Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 5b2dbea8

History | View | Annotate | Download (116.3 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

    
28
- KVM hypervisors can now access RBD storage directly without having to
29
  go through a block device.
30
- A new command 'gnt-cluster upgrade' was added that automates the upgrade
31
  procedure between two Ganeti versions that are both 2.10 or higher.
32
- The move-instance command can now change disk templates when moving
33
  instances, and does not require any node placement options to be
34
  specified if the destination cluster has a default iallocator.
35
- Users can now change the soundhw and cpuid settings for XEN hypervisors.
36
- Hail and hbal now have the (optional) capability of accessing average CPU
37
  load information through the monitoring deamon, and to use it to dynamically
38
  adapt the allocation of instances.
39

    
40
Misc changes
41
~~~~~~~~~~~~
42

    
43
- A new test framework for logical units was introduced and the test
44
  coverage for logical units was improved significantly.
45
- Opcodes are entirely generated from Haskell using the tool 'hs2py' and
46
  the module 'src/Ganeti/OpCodes.hs'.
47
- Constants are also generated from Haskell using the tool
48
  'hs2py-constants' and the module 'src/Ganeti/HsConstants.hs', with the
49
  exception of socket related constants, which require changing the
50
  cluster configuration file, and HVS related constants, because they
51
  are part of a port of instance queries to Haskell.  As a result, these
52
  changes will be part of the next release of Ganeti.
53

    
54
New dependencies
55
~~~~~~~~~~~~~~~~
56

    
57
The following new dependencies have been added/updated.
58

    
59
Python
60

    
61
- The version requirements for ``python-mock`` have increased to at least
62
  version 1.0.1. It is still used for testing only.
63

    
64

    
65
Version 2.9.0 rc3
66
-----------------
67

    
68
*(Released Tue, 15 Oct 2013)*
69

    
70
Incompatible/important changes
71
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
72

    
73
- hroller now also plans for capacity to move non-redundant instances off
74
  any node to be rebooted; the old behavior of completely ignoring any
75
  non-redundant instances can be restored by adding the --ignore-non-redundant
76
  option.
77
- The cluster option '--no-lvm-storage' was removed in favor of the new option
78
  '--enabled-disk-templates'.
79
- On instance creation, disk templates no longer need to be specified
80
  with '-t'. The default disk template will be taken from the list of
81
  enabled disk templates.
82
- The monitoring daemon is now running as root, in order to be able to collect
83
  information only available to root (such as the state of Xen instances).
84
- The ConfD client is now IPv6 compatible.
85
- File and shared file storage is no longer dis/enabled at configure time,
86
  but using the option '--enabled-disk-templates' at cluster initialization and
87
  modification.
88
- The default directories for file and shared file storage are not anymore
89
  specified at configure time, but taken from the cluster's configuration.
90
  They can be set at cluster initialization and modification with
91
  '--file-storage-dir' and '--shared-file-storage-dir'.
92
- Cluster verification now includes stricter checks regarding the
93
  default file and shared file storage directories. It now checks that
94
  the directories are explicitely allowed in the 'file-storage-paths' file and
95
  that the directories exist on all nodes.
96
- The list of allowed disk templates in the instance policy and the list
97
  of cluster-wide enabled disk templates is now checked for consistency
98
  on cluster or group modification. On cluster initialization, the ipolicy
99
  disk templates are ensured to be a subset of the cluster-wide enabled
100
  disk templates.
101

    
102
New features
103
~~~~~~~~~~~~
104

    
105
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
106
  the correct command syntax. It is possible to use different DRBD versions
107
  on different nodes as long as they are compatible to each other. This
108
  enables rolling upgrades of DRBD with no downtime. As permanent operation
109
  of different DRBD versions within a node group is discouraged,
110
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
111
- New "inst-status-xen" data collector for the monitoring daemon, providing
112
  information about the state of the xen instances on the nodes.
113
- New "lv" data collector for the monitoring daemon, collecting data about the
114
  logical volumes on the nodes, and pairing them with the name of the instances
115
  they belong to.
116
- New "diskstats" data collector, collecting the data from /proc/diskstats and
117
  presenting them over the monitoring daemon interface.
118
- The ConfD client is now IPv6 compatible.
119

    
120
New dependencies
121
~~~~~~~~~~~~~~~~
122
The following new dependencies have been added.
123

    
124
Python
125

    
126
- ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
127
  for the unit tests (and only used for testing).
128

    
129
Haskell
130

    
131
- ``hslogger`` (http://software.complete.org/hslogger) is now always
132
  required, even if confd is not enabled.
133

    
134
Since 2.9.0 rc2
135
~~~~~~~~~~~~~~~
136

    
137
- in implicit configuration upgrade, match ipolicy with enabled disk templates
138
- improved harep documentation (inherited from stable-2.8)
139

    
140

    
141
Version 2.9.0 rc2
142
-----------------
143

    
144
*(Released Wed, 9 Oct 2013)*
145

    
146
The second release candidate in the 2.9 series. Since 2.9.0 rc1:
147

    
148
- Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
149
  at least one DRBD instance.
150
- Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
151
  downgrading.
152

    
153

    
154
Version 2.9.0 rc1
155
-----------------
156

    
157
*(Released Tue, 1 Oct 2013)*
158

    
159
The first release candidate in the 2.9 series. Since 2.9.0 beta1:
160

    
161
- various bug fixes
162
- update of the documentation, in particular installation instructions
163
- merging of LD_* constants into DT_* constants
164
- python style changes to be compatible with newer versions of pylint
165

    
166

    
167
Version 2.9.0 beta1
168
-------------------
169

    
170
*(Released Thu, 29 Aug 2013)*
171

    
172
This was the first beta release of the 2.9 series. All important changes
173
are listed in the latest 2.9 entry.
174

    
175

    
176
Version 2.8.0
177
-------------
178

    
179
*(Released Mon, 30 Sep 2013)*
180

    
181
Incompatible/important changes
182
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
183

    
184
- Instance policy can contain multiple instance specs, as described in
185
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
186
  <design-partitioned>`. As a consequence, it's not possible to partially change
187
  or override instance specs. Bounding specs (min and max) can be specified as a
188
  whole using the new option ``--ipolicy-bounds-specs``, while standard
189
  specs use the new option ``--ipolicy-std-specs``.
190
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
191
  gnt-instance is a valid YAML object.
192
- hail now honors network restrictions when allocating nodes. This led to an
193
  update of the IAllocator protocol. See the IAllocator documentation for
194
  details.
195
- confd now only answers static configuration request over the network. luxid
196
  was extracted, listens on the local LUXI socket and responds to live queries.
197
  This allows finer grained permissions if using separate users.
198

    
199
New features
200
~~~~~~~~~~~~
201

    
202
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
203
  to always require authentication, ``--require-authentication``. It can
204
  be specified in ``$sysconfdir/default/ganeti``.
205
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
206
  be used to manage the disk templates to be used by instances in the cluster.
207
  Initially, it will be set to a list that includes plain, drbd, if they were
208
  enabled by specifying a volume group name, and file and sharedfile, if those
209
  were enabled at configure time. Additionally, it will include all disk
210
  templates that are currently used by instances. The order of disk templates
211
  will be based on Ganeti's history of supporting them. In the future, the
212
  first entry of the list will be used as a default disk template on instance
213
  creation.
214
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
215
  configuration back to the previous stable version.
216
- Disk templates in group ipolicy can be restored to the default value.
217
- Initial support for diskless instances and virtual clusters in QA.
218
- More QA and unit tests for instance policies.
219
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
220
  describing why the opcode itself was executed.
221
- The monitoring daemon is now available. It allows users to query the cluster
222
  for obtaining information about the status of the system. The daemon is only
223
  responsible for providing the information over the network: the actual data
224
  gathering is performed by data collectors (currently, only the DRBD status
225
  collector is available).
226
- In order to help developers work on Ganeti, a new script
227
  (``devel/build_chroot``) is provided, for building a chroot that contains all
228
  the required development libraries and tools for compiling Ganeti on a Debian
229
  Squeeze system.
230
- A new tool, ``harep``, for performing self-repair and recreation of instances
231
  in Ganeti has been added.
232
- Split queries are enabled for tags, network, exports, cluster info, groups,
233
  jobs, nodes.
234
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
235
  It prints the command line to set the current policies, to ease
236
  changing them.
237
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
238
  devices for KVM virtio-net interfaces will get created with VNET_HDR
239
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
240
  virtio-net interfaces, which prevents host kernel tainting and log
241
  flooding, when dealing with broken or malicious virtio-net drivers.
242
  It's set to true by default.
243
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
244
  failures.
245
- Support 'viridian' parameter in Xen HVM
246
- Support DSA SSH keys in bootstrap
247
- To simplify the work of packaging frameworks that want to add the needed users
248
  and groups in a split-user setup themselves, at build time three files in
249
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
250
  the groups to be generated, the ``users`` file contains, one per line, the
251
  users to be generated, optionally followed by their primary group, where
252
  important. The ``groupmemberships`` file contains, one per line, additional
253
  user-group membership relations that need to be established. The syntax of
254
  these files will remain stable in all future versions.
255

    
256

    
257
New dependencies
258
~~~~~~~~~~~~~~~~
259
The following new dependencies have been added:
260

    
261
For Haskell:
262
- The ``curl`` library is not optional anymore for compiling the Haskell code.
263
- ``snap-server`` library (if monitoring is enabled).
264

    
265
For Python:
266
- The minimum Python version needed to run Ganeti is now 2.6.
267
- ``yaml`` library (only for running the QA).
268

    
269
Since 2.8.0 rc3
270
~~~~~~~~~~~~~~~
271
- Perform proper cleanup on termination of Haskell daemons
272
- Fix corner-case in handling of remaining retry time
273

    
274

    
275
Version 2.8.0 rc3
276
-----------------
277

    
278
*(Released Tue, 17 Sep 2013)*
279

    
280
- To simplify the work of packaging frameworks that want to add the needed users
281
  and groups in a split-user setup themselves, at build time three files in
282
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
283
  the groups to be generated, the ``users`` file contains, one per line, the
284
  users to be generated, optionally followed by their primary group, where
285
  important. The ``groupmemberships`` file contains, one per line, additional
286
  user-group membership relations that need to be established. The syntax of
287
  these files will remain stable in all future versions.
288
- Add a default to file-driver when unspecified over RAPI (Issue 571)
289
- Mark the DSA host pubkey as optional, and remove it during config downgrade
290
  (Issue 560)
291
- Some documentation fixes
292

    
293

    
294
Version 2.8.0 rc2
295
-----------------
296

    
297
*(Released Tue, 27 Aug 2013)*
298

    
299
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
300

    
301
- Support 'viridian' parameter in Xen HVM (Issue 233)
302
- Include VCS version in ``gnt-cluster version``
303
- Support DSA SSH keys in bootstrap (Issue 338)
304
- Fix batch creation of instances
305
- Use FQDN to check master node status (Issue 551)
306
- Make the DRBD collector more failure-resilient
307

    
308

    
309
Version 2.8.0 rc1
310
-----------------
311

    
312
*(Released Fri, 2 Aug 2013)*
313

    
314
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
315

    
316
- Fix upgrading/downgrading from 2.7
317
- Increase maximum RAPI message size
318
- Documentation updates
319
- Split ``confd`` between ``luxid`` and ``confd``
320
- Merge 2.7 series up to the 2.7.1 release
321
- Allow the ``modify_etc_hosts`` option to be changed
322
- Add better debugging for ``luxid`` queries
323
- Expose bulk parameter for GetJobs in RAPI client
324
- Expose missing ``network`` fields in RAPI
325
- Add some ``cluster verify`` tests
326
- Some unittest fixes
327
- Fix a malfunction in ``hspace``'s tiered allocation
328
- Fix query compatibility between haskell and python implementations
329
- Add the ``vnet_hdr`` HV parameter for KVM
330
- Add ``--cleanup`` to instance failover
331
- Change the connected groups format in ``gnt-network info`` output; it
332
  was previously displayed as a raw list by mistake. (Merged from 2.7)
333

    
334

    
335
Version 2.8.0 beta1
336
-------------------
337

    
338
*(Released Mon, 24 Jun 2013)*
339

    
340
This was the first beta release of the 2.8 series. All important changes
341
are listed in the latest 2.8 entry.
342

    
343

    
344
Version 2.7.2
345
-------------
346

    
347
*(Released Thu, 26 Sep 2013)*
348

    
349
- Change the connected groups format in ``gnt-network info`` output; it
350
  was previously displayed as a raw list by mistake
351
- Check disk template in right dict when copying
352
- Support multi-instance allocs without iallocator
353
- Fix some errors in the documentation
354
- Fix formatting of tuple in an error message
355

    
356

    
357
Version 2.7.1
358
-------------
359

    
360
*(Released Thu, 25 Jul 2013)*
361

    
362
- Add logrotate functionality in daemon-util
363
- Add logrotate example file
364
- Add missing fields to network queries over rapi
365
- Fix network object timestamps
366
- Add support for querying network timestamps
367
- Fix a typo in the example crontab
368
- Fix a documentation typo
369

    
370

    
371
Version 2.7.0
372
-------------
373

    
374
*(Released Thu, 04 Jul 2013)*
375

    
376
Incompatible/important changes
377
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
378

    
379
- Instance policies for disk size were documented to be on a per-disk
380
  basis, but hail applied them to the sum of all disks. This has been
381
  fixed.
382
- ``hbal`` will now exit with status 0 if, during job execution over
383
  LUXI, early exit has been requested and all jobs are successful;
384
  before, exit status 1 was used, which cannot be differentiated from
385
  "job error" case
386
- Compatibility with newer versions of rbd has been fixed
387
- ``gnt-instance batch-create`` has been changed to use the bulk create
388
  opcode from Ganeti. This lead to incompatible changes in the format of
389
  the JSON file. It's now not a custom dict anymore but a dict
390
  compatible with the ``OpInstanceCreate`` opcode.
391
- Parent directories for file storage need to be listed in
392
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
393
  write the file automatically based on old configuration values, but it
394
  can not distribute it across all nodes and the file contents should be
395
  verified. Use ``gnt-cluster copyfile
396
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
397
  upgraded. The reason for requiring this list of paths now is that
398
  before it would have been possible to inject new paths via RPC,
399
  allowing files to be created in arbitrary locations. The RPC protocol
400
  is protected using SSL/X.509 certificates, but as a design principle
401
  Ganeti does not permit arbitrary paths to be passed.
402
- The parsing of the variants file for OSes (see
403
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
404
  lines and comment lines (starting with ``#``) are ignored for better
405
  readability.
406
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
407
  longer available. ``gnt-node add`` now invokes a new tool on the
408
  destination node, named ``prepare-node-join``, to configure the SSH
409
  daemon. Paramiko is no longer necessary to configure nodes' SSH
410
  daemons via ``gnt-node add``.
411
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
412
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
413
  the flag is not reset after a master failover.
414
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
415
  is recommended. 2.6 will be mandatory from the 2.8 series.
416

    
417

    
418
New features
419
~~~~~~~~~~~~
420

    
421
- New network management functionality to support automatic allocation
422
  of IP addresses and managing of network parameters. See
423
  :manpage:`gnt-network(8)` for more details.
424
- New external storage backend, to allow managing arbitrary storage
425
  systems external to the cluster. See
426
  :manpage:`ganeti-extstorage-interface(7)`.
427
- New ``exclusive-storage`` node parameter added, restricted to
428
  nodegroup level. When it's set to true, physical disks are assigned in
429
  an exclusive fashion to instances, as documented in :doc:`Partitioned
430
  Ganeti <design-partitioned>`.  Currently, only instances using the
431
  ``plain`` disk template are supported.
432
- The KVM hypervisor has been updated with many new hypervisor
433
  parameters, including a generic one for passing arbitrary command line
434
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
435
  compatible up to qemu 1.4.
436
- A new tool, called ``mon-collector``, is the stand-alone executor of
437
  the data collectors for a monitoring system. As of this version, it
438
  just includes the DRBD data collector, that can be executed by calling
439
  ``mon-collector`` using the ``drbd`` parameter. See
440
  :manpage:`mon-collector(7)`.
441
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
442
  for RAPI users. It allows granting permissions to query for
443
  information to a specific user without giving
444
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
445
- A new tool named ``node-cleanup`` has been added. It cleans remains of
446
  a cluster from a machine by stopping all daemons, removing
447
  certificates and ssconf files. Unless the ``--no-backup`` option is
448
  given, copies of the certificates are made.
449
- Instance creations now support the use of opportunistic locking,
450
  potentially speeding up the (parallel) creation of multiple instances.
451
  This feature is currently only available via the :doc:`RAPI
452
  <rapi>` interface and when an instance allocator is used. If the
453
  ``opportunistic_locking`` parameter is set the opcode will try to
454
  acquire as many locks as possible, but will not wait for any locks
455
  held by other opcodes. If not enough resources can be found to
456
  allocate the instance, the temporary error code
457
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
458
  retried thereafter, with or without opportunistic locking.
459
- New experimental linux-ha resource scripts.
460
- Restricted-commands support: ganeti can now be asked (via command line
461
  or rapi) to perform commands on a node. These are passed via ganeti
462
  RPC rather than ssh. This functionality is restricted to commands
463
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
464
  reasons. The file is not copied automatically.
465

    
466

    
467
Misc changes
468
~~~~~~~~~~~~
469

    
470
- Diskless instances are now externally mirrored (Issue 237). This for
471
  now has only been tested in conjunction with explicit target nodes for
472
  migration/failover.
473
- Queries not needing locks or RPC access to the node can now be
474
  performed by the confd daemon, making them independent from jobs, and
475
  thus faster to execute. This is selectable at configure time.
476
- The functionality for allocating multiple instances at once has been
477
  overhauled and is now also available through :doc:`RAPI <rapi>`.
478

    
479
There are no significant changes from version 2.7.0~rc3.
480

    
481

    
482
Version 2.7.0 rc3
483
-----------------
484

    
485
*(Released Tue, 25 Jun 2013)*
486

    
487
- Fix permissions on the confd query socket (Issue 477)
488
- Fix permissions on the job archive dir (Issue 498)
489
- Fix handling of an internal exception in replace-disks (Issue 472)
490
- Fix gnt-node info handling of shortened names (Issue 497)
491
- Fix gnt-instance grow-disk when wiping is enabled
492
- Documentation improvements, and support for newer pandoc
493
- Fix hspace honoring ipolicy for disks (Issue 484)
494
- Improve handling of the ``kvm_extra`` HV parameter
495

    
496

    
497
Version 2.7.0 rc2
498
-----------------
499

    
500
*(Released Fri, 24 May 2013)*
501

    
502
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
503
  symlink.
504
- Disks added through ``gnt-instance modify`` or created through
505
  ``gnt-instance recreate-disks`` are wiped, if the
506
  ``prealloc_wipe_disks`` flag is set.
507
- If wiping newly created disks fails, the disks are removed. Also,
508
  partial failures in creating disks through ``gnt-instance modify``
509
  triggers a cleanup of the partially-created disks.
510
- Removing the master IP address doesn't fail if the address has been
511
  already removed.
512
- Fix ownership of the OS log dir
513
- Workaround missing SO_PEERCRED constant (Issue 191)
514

    
515

    
516
Version 2.7.0 rc1
517
-----------------
518

    
519
*(Released Fri, 3 May 2013)*
520

    
521
This was the first release candidate of the 2.7 series. Since beta3:
522

    
523
- Fix kvm compatibility with qemu 1.4 (Issue 389)
524
- Documentation updates (admin guide, upgrade notes, install
525
  instructions) (Issue 372)
526
- Fix gnt-group list nodes and instances count (Issue 436)
527
- Fix compilation without non-mandatory libraries (Issue 441)
528
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
529
- Make confd logging more verbose at INFO level (Issue 435)
530
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
531
- Fix failure path for instance storage type conversion (Issue 229)
532
- Update htools text backend documentation
533
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
534
- Disable inter-cluster instance move for file-based instances, because
535
  it is dependant on instance export, which is not supported for
536
  file-based instances. (Issue 414)
537
- Fix gnt-job crashes on non-ascii characters (Issue 427)
538
- Fix volume group checks on non-vm-capable nodes (Issue 432)
539

    
540

    
541
Version 2.7.0 beta3
542
-------------------
543

    
544
*(Released Mon, 22 Apr 2013)*
545

    
546
This was the third beta release of the 2.7 series. Since beta2:
547

    
548
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
549
- Fix data loss on wrong usage of ``gnt-instance move``
550
- Properly export errors in confd-based job queries
551
- Add ``users-setup`` tool
552
- Fix iallocator protocol to report 0 as a disk size for diskless
553
  instances. This avoids hail breaking when a diskless instance is
554
  present.
555
- Fix job queue directory permission problem that made confd job queries
556
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
557
  for access to archived jobs (Issue 406).
558
- Limit the sizes of networks supported by ``gnt-network`` to something
559
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
560
- Fix bugs in instance disk template conversion
561
- Fix GHC 7 compatibility
562
- Fix ``burnin`` install path (Issue 426).
563
- Allow very small disk grows (Issue 347).
564
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
565
  that noded doesn't import masterd code (Issue 419).
566
- Make sure the default metavg at cluster init is the same as the vg, if
567
  unspecified (Issue 358).
568
- Fix cleanup of partially created disks (part of Issue 416)
569

    
570

    
571
Version 2.7.0 beta2
572
-------------------
573

    
574
*(Released Tue, 2 Apr 2013)*
575

    
576
This was the second beta release of the 2.7 series. Since beta1:
577

    
578
- Networks no longer have a "type" slot, since this information was
579
  unused in Ganeti: instead of it tags should be used.
580
- The rapi client now has a ``target_node`` option to MigrateInstance.
581
- Fix early exit return code for hbal (Issue 386).
582
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
583
- Fix ``rbd showmapped`` output parsing (Issue 312).
584
- Networks are now referenced indexed by UUID, rather than name. This
585
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
586
  use.
587
- The OS environment now includes network information.
588
- Deleting of a network is now disallowed if any instance nic is using
589
  it, to prevent dangling references.
590
- External storage is now documented in man pages.
591
- The exclusive_storage flag can now only be set at nodegroup level.
592
- Hbal can now submit an explicit priority with its jobs.
593
- Many network related locking fixes.
594
- Bump up the required pylint version to 0.25.1.
595
- Fix the ``no_remember`` option in RAPI client.
596
- Many ipolicy related tests, qa, and fixes.
597
- Many documentation improvements and fixes.
598
- Fix building with ``--disable-file-storage``.
599
- Fix ``-q`` option in htools, which was broken if passed more than
600
  once.
601
- Some haskell/python interaction improvements and fixes.
602
- Fix iallocator in case of missing LVM storage.
603
- Fix confd config load in case of ``--no-lvm-storage``.
604
- The confd/query functionality is now mentioned in the security
605
  documentation.
606

    
607

    
608
Version 2.7.0 beta1
609
-------------------
610

    
611
*(Released Wed, 6 Feb 2013)*
612

    
613
This was the first beta release of the 2.7 series. All important changes
614
are listed in the latest 2.7 entry.
615

    
616

    
617
Version 2.6.2
618
-------------
619

    
620
*(Released Fri, 21 Dec 2012)*
621

    
622
Important behaviour change: hbal won't rebalance anymore instances which
623
have the ``auto_balance`` attribute set to false. This was the intention
624
all along, but until now it only skipped those from the N+1 memory
625
reservation (DRBD-specific).
626

    
627
A significant number of bug fixes in this release:
628

    
629
- Fixed disk adoption interaction with ipolicy checks.
630
- Fixed networking issues when instances are started, stopped or
631
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
632
- Fixed the warning in cluster verify for shared storage instances not
633
  being redundant.
634
- Fixed removal of storage directory on shared file storage (issue 262).
635
- Fixed validation of LVM volume group name in OpClusterSetParams
636
  (``gnt-cluster modify``) (issue 285).
637
- Fixed runtime memory increases (``gnt-instance modify -m``).
638
- Fixed live migration under Xen's ``xl`` mode.
639
- Fixed ``gnt-instance console`` with ``xl``.
640
- Fixed building with newer Haskell compiler/libraries.
641
- Fixed PID file writing in Haskell daemons (confd); this prevents
642
  restart issues if confd was launched manually (outside of
643
  ``daemon-util``) while another copy of it was running
644
- Fixed a type error when doing live migrations with KVM (issue 297) and
645
  the error messages for failing migrations have been improved.
646
- Fixed opcode validation for the out-of-band commands (``gnt-node
647
  power``).
648
- Fixed a type error when unsetting OS hypervisor parameters (issue
649
  311); now it's possible to unset all OS-specific hypervisor
650
  parameters.
651
- Fixed the ``dry-run`` mode for many operations: verification of
652
  results was over-zealous but didn't take into account the ``dry-run``
653
  operation, resulting in "wrong" failures.
654
- Fixed bash completion in ``gnt-job list`` when the job queue has
655
  hundreds of entries; especially with older ``bash`` versions, this
656
  results in significant CPU usage.
657

    
658
And lastly, a few other improvements have been made:
659

    
660
- Added option to force master-failover without voting (issue 282).
661
- Clarified error message on lock conflict (issue 287).
662
- Logging of newly submitted jobs has been improved (issue 290).
663
- Hostname checks have been made uniform between instance rename and
664
  create (issue 291).
665
- The ``--submit`` option is now supported by ``gnt-debug delay``.
666
- Shutting down the master daemon by sending SIGTERM now stops it from
667
  processing jobs waiting for locks; instead, those jobs will be started
668
  once again after the master daemon is started the next time (issue
669
  296).
670
- Support for Xen's ``xl`` program has been improved (besides the fixes
671
  above).
672
- Reduced logging noise in the Haskell confd daemon (only show one log
673
  entry for each config reload, instead of two).
674
- Several man page updates and typo fixes.
675

    
676

    
677
Version 2.6.1
678
-------------
679

    
680
*(Released Fri, 12 Oct 2012)*
681

    
682
A small bugfix release. Among the bugs fixed:
683

    
684
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
685
  made the command unusable.
686
- Commands that issue many jobs don't fail anymore just because some jobs
687
  take so long that other jobs are archived.
688
- Failures during ``gnt-instance reinstall`` are reflected by the exit
689
  status.
690
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
691
  DRBD is enabled.
692
- When ``always_failover`` is set, ``--allow-failover`` is not required
693
  in migrate commands anymore.
694
- ``bash_completion`` works even if extglob is disabled.
695
- Fixed bug with locks that made failover for RDB-based instances fail.
696
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
697
  a random node instead of one based on the allocator metric.
698
- Support for newer versions of pylint and pep8.
699
- Hail doesn't fail anymore when trying to add an instance of type
700
  ``file``, ``sharedfile`` or ``rbd``.
701
- Added new Makefile target to rebuild the whole distribution, so that
702
  all files are included.
703

    
704

    
705
Version 2.6.0
706
-------------
707

    
708
*(Released Fri, 27 Jul 2012)*
709

    
710

    
711
.. attention:: The ``LUXI`` protocol has been made more consistent
712
   regarding its handling of command arguments. This, however, leads to
713
   incompatibility issues with previous versions. Please ensure that you
714
   restart Ganeti daemons soon after the upgrade, otherwise most
715
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
716
   pausing/resuming the watcher, cancelling and archiving jobs, querying
717
   the cluster configuration) will fail.
718

    
719

    
720
New features
721
~~~~~~~~~~~~
722

    
723
Instance run status
724
+++++++++++++++++++
725

    
726
The current ``admin_up`` field, which used to denote whether an instance
727
should be running or not, has been removed. Instead, ``admin_state`` is
728
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
729

    
730
The rational behind this is that an instance being “down” can have
731
different meanings:
732

    
733
- it could be down during a reboot
734
- it could be temporarily be down for a reinstall
735
- or it could be down because it is deprecated and kept just for its
736
  disk
737

    
738
The previous Boolean state was making it difficult to do capacity
739
calculations: should Ganeti reserve memory for a down instance? Now, the
740
tri-state field makes it clear:
741

    
742
- in ``up`` and ``down`` state, all resources are reserved for the
743
  instance, and it can be at any time brought up if it is down
744
- in ``offline`` state, only disk space is reserved for it, but not
745
  memory or CPUs
746

    
747
The field can have an extra use: since the transition between ``up`` and
748
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
749
transition between ``offline`` and ``down`` is done via ``gnt-instance
750
modify``, it is possible to given different rights to users. For
751
example, owners of an instance could be allowed to start/stop it, but
752
not transition it out of the offline state.
753

    
754
Instance policies and specs
755
+++++++++++++++++++++++++++
756

    
757
In previous Ganeti versions, an instance creation request was not
758
limited on the minimum size and on the maximum size just by the cluster
759
resources. As such, any policy could be implemented only in third-party
760
clients (RAPI clients, or shell wrappers over ``gnt-*``
761
tools). Furthermore, calculating cluster capacity via ``hspace`` again
762
required external input with regards to instance sizes.
763

    
764
In order to improve these workflows and to allow for example better
765
per-node group differentiation, we introduced instance specs, which
766
allow declaring:
767

    
768
- minimum instance disk size, disk count, memory size, cpu count
769
- maximum values for the above metrics
770
- and “standard” values (used in ``hspace`` to calculate the standard
771
  sized instances)
772

    
773
The minimum/maximum values can be also customised at node-group level,
774
for example allowing more powerful hardware to support bigger instance
775
memory sizes.
776

    
777
Beside the instance specs, there are a few other settings belonging to
778
the instance policy framework. It is possible now to customise, per
779
cluster and node-group:
780

    
781
- the list of allowed disk templates
782
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
783
- the maximum ratio of instance to spindles (see below for more
784
  information) for local storage
785

    
786
All these together should allow all tools that talk to Ganeti to know
787
what are the ranges of allowed values for instances and the
788
over-subscription that is allowed.
789

    
790
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
791
instance configuration, and the physical CPU configuration from the
792
node. For the spindle ratios however, we didn't track before these
793
values, so new parameters have been added:
794

    
795
- a new node parameter ``spindle_count``, defaults to 1, customisable at
796
  node group or node level
797
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
798

    
799
Note that spindles in this context doesn't need to mean actual
800
mechanical hard-drives; it's just a relative number for both the node
801
I/O capacity and instance I/O consumption.
802

    
803
Instance migration behaviour
804
++++++++++++++++++++++++++++
805

    
806
While live-migration is in general desirable over failover, it is
807
possible that for some workloads it is actually worse, due to the
808
variable time of the “suspend” phase during live migration.
809

    
810
To allow the tools to work consistently over such instances (without
811
having to hard-code instance names), a new backend parameter
812
``always_failover`` has been added to control the migration/failover
813
behaviour. When set to True, all migration requests for an instance will
814
instead fall-back to failover.
815

    
816
Instance memory ballooning
817
++++++++++++++++++++++++++
818

    
819
Initial support for memory ballooning has been added. The memory for an
820
instance is no longer fixed (backend parameter ``memory``), but instead
821
can vary between minimum and maximum values (backend parameters
822
``minmem`` and ``maxmem``). Currently we only change an instance's
823
memory when:
824

    
825
- live migrating or failing over and instance and the target node
826
  doesn't have enough memory
827
- user requests changing the memory via ``gnt-instance modify
828
  --runtime-memory``
829

    
830
Instance CPU pinning
831
++++++++++++++++++++
832

    
833
In order to control the use of specific CPUs by instance, support for
834
controlling CPU pinning has been added for the Xen, HVM and LXC
835
hypervisors. This is controlled by a new hypervisor parameter
836
``cpu_mask``; details about possible values for this are in the
837
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
838
VCPU-to-CPU mapping) form will work well only when all nodes in your
839
cluster have the same amount of CPUs.
840

    
841
Disk parameters
842
+++++++++++++++
843

    
844
Another area in which Ganeti was not customisable were the parameters
845
used for storage configuration, e.g. how many stripes to use for LVM,
846
DRBD resync configuration, etc.
847

    
848
To improve this area, we've added disks parameters, which are
849
customisable at cluster and node group level, and which allow to
850
specify various parameters for disks (DRBD has the most parameters
851
currently), for example:
852

    
853
- DRBD resync algorithm and parameters (e.g. speed)
854
- the default VG for meta-data volumes for DRBD
855
- number of stripes for LVM (plain disk template)
856
- the RBD pool
857

    
858
These parameters can be modified via ``gnt-cluster modify -D …`` and
859
``gnt-group modify -D …``, and are used at either instance creation (in
860
case of LVM stripes, for example) or at disk “activation” time
861
(e.g. resync speed).
862

    
863
Rados block device support
864
++++++++++++++++++++++++++
865

    
866
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
867
denoted by the ``rbd`` disk template type. This is considered
868
experimental, feedback is welcome. For details on configuring it, see
869
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
870

    
871
Master IP setup
872
+++++++++++++++
873

    
874
The existing master IP functionality works well only in simple setups (a
875
single network shared by all nodes); however, if nodes belong to
876
different networks, then the ``/32`` setup and lack of routing
877
information is not enough.
878

    
879
To allow the master IP to function well in more complex cases, the
880
system was reworked as follows:
881

    
882
- a master IP netmask setting has been added
883
- the master IP activation/turn-down code was moved from the node daemon
884
  to a separate script
885
- whether to run the Ganeti-supplied master IP script or a user-supplied
886
  on is a ``gnt-cluster init`` setting
887

    
888
Details about the location of the standard and custom setup scripts are
889
in the man page :manpage:`gnt-cluster(8)`; for information about the
890
setup script protocol, look at the Ganeti-supplied script.
891

    
892
SPICE support
893
+++++++++++++
894

    
895
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
896
improved.
897

    
898
It is now possible to use TLS-protected connections, and when renewing
899
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
900
it is now possible to specify spice or spice CA certificates. Also, it
901
is possible to configure a password for SPICE sessions via the
902
hypervisor parameter ``spice_password_file``.
903

    
904
There are also new parameters to control the compression and streaming
905
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
906
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
907
for the spice parameters.
908

    
909
Lastly, it is now possible to see the SPICE connection information via
910
``gnt-instance console``.
911

    
912
OVF converter
913
+++++++++++++
914

    
915
A new tool (``tools/ovfconverter``) has been added that supports
916
conversion between Ganeti and the `Open Virtualization Format
917
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
918
from).
919

    
920
This relies on the ``qemu-img`` tool to convert the disk formats, so the
921
actual compatibility with other virtualization solutions depends on it.
922

    
923
Confd daemon changes
924
++++++++++++++++++++
925

    
926
The configuration query daemon (``ganeti-confd``) is now optional, and
927
has been rewritten in Haskell; whether to use the daemon at all, use the
928
Python (default) or the Haskell version is selectable at configure time
929
via the ``--enable-confd`` parameter, which can take one of the
930
``haskell``, ``python`` or ``no`` values. If not used, disabling the
931
daemon will result in a smaller footprint; for larger systems, we
932
welcome feedback on the Haskell version which might become the default
933
in future versions.
934

    
935
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
936
daemon running. The Python version doesn't implement the new call.
937

    
938

    
939
User interface changes
940
~~~~~~~~~~~~~~~~~~~~~~
941

    
942
We have replaced the ``--disks`` option of ``gnt-instance
943
replace-disks`` with a more flexible ``--disk`` option, which allows
944
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
945
disk size and mode can be changed upon recreation (via ``gnt-instance
946
recreate-disks``, which accepts the same ``--disk`` option).
947

    
948
As many people are used to a ``show`` command, we have added that as an
949
alias to ``info`` on all ``gnt-*`` commands.
950

    
951
The ``gnt-instance grow-disk`` command has a new mode in which it can
952
accept the target size of the disk, instead of the delta; this can be
953
more safe since two runs in absolute mode will be idempotent, and
954
sometimes it's also easier to specify the desired size directly.
955

    
956
Also the handling of instances with regard to offline secondaries has
957
been improved. Instance operations should not fail because one of it's
958
secondary nodes is offline, even though it's safe to proceed.
959

    
960
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
961
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
962
minors to instance name.
963

    
964
API changes
965
~~~~~~~~~~~
966

    
967
RAPI coverage has improved, with (for example) new resources for
968
recreate-disks, node power-cycle, etc.
969

    
970
Compatibility
971
~~~~~~~~~~~~~
972

    
973
There is partial support for ``xl`` in the Xen hypervisor; feedback is
974
welcome.
975

    
976
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
977
whether to still support Python 2.4 or move to Python 2.6 as minimum
978
required version.
979

    
980
Support for Fedora has been slightly improved; the provided example
981
init.d script should work better on it and the INSTALL file should
982
document the needed dependencies.
983

    
984
Internal changes
985
~~~~~~~~~~~~~~~~
986

    
987
The deprecated ``QueryLocks`` LUXI request has been removed. Use
988
``Query(what=QR_LOCK, ...)`` instead.
989

    
990
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
991
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
992
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
993
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
994
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
995

    
996
RAPI client: ``CertificateError`` now derives from
997
``GanetiApiError``. This should make it more easy to handle Ganeti
998
errors.
999

    
1000
Deprecation warnings due to PyCrypto/paramiko import in
1001
``tools/setup-ssh`` have been silenced, as usually they are safe; please
1002
make sure to run an up-to-date paramiko version, if you use this tool.
1003

    
1004
The QA scripts now depend on Python 2.5 or above (the main code base
1005
still works with Python 2.4).
1006

    
1007
The configuration file (``config.data``) is now written without
1008
indentation for performance reasons; if you want to edit it, it can be
1009
re-formatted via ``tools/fmtjson``.
1010

    
1011
A number of bugs has been fixed in the cluster merge tool.
1012

    
1013
``x509`` certification verification (used in import-export) has been
1014
changed to allow the same clock skew as permitted by the cluster
1015
verification. This will remove some rare but hard to diagnose errors in
1016
import-export.
1017

    
1018

    
1019
Version 2.6.0 rc4
1020
-----------------
1021

    
1022
*(Released Thu, 19 Jul 2012)*
1023

    
1024
Very few changes from rc4 to the final release, only bugfixes:
1025

    
1026
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
1027
  instance, fix CDROM booting for KVM instances)
1028
- fixed node group modification of node parameters
1029
- fixed issue in LUClusterVerifyGroup with multi-group clusters
1030
- fixed generation of bash completion to ensure a stable ordering
1031
- fixed a few typos
1032

    
1033

    
1034
Version 2.6.0 rc3
1035
-----------------
1036

    
1037
*(Released Fri, 13 Jul 2012)*
1038

    
1039
Third release candidate for 2.6. The following changes were done from
1040
rc3 to rc4:
1041

    
1042
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1043
- Fixed an inconsistency in the LUXI protocol with the provided
1044
  arguments (NOT backwards compatible)
1045
- Fixed a bug with node groups ipolicy where ``min`` was greater than
1046
  the cluster ``std`` value
1047
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1048
  easier instance debugging on nodes (requires ``hconfd`` to work)
1049

    
1050

    
1051
Version 2.6.0 rc2
1052
-----------------
1053

    
1054
*(Released Tue, 03 Jul 2012)*
1055

    
1056
Second release candidate for 2.6. The following changes were done from
1057
rc2 to rc3:
1058

    
1059
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1060
  master candidates
1061
- Fixed a RAPI regression on missing beparams/memory
1062
- Fixed redistribution of files on offline nodes
1063
- Added possibility to run activate-disks even though secondaries are
1064
  offline. With this change it relaxes also the strictness on some other
1065
  commands which use activate disks internally:
1066
  * ``gnt-instance start|reboot|rename|backup|export``
1067
- Made it possible to remove safely an instance if its secondaries are
1068
  offline
1069
- Made it possible to reinstall even though secondaries are offline
1070

    
1071

    
1072
Version 2.6.0 rc1
1073
-----------------
1074

    
1075
*(Released Mon, 25 Jun 2012)*
1076

    
1077
First release candidate for 2.6. The following changes were done from
1078
rc1 to rc2:
1079

    
1080
- Fixed bugs with disk parameters and ``rbd`` templates as well as
1081
  ``instance_os_add``
1082
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1083
  behaviour. It supports now the modify operation
1084
- ``hcheck`` implemented to analyze cluster health and possibility of
1085
  improving health by rebalance
1086
- ``hbal`` has been improved in dealing with split instances
1087

    
1088

    
1089
Version 2.6.0 beta2
1090
-------------------
1091

    
1092
*(Released Mon, 11 Jun 2012)*
1093

    
1094
Second beta release of 2.6. The following changes were done from beta2
1095
to rc1:
1096

    
1097
- Fixed ``daemon-util`` with non-root user models
1098
- Fixed creation of plain instances with ``--no-wait-for-sync``
1099
- Fix wrong iv_names when running ``cfgupgrade``
1100
- Export more information in RAPI group queries
1101
- Fixed bug when changing instance network interfaces
1102
- Extended burnin to do NIC changes
1103
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1104
- Changed default for DRBD barriers
1105
- Fixed DRBD error reporting for syncer rate
1106
- Verify the options on disk parameters
1107

    
1108
And of course various fixes to documentation and improved unittests and
1109
QA.
1110

    
1111

    
1112
Version 2.6.0 beta1
1113
-------------------
1114

    
1115
*(Released Wed, 23 May 2012)*
1116

    
1117
First beta release of 2.6. The following changes were done from beta1 to
1118
beta2:
1119

    
1120
- integrated patch for distributions without ``start-stop-daemon``
1121
- adapted example init.d script to work on Fedora
1122
- fixed log handling in Haskell daemons
1123
- adapted checks in the watcher for pycurl linked against libnss
1124
- add partial support for ``xl`` instead of ``xm`` for Xen
1125
- fixed a type issue in cluster verification
1126
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1127
  clusters)
1128

    
1129
Plus integrated fixes from the 2.5 branch:
1130

    
1131
- fixed ``kvm-ifup`` to use ``/bin/bash``
1132
- fixed parallel build failures
1133
- KVM live migration when using a custom keymap
1134

    
1135

    
1136
Version 2.5.2
1137
-------------
1138

    
1139
*(Released Tue, 24 Jul 2012)*
1140

    
1141
A small bugfix release, with no new features:
1142

    
1143
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1144
  different default shell (e.g. Debian, Ubuntu)
1145
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1146
  243 and Debian bug #650664)
1147
- fixed compatibility with KVM versions that don't support multiple boot
1148
  devices (fixes Issue 230 and Debian bug #624256)
1149

    
1150
Additionally, a few fixes were done to the build system (fixed parallel
1151
build failures) and to the unittests (fixed race condition in test for
1152
FileID functions, and the default enable/disable mode for QA test is now
1153
customisable).
1154

    
1155

    
1156
Version 2.5.1
1157
-------------
1158

    
1159
*(Released Fri, 11 May 2012)*
1160

    
1161
A small bugfix release.
1162

    
1163
The main issues solved are on the topic of compatibility with newer LVM
1164
releases:
1165

    
1166
- fixed parsing of ``lv_attr`` field
1167
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1168
  the ``--force`` flag is needed
1169

    
1170
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1171
to accept kernel 3.x too (was hardcoded to 2.6.*).
1172

    
1173
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1174
of hook results and that also made display of other errors suboptimal
1175
was fixed; the code behaves now like 2.4 and earlier.
1176

    
1177
Another change in 2.5, the cleanup of the OS scripts environment, is too
1178
aggressive: it removed even the ``PATH`` variable, which requires the OS
1179
scripts to *always* need to export it. Since this is a bit too strict,
1180
we now export a minimal PATH, the same that we export for hooks.
1181

    
1182
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1183
integrated into this release.
1184

    
1185
Finally, a few other miscellaneous changes were done (no new features,
1186
just small improvements):
1187

    
1188
- Fix ``gnt-group --help`` display
1189
- Fix hardcoded Xen kernel path
1190
- Fix grow-disk handling of invalid units
1191
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1192
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1193

    
1194

    
1195
Version 2.5.0
1196
-------------
1197

    
1198
*(Released Thu, 12 Apr 2012)*
1199

    
1200
Incompatible/important changes and bugfixes
1201
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1202

    
1203
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1204
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1205
  to match the underlying LUXI interface.
1206
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1207
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1208
  not maintain backwards-compatibility as the underlying operation
1209
  changed in an incompatible way. The RAPI client can talk to old
1210
  servers, but it needs to be told so as the return value changed.
1211
- When creating file-based instances via RAPI, the ``file_driver``
1212
  parameter no longer defaults to ``loop`` and must be specified.
1213
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1214
  ``link`` instead.
1215
- Support for the undocumented and deprecated RAPI instance creation
1216
  request format version 0 has been dropped. Use version 1, supported
1217
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1218
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1219
  documentation <install>`.
1220
- The "cluster-verify" hooks are now executed per group by the
1221
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1222
  if you just run ``gnt-cluster verify``, which generates one opcode per
1223
  group.
1224
- The environment as passed to the OS scripts is cleared, and thus no
1225
  environment variables defined in the node daemon's environment will be
1226
  inherited by the scripts.
1227
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1228
  deprecated.
1229
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1230
  support operations involving multiple node groups.
1231
- Offline nodes are ignored when failing over an instance.
1232
- Support for KVM version 1.0, which changed the version reporting format
1233
  from 3 to 2 digits.
1234
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1235
  removal.
1236
- ``Makefile`` is now compatible with Automake 1.11.2
1237
- Includes all bugfixes made in the 2.4 series
1238

    
1239
New features
1240
~~~~~~~~~~~~
1241

    
1242
- The ganeti-htools project has been merged into the ganeti-core source
1243
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1244
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1245
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1246
  GPT-style partition tables (requires `parted
1247
  <http://www.gnu.org/s/parted/>`_).
1248
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1249
- Allowed adding tags on instance creation.
1250
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1251
- Allow instances to be started in a paused state, enabling the user to
1252
  see the complete console output on boot using the console.
1253
- Added new hypervisor flag to control default reboot behaviour
1254
  (``reboot_behavior``).
1255
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1256
- Improved out-of-band management support:
1257

    
1258
  - Added ``gnt-node health`` command reporting the health status of
1259
    nodes.
1260
  - Added ``gnt-node power`` command to manage power status of nodes.
1261
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1262

    
1263
- Instance migration can fall back to failover if instance is not
1264
  running.
1265
- Filters can be used when listing nodes, instances, groups and locks;
1266
  see :manpage:`ganeti(7)` manpage.
1267
- Added post-execution status as variables to :doc:`hooks <hooks>`
1268
  environment.
1269
- Instance tags are exported/imported together with the instance.
1270
- When given an explicit job ID, ``gnt-job info`` will work for archived
1271
  jobs.
1272
- Jobs can define dependencies on other jobs (not yet supported via
1273
  RAPI or command line, but used by internal commands and usable via
1274
  LUXI).
1275

    
1276
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1277
    dependencies.
1278

    
1279
- Instance failover is now available as a RAPI resource
1280
  (``/2/instances/[instance_name]/failover``).
1281
- ``gnt-instance info`` defaults to static information if primary node
1282
  is offline.
1283
- Opcodes have a new ``comment`` attribute.
1284
- Added basic SPICE support to KVM hypervisor.
1285
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1286

    
1287
Node group improvements
1288
~~~~~~~~~~~~~~~~~~~~~~~
1289

    
1290
- ``gnt-cluster verify`` has been modified to check groups separately,
1291
  thereby improving performance.
1292
- Node group support has been added to ``gnt-cluster verify-disks``,
1293
  which now operates per node group.
1294
- Watcher has been changed to work better with node groups.
1295

    
1296
  - One process and state file per node group.
1297
  - Slow watcher in one group doesn't block other group's watcher.
1298

    
1299
- Added new command, ``gnt-group evacuate``, to move all instances in a
1300
  node group to other groups.
1301
- Added ``gnt-instance change-group`` to move an instance to another
1302
  node group.
1303
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1304
  per-group operations.
1305
- Node groups can be tagged.
1306
- Some operations switch from an exclusive to a shared lock as soon as
1307
  possible.
1308
- Instance's primary and secondary nodes' groups are now available as
1309
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1310
  and ``snodes.group.uuid``).
1311

    
1312
Misc
1313
~~~~
1314

    
1315
- Numerous updates to documentation and manpages.
1316

    
1317
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1318
    descriptions.
1319
  - Some opcode/job results are now also documented, see :doc:`RAPI
1320
    <rapi>`.
1321

    
1322
- A lockset's internal lock is now also visible in lock monitor.
1323
- Log messages from job queue workers now contain information about the
1324
  opcode they're processing.
1325
- ``gnt-instance console`` no longer requires the instance lock.
1326
- A short delay when waiting for job changes reduces the number of LUXI
1327
  requests significantly.
1328
- DRBD metadata volumes are overwritten with zeros during disk creation.
1329
- Out-of-band commands no longer acquire the cluster lock in exclusive
1330
  mode.
1331
- ``devel/upload`` now uses correct permissions for directories.
1332

    
1333

    
1334
Version 2.5.0 rc6
1335
-----------------
1336

    
1337
*(Released Fri, 23 Mar 2012)*
1338

    
1339
This was the sixth release candidate of the 2.5 series.
1340

    
1341

    
1342
Version 2.5.0 rc5
1343
-----------------
1344

    
1345
*(Released Mon, 9 Jan 2012)*
1346

    
1347
This was the fifth release candidate of the 2.5 series.
1348

    
1349

    
1350
Version 2.5.0 rc4
1351
-----------------
1352

    
1353
*(Released Thu, 27 Oct 2011)*
1354

    
1355
This was the fourth release candidate of the 2.5 series.
1356

    
1357

    
1358
Version 2.5.0 rc3
1359
-----------------
1360

    
1361
*(Released Wed, 26 Oct 2011)*
1362

    
1363
This was the third release candidate of the 2.5 series.
1364

    
1365

    
1366
Version 2.5.0 rc2
1367
-----------------
1368

    
1369
*(Released Tue, 18 Oct 2011)*
1370

    
1371
This was the second release candidate of the 2.5 series.
1372

    
1373

    
1374
Version 2.5.0 rc1
1375
-----------------
1376

    
1377
*(Released Tue, 4 Oct 2011)*
1378

    
1379
This was the first release candidate of the 2.5 series.
1380

    
1381

    
1382
Version 2.5.0 beta3
1383
-------------------
1384

    
1385
*(Released Wed, 31 Aug 2011)*
1386

    
1387
This was the third beta release of the 2.5 series.
1388

    
1389

    
1390
Version 2.5.0 beta2
1391
-------------------
1392

    
1393
*(Released Mon, 22 Aug 2011)*
1394

    
1395
This was the second beta release of the 2.5 series.
1396

    
1397

    
1398
Version 2.5.0 beta1
1399
-------------------
1400

    
1401
*(Released Fri, 12 Aug 2011)*
1402

    
1403
This was the first beta release of the 2.5 series.
1404

    
1405

    
1406
Version 2.4.5
1407
-------------
1408

    
1409
*(Released Thu, 27 Oct 2011)*
1410

    
1411
- Fixed bug when parsing command line parameter values ending in
1412
  backslash
1413
- Fixed assertion error after unclean master shutdown
1414
- Disable HTTP client pool for RPC, significantly reducing memory usage
1415
  of master daemon
1416
- Fixed queue archive creation with wrong permissions
1417

    
1418

    
1419
Version 2.4.4
1420
-------------
1421

    
1422
*(Released Tue, 23 Aug 2011)*
1423

    
1424
Small bug-fixes:
1425

    
1426
- Fixed documentation for importing with ``--src-dir`` option
1427
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1428
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1429

    
1430

    
1431
Version 2.4.3
1432
-------------
1433

    
1434
*(Released Fri, 5 Aug 2011)*
1435

    
1436
Many bug-fixes and a few small features:
1437

    
1438
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1439
  issues when you tried to add an instance with two MAC addresses in one
1440
  request
1441
- KVM: fixed per-instance stored UID value
1442
- KVM: configure bridged NICs at migration start
1443
- KVM: Fix a bug where instance will not start with never KVM versions
1444
  (>= 0.14)
1445
- Added OS search path to ``gnt-cluster info``
1446
- Fixed an issue with ``file_storage_dir`` where you were forced to
1447
  provide an absolute path, but the documentation states it is a
1448
  relative path, the documentation was right
1449
- Added a new parameter to instance stop/start called ``--no-remember``
1450
  that will make the state change to not be remembered
1451
- Implemented ``no_remember`` at RAPI level
1452
- Improved the documentation
1453
- Node evacuation: don't call IAllocator if node is already empty
1454
- Fixed bug in DRBD8 replace disks on current nodes
1455
- Fixed bug in recreate-disks for DRBD instances
1456
- Moved assertion checking locks in ``gnt-instance replace-disks``
1457
  causing it to abort with not owning the right locks for some situation
1458
- Job queue: Fixed potential race condition when cancelling queued jobs
1459
- Fixed off-by-one bug in job serial generation
1460
- ``gnt-node volumes``: Fix instance names
1461
- Fixed aliases in bash completion
1462
- Fixed a bug in reopening log files after being sent a SIGHUP
1463
- Added a flag to burnin to allow specifying VCPU count
1464
- Bugfixes to non-root Ganeti configuration
1465

    
1466

    
1467
Version 2.4.2
1468
-------------
1469

    
1470
*(Released Thu, 12 May 2011)*
1471

    
1472
Many bug-fixes and a few new small features:
1473

    
1474
- Fixed a bug related to log opening failures
1475
- Fixed a bug in instance listing with orphan instances
1476
- Fixed a bug which prevented resetting the cluster-level node parameter
1477
  ``oob_program`` to the default
1478
- Many fixes related to the ``cluster-merge`` tool
1479
- Fixed a race condition in the lock monitor, which caused failures
1480
  during (at least) creation of many instances in parallel
1481
- Improved output for gnt-job info
1482
- Removed the quiet flag on some ssh calls which prevented debugging
1483
  failures
1484
- Improved the N+1 failure messages in cluster verify by actually
1485
  showing the memory values (needed and available)
1486
- Increased lock attempt timeouts so that when executing long operations
1487
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1488
  too early and thus prevent the use of the 'fair' mechanism
1489
- Changed instance query data (``gnt-instance info``) to not acquire
1490
  locks unless needed, thus allowing its use on locked instance if only
1491
  static information is asked for
1492
- Improved behaviour with filesystems that do not support rename on an
1493
  opened file
1494
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1495
  kept locks on all nodes during the wipe, which is unneeded
1496
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1497
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1498
  10GiB) which caused the wipe to fail right at the end in some cases
1499
- Fixed master IP activation when doing master failover with no-voting
1500
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1501
  the master node itself
1502
- Fixed potential data-loss in under disk full conditions, where Ganeti
1503
  wouldn't check correctly the return code and would consider
1504
  partially-written files 'correct'
1505
- Fixed bug related to multiple VGs and DRBD disk replacing
1506
- Added new disk parameter ``metavg`` that allows placement of the meta
1507
  device for DRBD in a different volume group
1508
- Fixed error handling in the node daemon when the system libc doesn't
1509
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1510
- Fixed lock release during replace-disks, which kept cluster-wide locks
1511
  when doing disk replaces with an iallocator script
1512
- Added check for missing bridges in cluster verify
1513
- Handle EPIPE errors while writing to the terminal better, so that
1514
  piping the output to e.g. ``less`` doesn't cause a backtrace
1515
- Fixed rare case where a ^C during Luxi calls could have been
1516
  interpreted as server errors, instead of simply terminating
1517
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1518
  assign-nodes``)
1519
- Added a few more parameters to the KVM hypervisor, allowing a second
1520
  CDROM, custom disk type for CDROMs and a floppy image
1521
- Removed redundant message in instance rename when the name is given
1522
  already as a FQDN
1523
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1524
  disks on new nodes, allowing recreation when the original instance
1525
  nodes are completely gone
1526
- Added option when converting disk templates to DRBD to skip waiting
1527
  for the resync, in order to make the instance available sooner
1528
- Added two new variables to the OS scripts environment (containing the
1529
  instance's nodes)
1530
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1531
  to allow use of ``pvgrub`` as bootloader
1532
- Changed the instance memory modifications to only check out-of-memory
1533
  conditions on memory increases, and turned the secondary node warnings
1534
  into errors (they can still be overridden via ``--force``)
1535
- Fixed the handling of a corner case when the Python installation gets
1536
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1537
  to execute a command that doesn't exist
1538
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1539
  node of the instance returned failures during instance shutdown; this
1540
  adds the option ``--ignore-consistency`` to gnt-instance move
1541

    
1542
And as usual, various improvements to the error messages, documentation
1543
and man pages.
1544

    
1545

    
1546
Version 2.4.1
1547
-------------
1548

    
1549
*(Released Wed, 09 Mar 2011)*
1550

    
1551
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1552
the RAPI users file if run twice (even with ``--dry-run``).
1553

    
1554
The release fixes that bug (nothing else changed).
1555

    
1556

    
1557
Version 2.4.0
1558
-------------
1559

    
1560
*(Released Mon, 07 Mar 2011)*
1561

    
1562
Final 2.4.0 release. Just a few small fixes:
1563

    
1564
- Fixed RAPI node evacuate
1565
- Fixed the kvm-ifup script
1566
- Fixed internal error handling for special job cases
1567
- Updated man page to specify the escaping feature for options
1568

    
1569

    
1570
Version 2.4.0 rc3
1571
-----------------
1572

    
1573
*(Released Mon, 28 Feb 2011)*
1574

    
1575
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1576
that this feature wiped the disks of the wrong instance, leading to loss
1577
of data.
1578

    
1579
Other changes:
1580

    
1581
- Fixed title of query field containing instance name
1582
- Expanded the glossary in the documentation
1583
- Fixed one unittest (internal issue)
1584

    
1585

    
1586
Version 2.4.0 rc2
1587
-----------------
1588

    
1589
*(Released Mon, 21 Feb 2011)*
1590

    
1591
A number of bug fixes plus just a couple functionality changes.
1592

    
1593
On the user-visible side, the ``gnt-* list`` command output has changed
1594
with respect to "special" field states. The current rc1 style of display
1595
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1596
the default output mode special fields states are displayed as follows:
1597

    
1598
- Offline resource: ``*``
1599
- Unavailable/not applicable: ``-``
1600
- Data missing (RPC failure): ``?``
1601
- Unknown field: ``??``
1602

    
1603
Another user-visible change is the addition of ``--force-join`` to
1604
``gnt-node add``.
1605

    
1606
As for bug fixes:
1607

    
1608
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1609
- Fixed regression in RAPI/instance reinstall where all parameters were
1610
  required (instead of optional)
1611
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1612
- Fixed iallocator usage (offline nodes were not considered offline)
1613
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1614
- Fixed hypervisor and OS parameter validation with respect to
1615
  non-vm_capable nodes
1616
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1617
  cosmetic)
1618
- Fixed ``tools/listrunner`` with respect to agent-based usage
1619

    
1620

    
1621
Version 2.4.0 rc1
1622
-----------------
1623

    
1624
*(Released Fri,  4 Feb 2011)*
1625

    
1626
Many changes and fixes since the beta1 release. While there were some
1627
internal changes, the code has been mostly stabilised for the RC
1628
release.
1629

    
1630
Note: the dumb allocator was removed in this release, as it was not kept
1631
up-to-date with the IAllocator protocol changes. It is recommended to
1632
use the ``hail`` command from the ganeti-htools package.
1633

    
1634
Note: the 2.4 and up versions of Ganeti are not compatible with the
1635
0.2.x branch of ganeti-htools. You need to upgrade to
1636
ganeti-htools-0.3.0 (or later).
1637

    
1638
Regressions fixed from 2.3
1639
~~~~~~~~~~~~~~~~~~~~~~~~~~
1640

    
1641
- Fixed the ``gnt-cluster verify-disks`` command
1642
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1643
  serially on nodes)
1644
- Fixed disk adoption breakage
1645
- Fixed wrong headers in instance listing for field aliases
1646

    
1647
Other bugs fixed
1648
~~~~~~~~~~~~~~~~
1649

    
1650
- Fixed corner case in KVM handling of NICs
1651
- Fixed many cases of wrong handling of non-vm_capable nodes
1652
- Fixed a bug where a missing instance symlink was not possible to
1653
  recreate with any ``gnt-*`` command (now ``gnt-instance
1654
  activate-disks`` does it)
1655
- Fixed the volume group name as reported by ``gnt-cluster
1656
  verify-disks``
1657
- Increased timeouts for the import-export code, hopefully leading to
1658
  fewer aborts due network or instance timeouts
1659
- Fixed bug in ``gnt-node list-storage``
1660
- Fixed bug where not all daemons were started on cluster
1661
  initialisation, but only at the first watcher run
1662
- Fixed many bugs in the OOB implementation
1663
- Fixed watcher behaviour in presence of instances with offline
1664
  secondaries
1665
- Fixed instance list output for instances running on the wrong node
1666
- a few fixes to the cluster-merge tool, but it still cannot merge
1667
  multi-node groups (currently it is not recommended to use this tool)
1668

    
1669

    
1670
Improvements
1671
~~~~~~~~~~~~
1672

    
1673
- Improved network configuration for the KVM hypervisor
1674
- Added e1000 as a supported NIC for Xen-HVM
1675
- Improved the lvmstrap tool to also be able to use partitions, as
1676
  opposed to full disks
1677
- Improved speed of disk wiping (the cluster parameter
1678
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1679
  of instance creations
1680
- Added documentation for the OS parameters
1681
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1682
  hypervisor is not responding
1683
- Added display of blacklisted and hidden OS information in
1684
  ``gnt-cluster info``
1685
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1686
  NIC and node parameters, which might create problems with currently
1687
  invalid (but undetected) configuration files, but prevents validation
1688
  failures when unrelated parameters are modified
1689
- Changed cluster initialisation to wait for the master daemon to become
1690
  available
1691
- Expanded the RAPI interface:
1692

    
1693
  - Added config redistribution resource
1694
  - Added activation/deactivation of instance disks
1695
  - Added export of console information
1696

    
1697
- Implemented log file reopening on SIGHUP, which allows using
1698
  logrotate(8) for the Ganeti log files
1699
- Added a basic OOB helper script as an example
1700

    
1701

    
1702
Version 2.4.0 beta1
1703
-------------------
1704

    
1705
*(Released Fri, 14 Jan 2011)*
1706

    
1707
User-visible
1708
~~~~~~~~~~~~
1709

    
1710
- Fixed timezone issues when formatting timestamps
1711
- Added support for node groups, available via ``gnt-group`` and other
1712
  commands
1713
- Added out-of-band framework and management, see :doc:`design
1714
  document <design-oob>`
1715
- Removed support for roman numbers from ``gnt-node list`` and
1716
  ``gnt-instance list``.
1717
- Allowed modification of master network interface via ``gnt-cluster
1718
  modify --master-netdev``
1719
- Accept offline secondaries while shutting down instance disks
1720
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1721
- Added support for multiple LVM volume groups
1722
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1723
  requested
1724
- Added commands to list available fields:
1725

    
1726
  - ``gnt-node list-fields``
1727
  - ``gnt-group list-fields``
1728
  - ``gnt-instance list-fields``
1729

    
1730
- Updated documentation and man pages
1731

    
1732
Integration
1733
~~~~~~~~~~~
1734

    
1735
- Moved ``rapi_users`` file into separate directory, now named
1736
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1737
  symlink
1738
- Added new tool for running commands on many machines,
1739
  ``tools/ganeti-listrunner``
1740
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1741
  improving the ``gnt-instance console`` output
1742
- Allowed customisation of disk index separator at ``configure`` time
1743
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1744
- Added support for non-partitioned md disks in ``lvmstrap``
1745
- Added script to gracefully power off KVM instances
1746
- Split ``utils`` module into smaller parts
1747
- Changed query operations to return more detailed information, e.g.
1748
  whether an information is unavailable due to an offline node. To use
1749
  this new functionality, the LUXI call ``Query`` must be used. Field
1750
  information is now stored by the master daemon and can be retrieved
1751
  using ``QueryFields``. Instances, nodes and groups can also be queried
1752
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1753
  exposed via RAPI). The following commands make use of this
1754
  infrastructure change:
1755

    
1756
  - ``gnt-group list``
1757
  - ``gnt-group list-fields``
1758
  - ``gnt-node list``
1759
  - ``gnt-node list-fields``
1760
  - ``gnt-instance list``
1761
  - ``gnt-instance list-fields``
1762
  - ``gnt-debug locks``
1763

    
1764
Remote API
1765
~~~~~~~~~~
1766

    
1767
- New RAPI resources (see :doc:`rapi`):
1768

    
1769
  - ``/2/modify``
1770
  - ``/2/groups``
1771
  - ``/2/groups/[group_name]``
1772
  - ``/2/groups/[group_name]/assign-nodes``
1773
  - ``/2/groups/[group_name]/modify``
1774
  - ``/2/groups/[group_name]/rename``
1775
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1776

    
1777
- RAPI changes:
1778

    
1779
  - Implemented ``no_install`` for instance creation
1780
  - Implemented OS parameters for instance reinstallation, allowing
1781
    use of special settings on reinstallation (e.g. for preserving data)
1782

    
1783
Misc
1784
~~~~
1785

    
1786
- Added IPv6 support in import/export
1787
- Pause DRBD synchronization while wiping disks on instance creation
1788
- Updated unittests and QA scripts
1789
- Improved network parameters passed to KVM
1790
- Converted man pages from docbook to reStructuredText
1791

    
1792

    
1793
Version 2.3.1
1794
-------------
1795

    
1796
*(Released Mon, 20 Dec 2010)*
1797

    
1798
Released version 2.3.1~rc1 without any changes.
1799

    
1800

    
1801
Version 2.3.1 rc1
1802
-----------------
1803

    
1804
*(Released Wed, 1 Dec 2010)*
1805

    
1806
- impexpd: Disable OpenSSL compression in socat if possible (backport
1807
  from master, commit e90739d625b, see :doc:`installation guide
1808
  <install-quick>` for details)
1809
- Changed unittest coverage report to exclude test scripts
1810
- Added script to check version format
1811

    
1812

    
1813
Version 2.3.0
1814
-------------
1815

    
1816
*(Released Wed, 1 Dec 2010)*
1817

    
1818
Released version 2.3.0~rc1 without any changes.
1819

    
1820

    
1821
Version 2.3.0 rc1
1822
-----------------
1823

    
1824
*(Released Fri, 19 Nov 2010)*
1825

    
1826
A number of bugfixes and documentation updates:
1827

    
1828
- Update ganeti-os-interface documentation
1829
- Fixed a bug related to duplicate MACs or similar items which should be
1830
  unique
1831
- Fix breakage in OS state modify
1832
- Reinstall instance: disallow offline secondaries (fixes bug related to
1833
  OS changing but reinstall failing)
1834
- plus all the other fixes between 2.2.1 and 2.2.2
1835

    
1836

    
1837
Version 2.3.0 rc0
1838
-----------------
1839

    
1840
*(Released Tue, 2 Nov 2010)*
1841

    
1842
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1843
- Fixed master failover race with watcher
1844
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1845
  configuration
1846
- Accept previously stopped instance for export with instance removal
1847
- Simplify and extend the environment variables for instance OS scripts
1848
- Added new node flags, ``master_capable`` and ``vm_capable``
1849
- Added optional instance disk wiping prior during allocation. This is a
1850
  cluster-wide option and can be set/modified using
1851
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1852
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1853
  :doc:`install-quick`
1854
- Added a new watcher option (``--ignore-pause``)
1855
- Added option to ignore offline node on instance start/stop
1856
  (``--ignore-offline``)
1857
- Allow overriding OS parameters with ``gnt-instance reinstall``
1858
- Added ability to change node's secondary IP address using ``gnt-node
1859
  modify``
1860
- Implemented privilege separation for all daemons except
1861
  ``ganeti-noded``, see ``configure`` options
1862
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1863
  verify``
1864
- Implemented job priorities (see ``ganeti(7)`` manpage)
1865
- Ignore failures while shutting down instances during failover from
1866
  offline node
1867
- Exit daemon's bootstrap process only once daemon is ready
1868
- Export more information via ``LUInstanceQuery``/remote API
1869
- Improved documentation, QA and unittests
1870
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1871
  restart if the file was created or changed
1872
- Added LUXI protocol version sent with each request and response,
1873
  allowing detection of server/client mismatches
1874
- Moved the Python scripts among gnt-* and ganeti-* into modules
1875
- Moved all code related to setting up SSH to an external script,
1876
  ``setup-ssh``
1877
- Infrastructure changes for node group support in future versions
1878

    
1879

    
1880
Version 2.2.2
1881
-------------
1882

    
1883
*(Released Fri, 19 Nov 2010)*
1884

    
1885
A few small bugs fixed, and some improvements to the build system:
1886

    
1887
- Fix documentation regarding conversion to drbd
1888
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1889
  -B``)
1890
- Fix error handling in node modify with multiple changes
1891
- Allow remote imports without checked names
1892

    
1893

    
1894
Version 2.2.1
1895
-------------
1896

    
1897
*(Released Tue, 19 Oct 2010)*
1898

    
1899
- Disable SSL session ID cache in RPC client
1900

    
1901

    
1902
Version 2.2.1 rc1
1903
-----------------
1904

    
1905
*(Released Thu, 14 Oct 2010)*
1906

    
1907
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1908
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1909
  with GnuTLS
1910
- Fix problems with interaction between Curl and Python's HTTP server,
1911
  resulting in increased speed in many RPC calls
1912
- Improve our release script to prevent breakage with older aclocal and
1913
  Python 2.6
1914

    
1915

    
1916
Version 2.2.1 rc0
1917
-----------------
1918

    
1919
*(Released Thu, 7 Oct 2010)*
1920

    
1921
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1922
  value retrieved from master
1923
- Added support for blacklisted or hidden OS definitions
1924
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1925
- Added support for -mem-path in KVM hypervisor abstraction layer
1926
- Allow overriding instance parameters in tool for inter-cluster
1927
  instance moves (``tools/move-instance``)
1928
- Improved opcode summaries (e.g. in ``gnt-job list``)
1929
- Improve consistency of OS listing by sorting it
1930
- Documentation updates
1931

    
1932

    
1933
Version 2.2.0.1
1934
---------------
1935

    
1936
*(Released Fri, 8 Oct 2010)*
1937

    
1938
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1939

    
1940

    
1941
Version 2.2.0
1942
-------------
1943

    
1944
*(Released Mon, 4 Oct 2010)*
1945

    
1946
- Fixed regression in ``gnt-instance rename``
1947

    
1948

    
1949
Version 2.2.0 rc2
1950
-----------------
1951

    
1952
*(Released Wed, 22 Sep 2010)*
1953

    
1954
- Fixed OS_VARIANT variable for OS scripts
1955
- Fixed cluster tag operations via RAPI
1956
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1957
- Disabled RAPI CA checks in watcher
1958

    
1959

    
1960
Version 2.2.0 rc1
1961
-----------------
1962

    
1963
*(Released Mon, 23 Aug 2010)*
1964

    
1965
- Support DRBD versions of the format "a.b.c.d"
1966
- Updated manpages
1967
- Re-introduce support for usage from multiple threads in RAPI client
1968
- Instance renames and modify via RAPI
1969
- Work around race condition between processing and archival in job
1970
  queue
1971
- Mark opcodes following failed one as failed, too
1972
- Job field ``lock_status`` was removed due to difficulties making it
1973
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1974
  locks is expected for a later 2.2.x release
1975
- Fixed dry-run behaviour with many commands
1976
- Support ``ssh-agent`` again when adding nodes
1977
- Many additional bugfixes
1978

    
1979

    
1980
Version 2.2.0 rc0
1981
-----------------
1982

    
1983
*(Released Fri, 30 Jul 2010)*
1984

    
1985
Important change: the internal RPC mechanism between Ganeti nodes has
1986
changed from using a home-grown http library (based on the Python base
1987
libraries) to use the PycURL library. This requires that PycURL is
1988
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1989
against GnuTLS by default. cURL's support for GnuTLS had known issues
1990
before cURL 7.21.0 and we recommend using the latest cURL release or
1991
linking against OpenSSL. Most other distributions already link PycURL
1992
and cURL against OpenSSL. The command::
1993

    
1994
  python -c 'import pycurl; print pycurl.version'
1995

    
1996
can be used to determine the libraries PycURL and cURL are linked
1997
against.
1998

    
1999
Other significant changes:
2000

    
2001
- Rewrote much of the internals of the job queue, in order to achieve
2002
  better parallelism; this decouples job query operations from the job
2003
  processing, and it should allow much nicer behaviour of the master
2004
  daemon under load, and it also has uncovered some long-standing bugs
2005
  related to the job serialisation (now fixed)
2006
- Added a default iallocator setting to the cluster parameters,
2007
  eliminating the need to always pass nodes or an iallocator for
2008
  operations that require selection of new node(s)
2009
- Added experimental support for the LXC virtualization method
2010
- Added support for OS parameters, which allows the installation of
2011
  instances to pass parameter to OS scripts in order to customise the
2012
  instance
2013
- Added a hypervisor parameter controlling the migration type (live or
2014
  non-live), since hypervisors have various levels of reliability; this
2015
  has renamed the 'live' parameter to 'mode'
2016
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
2017
  logical volumes, meaning that cluster verify will ignore them and not
2018
  flag their presence as errors
2019
- The watcher will now reset the error count for failed instances after
2020
  8 hours, thus allowing self-healing if the problem that caused the
2021
  instances to be down/fail to start has cleared in the meantime
2022
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2023
  check for, and warn, if the drbd module parameter ``usermode_helper``
2024
  is not consistent with the cluster-wide setting; this is needed to
2025
  make diagnose easier of failed drbd creations
2026
- Started adding base IPv6 support, but this is not yet
2027
  enabled/available for use
2028
- Rename operations (cluster, instance) will now return the new name,
2029
  which is especially useful if a short name was passed in
2030
- Added support for instance migration in RAPI
2031
- Added a tool to pre-configure nodes for the SSH setup, before joining
2032
  them to the cluster; this will allow in the future a simplified model
2033
  for node joining (but not yet fully enabled in 2.2); this needs the
2034
  paramiko python library
2035
- Fixed handling of name-resolving errors
2036
- Fixed consistency of job results on the error path
2037
- Fixed master-failover race condition when executed multiple times in
2038
  sequence
2039
- Fixed many bugs related to the job queue (mostly introduced during the
2040
  2.2 development cycle, so not all are impacting 2.1)
2041
- Fixed instance migration with missing disk symlinks
2042
- Fixed handling of unknown jobs in ``gnt-job archive``
2043
- And many other small fixes/improvements
2044

    
2045
Internal changes:
2046

    
2047
- Enhanced both the unittest and the QA coverage
2048
- Switched the opcode validation to a generic model, and extended the
2049
  validation to all opcode parameters
2050
- Changed more parts of the code that write shell scripts to use the
2051
  same class for this
2052
- Switched the master daemon to use the asyncore library for the Luxi
2053
  server endpoint
2054

    
2055

    
2056
Version 2.2.0 beta0
2057
-------------------
2058

    
2059
*(Released Thu, 17 Jun 2010)*
2060

    
2061
- Added tool (``move-instance``) and infrastructure to move instances
2062
  between separate clusters (see :doc:`separate documentation
2063
  <move-instance>` and :doc:`design document <design-2.2>`)
2064
- Added per-request RPC timeout
2065
- RAPI now requires a Content-Type header for requests with a body (e.g.
2066
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2067
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
2068
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2069
  reachable
2070
- Implemented initial support for running Ganeti daemons as separate
2071
  users, see configure-time flags ``--with-user-prefix`` and
2072
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2073
  time)
2074
- Instances can be removed after export (``gnt-backup export
2075
  --remove-instance``)
2076
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2077
  key (instead of 1024 bit)
2078
- Added new cluster configuration file for cluster domain secret
2079
- Import/export now use SSL instead of SSH
2080
- Added support for showing estimated time when exporting an instance,
2081
  see the ``ganeti-os-interface(7)`` manpage and look for
2082
  ``EXP_SIZE_FD``
2083

    
2084

    
2085
Version 2.1.8
2086
-------------
2087

    
2088
*(Released Tue, 16 Nov 2010)*
2089

    
2090
Some more bugfixes. Unless critical bugs occur, this will be the last
2091
2.1 release:
2092

    
2093
- Fix case of MAC special-values
2094
- Fix mac checker regex
2095
- backend: Fix typo causing "out of range" error
2096
- Add missing --units in gnt-instance list man page
2097

    
2098

    
2099
Version 2.1.7
2100
-------------
2101

    
2102
*(Released Tue, 24 Aug 2010)*
2103

    
2104
Bugfixes only:
2105
  - Don't ignore secondary node silently on non-mirrored disk templates
2106
    (issue 113)
2107
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2108
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2109
  - Properly document the usb_mouse parameter
2110
  - Fix path in ganeti-rapi(8) (issue 116)
2111
  - Adjust error message when the ganeti user's .ssh directory is
2112
    missing
2113
  - Add same-node-check when changing the disk template to drbd
2114

    
2115

    
2116
Version 2.1.6
2117
-------------
2118

    
2119
*(Released Fri, 16 Jul 2010)*
2120

    
2121
Bugfixes only:
2122
  - Add an option to only select some reboot types during qa/burnin.
2123
    (on some hypervisors consequent reboots are not supported)
2124
  - Fix infrequent race condition in master failover. Sometimes the old
2125
    master ip address would be still detected as up for a short time
2126
    after it was removed, causing failover to fail.
2127
  - Decrease mlockall warnings when the ctypes module is missing. On
2128
    Python 2.4 we support running even if no ctypes module is installed,
2129
    but we were too verbose about this issue.
2130
  - Fix building on old distributions, on which man doesn't have a
2131
    --warnings option.
2132
  - Fix RAPI not to ignore the MAC address on instance creation
2133
  - Implement the old instance creation format in the RAPI client.
2134

    
2135

    
2136
Version 2.1.5
2137
-------------
2138

    
2139
*(Released Thu, 01 Jul 2010)*
2140

    
2141
A small bugfix release:
2142
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2143
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2144
    a non-existing option
2145
  - Fix instance create: the --force-variant option was ignored
2146
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2147
  - Fix modify node storage with non-FQDN arguments
2148
  - Fix RAPI client to authenticate under Python 2.6 when used
2149
    for more than 5 requests needing authentication
2150
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2151
    when converting a non-shutdown drbd instance to plain
2152

    
2153

    
2154
Version 2.1.4
2155
-------------
2156

    
2157
*(Released Fri, 18 Jun 2010)*
2158

    
2159
A small bugfix release:
2160

    
2161
  - Fix live migration of KVM instances started with older Ganeti
2162
    versions which had fewer hypervisor parameters
2163
  - Fix gnt-instance grow-disk on down instances
2164
  - Fix an error-reporting bug during instance migration
2165
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2166
    silently ignoring broken ones
2167
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2168
    users
2169
  - Fix bug triggered by different API version os-es on different nodes
2170
  - Fix a bug in instance startup with custom hvparams: OS level
2171
    parameters would fail to be applied.
2172
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2173
    make it work completely well with OpenSSL)
2174
  - Fix handling of errors when resolving names from DNS
2175

    
2176

    
2177
Version 2.1.3
2178
-------------
2179

    
2180
*(Released Thu, 3 Jun 2010)*
2181

    
2182
A medium sized development cycle. Some new features, and some
2183
fixes/small improvements/cleanups.
2184

    
2185
Significant features
2186
~~~~~~~~~~~~~~~~~~~~
2187

    
2188
The node deamon now tries to mlock itself into memory, unless the
2189
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2190
its logs, and falls back to console logging. This allows emergency
2191
features such as ``gnt-node powercycle`` to work even in the event of a
2192
broken node disk (tested offlining the disk hosting the node's
2193
filesystem and dropping its memory caches; don't try this at home)
2194

    
2195
KVM: add vhost-net acceleration support. It can be tested with a new
2196
enough version of the kernel and of qemu-kvm.
2197

    
2198
KVM: Add instance chrooting feature. If you use privilege dropping for
2199
your VMs you can also now force them to chroot to an empty directory,
2200
before starting the emulated guest.
2201

    
2202
KVM: Add maximum migration bandwith and maximum downtime tweaking
2203
support (requires a new-enough version of qemu-kvm).
2204

    
2205
Cluster verify will now warn if the master node doesn't have the master
2206
ip configured on it.
2207

    
2208
Add a new (incompatible) instance creation request format to RAPI which
2209
supports all parameters (previously only a subset was supported, and it
2210
wasn't possible to extend the old format to accomodate all the new
2211
features. The old format is still supported, and a client can check for
2212
this feature, before using it, by checking for its presence in the
2213
``features`` RAPI resource.
2214

    
2215
Now with ancient latin support. Try it passing the ``--roman`` option to
2216
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2217
(requires the python-roman module to be installed, in order to work).
2218

    
2219
Other changes
2220
~~~~~~~~~~~~~
2221

    
2222
As usual many internal code refactorings, documentation updates, and
2223
such. Among others:
2224

    
2225
  - Lots of improvements and cleanups to the experimental Remote API
2226
    (RAPI) client library.
2227
  - A new unit test suite for the core daemon libraries.
2228
  - A fix to creating missing directories makes sure the umask is not
2229
    applied anymore. This enforces the same directory permissions
2230
    everywhere.
2231
  - Better handling terminating daemons with ctrl+c (used when running
2232
    them in debugging mode).
2233
  - Fix a race condition in live migrating a KVM instance, when stat()
2234
    on the old proc status file returned EINVAL, which is an unexpected
2235
    value.
2236
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2237
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2238

    
2239

    
2240
Version 2.1.2.1
2241
---------------
2242

    
2243
*(Released Fri, 7 May 2010)*
2244

    
2245
Fix a bug which prevented untagged KVM instances from starting.
2246

    
2247

    
2248
Version 2.1.2
2249
-------------
2250

    
2251
*(Released Fri, 7 May 2010)*
2252

    
2253
Another release with a long development cycle, during which many
2254
different features were added.
2255

    
2256
Significant features
2257
~~~~~~~~~~~~~~~~~~~~
2258

    
2259
The KVM hypervisor now can run the individual instances as non-root, to
2260
reduce the impact of a VM being hijacked due to bugs in the
2261
hypervisor. It is possible to run all instances as a single (non-root)
2262
user, to manually specify a user for each instance, or to dynamically
2263
allocate a user out of a cluster-wide pool to each instance, with the
2264
guarantee that no two instances will run under the same user ID on any
2265
given node.
2266

    
2267
An experimental RAPI client library, that can be used standalone
2268
(without the other Ganeti libraries), is provided in the source tree as
2269
``lib/rapi/client.py``. Note this client might change its interface in
2270
the future, as we iterate on its capabilities.
2271

    
2272
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2273
replace the cluster's certificates and crypto keys. This might help in
2274
case they have been compromised, or have simply expired.
2275

    
2276
A new disk option for instance creation has been added that allows one
2277
to "adopt" currently existing logical volumes, with data
2278
preservation. This should allow easier migration to Ganeti from
2279
unmanaged (or managed via other software) instances.
2280

    
2281
Another disk improvement is the possibility to convert between redundant
2282
(DRBD) and plain (LVM) disk configuration for an instance. This should
2283
allow better scalability (starting with one node and growing the
2284
cluster, or shrinking a two-node cluster to one node).
2285

    
2286
A new feature that could help with automated node failovers has been
2287
implemented: if a node sees itself as offline (by querying the master
2288
candidates), it will try to shutdown (hard) all instances and any active
2289
DRBD devices. This reduces the risk of duplicate instances if an
2290
external script automatically failovers the instances on such nodes. To
2291
enable this, the cluster parameter ``maintain_node_health`` should be
2292
enabled; in the future this option (per the name) will enable other
2293
automatic maintenance features.
2294

    
2295
Instance export/import now will reuse the original instance
2296
specifications for all parameters; that means exporting an instance,
2297
deleting it and the importing it back should give an almost identical
2298
instance. Note that the default import behaviour has changed from
2299
before, where it created only one NIC; now it recreates the original
2300
number of NICs.
2301

    
2302
Cluster verify has added a few new checks: SSL certificates validity,
2303
/etc/hosts consistency across the cluster, etc.
2304

    
2305
Other changes
2306
~~~~~~~~~~~~~
2307

    
2308
As usual, many internal changes were done, documentation fixes,
2309
etc. Among others:
2310

    
2311
- Fixed cluster initialization with disabled cluster storage (regression
2312
  introduced in 2.1.1)
2313
- File-based storage supports growing the disks
2314
- Fixed behaviour of node role changes
2315
- Fixed cluster verify for some corner cases, plus a general rewrite of
2316
  cluster verify to allow future extension with more checks
2317
- Fixed log spamming by watcher and node daemon (regression introduced
2318
  in 2.1.1)
2319
- Fixed possible validation issues when changing the list of enabled
2320
  hypervisors
2321
- Fixed cleanup of /etc/hosts during node removal
2322
- Fixed RAPI response for invalid methods
2323
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2324
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2325
  from ide disks, etc.)
2326
- Allow OS changes without re-installation (to record a changed OS
2327
  outside of Ganeti, or to allow OS renames)
2328
- Allow instance creation without OS installation (useful for example if
2329
  the OS will be installed manually, or restored from a backup not in
2330
  Ganeti format)
2331
- Implemented option to make cluster ``copyfile`` use the replication
2332
  network
2333
- Added list of enabled hypervisors to ssconf (possibly useful for
2334
  external scripts)
2335
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2336
  1.2 clusters
2337
- A partial form of node re-IP is possible via node readd, which now
2338
  allows changed node primary IP
2339
- Command line utilities now show an informational message if the job is
2340
  waiting for a lock
2341
- The logs of the master daemon now show the PID/UID/GID of the
2342
  connected client
2343

    
2344

    
2345
Version 2.1.1
2346
-------------
2347

    
2348
*(Released Fri, 12 Mar 2010)*
2349

    
2350
During the 2.1.0 long release candidate cycle, a lot of improvements and
2351
changes have accumulated with were released later as 2.1.1.
2352

    
2353
Major changes
2354
~~~~~~~~~~~~~
2355

    
2356
The node evacuate command (``gnt-node evacuate``) was significantly
2357
rewritten, and as such the IAllocator protocol was changed - a new
2358
request type has been added. This unfortunate change during a stable
2359
series is designed to improve performance of node evacuations; on
2360
clusters with more than about five nodes and which are well-balanced,
2361
evacuation should proceed in parallel for all instances of the node
2362
being evacuated. As such, any existing IAllocator scripts need to be
2363
updated, otherwise the above command will fail due to the unknown
2364
request. The provided "dumb" allocator has not been updated; but the
2365
ganeti-htools package supports the new protocol since version 0.2.4.
2366

    
2367
Another important change is increased validation of node and instance
2368
names. This might create problems in special cases, if invalid host
2369
names are being used.
2370

    
2371
Also, a new layer of hypervisor parameters has been added, that sits at
2372
OS level between the cluster defaults and the instance ones. This allows
2373
customisation of virtualization parameters depending on the installed
2374
OS. For example instances with OS 'X' may have a different KVM kernel
2375
(or any other parameter) than the cluster defaults. This is intended to
2376
help managing a multiple OSes on the same cluster, without manual
2377
modification of each instance's parameters.
2378

    
2379
A tool for merging clusters, ``cluster-merge``, has been added in the
2380
tools sub-directory.
2381

    
2382
Bug fixes
2383
~~~~~~~~~
2384

    
2385
- Improved the int/float conversions that should make the code more
2386
  robust in face of errors from the node daemons
2387
- Fixed the remove node code in case of internal configuration errors
2388
- Fixed the node daemon behaviour in face of inconsistent queue
2389
  directory (e.g. read-only file-system where we can't open the files
2390
  read-write, etc.)
2391
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2392
  now it either aborts cleanly or, if given the new "auto_promote"
2393
  parameter, will automatically promote other nodes as needed
2394
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2395
  completely prevent Ganeti from working
2396
- Fixed bug for instance export when not all disks were successfully
2397
  exported
2398
- Fixed behaviour of node add when the new node is slow in starting up
2399
  the node daemon
2400
- Fixed handling of signals in the LUXI client, which should improve
2401
  behaviour of command-line scripts
2402
- Added checks for invalid node/instance names in the configuration (now
2403
  flagged during cluster verify)
2404
- Fixed watcher behaviour for disk activation errors
2405
- Fixed two potentially endless loops in http library, which led to the
2406
  RAPI daemon hanging and consuming 100% CPU in some cases
2407
- Fixed bug in RAPI daemon related to hashed passwords
2408
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2409
  instances
2410
- Enhanced compatibility with non-Debian OSes, but not using absolute
2411
  path in some commands and allowing customisation of the ssh
2412
  configuration directory
2413
- Fixed possible future issue with new Python versions by abiding to the
2414
  proper use of ``__slots__`` attribute on classes
2415
- Added checks that should prevent directory traversal attacks
2416
- Many documentation fixes based on feedback from users
2417

    
2418
New features
2419
~~~~~~~~~~~~
2420

    
2421
- Added an "early_release" more for instance replace disks and node
2422
  evacuate, where we release locks earlier and thus allow higher
2423
  parallelism within the cluster
2424
- Added watcher hooks, intended to allow the watcher to restart other
2425
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2426
  course for any other purpose
2427
- Added a compile-time disable for DRBD barriers, to increase
2428
  performance if the administrator trusts the power supply or the
2429
  storage system to not lose writes
2430
- Added the option of using syslog for logging instead of, or in
2431
  addition to, Ganeti's own log files
2432
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2433
  can indeed boot from a paravirtual NIC
2434
- Added a generic debug level for many operations; while this is not
2435
  used widely yet, it allows one to pass the debug value all the way to
2436
  the OS scripts
2437
- Enhanced the hooks environment for instance moves (failovers,
2438
  migrations) where the primary/secondary nodes changed during the
2439
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2440
- Enhanced data validations for many user-supplied values; one important
2441
  item is the restrictions imposed on instance and node names, which
2442
  might reject some (invalid) host names
2443
- Add a configure-time option to disable file-based storage, if it's not
2444
  needed; this allows greater security separation between the master
2445
  node and the other nodes from the point of view of the inter-node RPC
2446
  protocol
2447
- Added user notification in interactive tools if job is waiting in the
2448
  job queue or trying to acquire locks
2449
- Added log messages when a job is waiting for locks
2450
- Added filtering by node tags in instance operations which admit
2451
  multiple instances (start, stop, reboot, reinstall)
2452
- Added a new tool for cluster mergers, ``cluster-merge``
2453
- Parameters from command line which are of the form ``a=b,c=d`` can now
2454
  use backslash escapes to pass in values which contain commas,
2455
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2456
  ``b,c``
2457
- For KVM, the instance name is the first parameter passed to KVM, so
2458
  that it's more visible in the process list
2459

    
2460

    
2461
Version 2.1.0
2462
-------------
2463

    
2464
*(Released Tue, 2 Mar 2010)*
2465

    
2466
Ganeti 2.1 brings many improvements with it. Major changes:
2467

    
2468
- Added infrastructure to ease automated disk repairs
2469
- Added new daemon to export configuration data in a cheaper way than
2470
  using the remote API
2471
- Instance NICs can now be routed instead of being associated with a
2472
  networking bridge
2473
- Improved job locking logic to reduce impact of jobs acquiring multiple
2474
  locks waiting for other long-running jobs
2475

    
2476
In-depth implementation details can be found in the Ganeti 2.1 design
2477
document.
2478

    
2479
Details
2480
~~~~~~~
2481

    
2482
- Added chroot hypervisor
2483
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2484
  ``device_model``)
2485
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2486
  ``bootloader_path`` and ``bootloader_args``)
2487
- Added the ``use_localtime`` option for the xen-hvm and kvm
2488
  hypervisors, and the default value for this has changed to false (in
2489
  2.0 xen-hvm always enabled it)
2490
- Added luxi call to submit multiple jobs in one go
2491
- Added cluster initialization option to not modify ``/etc/hosts``
2492
  file on nodes
2493
- Added network interface parameters
2494
- Added dry run mode to some LUs
2495
- Added RAPI resources:
2496

    
2497
  - ``/2/instances/[instance_name]/info``
2498
  - ``/2/instances/[instance_name]/replace-disks``
2499
  - ``/2/nodes/[node_name]/evacuate``
2500
  - ``/2/nodes/[node_name]/migrate``
2501
  - ``/2/nodes/[node_name]/role``
2502
  - ``/2/nodes/[node_name]/storage``
2503
  - ``/2/nodes/[node_name]/storage/modify``
2504
  - ``/2/nodes/[node_name]/storage/repair``
2505

    
2506
- Added OpCodes to evacuate or migrate all instances on a node
2507
- Added new command to list storage elements on nodes (``gnt-node
2508
  list-storage``) and modify them (``gnt-node modify-storage``)
2509
- Added new ssconf files with master candidate IP address
2510
  (``ssconf_master_candidates_ips``), node primary IP address
2511
  (``ssconf_node_primary_ips``) and node secondary IP address
2512
  (``ssconf_node_secondary_ips``)
2513
- Added ``ganeti-confd`` and a client library to query the Ganeti
2514
  configuration via UDP
2515
- Added ability to run hooks after cluster initialization and before
2516
  cluster destruction
2517
- Added automatic mode for disk replace (``gnt-instance replace-disks
2518
  --auto``)
2519
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2520
  after catastrophic data-loss
2521
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2522
  groups
2523
- Added ``gnt-instance move`` command to move instances
2524
- Added ``gnt-cluster watcher`` command to control watcher
2525
- Added ``gnt-node powercycle`` command to powercycle nodes
2526
- Added new job status field ``lock_status``
2527
- Added parseable error codes to cluster verification (``gnt-cluster
2528
  verify --error-codes``) and made output less verbose (use
2529
  ``--verbose`` to restore previous behaviour)
2530
- Added UUIDs to the main config entities (cluster, nodes, instances)
2531
- Added support for OS variants
2532
- Added support for hashed passwords in the Ganeti remote API users file
2533
  (``rapi_users``)
2534
- Added option to specify maximum timeout on instance shutdown
2535
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2536
- Added new helper script to start and stop Ganeti daemons
2537
  (``daemon-util``), with the intent to reduce the work necessary to
2538
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2539
  from one place
2540
- Added more unittests
2541
- Fixed critical bug in ganeti-masterd startup
2542
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2543
  now customisable at the cluster level for both the KVM and Xen
2544
  hypervisors using the new ``migration_port`` parameter
2545
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2546
  reinstalling an instance
2547
- Allowed ``@`` in tag names
2548
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2549
- Many documentation updates
2550
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2551
- ``gnt-instance reinstall`` can now reinstall multiple instances
2552
- Updated many command line parameters
2553
- Introduced new OS API version 15
2554
- No longer support a default hypervisor
2555
- Treat virtual LVs as inexistent
2556
- Improved job locking logic to reduce lock contention
2557
- Match instance and node names case insensitively
2558
- Reimplemented bash completion script to be more complete
2559
- Improved burnin
2560

    
2561

    
2562
Version 2.0.6
2563
-------------
2564

    
2565
*(Released Thu, 4 Feb 2010)*
2566

    
2567
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2568
- Fix a string formatting bug
2569
- Improve safety of the code in some error paths
2570
- Improve data validation in the master of values returned from nodes
2571

    
2572

    
2573
Version 2.0.5
2574
-------------
2575

    
2576
*(Released Thu, 17 Dec 2009)*
2577

    
2578
- Fix security issue due to missing validation of iallocator names; this
2579
  allows local and remote execution of arbitrary executables
2580
- Fix failure of gnt-node list during instance removal
2581
- Ship the RAPI documentation in the archive
2582

    
2583

    
2584
Version 2.0.4
2585
-------------
2586

    
2587
*(Released Wed, 30 Sep 2009)*
2588

    
2589
- Fixed many wrong messages
2590
- Fixed a few bugs related to the locking library
2591
- Fixed MAC checking at instance creation time
2592
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2593
- Fixed a few issues related to signal handling in both daemons and
2594
  scripts
2595
- Fixed the example startup script provided
2596
- Fixed insserv dependencies in the example startup script (patch from
2597
  Debian)
2598
- Fixed handling of drained nodes in the iallocator framework
2599
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2600
  #528618)
2601
- Fixed error related to invalid job IDs in job polling
2602
- Fixed job/opcode persistence on unclean master shutdown
2603
- Fixed handling of partial job processing after unclean master
2604
  shutdown
2605
- Fixed error reporting from LUs, previously all errors were converted
2606
  into execution errors
2607
- Fixed error reporting from burnin
2608
- Decreased significantly the memory usage of the job queue
2609
- Optimised slightly multi-job submission
2610
- Optimised slightly opcode loading
2611
- Backported the multi-job submit framework from the development
2612
  branch; multi-instance start and stop should be faster
2613
- Added script to clean archived jobs after 21 days; this will reduce
2614
  the size of the queue directory
2615
- Added some extra checks in disk size tracking
2616
- Added an example ethers hook script
2617
- Added a cluster parameter that prevents Ganeti from modifying of
2618
  /etc/hosts
2619
- Added more node information to RAPI responses
2620
- Added a ``gnt-job watch`` command that allows following the ouput of a
2621
  job
2622
- Added a bind-address option to ganeti-rapi
2623
- Added more checks to the configuration verify
2624
- Enhanced the burnin script such that some operations can be retried
2625
  automatically
2626
- Converted instance reinstall to multi-instance model
2627

    
2628

    
2629
Version 2.0.3
2630
-------------
2631

    
2632
*(Released Fri, 7 Aug 2009)*
2633

    
2634
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2635
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2636
  instances have mismatched disk sizes in the configuration
2637
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2638
  any configuration mismatches for disk sizes
2639
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2640
  failover to work on two-node clusters
2641
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2642
  unusable
2643
- Fixed detection of OS script errors in ``gnt-backup export``
2644
- Fixed exit code of ``gnt-backup export``
2645

    
2646

    
2647
Version 2.0.2
2648
-------------
2649

    
2650
*(Released Fri, 17 Jul 2009)*
2651

    
2652
- Added experimental support for stripped logical volumes; this should
2653
  enhance performance but comes with a higher complexity in the block
2654
  device handling; stripping is only enabled when passing
2655
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2656
  affected even in the non-stripped mode
2657
- Improved resiliency against transient failures at the end of DRBD
2658
  resyncs, and in general of DRBD resync checks
2659
- Fixed a couple of issues with exports and snapshot errors
2660
- Fixed a couple of issues in instance listing
2661
- Added display of the disk size in ``gnt-instance info``
2662
- Fixed checking for valid OSes in instance creation
2663
- Fixed handling of the "vcpus" parameter in instance listing and in
2664
  general of invalid parameters
2665
- Fixed http server library, and thus RAPI, to handle invalid
2666
  username/password combinations correctly; this means that now they
2667
  report unauthorized for queries too, not only for modifications,
2668
  allowing earlier detect of configuration problems
2669
- Added a new "role" node list field, equivalent to the master/master
2670
  candidate/drained/offline flags combinations
2671
- Fixed cluster modify and changes of candidate pool size
2672
- Fixed cluster verify error messages for wrong files on regular nodes
2673
- Fixed a couple of issues with node demotion from master candidate role
2674
- Fixed node readd issues
2675
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2676
- Added a new ``--no-voting`` option for masterfailover to fix failover
2677
  on two-nodes clusters when the former master node is unreachable
2678
- Added instance reinstall over RAPI
2679

    
2680

    
2681
Version 2.0.1
2682
-------------
2683

    
2684
*(Released Tue, 16 Jun 2009)*
2685

    
2686
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2687
  allow re-adding the start in single-user option (regression from 1.2)
2688
- the watcher writes the instance status to a file, to allow monitoring
2689
  to report the instance status (from the master) based on cached
2690
  results of the watcher's queries; while this can get stale if the
2691
  watcher is being locked due to other work on the cluster, this is
2692
  still an improvement
2693
- the watcher now also restarts the node daemon and the rapi daemon if
2694
  they died
2695
- fixed the watcher to handle full and drained queue cases
2696
- hooks export more instance data in the environment, which helps if
2697
  hook scripts need to take action based on the instance's properties
2698
  (no longer need to query back into ganeti)
2699
- instance failovers when the instance is stopped do not check for free
2700
  RAM, so that failing over a stopped instance is possible in low memory
2701
  situations
2702
- rapi uses queries for tags instead of jobs (for less job traffic), and
2703
  for cluster tags it won't talk to masterd at all but read them from
2704
  ssconf
2705
- a couple of error handling fixes in RAPI
2706
- drbd handling: improved the error handling of inconsistent disks after
2707
  resync to reduce the frequency of "there are some degraded disks for
2708
  this instance" messages
2709
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2710
  error handling path called a wrong function name)
2711

    
2712

    
2713
Version 2.0.0
2714
-------------
2715

    
2716
*(Released Wed, 27 May 2009)*
2717

    
2718
- no changes from rc5
2719

    
2720

    
2721
Version 2.0 rc5
2722
---------------
2723

    
2724
*(Released Wed, 20 May 2009)*
2725

    
2726
- fix a couple of bugs (validation, argument checks)
2727
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2728
- some small improvements to RAPI and IAllocator
2729
- make watcher automatically start the master daemon if down
2730

    
2731

    
2732
Version 2.0 rc4
2733
---------------
2734

    
2735
*(Released Mon, 27 Apr 2009)*
2736

    
2737
- change the OS list to not require locks; this helps with big clusters
2738
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2739
  volume group is broken
2740
- ``gnt-instance info``, without any arguments, doesn't run for all
2741
  instances anymore; either pass ``--all`` or pass the desired
2742
  instances; this helps against mistakes on big clusters where listing
2743
  the information for all instances takes a long time
2744
- miscellaneous doc and man pages fixes
2745

    
2746

    
2747
Version 2.0 rc3
2748
---------------
2749

    
2750
*(Released Wed, 8 Apr 2009)*
2751

    
2752
- Change the internal locking model of some ``gnt-node`` commands, in
2753
  order to reduce contention (and blocking of master daemon) when
2754
  batching many creation/reinstall jobs
2755
- Fixes to Xen soft reboot
2756
- No longer build documentation at build time, instead distribute it in
2757
  the archive, in order to reduce the need for the whole docbook/rst
2758
  toolchains
2759

    
2760

    
2761
Version 2.0 rc2
2762
---------------
2763

    
2764
*(Released Fri, 27 Mar 2009)*
2765

    
2766
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2767
- Fix watcher startup sequence, improves the behaviour of busy clusters
2768
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2769
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2770
  bind address and other places
2771
- Some documentation fixes and updates
2772

    
2773

    
2774
Version 2.0 rc1
2775
---------------
2776

    
2777
*(Released Mon, 2 Mar 2009)*
2778

    
2779
- More documentation updates, now all docs should be more-or-less
2780
  up-to-date
2781
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2782
  etc.)
2783
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2784
  and KVM)
2785
- Fix an issue related to $libdir/run/ganeti and cluster creation
2786

    
2787

    
2788
Version 2.0 beta2
2789
-----------------
2790

    
2791
*(Released Thu, 19 Feb 2009)*
2792

    
2793
- Xen PVM and KVM have switched the default value for the instance root
2794
  disk to the first partition on the first drive, instead of the whole
2795
  drive; this means that the OS installation scripts must be changed
2796
  accordingly
2797
- Man pages have been updated
2798
- RAPI has been switched by default to HTTPS, and the exported functions
2799
  should all work correctly
2800
- RAPI v1 has been removed
2801
- Many improvements to the KVM hypervisor
2802
- Block device errors are now better reported
2803
- Many other bugfixes and small improvements
2804

    
2805

    
2806
Version 2.0 beta1
2807
-----------------
2808

    
2809
*(Released Mon, 26 Jan 2009)*
2810

    
2811
- Version 2 is a general rewrite of the code and therefore the
2812
  differences are too many to list, see the design document for 2.0 in
2813
  the ``doc/`` subdirectory for more details
2814
- In this beta version there is not yet a migration path from 1.2 (there
2815
  will be one in the final 2.0 release)
2816
- A few significant changes are:
2817

    
2818
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2819
    various ``gnt-*`` commands are just front-ends to it
2820
  - all the commands are entered into, and executed from a job queue,
2821
    see the ``gnt-job(8)`` manpage
2822
  - the RAPI daemon supports read-write operations, secured by basic
2823
    HTTP authentication on top of HTTPS
2824
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2825
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2826
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2827
  - DRBD devices are using statically allocated minor numbers, which
2828
    will be assigned to existing instances during the migration process
2829
  - there is support for both Xen PVM and Xen HVM instances running on
2830
    the same cluster
2831
  - KVM virtualization is supported too
2832
  - file-based storage has been implemented, which means that it is
2833
    possible to run the cluster without LVM and DRBD storage, for
2834
    example using a shared filesystem exported from shared storage (and
2835
    still have live migration)
2836

    
2837

    
2838
Version 1.2.7
2839
-------------
2840

    
2841
*(Released Tue, 13 Jan 2009)*
2842

    
2843
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2844
- Reuse the old instance mac address by default on instance import, if
2845
  the instance name is the same.
2846
- Handle situations in which the node info rpc returns incomplete
2847
  results (issue 46)
2848
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2849
- Improved version of batcher:
2850

    
2851
  - state file support
2852
  - instance mac address support
2853
  - support for HVM clusters/instances
2854

    
2855
- Add an option to show the number of cpu sockets and nodes in
2856
  ``gnt-node list``
2857
- Support OSes that handle more than one version of the OS api (but do
2858
  not change the current API in any other way)
2859
- Fix ``gnt-node migrate``
2860
- ``gnt-debug`` man page
2861
- Fixes various more typos and small issues
2862
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2863

    
2864

    
2865
Version 1.2.6
2866
-------------
2867

    
2868
*(Released Wed, 24 Sep 2008)*
2869

    
2870
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2871
  type of disk exported to fully virtualized instances.
2872
- provide access to the serial console of HVM instances
2873
- instance auto_balance flag, set by default. If turned off it will
2874
  avoid warnings on cluster verify if there is not enough memory to fail
2875
  over an instance. in the future it will prevent automatically failing
2876
  it over when we will support that.
2877
- batcher tool for instance creation, see ``tools/README.batcher``
2878
- ``gnt-instance reinstall --select-os`` to interactively select a new
2879
  operating system when reinstalling an instance.
2880
- when changing the memory amount on instance modify a check has been
2881
  added that the instance will be able to start. also warnings are
2882
  emitted if the instance will not be able to fail over, if auto_balance
2883
  is true.
2884
- documentation fixes
2885
- sync fields between ``gnt-instance list/modify/add/import``
2886
- fix a race condition in drbd when the sync speed was set after giving
2887
  the device a remote peer.
2888

    
2889

    
2890
Version 1.2.5
2891
-------------
2892

    
2893
*(Released Tue, 22 Jul 2008)*
2894

    
2895
- note: the allowed size and number of tags per object were reduced
2896
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2897
- fixed twisted 8.x compatibility
2898
- fixed ``gnt-instance replace-disks`` with iallocator
2899
- add TCP keepalives on twisted connections to detect restarted nodes
2900
- disk increase support, see ``gnt-instance grow-disk``
2901
- implement bulk node/instance query for RAPI
2902
- add tags in node/instance listing (optional)
2903
- experimental migration (and live migration) support, read the man page
2904
  for ``gnt-instance migrate``
2905
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2906
  has some small improvements in handling its state file
2907

    
2908

    
2909
Version 1.2.4
2910
-------------
2911

    
2912
*(Released Fri, 13 Jun 2008)*
2913

    
2914
- Experimental readonly, REST-based remote API implementation;
2915
  automatically started on master node, TCP port 5080, if enabled by
2916
  ``--enable-rapi`` parameter to configure script.
2917
- Instance allocator support. Add and import instance accept a
2918
  ``--iallocator`` parameter, and call that instance allocator to decide
2919
  which node to use for the instance. The iallocator document describes
2920
  what's expected from an allocator script.
2921
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2922
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2923
  node is lost there is still enough memory to fail over the instances
2924
  that reside on it.
2925
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2926
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2927
  hooks will run, and their output, if any, will be displayed. Any
2928
  failing hook will make the verification return an error value.
2929
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2930
  primary and secondary interfaces
2931
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2932
  that is still declared as part of the cluster and has failed.
2933
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2934
  specifying output fields, that just adds the chosen fields to the
2935
  default ones.
2936
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2937
  export from the filesystem.
2938
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2939
  have been added. Using them you can enable/disable acpi and pae
2940
  support, and specify a path for a cd image to be exported to the
2941
  instance. These parameters as the name suggest only work on HVM
2942
  clusters.
2943
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2944
  PAE support will be set to the previously hardcoded values, but the
2945
  (previously hardcoded) path to the CDROM ISO image will be unset and
2946
  if required, needs to be set manually with ``gnt-instance modify``
2947
  after the upgrade.
2948
- The address to which an instance's VNC console is bound is now
2949
  selectable per-instance, rather than being cluster wide. Of course
2950
  this only applies to instances controlled via VNC, so currently just
2951
  applies to HVM clusters.
2952

    
2953

    
2954
Version 1.2.3
2955
-------------
2956

    
2957
*(Released Mon, 18 Feb 2008)*
2958

    
2959
- more tweaks to the disk activation code (especially helpful for DRBD)
2960
- change the default ``gnt-instance list`` output format, now there is
2961
  one combined status field (see the manpage for the exact values this
2962
  field will have)
2963
- some more fixes for the mac export to hooks change
2964
- make Ganeti not break with DRBD 8.2.x (which changed the version
2965
  format in ``/proc/drbd``) (issue 24)
2966
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2967
  template, allowing migration from DRBD0.7+MD to DRBD8
2968

    
2969

    
2970
Version 1.2.2
2971
-------------
2972

    
2973
*(Released Wed, 30 Jan 2008)*
2974

    
2975
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2976
  support (issue 23)
2977
- add command aliases infrastructure and a few aliases
2978
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2979
  man pages and the ``--help`` option of ``gnt-node
2980
  list``/``gnt-instance list``
2981
- fix ``gnt-backup list`` with down nodes (issue 21)
2982
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2983
- fix the dist archive and add a check for including svn/git files in
2984
  the future
2985
- some developer-related changes: improve the burnin and the QA suite,
2986
  add an upload script for testing during development
2987

    
2988

    
2989
Version 1.2.1
2990
-------------
2991

    
2992
*(Released Wed, 16 Jan 2008)*
2993

    
2994
- experimental HVM support, read the install document, section
2995
  "Initializing the cluster"
2996
- allow for the PVM hypervisor per-instance kernel and initrd paths
2997
- add a new command ``gnt-cluster verify-disks`` which uses a new
2998
  algorithm to improve the reconnection of the DRBD pairs if the device
2999
  on the secondary node has gone away
3000
- make logical volume code auto-activate LVs at disk activation time
3001
- slightly improve the speed of activating disks
3002
- allow specification of the MAC address at instance creation time, and
3003
  changing it later via ``gnt-instance modify``
3004
- fix handling of external commands that generate lots of output on
3005
  stderr
3006
- update documentation with regard to minimum version of DRBD8 supported
3007

    
3008

    
3009
Version 1.2.0
3010
-------------
3011

    
3012
*(Released Tue, 4 Dec 2007)*
3013

    
3014
- Log the ``xm create`` output to the node daemon log on failure (to
3015
  help diagnosing the error)
3016
- In debug mode, log all external commands output if failed to the logs
3017
- Change parsing of lvm commands to ignore stderr
3018

    
3019

    
3020
Version 1.2 beta3
3021
-----------------
3022

    
3023
*(Released Wed, 28 Nov 2007)*
3024

    
3025
- Another round of updates to the DRBD 8 code to deal with more failures
3026
  in the replace secondary node operation
3027
- Some more logging of failures in disk operations (lvm, drbd)
3028
- A few documentation updates
3029
- QA updates
3030

    
3031

    
3032
Version 1.2 beta2
3033
-----------------
3034

    
3035
*(Released Tue, 13 Nov 2007)*
3036

    
3037
- Change configuration file format from Python's Pickle to JSON.
3038
  Upgrading is possible using the cfgupgrade utility.
3039
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3040
  faster replace disks and is more stable (DRBD 8 has many improvements
3041
  compared to DRBD 0.7)
3042
- Added command line tags support (see man pages for ``gnt-instance``,
3043
  ``gnt-node``, ``gnt-cluster``)
3044
- Added instance rename support
3045
- Added multi-instance startup/shutdown
3046
- Added cluster rename support
3047
- Added ``gnt-node evacuate`` to simplify some node operations
3048
- Added instance reboot operation that can speedup reboot as compared to
3049
  stop and start
3050
- Soften the requirement that hostnames are in FQDN format
3051
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
3052
  reboots
3053
- Removed dependency on debian's patched fping that uses the
3054
  non-standard ``-S`` option
3055
- Now the OS definitions are searched for in multiple, configurable
3056
  paths (easier for distros to package)
3057
- Some changes to the hooks infrastructure (especially the new
3058
  post-configuration update hook)
3059
- Other small bugfixes
3060

    
3061
.. vim: set textwidth=72 syntax=rst :
3062
.. Local Variables:
3063
.. mode: rst
3064
.. fill-column: 72
3065
.. End: