fixed bug with locks that made failover for RDB-based instances fail.
679
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
680
  a random node instead of one based on the allocator metric.
681
- Support for newer versions of pylint and pep8.
682
- Hail doesn't fail anymore when trying to add an instance of type
683
  ``file``, ``sharedfile`` or ``rbd``.
684
- Added new Makefile target to rebuild the whole distribution, so that
685
  all files are included.
686

    
687

    
688
Version 2.6.0
689
-------------
690

    
691
*(Released Fri, 27 Jul 2012)*
692

    
693

    
694
.. attention:: The ``LUXI`` protocol has been made more consistent
695
   regarding its handling of command arguments. This, however, leads to
696
   incompatibility issues with previous versions. Please ensure that you
697
   restart Ganeti daemons soon after the upgrade, otherwise most
698
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
699
   pausing/resuming the watcher, cancelling and archiving jobs, querying
700
   the cluster configuration) will fail.
701

    
702

    
703
New features
704
~~~~~~~~~~~~
705

    
706
Instance run status
707
+++++++++++++++++++
708

    
709
The current ``admin_up`` field, which used to denote whether an instance
710
should be running or not, has been removed. Instead, ``admin_state`` is
711
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
712

    
713
The rational behind this is that an instance being “down” can have
714
different meanings:
715

    
716
- it could be down during a reboot
717
- it could be temporarily be down for a reinstall
718
- or it could be down because it is deprecated and kept just for its
719
  disk
720

    
721
The previous Boolean state was making it difficult to do capacity
722
calculations: should Ganeti reserve memory for a down instance? Now, the
723
tri-state field makes it clear:
724

    
725
- in ``up`` and ``down`` state, all resources are reserved for the
726
  instance, and it can be at any time brought up if it is down
727
- in ``offline`` state, only disk space is reserved for it, but not
728
  memory or CPUs
729

    
730
The field can have an extra use: since the transition between ``up`` and
731
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
732
transition between ``offline`` and ``down`` is done via ``gnt-instance
733
modify``, it is possible to given different rights to users. For
734
example, owners of an instance could be allowed to start/stop it, but
735
not transition it out of the offline state.
736

    
737
Instance policies and specs
738
+++++++++++++++++++++++++++
739

    
740
In previous Ganeti versions, an instance creation request was not
741
limited on the minimum size and on the maximum size just by the cluster
742
resources. As such, any policy could be implemented only in third-party
743
clients (RAPI clients, or shell wrappers over ``gnt-*``
744
tools). Furthermore, calculating cluster capacity via ``hspace`` again
745
required external input with regards to instance sizes.
746

    
747
In order to improve these workflows and to allow for example better
748
per-node group differentiation, we introduced instance specs, which
749
allow declaring:
750

    
751
- minimum instance disk size, disk count, memory size, cpu count
752
- maximum values for the above metrics
753
- and “standard” values (used in ``hspace`` to calculate the standard
754
  sized instances)
755

    
756
The minimum/maximum values can be also customised at node-group level,
757
for example allowing more powerful hardware to support bigger instance
758
memory sizes.
759

    
760
Beside the instance specs, there are a few other settings belonging to
761
the instance policy framework. It is possible now to customise, per
762
cluster and node-group:
763

    
764
- the list of allowed disk templates
765
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
766
- the maximum ratio of instance to spindles (see below for more
767
  information) for local storage
768

    
769
All these together should allow all tools that talk to Ganeti to know
770
what are the ranges of allowed values for instances and the
771
over-subscription that is allowed.
772

    
773
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
774
instance configuration, and the physical CPU configuration from the
775
node. For the spindle ratios however, we didn't track before these
776
values, so new parameters have been added:
777

    
778
- a new node parameter ``spindle_count``, defaults to 1, customisable at
779
  node group or node level
780
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
781

    
782
Note that spindles in this context doesn't need to mean actual
783
mechanical hard-drives; it's just a relative number for both the node
784
I/O capacity and instance I/O consumption.
785

    
786
Instance migration behaviour
787
++++++++++++++++++++++++++++
788

    
789
While live-migration is in general desirable over failover, it is
790
possible that for some workloads it is actually worse, due to the
791
variable time of the “suspend” phase during live migration.
792

    
793
To allow the tools to work consistently over such instances (without
794
having to hard-code instance names), a new backend parameter
795
``always_failover`` has been added to control the migration/failover
796
behaviour. When set to True, all migration requests for an instance will
797
instead fall-back to failover.
798

    
799
Instance memory ballooning
800
++++++++++++++++++++++++++
801

    
802
Initial support for memory ballooning has been added. The memory for an
803
instance is no longer fixed (backend parameter ``memory``), but instead
804
can vary between minimum and maximum values (backend parameters
805
``minmem`` and ``maxmem``). Currently we only change an instance's
806
memory when:
807

    
808
- live migrating or failing over and instance and the target node
809
  doesn't have enough memory
810
- user requests changing the memory via ``gnt-instance modify
811
  --runtime-memory``
812

    
813
Instance CPU pinning
814
++++++++++++++++++++
815

    
816
In order to control the use of specific CPUs by instance, support for
817
controlling CPU pinning has been added for the Xen, HVM and LXC
818
hypervisors. This is controlled by a new hypervisor parameter
819
``cpu_mask``; details about possible values for this are in the
820
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
821
VCPU-to-CPU mapping) form will work well only when all nodes in your
822
cluster have the same amount of CPUs.
823

    
824
Disk parameters
825
+++++++++++++++
826

    
827
Another area in which Ganeti was not customisable were the parameters
828
used for storage configuration, e.g. how many stripes to use for LVM,
829
DRBD resync configuration, etc.
830

    
831
To improve this area, we've added disks parameters, which are
832
customisable at cluster and node group level, and which allow to
833
specify various parameters for disks (DRBD has the most parameters
834
currently), for example:
835

    
836
- DRBD resync algorithm and parameters (e.g. speed)
837
- the default VG for meta-data volumes for DRBD
838
- number of stripes for LVM (plain disk template)
839
- the RBD pool
840

    
841
These parameters can be modified via ``gnt-cluster modify -D …`` and
842
``gnt-group modify -D …``, and are used at either instance creation (in
843
case of LVM stripes, for example) or at disk “activation” time
844
(e.g. resync speed).
845

    
846
Rados block device support
847
++++++++++++++++++++++++++
848

    
849
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
850
denoted by the ``rbd`` disk template type. This is considered
851
experimental, feedback is welcome. For details on configuring it, see
852
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
853

    
854
Master IP setup
855
+++++++++++++++
856

    
857
The existing master IP functionality works well only in simple setups (a
858
single network shared by all nodes); however, if nodes belong to
859
different networks, then the ``/32`` setup and lack of routing
860
information is not enough.
861

    
862
To allow the master IP to function well in more complex cases, the
863
system was reworked as follows:
864

    
865
- a master IP netmask setting has been added
866
- the master IP activation/turn-down code was moved from the node daemon
867
  to a separate script
868
- whether to run the Ganeti-supplied master IP script or a user-supplied
869
  on is a ``gnt-cluster init`` setting
870

    
871
Details about the location of the standard and custom setup scripts are
872
in the man page :manpage:`gnt-cluster(8)`; for information about the
873
setup script protocol, look at the Ganeti-supplied script.
874

    
875
SPICE support
876
+++++++++++++
877

    
878
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
879
improved.
880

    
881
It is now possible to use TLS-protected connections, and when renewing
882
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
883
it is now possible to specify spice or spice CA certificates. Also, it
884
is possible to configure a password for SPICE sessions via the
885
hypervisor parameter ``spice_password_file``.
886

    
887
There are also new parameters to control the compression and streaming
888
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
889
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
890
for the spice parameters.
891

    
892
Lastly, it is now possible to see the SPICE connection information via
893
``gnt-instance console``.
894

    
895
OVF converter
896
+++++++++++++
897

    
898
A new tool (``tools/ovfconverter``) has been added that supports
899
conversion between Ganeti and the `Open Virtualization Format
900
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
901
from).
902

    
903
This relies on the ``qemu-img`` tool to convert the disk formats, so the
904
actual compatibility with other virtualization solutions depends on it.
905

    
906
Confd daemon changes
907
++++++++++++++++++++
908

    
909
The configuration query daemon (``ganeti-confd``) is now optional, and
910
has been rewritten in Haskell; whether to use the daemon at all, use the
911
Python (default) or the Haskell version is selectable at configure time
912
via the ``--enable-confd`` parameter, which can take one of the
913
``haskell``, ``python`` or ``no`` values. If not used, disabling the
914
daemon will result in a smaller footprint; for larger systems, we
915
welcome feedback on the Haskell version which might become the default
916
in future versions.
917

    
918
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
919
daemon running. The Python version doesn't implement the new call.
920

    
921

    
922
User interface changes
923
~~~~~~~~~~~~~~~~~~~~~~
924

    
925
We have replaced the ``--disks`` option of ``gnt-instance
926
replace-disks`` with a more flexible ``--disk`` option, which allows
927
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
928
disk size and mode can be changed upon recreation (via ``gnt-instance
929
recreate-disks``, which accepts the same ``--disk`` option).
930

    
931
As many people are used to a ``show`` command, we have added that as an
932
alias to ``info`` on all ``gnt-*`` commands.
933

    
934
The ``gnt-instance grow-disk`` command has a new mode in which it can
935
accept the target size of the disk, instead of the delta; this can be
936
more safe since two runs in absolute mode will be idempotent, and
937
sometimes it's also easier to specify the desired size directly.
938

    
939
Also the handling of instances with regard to offline secondaries has
940
been improved. Instance operations should not fail because one of it's
941
secondary nodes is offline, even though it's safe to proceed.
942

    
943
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
944
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
945
minors to instance name.
946

    
947
API changes
948
~~~~~~~~~~~
949

    
950
RAPI coverage has improved, with (for example) new resources for
951
recreate-disks, node power-cycle, etc.
952

    
953
Compatibility
954
~~~~~~~~~~~~~
955

    
956
There is partial support for ``xl`` in the Xen hypervisor; feedback is
957
welcome.
958

    
959
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
960
whether to still support Python 2.4 or move to Python 2.6 as minimum
961
required version.
962

    
963
Support for Fedora has been slightly improved; the provided example
964
init.d script should work better on it and the INSTALL file should
965
document the needed dependencies.
966

    
967
Internal changes
968
~~~~~~~~~~~~~~~~
969

    
970
The deprecated ``QueryLocks`` LUXI request has been removed. Use
971
``Query(what=QR_LOCK, ...)`` instead.
972

    
973
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
974
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
975
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
976
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
977
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
978

    
979
RAPI client: ``CertificateError`` now derives from
980
``GanetiApiError``. This should make it more easy to handle Ganeti
981
errors.
982

    
983
Deprecation warnings due to PyCrypto/paramiko import in
984
``tools/setup-ssh`` have been silenced, as usually they are safe; please
985
make sure to run an up-to-date paramiko version, if you use this tool.
986

    
987
The QA scripts now depend on Python 2.5 or above (the main code base
988
still works with Python 2.4).
989

    
990
The configuration file (``config.data``) is now written without
991
indentation for performance reasons; if you want to edit it, it can be
992
re-formatted via ``tools/fmtjson``.
993

    
994
A number of bugs has been fixed in the cluster merge tool.
995

    
996
``x509`` certification verification (used in import-export) has been
997
changed to allow the same clock skew as permitted by the cluster
998
verification. This will remove some rare but hard to diagnose errors in
999
import-export.
1000

    
1001

    
1002
Version 2.6.0 rc4
1003
-----------------
1004

    
1005
*(Released Thu, 19 Jul 2012)*
1006

    
1007
Very few changes from rc4 to the final release, only bugfixes:
1008

    
1009
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
1010
  instance, fix CDROM booting for KVM instances)
1011
- fixed node group modification of node parameters
1012
- fixed issue in LUClusterVerifyGroup with multi-group clusters
1013
- fixed generation of bash completion to ensure a stable ordering
1014
- fixed a few typos
1015

    
1016

    
1017
Version 2.6.0 rc3
1018
-----------------
1019

    
1020
*(Released Fri, 13 Jul 2012)*
1021

    
1022
Third release candidate for 2.6. The following changes were done from
1023
rc3 to rc4:
1024

    
1025
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1026
- Fixed an inconsistency in the LUXI protocol with the provided
1027
  arguments (NOT backwards compatible)
1028
- Fixed a bug with node groups ipolicy where ``min`` was greater than
1029
  the cluster ``std`` value
1030
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1031
  easier instance debugging on nodes (requires ``hconfd`` to work)
1032

    
1033

    
1034
Version 2.6.0 rc2
1035
-----------------
1036

    
1037
*(Released Tue, 03 Jul 2012)*
1038

    
1039
Second release candidate for 2.6. The following changes were done from
1040
rc2 to rc3:
1041

    
1042
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1043
  master candidates
1044
- Fixed a RAPI regression on missing beparams/memory
1045
- Fixed redistribution of files on offline nodes
1046
- Added possibility to run activate-disks even though secondaries are
1047
  offline. With this change it relaxes also the strictness on some other
1048
  commands which use activate disks internally:
1049
  * ``gnt-instance start|reboot|rename|backup|export``
1050
- Made it possible to remove safely an instance if its secondaries are
1051
  offline
1052
- Made it possible to reinstall even though secondaries are offline
1053

    
1054

    
1055
Version 2.6.0 rc1
1056
-----------------
1057

    
1058
*(Released Mon, 25 Jun 2012)*
1059

    
1060
First release candidate for 2.6. The following changes were done from
1061
rc1 to rc2:
1062

    
1063
- Fixed bugs with disk parameters and ``rbd`` templates as well as
1064
  ``instance_os_add``
1065
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1066
  behaviour. It supports now the modify operation
1067
- ``hcheck`` implemented to analyze cluster health and possibility of
1068
  improving health by rebalance
1069
- ``hbal`` has been improved in dealing with split instances
1070

    
1071

    
1072
Version 2.6.0 beta2
1073
-------------------
1074

    
1075
*(Released Mon, 11 Jun 2012)*
1076

    
1077
Second beta release of 2.6. The following changes were done from beta2
1078
to rc1:
1079

    
1080
- Fixed ``daemon-util`` with non-root user models
1081
- Fixed creation of plain instances with ``--no-wait-for-sync``
1082
- Fix wrong iv_names when running ``cfgupgrade``
1083
- Export more information in RAPI group queries
1084
- Fixed bug when changing instance network interfaces
1085
- Extended burnin to do NIC changes
1086
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1087
- Changed default for DRBD barriers
1088
- Fixed DRBD error reporting for syncer rate
1089
- Verify the options on disk parameters
1090

    
1091
And of course various fixes to documentation and improved unittests and
1092
QA.
1093

    
1094

    
1095
Version 2.6.0 beta1
1096
-------------------
1097

    
1098
*(Released Wed, 23 May 2012)*
1099

    
1100
First beta release of 2.6. The following changes were done from beta1 to
1101
beta2:
1102

    
1103
- integrated patch for distributions without ``start-stop-daemon``
1104
- adapted example init.d script to work on Fedora
1105
- fixed log handling in Haskell daemons
1106
- adapted checks in the watcher for pycurl linked against libnss
1107
- add partial support for ``xl`` instead of ``xm`` for Xen
1108
- fixed a type issue in cluster verification
1109
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1110
  clusters)
1111

    
1112
Plus integrated fixes from the 2.5 branch:
1113

    
1114
- fixed ``kvm-ifup`` to use ``/bin/bash``
1115
- fixed parallel build failures
1116
- KVM live migration when using a custom keymap
1117

    
1118

    
1119
Version 2.5.2
1120
-------------
1121

    
1122
*(Released Tue, 24 Jul 2012)*
1123

    
1124
A small bugfix release, with no new features:
1125

    
1126
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1127
  different default shell (e.g. Debian, Ubuntu)
1128
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1129
  243 and Debian bug #650664)
1130
- fixed compatibility with KVM versions that don't support multiple boot
1131
  devices (fixes Issue 230 and Debian bug #624256)
1132

    
1133
Additionally, a few fixes were done to the build system (fixed parallel
1134
build failures) and to the unittests (fixed race condition in test for
1135
FileID functions, and the default enable/disable mode for QA test is now
1136
customisable).
1137

    
1138

    
1139
Version 2.5.1
1140
-------------
1141

    
1142
*(Released Fri, 11 May 2012)*
1143

    
1144
A small bugfix release.
1145

    
1146
The main issues solved are on the topic of compatibility with newer LVM
1147
releases:
1148

    
1149
- fixed parsing of ``lv_attr`` field
1150
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1151
  the ``--force`` flag is needed
1152

    
1153
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1154
to accept kernel 3.x too (was hardcoded to 2.6.*).
1155

    
1156
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1157
of hook results and that also made display of other errors suboptimal
1158
was fixed; the code behaves now like 2.4 and earlier.
1159

    
1160
Another change in 2.5, the cleanup of the OS scripts environment, is too
1161
aggressive: it removed even the ``PATH`` variable, which requires the OS
1162
scripts to *always* need to export it. Since this is a bit too strict,
1163
we now export a minimal PATH, the same that we export for hooks.
1164

    
1165
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1166
integrated into this release.
1167

    
1168
Finally, a few other miscellaneous changes were done (no new features,
1169
just small improvements):
1170

    
1171
- Fix ``gnt-group --help`` display
1172
- Fix hardcoded Xen kernel path
1173
- Fix grow-disk handling of invalid units
1174
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1175
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1176

    
1177

    
1178
Version 2.5.0
1179
-------------
1180

    
1181
*(Released Thu, 12 Apr 2012)*
1182

    
1183
Incompatible/important changes and bugfixes
1184
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1185

    
1186
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1187
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1188
  to match the underlying LUXI interface.
1189
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1190
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1191
  not maintain backwards-compatibility as the underlying operation
1192
  changed in an incompatible way. The RAPI client can talk to old
1193
  servers, but it needs to be told so as the return value changed.
1194
- When creating file-based instances via RAPI, the ``file_driver``
1195
  parameter no longer defaults to ``loop`` and must be specified.
1196
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1197
  ``link`` instead.
1198
- Support for the undocumented and deprecated RAPI instance creation
1199
  request format version 0 has been dropped. Use version 1, supported
1200
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1201
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1202
  documentation <install>`.
1203
- The "cluster-verify" hooks are now executed per group by the
1204
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1205
  if you just run ``gnt-cluster verify``, which generates one opcode per
1206
  group.
1207
- The environment as passed to the OS scripts is cleared, and thus no
1208
  environment variables defined in the node daemon's environment will be
1209
  inherited by the scripts.
1210
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1211
  deprecated.
1212
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1213
  support operations involving multiple node groups.
1214
- Offline nodes are ignored when failing over an instance.
1215
- Support for KVM version 1.0, which changed the version reporting format
1216
  from 3 to 2 digits.
1217
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1218
  removal.
1219
- ``Makefile`` is now compatible with Automake 1.11.2
1220
- Includes all bugfixes made in the 2.4 series
1221

    
1222
New features
1223
~~~~~~~~~~~~
1224

    
1225
- The ganeti-htools project has been merged into the ganeti-core source
1226
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1227
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1228
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1229
  GPT-style partition tables (requires `parted
1230
  <http://www.gnu.org/s/parted/>`_).
1231
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1232
- Allowed adding tags on instance creation.
1233
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1234
- Allow instances to be started in a paused state, enabling the user to
1235
  see the complete console output on boot using the console.
1236
- Added new hypervisor flag to control default reboot behaviour
1237
  (``reboot_behavior``).
1238
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1239
- Improved out-of-band management support:
1240

    
1241
  - Added ``gnt-node health`` command reporting the health status of
1242
    nodes.
1243
  - Added ``gnt-node power`` command to manage power status of nodes.
1244
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1245

    
1246
- Instance migration can fall back to failover if instance is not
1247
  running.
1248
- Filters can be used when listing nodes, instances, groups and locks;
1249
  see :manpage:`ganeti(7)` manpage.
1250
- Added post-execution status as variables to :doc:`hooks <hooks>`
1251
  environment.
1252
- Instance tags are exported/imported together with the instance.
1253
- When given an explicit job ID, ``gnt-job info`` will work for archived
1254
  jobs.
1255
- Jobs can define dependencies on other jobs (not yet supported via
1256
  RAPI or command line, but used by internal commands and usable via
1257
  LUXI).
1258

    
1259
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1260
    dependencies.
1261

    
1262
- Instance failover is now available as a RAPI resource
1263
  (``/2/instances/[instance_name]/failover``).
1264
- ``gnt-instance info`` defaults to static information if primary node
1265
  is offline.
1266
- Opcodes have a new ``comment`` attribute.
1267
- Added basic SPICE support to KVM hypervisor.
1268
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1269

    
1270
Node group improvements
1271
~~~~~~~~~~~~~~~~~~~~~~~
1272

    
1273
- ``gnt-cluster verify`` has been modified to check groups separately,
1274
  thereby improving performance.
1275
- Node group support has been added to ``gnt-cluster verify-disks``,
1276
  which now operates per node group.
1277
- Watcher has been changed to work better with node groups.
1278

    
1279
  - One process and state file per node group.
1280
  - Slow watcher in one group doesn't block other group's watcher.
1281

    
1282
- Added new command, ``gnt-group evacuate``, to move all instances in a
1283
  node group to other groups.
1284
- Added ``gnt-instance change-group`` to move an instance to another
1285
  node group.
1286
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1287
  per-group operations.
1288
- Node groups can be tagged.
1289
- Some operations switch from an exclusive to a shared lock as soon as
1290
  possible.
1291
- Instance's primary and secondary nodes' groups are now available as
1292
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1293
  and ``snodes.group.uuid``).
1294

    
1295
Misc
1296
~~~~
1297

    
1298
- Numerous updates to documentation and manpages.
1299

    
1300
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1301
    descriptions.
1302
  - Some opcode/job results are now also documented, see :doc:`RAPI
1303
    <rapi>`.
1304

    
1305
- A lockset's internal lock is now also visible in lock monitor.
1306
- Log messages from job queue workers now contain information about the
1307
  opcode they're processing.
1308
- ``gnt-instance console`` no longer requires the instance lock.
1309
- A short delay when waiting for job changes reduces the number of LUXI
1310
  requests significantly.
1311
- DRBD metadata volumes are overwritten with zeros during disk creation.
1312
- Out-of-band commands no longer acquire the cluster lock in exclusive
1313
  mode.
1314
- ``devel/upload`` now uses correct permissions for directories.
1315

    
1316

    
1317
Version 2.5.0 rc6
1318
-----------------
1319

    
1320
*(Released Fri, 23 Mar 2012)*
1321

    
1322
This was the sixth release candidate of the 2.5 series.
1323

    
1324

    
1325
Version 2.5.0 rc5
1326
-----------------
1327

    
1328
*(Released Mon, 9 Jan 2012)*
1329

    
1330
This was the fifth release candidate of the 2.5 series.
1331

    
1332

    
1333
Version 2.5.0 rc4
1334
-----------------
1335

    
1336
*(Released Thu, 27 Oct 2011)*
1337

    
1338
This was the fourth release candidate of the 2.5 series.
1339

    
1340

    
1341
Version 2.5.0 rc3
1342
-----------------
1343

    
1344
*(Released Wed, 26 Oct 2011)*
1345

    
1346
This was the third release candidate of the 2.5 series.
1347

    
1348

    
1349
Version 2.5.0 rc2
1350
-----------------
1351

    
1352
*(Released Tue, 18 Oct 2011)*
1353

    
1354
This was the second release candidate of the 2.5 series.
1355

    
1356

    
1357
Version 2.5.0 rc1
1358
-----------------
1359

    
1360
*(Released Tue, 4 Oct 2011)*
1361

    
1362
This was the first release candidate of the 2.5 series.
1363

    
1364

    
1365
Version 2.5.0 beta3
1366
-------------------
1367

    
1368
*(Released Wed, 31 Aug 2011)*
1369

    
1370
This was the third beta release of the 2.5 series.
1371

    
1372

    
1373
Version 2.5.0 beta2
1374
-------------------
1375

    
1376
*(Released Mon, 22 Aug 2011)*
1377

    
1378
This was the second beta release of the 2.5 series.
1379

    
1380

    
1381
Version 2.5.0 beta1
1382
-------------------
1383

    
1384
*(Released Fri, 12 Aug 2011)*
1385

    
1386
This was the first beta release of the 2.5 series.
1387

    
1388

    
1389
Version 2.4.5
1390
-------------
1391

    
1392
*(Released Thu, 27 Oct 2011)*
1393

    
1394
- Fixed bug when parsing command line parameter values ending in
1395
  backslash
1396
- Fixed assertion error after unclean master shutdown
1397
- Disable HTTP client pool for RPC, significantly reducing memory usage
1398
  of master daemon
1399
- Fixed queue archive creation with wrong permissions
1400

    
1401

    
1402
Version 2.4.4
1403
-------------
1404

    
1405
*(Released Tue, 23 Aug 2011)*
1406

    
1407
Small bug-fixes:
1408

    
1409
- Fixed documentation for importing with ``--src-dir`` option
1410
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1411
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1412

    
1413

    
1414
Version 2.4.3
1415
-------------
1416

    
1417
*(Released Fri, 5 Aug 2011)*
1418

    
1419
Many bug-fixes and a few small features:
1420

    
1421
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1422
  issues when you tried to add an instance with two MAC addresses in one
1423
  request
1424
- KVM: fixed per-instance stored UID value
1425
- KVM: configure bridged NICs at migration start
1426
- KVM: Fix a bug where instance will not start with never KVM versions
1427
  (>= 0.14)
1428
- Added OS search path to ``gnt-cluster info``
1429
- Fixed an issue with ``file_storage_dir`` where you were forced to
1430
  provide an absolute path, but the documentation states it is a
1431
  relative path, the documentation was right
1432
- Added a new parameter to instance stop/start called ``--no-remember``
1433
  that will make the state change to not be remembered
1434
- Implemented ``no_remember`` at RAPI level
1435
- Improved the documentation
1436
- Node evacuation: don't call IAllocator if node is already empty
1437
- Fixed bug in DRBD8 replace disks on current nodes
1438
- Fixed bug in recreate-disks for DRBD instances
1439
- Moved assertion checking locks in ``gnt-instance replace-disks``
1440
  causing it to abort with not owning the right locks for some situation
1441
- Job queue: Fixed potential race condition when cancelling queued jobs
1442
- Fixed off-by-one bug in job serial generation
1443
- ``gnt-node volumes``: Fix instance names
1444
- Fixed aliases in bash completion
1445
- Fixed a bug in reopening log files after being sent a SIGHUP
1446
- Added a flag to burnin to allow specifying VCPU count
1447
- Bugfixes to non-root Ganeti configuration
1448

    
1449

    
1450
Version 2.4.2
1451
-------------
1452

    
1453
*(Released Thu, 12 May 2011)*
1454

    
1455
Many bug-fixes and a few new small features:
1456

    
1457
- Fixed a bug related to log opening failures
1458
- Fixed a bug in instance listing with orphan instances
1459
- Fixed a bug which prevented resetting the cluster-level node parameter
1460
  ``oob_program`` to the default
1461
- Many fixes related to the ``cluster-merge`` tool
1462
- Fixed a race condition in the lock monitor, which caused failures
1463
  during (at least) creation of many instances in parallel
1464
- Improved output for gnt-job info
1465
- Removed the quiet flag on some ssh calls which prevented debugging
1466
  failures
1467
- Improved the N+1 failure messages in cluster verify by actually
1468
  showing the memory values (needed and available)
1469
- Increased lock attempt timeouts so that when executing long operations
1470
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1471
  too early and thus prevent the use of the 'fair' mechanism
1472
- Changed instance query data (``gnt-instance info``) to not acquire
1473
  locks unless needed, thus allowing its use on locked instance if only
1474
  static information is asked for
1475
- Improved behaviour with filesystems that do not support rename on an
1476
  opened file
1477
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1478
  kept locks on all nodes during the wipe, which is unneeded
1479
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1480
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1481
  10GiB) which caused the wipe to fail right at the end in some cases
1482
- Fixed master IP activation when doing master failover with no-voting
1483
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1484
  the master node itself
1485
- Fixed potential data-loss in under disk full conditions, where Ganeti
1486
  wouldn't check correctly the return code and would consider
1487
  partially-written files 'correct'
1488
- Fixed bug related to multiple VGs and DRBD disk replacing
1489
- Added new disk parameter ``metavg`` that allows placement of the meta
1490
  device for DRBD in a different volume group
1491
- Fixed error handling in the node daemon when the system libc doesn't
1492
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1493
- Fixed lock release during replace-disks, which kept cluster-wide locks
1494
  when doing disk replaces with an iallocator script
1495
- Added check for missing bridges in cluster verify
1496
- Handle EPIPE errors while writing to the terminal better, so that
1497
  piping the output to e.g. ``less`` doesn't cause a backtrace
1498
- Fixed rare case where a ^C during Luxi calls could have been
1499
  interpreted as server errors, instead of simply terminating
1500
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1501
  assign-nodes``)
1502
- Added a few more parameters to the KVM hypervisor, allowing a second
1503
  CDROM, custom disk type for CDROMs and a floppy image
1504
- Removed redundant message in instance rename when the name is given
1505
  already as a FQDN
1506
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1507
  disks on new nodes, allowing recreation when the original instance
1508
  nodes are completely gone
1509
- Added option when converting disk templates to DRBD to skip waiting
1510
  for the resync, in order to make the instance available sooner
1511
- Added two new variables to the OS scripts environment (containing the
1512
  instance's nodes)
1513
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1514
  to allow use of ``pvgrub`` as bootloader
1515
- Changed the instance memory modifications to only check out-of-memory
1516
  conditions on memory increases, and turned the secondary node warnings
1517
  into errors (they can still be overridden via ``--force``)
1518
- Fixed the handling of a corner case when the Python installation gets
1519
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1520
  to execute a command that doesn't exist
1521
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1522
  node of the instance returned failures during instance shutdown; this
1523
  adds the option ``--ignore-consistency`` to gnt-instance move
1524

    
1525
And as usual, various improvements to the error messages, documentation
1526
and man pages.
1527

    
1528

    
1529
Version 2.4.1
1530
-------------
1531

    
1532
*(Released Wed, 09 Mar 2011)*
1533

    
1534
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1535
the RAPI users file if run twice (even with ``--dry-run``).
1536

    
1537
The release fixes that bug (nothing else changed).
1538

    
1539

    
1540
Version 2.4.0
1541
-------------
1542

    
1543
*(Released Mon, 07 Mar 2011)*
1544

    
1545
Final 2.4.0 release. Just a few small fixes:
1546

    
1547
- Fixed RAPI node evacuate
1548
- Fixed the kvm-ifup script
1549
- Fixed internal error handling for special job cases
1550
- Updated man page to specify the escaping feature for options
1551

    
1552

    
1553
Version 2.4.0 rc3
1554
-----------------
1555

    
1556
*(Released Mon, 28 Feb 2011)*
1557

    
1558
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1559
that this feature wiped the disks of the wrong instance, leading to loss
1560
of data.
1561

    
1562
Other changes:
1563

    
1564
- Fixed title of query field containing instance name
1565
- Expanded the glossary in the documentation
1566
- Fixed one unittest (internal issue)
1567

    
1568

    
1569
Version 2.4.0 rc2
1570
-----------------
1571

    
1572
*(Released Mon, 21 Feb 2011)*
1573

    
1574
A number of bug fixes plus just a couple functionality changes.
1575

    
1576
On the user-visible side, the ``gnt-* list`` command output has changed
1577
with respect to "special" field states. The current rc1 style of display
1578
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1579
the default output mode special fields states are displayed as follows:
1580

    
1581
- Offline resource: ``*``
1582
- Unavailable/not applicable: ``-``
1583
- Data missing (RPC failure): ``?``
1584
- Unknown field: ``??``
1585

    
1586
Another user-visible change is the addition of ``--force-join`` to
1587
``gnt-node add``.
1588

    
1589
As for bug fixes:
1590

    
1591
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1592
- Fixed regression in RAPI/instance reinstall where all parameters were
1593
  required (instead of optional)
1594
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1595
- Fixed iallocator usage (offline nodes were not considered offline)
1596
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1597
- Fixed hypervisor and OS parameter validation with respect to
1598
  non-vm_capable nodes
1599
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1600
  cosmetic)
1601
- Fixed ``tools/listrunner`` with respect to agent-based usage
1602

    
1603

    
1604
Version 2.4.0 rc1
1605
-----------------
1606

    
1607
*(Released Fri,  4 Feb 2011)*
1608

    
1609
Many changes and fixes since the beta1 release. While there were some
1610
internal changes, the code has been mostly stabilised for the RC
1611
release.
1612

    
1613
Note: the dumb allocator was removed in this release, as it was not kept
1614
up-to-date with the IAllocator protocol changes. It is recommended to
1615
use the ``hail`` command from the ganeti-htools package.
1616

    
1617
Note: the 2.4 and up versions of Ganeti are not compatible with the
1618
0.2.x branch of ganeti-htools. You need to upgrade to
1619
ganeti-htools-0.3.0 (or later).
1620

    
1621
Regressions fixed from 2.3
1622
~~~~~~~~~~~~~~~~~~~~~~~~~~
1623

    
1624
- Fixed the ``gnt-cluster verify-disks`` command
1625
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1626
  serially on nodes)
1627
- Fixed disk adoption breakage
1628
- Fixed wrong headers in instance listing for field aliases
1629

    
1630
Other bugs fixed
1631
~~~~~~~~~~~~~~~~
1632

    
1633
- Fixed corner case in KVM handling of NICs
1634
- Fixed many cases of wrong handling of non-vm_capable nodes
1635
- Fixed a bug where a missing instance symlink was not possible to
1636
  recreate with any ``gnt-*`` command (now ``gnt-instance
1637
  activate-disks`` does it)
1638
- Fixed the volume group name as reported by ``gnt-cluster
1639
  verify-disks``
1640
- Increased timeouts for the import-export code, hopefully leading to
1641
  fewer aborts due network or instance timeouts
1642
- Fixed bug in ``gnt-node list-storage``
1643
- Fixed bug where not all daemons were started on cluster
1644
  initialisation, but only at the first watcher run
1645
- Fixed many bugs in the OOB implementation
1646
- Fixed watcher behaviour in presence of instances with offline
1647
  secondaries
1648
- Fixed instance list output for instances running on the wrong node
1649
- a few fixes to the cluster-merge tool, but it still cannot merge
1650
  multi-node groups (currently it is not recommended to use this tool)
1651

    
1652

    
1653
Improvements
1654
~~~~~~~~~~~~
1655

    
1656
- Improved network configuration for the KVM hypervisor
1657
- Added e1000 as a supported NIC for Xen-HVM
1658
- Improved the lvmstrap tool to also be able to use partitions, as
1659
  opposed to full disks
1660
- Improved speed of disk wiping (the cluster parameter
1661
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1662
  of instance creations
1663
- Added documentation for the OS parameters
1664
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1665
  hypervisor is not responding
1666
- Added display of blacklisted and hidden OS information in
1667
  ``gnt-cluster info``
1668
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1669
  NIC and node parameters, which might create problems with currently
1670
  invalid (but undetected) configuration files, but prevents validation
1671
  failures when unrelated parameters are modified
1672
- Changed cluster initialisation to wait for the master daemon to become
1673
  available
1674
- Expanded the RAPI interface:
1675

    
1676
  - Added config redistribution resource
1677
  - Added activation/deactivation of instance disks
1678
  - Added export of console information
1679

    
1680
- Implemented log file reopening on SIGHUP, which allows using
1681
  logrotate(8) for the Ganeti log files
1682
- Added a basic OOB helper script as an example
1683

    
1684

    
1685
Version 2.4.0 beta1
1686
-------------------
1687

    
1688
*(Released Fri, 14 Jan 2011)*
1689

    
1690
User-visible
1691
~~~~~~~~~~~~
1692

    
1693
- Fixed timezone issues when formatting timestamps
1694
- Added support for node groups, available via ``gnt-group`` and other
1695
  commands
1696
- Added out-of-band framework and management, see :doc:`design
1697
  document <design-oob>`
1698
- Removed support for roman numbers from ``gnt-node list`` and
1699
  ``gnt-instance list``.
1700
- Allowed modification of master network interface via ``gnt-cluster
1701
  modify --master-netdev``
1702
- Accept offline secondaries while shutting down instance disks
1703
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1704
- Added support for multiple LVM volume groups
1705
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1706
  requested
1707
- Added commands to list available fields:
1708

    
1709
  - ``gnt-node list-fields``
1710
  - ``gnt-group list-fields``
1711
  - ``gnt-instance list-fields``
1712

    
1713
- Updated documentation and man pages
1714

    
1715
Integration
1716
~~~~~~~~~~~
1717

    
1718
- Moved ``rapi_users`` file into separate directory, now named
1719
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1720
  symlink
1721
- Added new tool for running commands on many machines,
1722
  ``tools/ganeti-listrunner``
1723
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1724
  improving the ``gnt-instance console`` output
1725
- Allowed customisation of disk index separator at ``configure`` time
1726
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1727
- Added support for non-partitioned md disks in ``lvmstrap``
1728
- Added script to gracefully power off KVM instances
1729
- Split ``utils`` module into smaller parts
1730
- Changed query operations to return more detailed information, e.g.
1731
  whether an information is unavailable due to an offline node. To use
1732
  this new functionality, the LUXI call ``Query`` must be used. Field
1733
  information is now stored by the master daemon and can be retrieved
1734
  using ``QueryFields``. Instances, nodes and groups can also be queried
1735
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1736
  exposed via RAPI). The following commands make use of this
1737
  infrastructure change:
1738

    
1739
  - ``gnt-group list``
1740
  - ``gnt-group list-fields``
1741
  - ``gnt-node list``
1742
  - ``gnt-node list-fields``
1743
  - ``gnt-instance list``
1744
  - ``gnt-instance list-fields``
1745
  - ``gnt-debug locks``
1746

    
1747
Remote API
1748
~~~~~~~~~~
1749

    
1750
- New RAPI resources (see :doc:`rapi`):
1751

    
1752
  - ``/2/modify``
1753
  - ``/2/groups``
1754
  - ``/2/groups/[group_name]``
1755
  - ``/2/groups/[group_name]/assign-nodes``
1756
  - ``/2/groups/[group_name]/modify``
1757
  - ``/2/groups/[group_name]/rename``
1758
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1759

    
1760
- RAPI changes:
1761

    
1762
  - Implemented ``no_install`` for instance creation
1763
  - Implemented OS parameters for instance reinstallation, allowing
1764
    use of special settings on reinstallation (e.g. for preserving data)
1765

    
1766
Misc
1767
~~~~
1768

    
1769
- Added IPv6 support in import/export
1770
- Pause DRBD synchronization while wiping disks on instance creation
1771
- Updated unittests and QA scripts
1772
- Improved network parameters passed to KVM
1773
- Converted man pages from docbook to reStructuredText
1774

    
1775

    
1776
Version 2.3.1
1777
-------------
1778

    
1779
*(Released Mon, 20 Dec 2010)*
1780

    
1781
Released version 2.3.1~rc1 without any changes.
1782

    
1783

    
1784
Version 2.3.1 rc1
1785
-----------------
1786

    
1787
*(Released Wed, 1 Dec 2010)*
1788

    
1789
- impexpd: Disable OpenSSL compression in socat if possible (backport
1790
  from master, commit e90739d625b, see :doc:`installation guide
1791
  <install-quick>` for details)
1792
- Changed unittest coverage report to exclude test scripts
1793
- Added script to check version format
1794

    
1795

    
1796
Version 2.3.0
1797
-------------
1798

    
1799
*(Released Wed, 1 Dec 2010)*
1800

    
1801
Released version 2.3.0~rc1 without any changes.
1802

    
1803

    
1804
Version 2.3.0 rc1
1805
-----------------
1806

    
1807
*(Released Fri, 19 Nov 2010)*
1808

    
1809
A number of bugfixes and documentation updates:
1810

    
1811
- Update ganeti-os-interface documentation
1812
- Fixed a bug related to duplicate MACs or similar items which should be
1813
  unique
1814
- Fix breakage in OS state modify
1815
- Reinstall instance: disallow offline secondaries (fixes bug related to
1816
  OS changing but reinstall failing)
1817
- plus all the other fixes between 2.2.1 and 2.2.2
1818

    
1819

    
1820
Version 2.3.0 rc0
1821
-----------------
1822

    
1823
*(Released Tue, 2 Nov 2010)*
1824

    
1825
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1826
- Fixed master failover race with watcher
1827
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1828
  configuration
1829
- Accept previously stopped instance for export with instance removal
1830
- Simplify and extend the environment variables for instance OS scripts
1831
- Added new node flags, ``master_capable`` and ``vm_capable``
1832
- Added optional instance disk wiping prior during allocation. This is a
1833
  cluster-wide option and can be set/modified using
1834
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1835
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1836
  :doc:`install-quick`
1837
- Added a new watcher option (``--ignore-pause``)
1838
- Added option to ignore offline node on instance start/stop
1839
  (``--ignore-offline``)
1840
- Allow overriding OS parameters with ``gnt-instance reinstall``
1841
- Added ability to change node's secondary IP address using ``gnt-node
1842
  modify``
1843
- Implemented privilege separation for all daemons except
1844
  ``ganeti-noded``, see ``configure`` options
1845
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1846
  verify``
1847
- Implemented job priorities (see ``ganeti(7)`` manpage)
1848
- Ignore failures while shutting down instances during failover from
1849
  offline node
1850
- Exit daemon's bootstrap process only once daemon is ready
1851
- Export more information via ``LUInstanceQuery``/remote API
1852
- Improved documentation, QA and unittests
1853
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1854
  restart if the file was created or changed
1855
- Added LUXI protocol version sent with each request and response,
1856
  allowing detection of server/client mismatches
1857
- Moved the Python scripts among gnt-* and ganeti-* into modules
1858
- Moved all code related to setting up SSH to an external script,
1859
  ``setup-ssh``
1860
- Infrastructure changes for node group support in future versions
1861

    
1862

    
1863
Version 2.2.2
1864
-------------
1865

    
1866
*(Released Fri, 19 Nov 2010)*
1867

    
1868
A few small bugs fixed, and some improvements to the build system:
1869

    
1870
- Fix documentation regarding conversion to drbd
1871
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1872
  -B``)
1873
- Fix error handling in node modify with multiple changes
1874
- Allow remote imports without checked names
1875

    
1876

    
1877
Version 2.2.1
1878
-------------
1879

    
1880
*(Released Tue, 19 Oct 2010)*
1881

    
1882
- Disable SSL session ID cache in RPC client
1883

    
1884

    
1885
Version 2.2.1 rc1
1886
-----------------
1887

    
1888
*(Released Thu, 14 Oct 2010)*
1889

    
1890
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1891
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1892
  with GnuTLS
1893
- Fix problems with interaction between Curl and Python's HTTP server,
1894
  resulting in increased speed in many RPC calls
1895
- Improve our release script to prevent breakage with older aclocal and
1896
  Python 2.6
1897

    
1898

    
1899
Version 2.2.1 rc0
1900
-----------------
1901

    
1902
*(Released Thu, 7 Oct 2010)*
1903

    
1904
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1905
  value retrieved from master
1906
- Added support for blacklisted or hidden OS definitions
1907
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1908
- Added support for -mem-path in KVM hypervisor abstraction layer
1909
- Allow overriding instance parameters in tool for inter-cluster
1910
  instance moves (``tools/move-instance``)
1911
- Improved opcode summaries (e.g. in ``gnt-job list``)
1912
- Improve consistency of OS listing by sorting it
1913
- Documentation updates
1914

    
1915

    
1916
Version 2.2.0.1
1917
---------------
1918

    
1919
*(Released Fri, 8 Oct 2010)*
1920

    
1921
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1922

    
1923

    
1924
Version 2.2.0
1925
-------------
1926

    
1927
*(Released Mon, 4 Oct 2010)*
1928

    
1929
- Fixed regression in ``gnt-instance rename``
1930

    
1931

    
1932
Version 2.2.0 rc2
1933
-----------------
1934

    
1935
*(Released Wed, 22 Sep 2010)*
1936

    
1937
- Fixed OS_VARIANT variable for OS scripts
1938
- Fixed cluster tag operations via RAPI
1939
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1940
- Disabled RAPI CA checks in watcher
1941

    
1942

    
1943
Version 2.2.0 rc1
1944
-----------------
1945

    
1946
*(Released Mon, 23 Aug 2010)*
1947

    
1948
- Support DRBD versions of the format "a.b.c.d"
1949
- Updated manpages
1950
- Re-introduce support for usage from multiple threads in RAPI client
1951
- Instance renames and modify via RAPI
1952
- Work around race condition between processing and archival in job
1953
  queue
1954
- Mark opcodes following failed one as failed, too
1955
- Job field ``lock_status`` was removed due to difficulties making it
1956
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1957
  locks is expected for a later 2.2.x release
1958
- Fixed dry-run behaviour with many commands
1959
- Support ``ssh-agent`` again when adding nodes
1960
- Many additional bugfixes
1961

    
1962

    
1963
Version 2.2.0 rc0
1964
-----------------
1965

    
1966
*(Released Fri, 30 Jul 2010)*
1967

    
1968
Important change: the internal RPC mechanism between Ganeti nodes has
1969
changed from using a home-grown http library (based on the Python base
1970
libraries) to use the PycURL library. This requires that PycURL is
1971
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1972
against GnuTLS by default. cURL's support for GnuTLS had known issues
1973
before cURL 7.21.0 and we recommend using the latest cURL release or
1974
linking against OpenSSL. Most other distributions already link PycURL
1975
and cURL against OpenSSL. The command::
1976

    
1977
  python -c 'import pycurl; print pycurl.version'
1978

    
1979
can be used to determine the libraries PycURL and cURL are linked
1980
against.
1981

    
1982
Other significant changes:
1983

    
1984
- Rewrote much of the internals of the job queue, in order to achieve
1985
  better parallelism; this decouples job query operations from the job
1986
  processing, and it should allow much nicer behaviour of the master
1987
  daemon under load, and it also has uncovered some long-standing bugs
1988
  related to the job serialisation (now fixed)
1989
- Added a default iallocator setting to the cluster parameters,
1990
  eliminating the need to always pass nodes or an iallocator for
1991
  operations that require selection of new node(s)
1992
- Added experimental support for the LXC virtualization method
1993
- Added support for OS parameters, which allows the installation of
1994
  instances to pass parameter to OS scripts in order to customise the
1995
  instance
1996
- Added a hypervisor parameter controlling the migration type (live or
1997
  non-live), since hypervisors have various levels of reliability; this
1998
  has renamed the 'live' parameter to 'mode'
1999
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
2000
  logical volumes, meaning that cluster verify will ignore them and not
2001
  flag their presence as errors
2002
- The watcher will now reset the error count for failed instances after
2003
  8 hours, thus allowing self-healing if the problem that caused the
2004
  instances to be down/fail to start has cleared in the meantime
2005
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2006
  check for, and warn, if the drbd module parameter ``usermode_helper``
2007
  is not consistent with the cluster-wide setting; this is needed to
2008
  make diagnose easier of failed drbd creations
2009
- Started adding base IPv6 support, but this is not yet
2010
  enabled/available for use
2011
- Rename operations (cluster, instance) will now return the new name,
2012
  which is especially useful if a short name was passed in
2013
- Added support for instance migration in RAPI
2014
- Added a tool to pre-configure nodes for the SSH setup, before joining
2015
  them to the cluster; this will allow in the future a simplified model
2016
  for node joining (but not yet fully enabled in 2.2); this needs the
2017
  paramiko python library
2018
- Fixed handling of name-resolving errors
2019
- Fixed consistency of job results on the error path
2020
- Fixed master-failover race condition when executed multiple times in
2021
  sequence
2022
- Fixed many bugs related to the job queue (mostly introduced during the
2023
  2.2 development cycle, so not all are impacting 2.1)
2024
- Fixed instance migration with missing disk symlinks
2025
- Fixed handling of unknown jobs in ``gnt-job archive``
2026
- And many other small fixes/improvements
2027

    
2028
Internal changes:
2029

    
2030
- Enhanced both the unittest and the QA coverage
2031
- Switched the opcode validation to a generic model, and extended the
2032
  validation to all opcode parameters
2033
- Changed more parts of the code that write shell scripts to use the
2034
  same class for this
2035
- Switched the master daemon to use the asyncore library for the Luxi
2036
  server endpoint
2037

    
2038

    
2039
Version 2.2.0 beta0
2040
-------------------
2041

    
2042
*(Released Thu, 17 Jun 2010)*
2043

    
2044
- Added tool (``move-instance``) and infrastructure to move instances
2045
  between separate clusters (see :doc:`separate documentation
2046
  <move-instance>` and :doc:`design document <design-2.2>`)
2047
- Added per-request RPC timeout
2048
- RAPI now requires a Content-Type header for requests with a body (e.g.
2049
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2050
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
2051
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2052
  reachable
2053
- Implemented initial support for running Ganeti daemons as separate
2054
  users, see configure-time flags ``--with-user-prefix`` and
2055
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2056
  time)
2057
- Instances can be removed after export (``gnt-backup export
2058
  --remove-instance``)
2059
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2060
  key (instead of 1024 bit)
2061
- Added new cluster configuration file for cluster domain secret
2062
- Import/export now use SSL instead of SSH
2063
- Added support for showing estimated time when exporting an instance,
2064
  see the ``ganeti-os-interface(7)`` manpage and look for
2065
  ``EXP_SIZE_FD``
2066

    
2067

    
2068
Version 2.1.8
2069
-------------
2070

    
2071
*(Released Tue, 16 Nov 2010)*
2072

    
2073
Some more bugfixes. Unless critical bugs occur, this will be the last
2074
2.1 release:
2075

    
2076
- Fix case of MAC special-values
2077
- Fix mac checker regex
2078
- backend: Fix typo causing "out of range" error
2079
- Add missing --units in gnt-instance list man page
2080

    
2081

    
2082
Version 2.1.7
2083
-------------
2084

    
2085
*(Released Tue, 24 Aug 2010)*
2086

    
2087
Bugfixes only:
2088
  - Don't ignore secondary node silently on non-mirrored disk templates
2089
    (issue 113)
2090
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2091
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2092
  - Properly document the usb_mouse parameter
2093
  - Fix path in ganeti-rapi(8) (issue 116)
2094
  - Adjust error message when the ganeti user's .ssh directory is
2095
    missing
2096
  - Add same-node-check when changing the disk template to drbd
2097

    
2098

    
2099
Version 2.1.6
2100
-------------
2101

    
2102
*(Released Fri, 16 Jul 2010)*
2103

    
2104
Bugfixes only:
2105
  - Add an option to only select some reboot types during qa/burnin.
2106
    (on some hypervisors consequent reboots are not supported)
2107
  - Fix infrequent race condition in master failover. Sometimes the old
2108
    master ip address would be still detected as up for a short time
2109
    after it was removed, causing failover to fail.
2110
  - Decrease mlockall warnings when the ctypes module is missing. On
2111
    Python 2.4 we support running even if no ctypes module is installed,
2112
    but we were too verbose about this issue.
2113
  - Fix building on old distributions, on which man doesn't have a
2114
    --warnings option.
2115
  - Fix RAPI not to ignore the MAC address on instance creation
2116
  - Implement the old instance creation format in the RAPI client.
2117

    
2118

    
2119
Version 2.1.5
2120
-------------
2121

    
2122
*(Released Thu, 01 Jul 2010)*
2123

    
2124
A small bugfix release:
2125
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2126
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2127
    a non-existing option
2128
  - Fix instance create: the --force-variant option was ignored
2129
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2130
  - Fix modify node storage with non-FQDN arguments
2131
  - Fix RAPI client to authenticate under Python 2.6 when used
2132
    for more than 5 requests needing authentication
2133
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2134
    when converting a non-shutdown drbd instance to plain
2135

    
2136

    
2137
Version 2.1.4
2138
-------------
2139

    
2140
*(Released Fri, 18 Jun 2010)*
2141

    
2142
A small bugfix release:
2143

    
2144
  - Fix live migration of KVM instances started with older Ganeti
2145
    versions which had fewer hypervisor parameters
2146
  - Fix gnt-instance grow-disk on down instances
2147
  - Fix an error-reporting bug during instance migration
2148
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2149
    silently ignoring broken ones
2150
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2151
    users
2152
  - Fix bug triggered by different API version os-es on different nodes
2153
  - Fix a bug in instance startup with custom hvparams: OS level
2154
    parameters would fail to be applied.
2155
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2156
    make it work completely well with OpenSSL)
2157
  - Fix handling of errors when resolving names from DNS
2158

    
2159

    
2160
Version 2.1.3
2161
-------------
2162

    
2163
*(Released Thu, 3 Jun 2010)*
2164

    
2165
A medium sized development cycle. Some new features, and some
2166
fixes/small improvements/cleanups.
2167

    
2168
Significant features
2169
~~~~~~~~~~~~~~~~~~~~
2170

    
2171
The node deamon now tries to mlock itself into memory, unless the
2172
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2173
its logs, and falls back to console logging. This allows emergency
2174
features such as ``gnt-node powercycle`` to work even in the event of a
2175
broken node disk (tested offlining the disk hosting the node's
2176
filesystem and dropping its memory caches; don't try this at home)
2177

    
2178
KVM: add vhost-net acceleration support. It can be tested with a new
2179
enough version of the kernel and of qemu-kvm.
2180

    
2181
KVM: Add instance chrooting feature. If you use privilege dropping for
2182
your VMs you can also now force them to chroot to an empty directory,
2183
before starting the emulated guest.
2184

    
2185
KVM: Add maximum migration bandwith and maximum downtime tweaking
2186
support (requires a new-enough version of qemu-kvm).
2187

    
2188
Cluster verify will now warn if the master node doesn't have the master
2189
ip configured on it.
2190

    
2191
Add a new (incompatible) instance creation request format to RAPI which
2192
supports all parameters (previously only a subset was supported, and it
2193
wasn't possible to extend the old format to accomodate all the new
2194
features. The old format is still supported, and a client can check for
2195
this feature, before using it, by checking for its presence in the
2196
``features`` RAPI resource.
2197

    
2198
Now with ancient latin support. Try it passing the ``--roman`` option to
2199
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2200
(requires the python-roman module to be installed, in order to work).
2201

    
2202
Other changes
2203
~~~~~~~~~~~~~
2204

    
2205
As usual many internal code refactorings, documentation updates, and
2206
such. Among others:
2207

    
2208
  - Lots of improvements and cleanups to the experimental Remote API
2209
    (RAPI) client library.
2210
  - A new unit test suite for the core daemon libraries.
2211
  - A fix to creating missing directories makes sure the umask is not
2212
    applied anymore. This enforces the same directory permissions
2213
    everywhere.
2214
  - Better handling terminating daemons with ctrl+c (used when running
2215
    them in debugging mode).
2216
  - Fix a race condition in live migrating a KVM instance, when stat()
2217
    on the old proc status file returned EINVAL, which is an unexpected
2218
    value.
2219
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2220
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2221

    
2222

    
2223
Version 2.1.2.1
2224
---------------
2225

    
2226
*(Released Fri, 7 May 2010)*
2227

    
2228
Fix a bug which prevented untagged KVM instances from starting.
2229

    
2230

    
2231
Version 2.1.2
2232
-------------
2233

    
2234
*(Released Fri, 7 May 2010)*
2235

    
2236
Another release with a long development cycle, during which many
2237
different features were added.
2238

    
2239
Significant features
2240
~~~~~~~~~~~~~~~~~~~~
2241

    
2242
The KVM hypervisor now can run the individual instances as non-root, to
2243
reduce the impact of a VM being hijacked due to bugs in the
2244
hypervisor. It is possible to run all instances as a single (non-root)
2245
user, to manually specify a user for each instance, or to dynamically
2246
allocate a user out of a cluster-wide pool to each instance, with the
2247
guarantee that no two instances will run under the same user ID on any
2248
given node.
2249

    
2250
An experimental RAPI client library, that can be used standalone
2251
(without the other Ganeti libraries), is provided in the source tree as
2252
``lib/rapi/client.py``. Note this client might change its interface in
2253
the future, as we iterate on its capabilities.
2254

    
2255
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2256
replace the cluster's certificates and crypto keys. This might help in
2257
case they have been compromised, or have simply expired.
2258

    
2259
A new disk option for instance creation has been added that allows one
2260
to "adopt" currently existing logical volumes, with data
2261
preservation. This should allow easier migration to Ganeti from
2262
unmanaged (or managed via other software) instances.
2263

    
2264
Another disk improvement is the possibility to convert between redundant
2265
(DRBD) and plain (LVM) disk configuration for an instance. This should
2266
allow better scalability (starting with one node and growing the
2267
cluster, or shrinking a two-node cluster to one node).
2268

    
2269
A new feature that could help with automated node failovers has been
2270
implemented: if a node sees itself as offline (by querying the master
2271
candidates), it will try to shutdown (hard) all instances and any active
2272
DRBD devices. This reduces the risk of duplicate instances if an
2273
external script automatically failovers the instances on such nodes. To
2274
enable this, the cluster parameter ``maintain_node_health`` should be
2275
enabled; in the future this option (per the name) will enable other
2276
automatic maintenance features.
2277

    
2278
Instance export/import now will reuse the original instance
2279
specifications for all parameters; that means exporting an instance,
2280
deleting it and the importing it back should give an almost identical
2281
instance. Note that the default import behaviour has changed from
2282
before, where it created only one NIC; now it recreates the original
2283
number of NICs.
2284

    
2285
Cluster verify has added a few new checks: SSL certificates validity,
2286
/etc/hosts consistency across the cluster, etc.
2287

    
2288
Other changes
2289
~~~~~~~~~~~~~
2290

    
2291
As usual, many internal changes were done, documentation fixes,
2292
etc. Among others:
2293

    
2294
- Fixed cluster initialization with disabled cluster storage (regression
2295
  introduced in 2.1.1)
2296
- File-based storage supports growing the disks
2297
- Fixed behaviour of node role changes
2298
- Fixed cluster verify for some corner cases, plus a general rewrite of
2299
  cluster verify to allow future extension with more checks
2300
- Fixed log spamming by watcher and node daemon (regression introduced
2301
  in 2.1.1)
2302
- Fixed possible validation issues when changing the list of enabled
2303
  hypervisors
2304
- Fixed cleanup of /etc/hosts during node removal
2305
- Fixed RAPI response for invalid methods
2306
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2307
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2308
  from ide disks, etc.)
2309
- Allow OS changes without re-installation (to record a changed OS
2310
  outside of Ganeti, or to allow OS renames)
2311
- Allow instance creation without OS installation (useful for example if
2312
  the OS will be installed manually, or restored from a backup not in
2313
  Ganeti format)
2314
- Implemented option to make cluster ``copyfile`` use the replication
2315
  network
2316
- Added list of enabled hypervisors to ssconf (possibly useful for
2317
  external scripts)
2318
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2319
  1.2 clusters
2320
- A partial form of node re-IP is possible via node readd, which now
2321
  allows changed node primary IP
2322
- Command line utilities now show an informational message if the job is
2323
  waiting for a lock
2324
- The logs of the master daemon now show the PID/UID/GID of the
2325
  connected client
2326

    
2327

    
2328
Version 2.1.1
2329
-------------
2330

    
2331
*(Released Fri, 12 Mar 2010)*
2332

    
2333
During the 2.1.0 long release candidate cycle, a lot of improvements and
2334
changes have accumulated with were released later as 2.1.1.
2335

    
2336
Major changes
2337
~~~~~~~~~~~~~
2338

    
2339
The node evacuate command (``gnt-node evacuate``) was significantly
2340
rewritten, and as such the IAllocator protocol was changed - a new
2341
request type has been added. This unfortunate change during a stable
2342
series is designed to improve performance of node evacuations; on
2343
clusters with more than about five nodes and which are well-balanced,
2344
evacuation should proceed in parallel for all instances of the node
2345
being evacuated. As such, any existing IAllocator scripts need to be
2346
updated, otherwise the above command will fail due to the unknown
2347
request. The provided "dumb" allocator has not been updated; but the
2348
ganeti-htools package supports the new protocol since version 0.2.4.
2349

    
2350
Another important change is increased validation of node and instance
2351
names. This might create problems in special cases, if invalid host
2352
names are being used.
2353

    
2354
Also, a new layer of hypervisor parameters has been added, that sits at
2355
OS level between the cluster defaults and the instance ones. This allows
2356
customisation of virtualization parameters depending on the installed
2357
OS. For example instances with OS 'X' may have a different KVM kernel
2358
(or any other parameter) than the cluster defaults. This is intended to
2359
help managing a multiple OSes on the same cluster, without manual
2360
modification of each instance's parameters.
2361

    
2362
A tool for merging clusters, ``cluster-merge``, has been added in the
2363
tools sub-directory.
2364

    
2365
Bug fixes
2366
~~~~~~~~~
2367

    
2368
- Improved the int/float conversions that should make the code more
2369
  robust in face of errors from the node daemons
2370
- Fixed the remove node code in case of internal configuration errors
2371
- Fixed the node daemon behaviour in face of inconsistent queue
2372
  directory (e.g. read-only file-system where we can't open the files
2373
  read-write, etc.)
2374
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2375
  now it either aborts cleanly or, if given the new "auto_promote"
2376
  parameter, will automatically promote other nodes as needed
2377
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2378
  completely prevent Ganeti from working
2379
- Fixed bug for instance export when not all disks were successfully
2380
  exported
2381
- Fixed behaviour of node add when the new node is slow in starting up
2382
  the node daemon
2383
- Fixed handling of signals in the LUXI client, which should improve
2384
  behaviour of command-line scripts
2385
- Added checks for invalid node/instance names in the configuration (now
2386
  flagged during cluster verify)
2387
- Fixed watcher behaviour for disk activation errors
2388
- Fixed two potentially endless loops in http library, which led to the
2389
  RAPI daemon hanging and consuming 100% CPU in some cases
2390
- Fixed bug in RAPI daemon related to hashed passwords
2391
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2392
  instances
2393
- Enhanced compatibility with non-Debian OSes, but not using absolute
2394
  path in some commands and allowing customisation of the ssh
2395
  configuration directory
2396
- Fixed possible future issue with new Python versions by abiding to the
2397
  proper use of ``__slots__`` attribute on classes
2398
- Added checks that should prevent directory traversal attacks
2399
- Many documentation fixes based on feedback from users
2400

    
2401
New features
2402
~~~~~~~~~~~~
2403

    
2404
- Added an "early_release" more for instance replace disks and node
2405
  evacuate, where we release locks earlier and thus allow higher
2406
  parallelism within the cluster
2407
- Added watcher hooks, intended to allow the watcher to restart other
2408
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2409
  course for any other purpose
2410
- Added a compile-time disable for DRBD barriers, to increase
2411
  performance if the administrator trusts the power supply or the
2412
  storage system to not lose writes
2413
- Added the option of using syslog for logging instead of, or in
2414
  addition to, Ganeti's own log files
2415
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2416
  can indeed boot from a paravirtual NIC
2417
- Added a generic debug level for many operations; while this is not
2418
  used widely yet, it allows one to pass the debug value all the way to
2419
  the OS scripts
2420
- Enhanced the hooks environment for instance moves (failovers,
2421
  migrations) where the primary/secondary nodes changed during the
2422
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2423
- Enhanced data validations for many user-supplied values; one important
2424
  item is the restrictions imposed on instance and node names, which
2425
  might reject some (invalid) host names
2426
- Add a configure-time option to disable file-based storage, if it's not
2427
  needed; this allows greater security separation between the master
2428
  node and the other nodes from the point of view of the inter-node RPC
2429
  protocol
2430
- Added user notification in interactive tools if job is waiting in the
2431
  job queue or trying to acquire locks
2432
- Added log messages when a job is waiting for locks
2433
- Added filtering by node tags in instance operations which admit
2434
  multiple instances (start, stop, reboot, reinstall)
2435
- Added a new tool for cluster mergers, ``cluster-merge``
2436
- Parameters from command line which are of the form ``a=b,c=d`` can now
2437
  use backslash escapes to pass in values which contain commas,
2438
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2439
  ``b,c``
2440
- For KVM, the instance name is the first parameter passed to KVM, so
2441
  that it's more visible in the process list
2442

    
2443

    
2444
Version 2.1.0
2445
-------------
2446

    
2447
*(Released Tue, 2 Mar 2010)*
2448

    
2449
Ganeti 2.1 brings many improvements with it. Major changes:
2450

    
2451
- Added infrastructure to ease automated disk repairs
2452
- Added new daemon to export configuration data in a cheaper way than
2453
  using the remote API
2454
- Instance NICs can now be routed instead of being associated with a
2455
  networking bridge
2456
- Improved job locking logic to reduce impact of jobs acquiring multiple
2457
  locks waiting for other long-running jobs
2458

    
2459
In-depth implementation details can be found in the Ganeti 2.1 design
2460
document.
2461

    
2462
Details
2463
~~~~~~~
2464

    
2465
- Added chroot hypervisor
2466
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2467
  ``device_model``)
2468
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2469
  ``bootloader_path`` and ``bootloader_args``)
2470
- Added the ``use_localtime`` option for the xen-hvm and kvm
2471
  hypervisors, and the default value for this has changed to false (in
2472
  2.0 xen-hvm always enabled it)
2473
- Added luxi call to submit multiple jobs in one go
2474
- Added cluster initialization option to not modify ``/etc/hosts``
2475
  file on nodes
2476
- Added network interface parameters
2477
- Added dry run mode to some LUs
2478
- Added RAPI resources:
2479

    
2480
  - ``/2/instances/[instance_name]/info``
2481
  - ``/2/instances/[instance_name]/replace-disks``
2482
  - ``/2/nodes/[node_name]/evacuate``
2483
  - ``/2/nodes/[node_name]/migrate``
2484
  - ``/2/nodes/[node_name]/role``
2485
  - ``/2/nodes/[node_name]/storage``
2486
  - ``/2/nodes/[node_name]/storage/modify``
2487
  - ``/2/nodes/[node_name]/storage/repair``
2488

    
2489
- Added OpCodes to evacuate or migrate all instances on a node
2490
- Added new command to list storage elements on nodes (``gnt-node
2491
  list-storage``) and modify them (``gnt-node modify-storage``)
2492
- Added new ssconf files with master candidate IP address
2493
  (``ssconf_master_candidates_ips``), node primary IP address
2494
  (``ssconf_node_primary_ips``) and node secondary IP address
2495
  (``ssconf_node_secondary_ips``)
2496
- Added ``ganeti-confd`` and a client library to query the Ganeti
2497
  configuration via UDP
2498
- Added ability to run hooks after cluster initialization and before
2499
  cluster destruction
2500
- Added automatic mode for disk replace (``gnt-instance replace-disks
2501
  --auto``)
2502
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2503
  after catastrophic data-loss
2504
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2505
  groups
2506
- Added ``gnt-instance move`` command to move instances
2507
- Added ``gnt-cluster watcher`` command to control watcher
2508
- Added ``gnt-node powercycle`` command to powercycle nodes
2509
- Added new job status field ``lock_status``
2510
- Added parseable error codes to cluster verification (``gnt-cluster
2511
  verify --error-codes``) and made output less verbose (use
2512
  ``--verbose`` to restore previous behaviour)
2513
- Added UUIDs to the main config entities (cluster, nodes, instances)
2514
- Added support for OS variants
2515
- Added support for hashed passwords in the Ganeti remote API users file
2516
  (``rapi_users``)
2517
- Added option to specify maximum timeout on instance shutdown
2518
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2519
- Added new helper script to start and stop Ganeti daemons
2520
  (``daemon-util``), with the intent to reduce the work necessary to
2521
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2522
  from one place
2523
- Added more unittests
2524
- Fixed critical bug in ganeti-masterd startup
2525
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2526
  now customisable at the cluster level for both the KVM and Xen
2527
  hypervisors using the new ``migration_port`` parameter
2528
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2529
  reinstalling an instance
2530
- Allowed ``@`` in tag names
2531
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2532
- Many documentation updates
2533
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2534
- ``gnt-instance reinstall`` can now reinstall multiple instances
2535
- Updated many command line parameters
2536
- Introduced new OS API version 15
2537
- No longer support a default hypervisor
2538
- Treat virtual LVs as inexistent
2539
- Improved job locking logic to reduce lock contention
2540
- Match instance and node names case insensitively
2541
- Reimplemented bash completion script to be more complete
2542
- Improved burnin
2543

    
2544

    
2545
Version 2.0.6
2546
-------------
2547

    
2548
*(Released Thu, 4 Feb 2010)*
2549

    
2550
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2551
- Fix a string formatting bug
2552
- Improve safety of the code in some error paths
2553
- Improve data validation in the master of values returned from nodes
2554

    
2555

    
2556
Version 2.0.5
2557
-------------
2558

    
2559
*(Released Thu, 17 Dec 2009)*
2560

    
2561
- Fix security issue due to missing validation of iallocator names; this
2562
  allows local and remote execution of arbitrary executables
2563
- Fix failure of gnt-node list during instance removal
2564
- Ship the RAPI documentation in the archive
2565

    
2566

    
2567
Version 2.0.4
2568
-------------
2569

    
2570
*(Released Wed, 30 Sep 2009)*
2571

    
2572
- Fixed many wrong messages
2573
- Fixed a few bugs related to the locking library
2574
- Fixed MAC checking at instance creation time
2575
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2576
- Fixed a few issues related to signal handling in both daemons and
2577
  scripts
2578
- Fixed the example startup script provided
2579
- Fixed insserv dependencies in the example startup script (patch from
2580
  Debian)
2581
- Fixed handling of drained nodes in the iallocator framework
2582
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2583
  #528618)
2584
- Fixed error related to invalid job IDs in job polling
2585
- Fixed job/opcode persistence on unclean master shutdown
2586
- Fixed handling of partial job processing after unclean master
2587
  shutdown
2588
- Fixed error reporting from LUs, previously all errors were converted
2589
  into execution errors
2590
- Fixed error reporting from burnin
2591
- Decreased significantly the memory usage of the job queue
2592
- Optimised slightly multi-job submission
2593
- Optimised slightly opcode loading
2594
- Backported the multi-job submit framework from the development
2595
  branch; multi-instance start and stop should be faster
2596
- Added script to clean archived jobs after 21 days; this will reduce
2597
  the size of the queue directory
2598
- Added some extra checks in disk size tracking
2599
- Added an example ethers hook script
2600
- Added a cluster parameter that prevents Ganeti from modifying of
2601
  /etc/hosts
2602
- Added more node information to RAPI responses
2603
- Added a ``gnt-job watch`` command that allows following the ouput of a
2604
  job
2605
- Added a bind-address option to ganeti-rapi
2606
- Added more checks to the configuration verify
2607
- Enhanced the burnin script such that some operations can be retried
2608
  automatically
2609
- Converted instance reinstall to multi-instance model
2610

    
2611

    
2612
Version 2.0.3
2613
-------------
2614

    
2615
*(Released Fri, 7 Aug 2009)*
2616

    
2617
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2618
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2619
  instances have mismatched disk sizes in the configuration
2620
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2621
  any configuration mismatches for disk sizes
2622
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2623
  failover to work on two-node clusters
2624
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2625
  unusable
2626
- Fixed detection of OS script errors in ``gnt-backup export``
2627
- Fixed exit code of ``gnt-backup export``
2628

    
2629

    
2630
Version 2.0.2
2631
-------------
2632

    
2633
*(Released Fri, 17 Jul 2009)*
2634

    
2635
- Added experimental support for stripped logical volumes; this should
2636
  enhance performance but comes with a higher complexity in the block
2637
  device handling; stripping is only enabled when passing
2638
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2639
  affected even in the non-stripped mode
2640
- Improved resiliency against transient failures at the end of DRBD
2641
  resyncs, and in general of DRBD resync checks
2642
- Fixed a couple of issues with exports and snapshot errors
2643
- Fixed a couple of issues in instance listing
2644
- Added display of the disk size in ``gnt-instance info``
2645
- Fixed checking for valid OSes in instance creation
2646
- Fixed handling of the "vcpus" parameter in instance listing and in
2647
  general of invalid parameters
2648
- Fixed http server library, and thus RAPI, to handle invalid
2649
  username/password combinations correctly; this means that now they
2650
  report unauthorized for queries too, not only for modifications,
2651
  allowing earlier detect of configuration problems
2652
- Added a new "role" node list field, equivalent to the master/master
2653
  candidate/drained/offline flags combinations
2654
- Fixed cluster modify and changes of candidate pool size
2655
- Fixed cluster verify error messages for wrong files on regular nodes
2656
- Fixed a couple of issues with node demotion from master candidate role
2657
- Fixed node readd issues
2658
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2659
- Added a new ``--no-voting`` option for masterfailover to fix failover
2660
  on two-nodes clusters when the former master node is unreachable
2661
- Added instance reinstall over RAPI
2662

    
2663

    
2664
Version 2.0.1
2665
-------------
2666

    
2667
*(Released Tue, 16 Jun 2009)*
2668

    
2669
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2670
  allow re-adding the start in single-user option (regression from 1.2)
2671
- the watcher writes the instance status to a file, to allow monitoring
2672
  to report the instance status (from the master) based on cached
2673
  results of the watcher's queries; while this can get stale if the
2674
  watcher is being locked due to other work on the cluster, this is
2675
  still an improvement
2676
- the watcher now also restarts the node daemon and the rapi daemon if
2677
  they died
2678
- fixed the watcher to handle full and drained queue cases
2679
- hooks export more instance data in the environment, which helps if
2680
  hook scripts need to take action based on the instance's properties
2681
  (no longer need to query back into ganeti)
2682
- instance failovers when the instance is stopped do not check for free
2683
  RAM, so that failing over a stopped instance is possible in low memory
2684
  situations
2685
- rapi uses queries for tags instead of jobs (for less job traffic), and
2686
  for cluster tags it won't talk to masterd at all but read them from
2687
  ssconf
2688
- a couple of error handling fixes in RAPI
2689
- drbd handling: improved the error handling of inconsistent disks after
2690
  resync to reduce the frequency of "there are some degraded disks for
2691
  this instance" messages
2692
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2693
  error handling path called a wrong function name)
2694

    
2695

    
2696
Version 2.0.0
2697
-------------
2698

    
2699
*(Released Wed, 27 May 2009)*
2700

    
2701
- no changes from rc5
2702

    
2703

    
2704
Version 2.0 rc5
2705
---------------
2706

    
2707
*(Released Wed, 20 May 2009)*
2708

    
2709
- fix a couple of bugs (validation, argument checks)
2710
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2711
- some small improvements to RAPI and IAllocator
2712
- make watcher automatically start the master daemon if down
2713

    
2714

    
2715
Version 2.0 rc4
2716
---------------
2717

    
2718
*(Released Mon, 27 Apr 2009)*
2719

    
2720
- change the OS list to not require locks; this helps with big clusters
2721
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2722
  volume group is broken
2723
- ``gnt-instance info``, without any arguments, doesn't run for all
2724
  instances anymore; either pass ``--all`` or pass the desired
2725
  instances; this helps against mistakes on big clusters where listing
2726
  the information for all instances takes a long time
2727
- miscellaneous doc and man pages fixes
2728

    
2729

    
2730
Version 2.0 rc3
2731
---------------
2732

    
2733
*(Released Wed, 8 Apr 2009)*
2734

    
2735
- Change the internal locking model of some ``gnt-node`` commands, in
2736
  order to reduce contention (and blocking of master daemon) when
2737
  batching many creation/reinstall jobs
2738
- Fixes to Xen soft reboot
2739
- No longer build documentation at build time, instead distribute it in
2740
  the archive, in order to reduce the need for the whole docbook/rst
2741
  toolchains
2742

    
2743

    
2744
Version 2.0 rc2
2745
---------------
2746

    
2747
*(Released Fri, 27 Mar 2009)*
2748

    
2749
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2750
- Fix watcher startup sequence, improves the behaviour of busy clusters
2751
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2752
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2753
  bind address and other places
2754
- Some documentation fixes and updates
2755

    
2756

    
2757
Version 2.0 rc1
2758
---------------
2759

    
2760
*(Released Mon, 2 Mar 2009)*
2761

    
2762
- More documentation updates, now all docs should be more-or-less
2763
  up-to-date
2764
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2765
  etc.)
2766
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2767
  and KVM)
2768
- Fix an issue related to $libdir/run/ganeti and cluster creation
2769

    
2770

    
2771
Version 2.0 beta2
2772
-----------------
2773

    
2774
*(Released Thu, 19 Feb 2009)*
2775

    
2776
- Xen PVM and KVM have switched the default value for the instance root
2777
  disk to the first partition on the first drive, instead of the whole
2778
  drive; this means that the OS installation scripts must be changed
2779
  accordingly
2780
- Man pages have been updated
2781
- RAPI has been switched by default to HTTPS, and the exported functions
2782
  should all work correctly
2783
- RAPI v1 has been removed
2784
- Many improvements to the KVM hypervisor
2785
- Block device errors are now better reported
2786
- Many other bugfixes and small improvements
2787

    
2788

    
2789
Version 2.0 beta1
2790
-----------------
2791

    
2792
*(Released Mon, 26 Jan 2009)*
2793

    
2794
- Version 2 is a general rewrite of the code and therefore the
2795
  differences are too many to list, see the design document for 2.0 in
2796
  the ``doc/`` subdirectory for more details
2797
- In this beta version there is not yet a migration path from 1.2 (there
2798
  will be one in the final 2.0 release)
2799
- A few significant changes are:
2800

    
2801
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2802
    various ``gnt-*`` commands are just front-ends to it
2803
  - all the commands are entered into, and executed from a job queue,
2804
    see the ``gnt-job(8)`` manpage
2805
  - the RAPI daemon supports read-write operations, secured by basic
2806
    HTTP authentication on top of HTTPS
2807
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2808
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2809
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2810
  - DRBD devices are using statically allocated minor numbers, which
2811
    will be assigned to existing instances during the migration process
2812
  - there is support for both Xen PVM and Xen HVM instances running on
2813
    the same cluster
2814
  - KVM virtualization is supported too
2815
  - file-based storage has been implemented, which means that it is
2816
    possible to run the cluster without LVM and DRBD storage, for
2817
    example using a shared filesystem exported from shared storage (and
2818
    still have live migration)
2819

    
2820

    
2821
Version 1.2.7
2822
-------------
2823

    
2824
*(Released Tue, 13 Jan 2009)*
2825

    
2826
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2827
- Reuse the old instance mac address by default on instance import, if
2828
  the instance name is the same.
2829
- Handle situations in which the node info rpc returns incomplete
2830
  results (issue 46)
2831
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2832
- Improved version of batcher:
2833

    
2834
  - state file support
2835
  - instance mac address support
2836
  - support for HVM clusters/instances
2837

    
2838
- Add an option to show the number of cpu sockets and nodes in
2839
  ``gnt-node list``
2840
- Support OSes that handle more than one version of the OS api (but do
2841
  not change the current API in any other way)
2842
- Fix ``gnt-node migrate``
2843
- ``gnt-debug`` man page
2844
- Fixes various more typos and small issues
2845
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2846

    
2847

    
2848
Version 1.2.6
2849
-------------
2850

    
2851
*(Released Wed, 24 Sep 2008)*
2852

    
2853
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2854
  type of disk exported to fully virtualized instances.
2855
- provide access to the serial console of HVM instances
2856
- instance auto_balance flag, set by default. If turned off it will
2857
  avoid warnings on cluster verify if there is not enough memory to fail
2858
  over an instance. in the future it will prevent automatically failing
2859
  it over when we will support that.
2860
- batcher tool for instance creation, see ``tools/README.batcher``
2861
- ``gnt-instance reinstall --select-os`` to interactively select a new
2862
  operating system when reinstalling an instance.
2863
- when changing the memory amount on instance modify a check has been
2864
  added that the instance will be able to start. also warnings are
2865
  emitted if the instance will not be able to fail over, if auto_balance
2866
  is true.
2867
- documentation fixes
2868
- sync fields between ``gnt-instance list/modify/add/import``
2869
- fix a race condition in drbd when the sync speed was set after giving
2870
  the device a remote peer.
2871

    
2872

    
2873
Version 1.2.5
2874
-------------
2875

    
2876
*(Released Tue, 22 Jul 2008)*
2877

    
2878
- note: the allowed size and number of tags per object were reduced
2879
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2880
- fixed twisted 8.x compatibility
2881
- fixed ``gnt-instance replace-disks`` with iallocator
2882
- add TCP keepalives on twisted connections to detect restarted nodes
2883
- disk increase support, see ``gnt-instance grow-disk``
2884
- implement bulk node/instance query for RAPI
2885
- add tags in node/instance listing (optional)
2886
- experimental migration (and live migration) support, read the man page
2887
  for ``gnt-instance migrate``
2888
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2889
  has some small improvements in handling its state file
2890

    
2891

    
2892
Version 1.2.4
2893
-------------
2894

    
2895
*(Released Fri, 13 Jun 2008)*
2896

    
2897
- Experimental readonly, REST-based remote API implementation;
2898
  automatically started on master node, TCP port 5080, if enabled by
2899
  ``--enable-rapi`` parameter to configure script.
2900
- Instance allocator support. Add and import instance accept a
2901
  ``--iallocator`` parameter, and call that instance allocator to decide
2902
  which node to use for the instance. The iallocator document describes
2903
  what's expected from an allocator script.
2904
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2905
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2906
  node is lost there is still enough memory to fail over the instances
2907
  that reside on it.
2908
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2909
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2910
  hooks will run, and their output, if any, will be displayed. Any
2911
  failing hook will make the verification return an error value.
2912
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2913
  primary and secondary interfaces
2914
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2915
  that is still declared as part of the cluster and has failed.
2916
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2917
  specifying output fields, that just adds the chosen fields to the
2918
  default ones.
2919
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2920
  export from the filesystem.
2921
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2922
  have been added. Using them you can enable/disable acpi and pae
2923
  support, and specify a path for a cd image to be exported to the
2924
  instance. These parameters as the name suggest only work on HVM
2925
  clusters.
2926
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2927
  PAE support will be set to the previously hardcoded values, but the
2928
  (previously hardcoded) path to the CDROM ISO image will be unset and
2929
  if required, needs to be set manually with ``gnt-instance modify``
2930
  after the upgrade.
2931
- The address to which an instance's VNC console is bound is now
2932
  selectable per-instance, rather than being cluster wide. Of course
2933
  this only applies to instances controlled via VNC, so currently just
2934
  applies to HVM clusters.
2935

    
2936

    
2937
Version 1.2.3
2938
-------------
2939

    
2940
*(Released Mon, 18 Feb 2008)*
2941

    
2942
- more tweaks to the disk activation code (especially helpful for DRBD)
2943
- change the default ``gnt-instance list`` output format, now there is
2944
  one combined status field (see the manpage for the exact values this
2945
  field will have)
2946
- some more fixes for the mac export to hooks change
2947
- make Ganeti not break with DRBD 8.2.x (which changed the version
2948
  format in ``/proc/drbd``) (issue 24)
2949
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2950
  template, allowing migration from DRBD0.7+MD to DRBD8
2951

    
2952

    
2953
Version 1.2.2
2954
-------------
2955

    
2956
*(Released Wed, 30 Jan 2008)*
2957

    
2958
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2959
  support (issue 23)
2960
- add command aliases infrastructure and a few aliases
2961
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2962
  man pages and the ``--help`` option of ``gnt-node
2963
  list``/``gnt-instance list``
2964
- fix ``gnt-backup list`` with down nodes (issue 21)
2965
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2966
- fix the dist archive and add a check for including svn/git files in
2967
  the future
2968
- some developer-related changes: improve the burnin and the QA suite,
2969
  add an upload script for testing during development
2970

    
2971

    
2972
Version 1.2.1
2973
-------------
2974

    
2975
*(Released Wed, 16 Jan 2008)*
2976

    
2977
- experimental HVM support, read the install document, section
2978
  "Initializing the cluster"
2979
- allow for the PVM hypervisor per-instance kernel and initrd paths
2980
- add a new command ``gnt-cluster verify-disks`` which uses a new
2981
  algorithm to improve the reconnection of the DRBD pairs if the device
2982
  on the secondary node has gone away
2983
- make logical volume code auto-activate LVs at disk activation time
2984
- slightly improve the speed of activating disks
2985
- allow specification of the MAC address at instance creation time, and
2986
  changing it later via ``gnt-instance modify``
2987
- fix handling of external commands that generate lots of output on
2988
  stderr
2989
- update documentation with regard to minimum version of DRBD8 supported
2990

    
2991

    
2992
Version 1.2.0
2993
-------------
2994

    
2995
*(Released Tue, 4 Dec 2007)*
2996

    
2997
- Log the ``xm create`` output to the node daemon log on failure (to
2998
  help diagnosing the error)
2999
- In debug mode, log all external commands output if failed to the logs
3000
- Change parsing of lvm commands to ignore stderr
3001

    
3002

    
3003
Version 1.2 beta3
3004
-----------------
3005

    
3006
*(Released Wed, 28 Nov 2007)*
3007

    
3008
- Another round of updates to the DRBD 8 code to deal with more failures
3009
  in the replace secondary node operation
3010
- Some more logging of failures in disk operations (lvm, drbd)
3011
- A few documentation updates
3012
- QA updates
3013

    
3014

    
3015
Version 1.2 beta2
3016
-----------------
3017

    
3018
*(Released Tue, 13 Nov 2007)*
3019

    
3020
- Change configuration file format from Python's Pickle to JSON.
3021
  Upgrading is possible using the cfgupgrade utility.
3022
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3023
  faster replace disks and is more stable (DRBD 8 has many improvements
3024
  compared to DRBD 0.7)
3025
- Added command line tags support (see man pages for ``gnt-instance``,
3026
  ``gnt-node``, ``gnt-cluster``)
3027
- Added instance rename support
3028
- Added multi-instance startup/shutdown
3029
- Added cluster rename support
3030
- Added ``gnt-node evacuate`` to simplify some node operations
3031
- Added instance reboot operation that can speedup reboot as compared to
3032
  stop and start
3033
- Soften the requirement that hostnames are in FQDN format
3034
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
3035
  reboots
3036
- Removed dependency on debian's patched fping that uses the
3037
  non-standard ``-S`` option
3038
- Now the OS definitions are searched for in multiple, configurable
3039
  paths (easier for distros to package)
3040
- Some changes to the hooks infrastructure (especially the new
3041
  post-configuration update hook)
3042
- Other small bugfixes
3043

    
3044
.. vim: set textwidth=72 syntax=rst :
3045
.. Local Variables:
3046
.. mode: rst
3047
.. fill-column: 72
3048
.. End: