Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 7afed8fe

History | View | Annotate | Download (118.7 kB)

1
News
2
====
3

    
4

    
5
Version 2.11.0 alpha1
6
---------------------
7

    
8
*(unreleased)*
9

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

    
13
- Over luxi, syntactially incorrect queries are now rejected as a whole;
14
  before, a 'SumbmitManyJobs' request was partially executed, if the outer
15
  structure of the request was syntactically correct. As the luxi protocol
16
  is internal (external applications are expected to use RAPI), the impact
17
  of this incompatible change should be limited.
18

    
19
New features
20
~~~~~~~~~~~~
21

    
22
- Instance moves, backups and imports can now use compression to transfer the
23
  instance data.
24
- Node groups can be configured to use an SSH port different than the
25
  default 22.
26

    
27
New dependencies
28
~~~~~~~~~~~~~~~~
29
The following new dependencies have been added:
30

    
31
For Haskell:
32
- ``zlib`` library (http://hackage.haskell.org/package/base64-bytestring)
33
- ``base64-bytestring`` library (http://hackage.haskell.org/package/zlib)
34

    
35

    
36
Version 2.10.0 alpha1
37
---------------------
38

    
39
*(unreleased)*
40

    
41
Incompatible/important changes
42
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
43

    
44
- Adding disks with 'gnt-instance modify' now waits for the disks to sync per
45
  default. Specify --no-wait-for-sync to override this behavior.
46
- The Ganeti python code now adheres to a private-module layout. In particular,
47
  the module 'ganeti' is no longer in the python search path.
48
- On instance allocation, the iallocator now considers non-LVM storage
49
  properly. In particular, actual file storage space information is used
50
  when allocating space for a file/sharedfile instance.
51
- When disabling disk templates cluster-wide, the cluster now first
52
  checks whether there are instances still using those templates.
53
- 'gnt-node list-storage' now also reports storage information about
54
  file-based storage types.
55

    
56
New features
57
~~~~~~~~~~~~
58

    
59
- KVM hypervisors can now access RBD storage directly without having to
60
  go through a block device.
61
- A new command 'gnt-cluster upgrade' was added that automates the upgrade
62
  procedure between two Ganeti versions that are both 2.10 or higher.
63
- The move-instance command can now change disk templates when moving
64
  instances, and does not require any node placement options to be
65
  specified if the destination cluster has a default iallocator.
66
- Users can now change the soundhw and cpuid settings for XEN hypervisors.
67
- Hail and hbal now have the (optional) capability of accessing average CPU
68
  load information through the monitoring deamon, and to use it to dynamically
69
  adapt the allocation of instances.
70
- Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
71
  so that disk and NIC modifications take effect without the need of actual
72
  reboot. There are a couple of constrains currently for this feature:
73

    
74
   - only KVM hypervisor (versions >= 1.0) supports it,
75
   - one can not (yet) hotplug a disk using userspace access mode for RBD
76
   - in case of a downgrade instances should suffer a reboot in order to
77
     be migratable (due to core change of runtime files)
78

    
79
Misc changes
80
~~~~~~~~~~~~
81

    
82
- A new test framework for logical units was introduced and the test
83
  coverage for logical units was improved significantly.
84
- Opcodes are entirely generated from Haskell using the tool 'hs2py' and
85
  the module 'src/Ganeti/OpCodes.hs'.
86
- Constants are also generated from Haskell using the tool
87
  'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
88
  exception of socket related constants, which require changing the
89
  cluster configuration file, and HVS related constants, because they
90
  are part of a port of instance queries to Haskell.  As a result, these
91
  changes will be part of the next release of Ganeti.
92

    
93
New dependencies
94
~~~~~~~~~~~~~~~~
95

    
96
The following new dependencies have been added/updated.
97

    
98
Python
99

    
100
- The version requirements for ``python-mock`` have increased to at least
101
  version 1.0.1. It is still used for testing only.
102

    
103

    
104
Version 2.9.1
105
-------------
106

    
107
*(Released Wed, 13 Nov 2013)*
108

    
109
- fix bug, that kept nodes offline when readding
110
- when verifying DRBD versions, ignore unavailable nodes
111
- fix bug that made the console unavailable on kvm in split-user
112
  setup (issue 608)
113
- DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
114

    
115

    
116
Version 2.9.0
117
-------------
118

    
119
*(Released Tue, 5 Nov 2013)*
120

    
121
Incompatible/important changes
122
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
123

    
124
- hroller now also plans for capacity to move non-redundant instances off
125
  any node to be rebooted; the old behavior of completely ignoring any
126
  non-redundant instances can be restored by adding the --ignore-non-redundant
127
  option.
128
- The cluster option '--no-lvm-storage' was removed in favor of the new option
129
  '--enabled-disk-templates'.
130
- On instance creation, disk templates no longer need to be specified
131
  with '-t'. The default disk template will be taken from the list of
132
  enabled disk templates.
133
- The monitoring daemon is now running as root, in order to be able to collect
134
  information only available to root (such as the state of Xen instances).
135
- The ConfD client is now IPv6 compatible.
136
- File and shared file storage is no longer dis/enabled at configure time,
137
  but using the option '--enabled-disk-templates' at cluster initialization and
138
  modification.
139
- The default directories for file and shared file storage are not anymore
140
  specified at configure time, but taken from the cluster's configuration.
141
  They can be set at cluster initialization and modification with
142
  '--file-storage-dir' and '--shared-file-storage-dir'.
143
- Cluster verification now includes stricter checks regarding the
144
  default file and shared file storage directories. It now checks that
145
  the directories are explicitely allowed in the 'file-storage-paths' file and
146
  that the directories exist on all nodes.
147
- The list of allowed disk templates in the instance policy and the list
148
  of cluster-wide enabled disk templates is now checked for consistency
149
  on cluster or group modification. On cluster initialization, the ipolicy
150
  disk templates are ensured to be a subset of the cluster-wide enabled
151
  disk templates.
152

    
153
New features
154
~~~~~~~~~~~~
155

    
156
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
157
  the correct command syntax. It is possible to use different DRBD versions
158
  on different nodes as long as they are compatible to each other. This
159
  enables rolling upgrades of DRBD with no downtime. As permanent operation
160
  of different DRBD versions within a node group is discouraged,
161
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
162
- New "inst-status-xen" data collector for the monitoring daemon, providing
163
  information about the state of the xen instances on the nodes.
164
- New "lv" data collector for the monitoring daemon, collecting data about the
165
  logical volumes on the nodes, and pairing them with the name of the instances
166
  they belong to.
167
- New "diskstats" data collector, collecting the data from /proc/diskstats and
168
  presenting them over the monitoring daemon interface.
169
- The ConfD client is now IPv6 compatible.
170

    
171
New dependencies
172
~~~~~~~~~~~~~~~~
173
The following new dependencies have been added.
174

    
175
Python
176

    
177
- ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
178
  for the unit tests (and only used for testing).
179

    
180
Haskell
181

    
182
- ``hslogger`` (http://software.complete.org/hslogger) is now always
183
  required, even if confd is not enabled.
184

    
185
Since 2.9.0 rc3
186
~~~~~~~~~~~~~~~
187

    
188
- Correctly start/stop luxid during gnt-cluster master-failover (inherited
189
  from stable-2.8)
190
- Improved error messsages (inherited from stable-2.8)
191

    
192

    
193
Version 2.9.0 rc3
194
-----------------
195

    
196
*(Released Tue, 15 Oct 2013)*
197

    
198
The third release candidate in the 2.9 series. Since 2.9.0 rc2:
199

    
200
- in implicit configuration upgrade, match ipolicy with enabled disk templates
201
- improved harep documentation (inherited from stable-2.8)
202

    
203

    
204
Version 2.9.0 rc2
205
-----------------
206

    
207
*(Released Wed, 9 Oct 2013)*
208

    
209
The second release candidate in the 2.9 series. Since 2.9.0 rc1:
210

    
211
- Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
212
  at least one DRBD instance.
213
- Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
214
  downgrading.
215

    
216

    
217
Version 2.9.0 rc1
218
-----------------
219

    
220
*(Released Tue, 1 Oct 2013)*
221

    
222
The first release candidate in the 2.9 series. Since 2.9.0 beta1:
223

    
224
- various bug fixes
225
- update of the documentation, in particular installation instructions
226
- merging of LD_* constants into DT_* constants
227
- python style changes to be compatible with newer versions of pylint
228

    
229

    
230
Version 2.9.0 beta1
231
-------------------
232

    
233
*(Released Thu, 29 Aug 2013)*
234

    
235
This was the first beta release of the 2.9 series. All important changes
236
are listed in the latest 2.9 entry.
237

    
238

    
239
Version 2.8.2
240
-------------
241

    
242
*(Released Thu, 07 Nov 2013)*
243

    
244
- DRBD: ensure peers are UpToDate for dual-primary
245
- Improve error message for replace-disks
246
- More dependency checks at configure time
247
- Placate warnings on ganeti.outils_unittest.py
248

    
249

    
250
Version 2.8.1
251
-------------
252

    
253
*(Released Thu, 17 Oct 2013)*
254

    
255
- Correctly start/stop luxid during gnt-cluster master-failover
256
- Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
257
- Fix path for the job queue serial file
258
- Improved harep man page
259
- Minor documentation improvements
260

    
261

    
262
Version 2.8.0
263
-------------
264

    
265
*(Released Mon, 30 Sep 2013)*
266

    
267
Incompatible/important changes
268
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
269

    
270
- Instance policy can contain multiple instance specs, as described in
271
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
272
  <design-partitioned>`. As a consequence, it's not possible to partially change
273
  or override instance specs. Bounding specs (min and max) can be specified as a
274
  whole using the new option ``--ipolicy-bounds-specs``, while standard
275
  specs use the new option ``--ipolicy-std-specs``.
276
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
277
  gnt-instance is a valid YAML object.
278
- hail now honors network restrictions when allocating nodes. This led to an
279
  update of the IAllocator protocol. See the IAllocator documentation for
280
  details.
281
- confd now only answers static configuration request over the network. luxid
282
  was extracted, listens on the local LUXI socket and responds to live queries.
283
  This allows finer grained permissions if using separate users.
284

    
285
New features
286
~~~~~~~~~~~~
287

    
288
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
289
  to always require authentication, ``--require-authentication``. It can
290
  be specified in ``$sysconfdir/default/ganeti``.
291
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
292
  be used to manage the disk templates to be used by instances in the cluster.
293
  Initially, it will be set to a list that includes plain, drbd, if they were
294
  enabled by specifying a volume group name, and file and sharedfile, if those
295
  were enabled at configure time. Additionally, it will include all disk
296
  templates that are currently used by instances. The order of disk templates
297
  will be based on Ganeti's history of supporting them. In the future, the
298
  first entry of the list will be used as a default disk template on instance
299
  creation.
300
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
301
  configuration back to the previous stable version.
302
- Disk templates in group ipolicy can be restored to the default value.
303
- Initial support for diskless instances and virtual clusters in QA.
304
- More QA and unit tests for instance policies.
305
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
306
  describing why the opcode itself was executed.
307
- The monitoring daemon is now available. It allows users to query the cluster
308
  for obtaining information about the status of the system. The daemon is only
309
  responsible for providing the information over the network: the actual data
310
  gathering is performed by data collectors (currently, only the DRBD status
311
  collector is available).
312
- In order to help developers work on Ganeti, a new script
313
  (``devel/build_chroot``) is provided, for building a chroot that contains all
314
  the required development libraries and tools for compiling Ganeti on a Debian
315
  Squeeze system.
316
- A new tool, ``harep``, for performing self-repair and recreation of instances
317
  in Ganeti has been added.
318
- Split queries are enabled for tags, network, exports, cluster info, groups,
319
  jobs, nodes.
320
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
321
  It prints the command line to set the current policies, to ease
322
  changing them.
323
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
324
  devices for KVM virtio-net interfaces will get created with VNET_HDR
325
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
326
  virtio-net interfaces, which prevents host kernel tainting and log
327
  flooding, when dealing with broken or malicious virtio-net drivers.
328
  It's set to true by default.
329
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
330
  failures.
331
- Support 'viridian' parameter in Xen HVM
332
- Support DSA SSH keys in bootstrap
333
- To simplify the work of packaging frameworks that want to add the needed users
334
  and groups in a split-user setup themselves, at build time three files in
335
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
336
  the groups to be generated, the ``users`` file contains, one per line, the
337
  users to be generated, optionally followed by their primary group, where
338
  important. The ``groupmemberships`` file contains, one per line, additional
339
  user-group membership relations that need to be established. The syntax of
340
  these files will remain stable in all future versions.
341

    
342

    
343
New dependencies
344
~~~~~~~~~~~~~~~~
345
The following new dependencies have been added:
346

    
347
For Haskell:
348
- The ``curl`` library is not optional anymore for compiling the Haskell code.
349
- ``snap-server`` library (if monitoring is enabled).
350

    
351
For Python:
352
- The minimum Python version needed to run Ganeti is now 2.6.
353
- ``yaml`` library (only for running the QA).
354

    
355
Since 2.8.0 rc3
356
~~~~~~~~~~~~~~~
357
- Perform proper cleanup on termination of Haskell daemons
358
- Fix corner-case in handling of remaining retry time
359

    
360

    
361
Version 2.8.0 rc3
362
-----------------
363

    
364
*(Released Tue, 17 Sep 2013)*
365

    
366
- To simplify the work of packaging frameworks that want to add the needed users
367
  and groups in a split-user setup themselves, at build time three files in
368
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
369
  the groups to be generated, the ``users`` file contains, one per line, the
370
  users to be generated, optionally followed by their primary group, where
371
  important. The ``groupmemberships`` file contains, one per line, additional
372
  user-group membership relations that need to be established. The syntax of
373
  these files will remain stable in all future versions.
374
- Add a default to file-driver when unspecified over RAPI (Issue 571)
375
- Mark the DSA host pubkey as optional, and remove it during config downgrade
376
  (Issue 560)
377
- Some documentation fixes
378

    
379

    
380
Version 2.8.0 rc2
381
-----------------
382

    
383
*(Released Tue, 27 Aug 2013)*
384

    
385
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
386

    
387
- Support 'viridian' parameter in Xen HVM (Issue 233)
388
- Include VCS version in ``gnt-cluster version``
389
- Support DSA SSH keys in bootstrap (Issue 338)
390
- Fix batch creation of instances
391
- Use FQDN to check master node status (Issue 551)
392
- Make the DRBD collector more failure-resilient
393

    
394

    
395
Version 2.8.0 rc1
396
-----------------
397

    
398
*(Released Fri, 2 Aug 2013)*
399

    
400
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
401

    
402
- Fix upgrading/downgrading from 2.7
403
- Increase maximum RAPI message size
404
- Documentation updates
405
- Split ``confd`` between ``luxid`` and ``confd``
406
- Merge 2.7 series up to the 2.7.1 release
407
- Allow the ``modify_etc_hosts`` option to be changed
408
- Add better debugging for ``luxid`` queries
409
- Expose bulk parameter for GetJobs in RAPI client
410
- Expose missing ``network`` fields in RAPI
411
- Add some ``cluster verify`` tests
412
- Some unittest fixes
413
- Fix a malfunction in ``hspace``'s tiered allocation
414
- Fix query compatibility between haskell and python implementations
415
- Add the ``vnet_hdr`` HV parameter for KVM
416
- Add ``--cleanup`` to instance failover
417
- Change the connected groups format in ``gnt-network info`` output; it
418
  was previously displayed as a raw list by mistake. (Merged from 2.7)
419

    
420

    
421
Version 2.8.0 beta1
422
-------------------
423

    
424
*(Released Mon, 24 Jun 2013)*
425

    
426
This was the first beta release of the 2.8 series. All important changes
427
are listed in the latest 2.8 entry.
428

    
429

    
430
Version 2.7.2
431
-------------
432

    
433
*(Released Thu, 26 Sep 2013)*
434

    
435
- Change the connected groups format in ``gnt-network info`` output; it
436
  was previously displayed as a raw list by mistake
437
- Check disk template in right dict when copying
438
- Support multi-instance allocs without iallocator
439
- Fix some errors in the documentation
440
- Fix formatting of tuple in an error message
441

    
442

    
443
Version 2.7.1
444
-------------
445

    
446
*(Released Thu, 25 Jul 2013)*
447

    
448
- Add logrotate functionality in daemon-util
449
- Add logrotate example file
450
- Add missing fields to network queries over rapi
451
- Fix network object timestamps
452
- Add support for querying network timestamps
453
- Fix a typo in the example crontab
454
- Fix a documentation typo
455

    
456

    
457
Version 2.7.0
458
-------------
459

    
460
*(Released Thu, 04 Jul 2013)*
461

    
462
Incompatible/important changes
463
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
464

    
465
- Instance policies for disk size were documented to be on a per-disk
466
  basis, but hail applied them to the sum of all disks. This has been
467
  fixed.
468
- ``hbal`` will now exit with status 0 if, during job execution over
469
  LUXI, early exit has been requested and all jobs are successful;
470
  before, exit status 1 was used, which cannot be differentiated from
471
  "job error" case
472
- Compatibility with newer versions of rbd has been fixed
473
- ``gnt-instance batch-create`` has been changed to use the bulk create
474
  opcode from Ganeti. This lead to incompatible changes in the format of
475
  the JSON file. It's now not a custom dict anymore but a dict
476
  compatible with the ``OpInstanceCreate`` opcode.
477
- Parent directories for file storage need to be listed in
478
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
479
  write the file automatically based on old configuration values, but it
480
  can not distribute it across all nodes and the file contents should be
481
  verified. Use ``gnt-cluster copyfile
482
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
483
  upgraded. The reason for requiring this list of paths now is that
484
  before it would have been possible to inject new paths via RPC,
485
  allowing files to be created in arbitrary locations. The RPC protocol
486
  is protected using SSL/X.509 certificates, but as a design principle
487
  Ganeti does not permit arbitrary paths to be passed.
488
- The parsing of the variants file for OSes (see
489
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
490
  lines and comment lines (starting with ``#``) are ignored for better
491
  readability.
492
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
493
  longer available. ``gnt-node add`` now invokes a new tool on the
494
  destination node, named ``prepare-node-join``, to configure the SSH
495
  daemon. Paramiko is no longer necessary to configure nodes' SSH
496
  daemons via ``gnt-node add``.
497
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
498
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
499
  the flag is not reset after a master failover.
500
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
501
  is recommended. 2.6 will be mandatory from the 2.8 series.
502

    
503

    
504
New features
505
~~~~~~~~~~~~
506

    
507
- New network management functionality to support automatic allocation
508
  of IP addresses and managing of network parameters. See
509
  :manpage:`gnt-network(8)` for more details.
510
- New external storage backend, to allow managing arbitrary storage
511
  systems external to the cluster. See
512
  :manpage:`ganeti-extstorage-interface(7)`.
513
- New ``exclusive-storage`` node parameter added, restricted to
514
  nodegroup level. When it's set to true, physical disks are assigned in
515
  an exclusive fashion to instances, as documented in :doc:`Partitioned
516
  Ganeti <design-partitioned>`.  Currently, only instances using the
517
  ``plain`` disk template are supported.
518
- The KVM hypervisor has been updated with many new hypervisor
519
  parameters, including a generic one for passing arbitrary command line
520
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
521
  compatible up to qemu 1.4.
522
- A new tool, called ``mon-collector``, is the stand-alone executor of
523
  the data collectors for a monitoring system. As of this version, it
524
  just includes the DRBD data collector, that can be executed by calling
525
  ``mon-collector`` using the ``drbd`` parameter. See
526
  :manpage:`mon-collector(7)`.
527
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
528
  for RAPI users. It allows granting permissions to query for
529
  information to a specific user without giving
530
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
531
- A new tool named ``node-cleanup`` has been added. It cleans remains of
532
  a cluster from a machine by stopping all daemons, removing
533
  certificates and ssconf files. Unless the ``--no-backup`` option is
534
  given, copies of the certificates are made.
535
- Instance creations now support the use of opportunistic locking,
536
  potentially speeding up the (parallel) creation of multiple instances.
537
  This feature is currently only available via the :doc:`RAPI
538
  <rapi>` interface and when an instance allocator is used. If the
539
  ``opportunistic_locking`` parameter is set the opcode will try to
540
  acquire as many locks as possible, but will not wait for any locks
541
  held by other opcodes. If not enough resources can be found to
542
  allocate the instance, the temporary error code
543
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
544
  retried thereafter, with or without opportunistic locking.
545
- New experimental linux-ha resource scripts.
546
- Restricted-commands support: ganeti can now be asked (via command line
547
  or rapi) to perform commands on a node. These are passed via ganeti
548
  RPC rather than ssh. This functionality is restricted to commands
549
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
550
  reasons. The file is not copied automatically.
551

    
552

    
553
Misc changes
554
~~~~~~~~~~~~
555

    
556
- Diskless instances are now externally mirrored (Issue 237). This for
557
  now has only been tested in conjunction with explicit target nodes for
558
  migration/failover.
559
- Queries not needing locks or RPC access to the node can now be
560
  performed by the confd daemon, making them independent from jobs, and
561
  thus faster to execute. This is selectable at configure time.
562
- The functionality for allocating multiple instances at once has been
563
  overhauled and is now also available through :doc:`RAPI <rapi>`.
564

    
565
There are no significant changes from version 2.7.0~rc3.
566

    
567

    
568
Version 2.7.0 rc3
569
-----------------
570

    
571
*(Released Tue, 25 Jun 2013)*
572

    
573
- Fix permissions on the confd query socket (Issue 477)
574
- Fix permissions on the job archive dir (Issue 498)
575
- Fix handling of an internal exception in replace-disks (Issue 472)
576
- Fix gnt-node info handling of shortened names (Issue 497)
577
- Fix gnt-instance grow-disk when wiping is enabled
578
- Documentation improvements, and support for newer pandoc
579
- Fix hspace honoring ipolicy for disks (Issue 484)
580
- Improve handling of the ``kvm_extra`` HV parameter
581

    
582

    
583
Version 2.7.0 rc2
584
-----------------
585

    
586
*(Released Fri, 24 May 2013)*
587

    
588
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
589
  symlink.
590
- Disks added through ``gnt-instance modify`` or created through
591
  ``gnt-instance recreate-disks`` are wiped, if the
592
  ``prealloc_wipe_disks`` flag is set.
593
- If wiping newly created disks fails, the disks are removed. Also,
594
  partial failures in creating disks through ``gnt-instance modify``
595
  triggers a cleanup of the partially-created disks.
596
- Removing the master IP address doesn't fail if the address has been
597
  already removed.
598
- Fix ownership of the OS log dir
599
- Workaround missing SO_PEERCRED constant (Issue 191)
600

    
601

    
602
Version 2.7.0 rc1
603
-----------------
604

    
605
*(Released Fri, 3 May 2013)*
606

    
607
This was the first release candidate of the 2.7 series. Since beta3:
608

    
609
- Fix kvm compatibility with qemu 1.4 (Issue 389)
610
- Documentation updates (admin guide, upgrade notes, install
611
  instructions) (Issue 372)
612
- Fix gnt-group list nodes and instances count (Issue 436)
613
- Fix compilation without non-mandatory libraries (Issue 441)
614
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
615
- Make confd logging more verbose at INFO level (Issue 435)
616
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
617
- Fix failure path for instance storage type conversion (Issue 229)
618
- Update htools text backend documentation
619
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
620
- Disable inter-cluster instance move for file-based instances, because
621
  it is dependant on instance export, which is not supported for
622
  file-based instances. (Issue 414)
623
- Fix gnt-job crashes on non-ascii characters (Issue 427)
624
- Fix volume group checks on non-vm-capable nodes (Issue 432)
625

    
626

    
627
Version 2.7.0 beta3
628
-------------------
629

    
630
*(Released Mon, 22 Apr 2013)*
631

    
632
This was the third beta release of the 2.7 series. Since beta2:
633

    
634
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
635
- Fix data loss on wrong usage of ``gnt-instance move``
636
- Properly export errors in confd-based job queries
637
- Add ``users-setup`` tool
638
- Fix iallocator protocol to report 0 as a disk size for diskless
639
  instances. This avoids hail breaking when a diskless instance is
640
  present.
641
- Fix job queue directory permission problem that made confd job queries
642
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
643
  for access to archived jobs (Issue 406).
644
- Limit the sizes of networks supported by ``gnt-network`` to something
645
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
646
- Fix bugs in instance disk template conversion
647
- Fix GHC 7 compatibility
648
- Fix ``burnin`` install path (Issue 426).
649
- Allow very small disk grows (Issue 347).
650
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
651
  that noded doesn't import masterd code (Issue 419).
652
- Make sure the default metavg at cluster init is the same as the vg, if
653
  unspecified (Issue 358).
654
- Fix cleanup of partially created disks (part of Issue 416)
655

    
656

    
657
Version 2.7.0 beta2
658
-------------------
659

    
660
*(Released Tue, 2 Apr 2013)*
661

    
662
This was the second beta release of the 2.7 series. Since beta1:
663

    
664
- Networks no longer have a "type" slot, since this information was
665
  unused in Ganeti: instead of it tags should be used.
666
- The rapi client now has a ``target_node`` option to MigrateInstance.
667
- Fix early exit return code for hbal (Issue 386).
668
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
669
- Fix ``rbd showmapped`` output parsing (Issue 312).
670
- Networks are now referenced indexed by UUID, rather than name. This
671
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
672
  use.
673
- The OS environment now includes network information.
674
- Deleting of a network is now disallowed if any instance nic is using
675
  it, to prevent dangling references.
676
- External storage is now documented in man pages.
677
- The exclusive_storage flag can now only be set at nodegroup level.
678
- Hbal can now submit an explicit priority with its jobs.
679
- Many network related locking fixes.
680
- Bump up the required pylint version to 0.25.1.
681
- Fix the ``no_remember`` option in RAPI client.
682
- Many ipolicy related tests, qa, and fixes.
683
- Many documentation improvements and fixes.
684
- Fix building with ``--disable-file-storage``.
685
- Fix ``-q`` option in htools, which was broken if passed more than
686
  once.
687
- Some haskell/python interaction improvements and fixes.
688
- Fix iallocator in case of missing LVM storage.
689
- Fix confd config load in case of ``--no-lvm-storage``.
690
- The confd/query functionality is now mentioned in the security
691
  documentation.
692

    
693

    
694
Version 2.7.0 beta1
695
-------------------
696

    
697
*(Released Wed, 6 Feb 2013)*
698

    
699
This was the first beta release of the 2.7 series. All important changes
700
are listed in the latest 2.7 entry.
701

    
702

    
703
Version 2.6.2
704
-------------
705

    
706
*(Released Fri, 21 Dec 2012)*
707

    
708
Important behaviour change: hbal won't rebalance anymore instances which
709
have the ``auto_balance`` attribute set to false. This was the intention
710
all along, but until now it only skipped those from the N+1 memory
711
reservation (DRBD-specific).
712

    
713
A significant number of bug fixes in this release:
714

    
715
- Fixed disk adoption interaction with ipolicy checks.
716
- Fixed networking issues when instances are started, stopped or
717
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
718
- Fixed the warning in cluster verify for shared storage instances not
719
  being redundant.
720
- Fixed removal of storage directory on shared file storage (issue 262).
721
- Fixed validation of LVM volume group name in OpClusterSetParams
722
  (``gnt-cluster modify``) (issue 285).
723
- Fixed runtime memory increases (``gnt-instance modify -m``).
724
- Fixed live migration under Xen's ``xl`` mode.
725
- Fixed ``gnt-instance console`` with ``xl``.
726
- Fixed building with newer Haskell compiler/libraries.
727
- Fixed PID file writing in Haskell daemons (confd); this prevents
728
  restart issues if confd was launched manually (outside of
729
  ``daemon-util``) while another copy of it was running
730
- Fixed a type error when doing live migrations with KVM (issue 297) and
731
  the error messages for failing migrations have been improved.
732
- Fixed opcode validation for the out-of-band commands (``gnt-node
733
  power``).
734
- Fixed a type error when unsetting OS hypervisor parameters (issue
735
  311); now it's possible to unset all OS-specific hypervisor
736
  parameters.
737
- Fixed the ``dry-run`` mode for many operations: verification of
738
  results was over-zealous but didn't take into account the ``dry-run``
739
  operation, resulting in "wrong" failures.
740
- Fixed bash completion in ``gnt-job list`` when the job queue has
741
  hundreds of entries; especially with older ``bash`` versions, this
742
  results in significant CPU usage.
743

    
744
And lastly, a few other improvements have been made:
745

    
746
- Added option to force master-failover without voting (issue 282).
747
- Clarified error message on lock conflict (issue 287).
748
- Logging of newly submitted jobs has been improved (issue 290).
749
- Hostname checks have been made uniform between instance rename and
750
  create (issue 291).
751
- The ``--submit`` option is now supported by ``gnt-debug delay``.
752
- Shutting down the master daemon by sending SIGTERM now stops it from
753
  processing jobs waiting for locks; instead, those jobs will be started
754
  once again after the master daemon is started the next time (issue
755
  296).
756
- Support for Xen's ``xl`` program has been improved (besides the fixes
757
  above).
758
- Reduced logging noise in the Haskell confd daemon (only show one log
759
  entry for each config reload, instead of two).
760
- Several man page updates and typo fixes.
761

    
762

    
763
Version 2.6.1
764
-------------
765

    
766
*(Released Fri, 12 Oct 2012)*
767

    
768
A small bugfix release. Among the bugs fixed:
769

    
770
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
771
  made the command unusable.
772
- Commands that issue many jobs don't fail anymore just because some jobs
773
  take so long that other jobs are archived.
774
- Failures during ``gnt-instance reinstall`` are reflected by the exit
775
  status.
776
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
777
  DRBD is enabled.
778
- When ``always_failover`` is set, ``--allow-failover`` is not required
779
  in migrate commands anymore.
780
- ``bash_completion`` works even if extglob is disabled.
781
- Fixed bug with locks that made failover for RDB-based instances fail.
782
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
783
  a random node instead of one based on the allocator metric.
784
- Support for newer versions of pylint and pep8.
785
- Hail doesn't fail anymore when trying to add an instance of type
786
  ``file``, ``sharedfile`` or ``rbd``.
787
- Added new Makefile target to rebuild the whole distribution, so that
788
  all files are included.
789

    
790

    
791
Version 2.6.0
792
-------------
793

    
794
*(Released Fri, 27 Jul 2012)*
795

    
796

    
797
.. attention:: The ``LUXI`` protocol has been made more consistent
798
   regarding its handling of command arguments. This, however, leads to
799
   incompatibility issues with previous versions. Please ensure that you
800
   restart Ganeti daemons soon after the upgrade, otherwise most
801
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
802
   pausing/resuming the watcher, cancelling and archiving jobs, querying
803
   the cluster configuration) will fail.
804

    
805

    
806
New features
807
~~~~~~~~~~~~
808

    
809
Instance run status
810
+++++++++++++++++++
811

    
812
The current ``admin_up`` field, which used to denote whether an instance
813
should be running or not, has been removed. Instead, ``admin_state`` is
814
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
815

    
816
The rational behind this is that an instance being “down” can have
817
different meanings:
818

    
819
- it could be down during a reboot
820
- it could be temporarily be down for a reinstall
821
- or it could be down because it is deprecated and kept just for its
822
  disk
823

    
824
The previous Boolean state was making it difficult to do capacity
825
calculations: should Ganeti reserve memory for a down instance? Now, the
826
tri-state field makes it clear:
827

    
828
- in ``up`` and ``down`` state, all resources are reserved for the
829
  instance, and it can be at any time brought up if it is down
830
- in ``offline`` state, only disk space is reserved for it, but not
831
  memory or CPUs
832

    
833
The field can have an extra use: since the transition between ``up`` and
834
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
835
transition between ``offline`` and ``down`` is done via ``gnt-instance
836
modify``, it is possible to given different rights to users. For
837
example, owners of an instance could be allowed to start/stop it, but
838
not transition it out of the offline state.
839

    
840
Instance policies and specs
841
+++++++++++++++++++++++++++
842

    
843
In previous Ganeti versions, an instance creation request was not
844
limited on the minimum size and on the maximum size just by the cluster
845
resources. As such, any policy could be implemented only in third-party
846
clients (RAPI clients, or shell wrappers over ``gnt-*``
847
tools). Furthermore, calculating cluster capacity via ``hspace`` again
848
required external input with regards to instance sizes.
849

    
850
In order to improve these workflows and to allow for example better
851
per-node group differentiation, we introduced instance specs, which
852
allow declaring:
853

    
854
- minimum instance disk size, disk count, memory size, cpu count
855
- maximum values for the above metrics
856
- and “standard” values (used in ``hspace`` to calculate the standard
857
  sized instances)
858

    
859
The minimum/maximum values can be also customised at node-group level,
860
for example allowing more powerful hardware to support bigger instance
861
memory sizes.
862

    
863
Beside the instance specs, there are a few other settings belonging to
864
the instance policy framework. It is possible now to customise, per
865
cluster and node-group:
866

    
867
- the list of allowed disk templates
868
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
869
- the maximum ratio of instance to spindles (see below for more
870
  information) for local storage
871

    
872
All these together should allow all tools that talk to Ganeti to know
873
what are the ranges of allowed values for instances and the
874
over-subscription that is allowed.
875

    
876
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
877
instance configuration, and the physical CPU configuration from the
878
node. For the spindle ratios however, we didn't track before these
879
values, so new parameters have been added:
880

    
881
- a new node parameter ``spindle_count``, defaults to 1, customisable at
882
  node group or node level
883
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
884

    
885
Note that spindles in this context doesn't need to mean actual
886
mechanical hard-drives; it's just a relative number for both the node
887
I/O capacity and instance I/O consumption.
888

    
889
Instance migration behaviour
890
++++++++++++++++++++++++++++
891

    
892
While live-migration is in general desirable over failover, it is
893
possible that for some workloads it is actually worse, due to the
894
variable time of the “suspend” phase during live migration.
895

    
896
To allow the tools to work consistently over such instances (without
897
having to hard-code instance names), a new backend parameter
898
``always_failover`` has been added to control the migration/failover
899
behaviour. When set to True, all migration requests for an instance will
900
instead fall-back to failover.
901

    
902
Instance memory ballooning
903
++++++++++++++++++++++++++
904

    
905
Initial support for memory ballooning has been added. The memory for an
906
instance is no longer fixed (backend parameter ``memory``), but instead
907
can vary between minimum and maximum values (backend parameters
908
``minmem`` and ``maxmem``). Currently we only change an instance's
909
memory when:
910

    
911
- live migrating or failing over and instance and the target node
912
  doesn't have enough memory
913
- user requests changing the memory via ``gnt-instance modify
914
  --runtime-memory``
915

    
916
Instance CPU pinning
917
++++++++++++++++++++
918

    
919
In order to control the use of specific CPUs by instance, support for
920
controlling CPU pinning has been added for the Xen, HVM and LXC
921
hypervisors. This is controlled by a new hypervisor parameter
922
``cpu_mask``; details about possible values for this are in the
923
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
924
VCPU-to-CPU mapping) form will work well only when all nodes in your
925
cluster have the same amount of CPUs.
926

    
927
Disk parameters
928
+++++++++++++++
929

    
930
Another area in which Ganeti was not customisable were the parameters
931
used for storage configuration, e.g. how many stripes to use for LVM,
932
DRBD resync configuration, etc.
933

    
934
To improve this area, we've added disks parameters, which are
935
customisable at cluster and node group level, and which allow to
936
specify various parameters for disks (DRBD has the most parameters
937
currently), for example:
938

    
939
- DRBD resync algorithm and parameters (e.g. speed)
940
- the default VG for meta-data volumes for DRBD
941
- number of stripes for LVM (plain disk template)
942
- the RBD pool
943

    
944
These parameters can be modified via ``gnt-cluster modify -D …`` and
945
``gnt-group modify -D …``, and are used at either instance creation (in
946
case of LVM stripes, for example) or at disk “activation” time
947
(e.g. resync speed).
948

    
949
Rados block device support
950
++++++++++++++++++++++++++
951

    
952
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
953
denoted by the ``rbd`` disk template type. This is considered
954
experimental, feedback is welcome. For details on configuring it, see
955
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
956

    
957
Master IP setup
958
+++++++++++++++
959

    
960
The existing master IP functionality works well only in simple setups (a
961
single network shared by all nodes); however, if nodes belong to
962
different networks, then the ``/32`` setup and lack of routing
963
information is not enough.
964

    
965
To allow the master IP to function well in more complex cases, the
966
system was reworked as follows:
967

    
968
- a master IP netmask setting has been added
969
- the master IP activation/turn-down code was moved from the node daemon
970
  to a separate script
971
- whether to run the Ganeti-supplied master IP script or a user-supplied
972
  on is a ``gnt-cluster init`` setting
973

    
974
Details about the location of the standard and custom setup scripts are
975
in the man page :manpage:`gnt-cluster(8)`; for information about the
976
setup script protocol, look at the Ganeti-supplied script.
977

    
978
SPICE support
979
+++++++++++++
980

    
981
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
982
improved.
983

    
984
It is now possible to use TLS-protected connections, and when renewing
985
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
986
it is now possible to specify spice or spice CA certificates. Also, it
987
is possible to configure a password for SPICE sessions via the
988
hypervisor parameter ``spice_password_file``.
989

    
990
There are also new parameters to control the compression and streaming
991
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
992
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
993
for the spice parameters.
994

    
995
Lastly, it is now possible to see the SPICE connection information via
996
``gnt-instance console``.
997

    
998
OVF converter
999
+++++++++++++
1000

    
1001
A new tool (``tools/ovfconverter``) has been added that supports
1002
conversion between Ganeti and the `Open Virtualization Format
1003
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1004
from).
1005

    
1006
This relies on the ``qemu-img`` tool to convert the disk formats, so the
1007
actual compatibility with other virtualization solutions depends on it.
1008

    
1009
Confd daemon changes
1010
++++++++++++++++++++
1011

    
1012
The configuration query daemon (``ganeti-confd``) is now optional, and
1013
has been rewritten in Haskell; whether to use the daemon at all, use the
1014
Python (default) or the Haskell version is selectable at configure time
1015
via the ``--enable-confd`` parameter, which can take one of the
1016
``haskell``, ``python`` or ``no`` values. If not used, disabling the
1017
daemon will result in a smaller footprint; for larger systems, we
1018
welcome feedback on the Haskell version which might become the default
1019
in future versions.
1020

    
1021
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1022
daemon running. The Python version doesn't implement the new call.
1023

    
1024

    
1025
User interface changes
1026
~~~~~~~~~~~~~~~~~~~~~~
1027

    
1028
We have replaced the ``--disks`` option of ``gnt-instance
1029
replace-disks`` with a more flexible ``--disk`` option, which allows
1030
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1031
disk size and mode can be changed upon recreation (via ``gnt-instance
1032
recreate-disks``, which accepts the same ``--disk`` option).
1033

    
1034
As many people are used to a ``show`` command, we have added that as an
1035
alias to ``info`` on all ``gnt-*`` commands.
1036

    
1037
The ``gnt-instance grow-disk`` command has a new mode in which it can
1038
accept the target size of the disk, instead of the delta; this can be
1039
more safe since two runs in absolute mode will be idempotent, and
1040
sometimes it's also easier to specify the desired size directly.
1041

    
1042
Also the handling of instances with regard to offline secondaries has
1043
been improved. Instance operations should not fail because one of it's
1044
secondary nodes is offline, even though it's safe to proceed.
1045

    
1046
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1047
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1048
minors to instance name.
1049

    
1050
API changes
1051
~~~~~~~~~~~
1052

    
1053
RAPI coverage has improved, with (for example) new resources for
1054
recreate-disks, node power-cycle, etc.
1055

    
1056
Compatibility
1057
~~~~~~~~~~~~~
1058

    
1059
There is partial support for ``xl`` in the Xen hypervisor; feedback is
1060
welcome.
1061

    
1062
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1063
whether to still support Python 2.4 or move to Python 2.6 as minimum
1064
required version.
1065

    
1066
Support for Fedora has been slightly improved; the provided example
1067
init.d script should work better on it and the INSTALL file should
1068
document the needed dependencies.
1069

    
1070
Internal changes
1071
~~~~~~~~~~~~~~~~
1072

    
1073
The deprecated ``QueryLocks`` LUXI request has been removed. Use
1074
``Query(what=QR_LOCK, ...)`` instead.
1075

    
1076
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1077
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1078
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1079
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1080
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1081

    
1082
RAPI client: ``CertificateError`` now derives from
1083
``GanetiApiError``. This should make it more easy to handle Ganeti
1084
errors.
1085

    
1086
Deprecation warnings due to PyCrypto/paramiko import in
1087
``tools/setup-ssh`` have been silenced, as usually they are safe; please
1088
make sure to run an up-to-date paramiko version, if you use this tool.
1089

    
1090
The QA scripts now depend on Python 2.5 or above (the main code base
1091
still works with Python 2.4).
1092

    
1093
The configuration file (``config.data``) is now written without
1094
indentation for performance reasons; if you want to edit it, it can be
1095
re-formatted via ``tools/fmtjson``.
1096

    
1097
A number of bugs has been fixed in the cluster merge tool.
1098

    
1099
``x509`` certification verification (used in import-export) has been
1100
changed to allow the same clock skew as permitted by the cluster
1101
verification. This will remove some rare but hard to diagnose errors in
1102
import-export.
1103

    
1104

    
1105
Version 2.6.0 rc4
1106
-----------------
1107

    
1108
*(Released Thu, 19 Jul 2012)*
1109

    
1110
Very few changes from rc4 to the final release, only bugfixes:
1111

    
1112
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
1113
  instance, fix CDROM booting for KVM instances)
1114
- fixed node group modification of node parameters
1115
- fixed issue in LUClusterVerifyGroup with multi-group clusters
1116
- fixed generation of bash completion to ensure a stable ordering
1117
- fixed a few typos
1118

    
1119

    
1120
Version 2.6.0 rc3
1121
-----------------
1122

    
1123
*(Released Fri, 13 Jul 2012)*
1124

    
1125
Third release candidate for 2.6. The following changes were done from
1126
rc3 to rc4:
1127

    
1128
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1129
- Fixed an inconsistency in the LUXI protocol with the provided
1130
  arguments (NOT backwards compatible)
1131
- Fixed a bug with node groups ipolicy where ``min`` was greater than
1132
  the cluster ``std`` value
1133
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1134
  easier instance debugging on nodes (requires ``hconfd`` to work)
1135

    
1136

    
1137
Version 2.6.0 rc2
1138
-----------------
1139

    
1140
*(Released Tue, 03 Jul 2012)*
1141

    
1142
Second release candidate for 2.6. The following changes were done from
1143
rc2 to rc3:
1144

    
1145
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1146
  master candidates
1147
- Fixed a RAPI regression on missing beparams/memory
1148
- Fixed redistribution of files on offline nodes
1149
- Added possibility to run activate-disks even though secondaries are
1150
  offline. With this change it relaxes also the strictness on some other
1151
  commands which use activate disks internally:
1152
  * ``gnt-instance start|reboot|rename|backup|export``
1153
- Made it possible to remove safely an instance if its secondaries are
1154
  offline
1155
- Made it possible to reinstall even though secondaries are offline
1156

    
1157

    
1158
Version 2.6.0 rc1
1159
-----------------
1160

    
1161
*(Released Mon, 25 Jun 2012)*
1162

    
1163
First release candidate for 2.6. The following changes were done from
1164
rc1 to rc2:
1165

    
1166
- Fixed bugs with disk parameters and ``rbd`` templates as well as
1167
  ``instance_os_add``
1168
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1169
  behaviour. It supports now the modify operation
1170
- ``hcheck`` implemented to analyze cluster health and possibility of
1171
  improving health by rebalance
1172
- ``hbal`` has been improved in dealing with split instances
1173

    
1174

    
1175
Version 2.6.0 beta2
1176
-------------------
1177

    
1178
*(Released Mon, 11 Jun 2012)*
1179

    
1180
Second beta release of 2.6. The following changes were done from beta2
1181
to rc1:
1182

    
1183
- Fixed ``daemon-util`` with non-root user models
1184
- Fixed creation of plain instances with ``--no-wait-for-sync``
1185
- Fix wrong iv_names when running ``cfgupgrade``
1186
- Export more information in RAPI group queries
1187
- Fixed bug when changing instance network interfaces
1188
- Extended burnin to do NIC changes
1189
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1190
- Changed default for DRBD barriers
1191
- Fixed DRBD error reporting for syncer rate
1192
- Verify the options on disk parameters
1193

    
1194
And of course various fixes to documentation and improved unittests and
1195
QA.
1196

    
1197

    
1198
Version 2.6.0 beta1
1199
-------------------
1200

    
1201
*(Released Wed, 23 May 2012)*
1202

    
1203
First beta release of 2.6. The following changes were done from beta1 to
1204
beta2:
1205

    
1206
- integrated patch for distributions without ``start-stop-daemon``
1207
- adapted example init.d script to work on Fedora
1208
- fixed log handling in Haskell daemons
1209
- adapted checks in the watcher for pycurl linked against libnss
1210
- add partial support for ``xl`` instead of ``xm`` for Xen
1211
- fixed a type issue in cluster verification
1212
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1213
  clusters)
1214

    
1215
Plus integrated fixes from the 2.5 branch:
1216

    
1217
- fixed ``kvm-ifup`` to use ``/bin/bash``
1218
- fixed parallel build failures
1219
- KVM live migration when using a custom keymap
1220

    
1221

    
1222
Version 2.5.2
1223
-------------
1224

    
1225
*(Released Tue, 24 Jul 2012)*
1226

    
1227
A small bugfix release, with no new features:
1228

    
1229
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1230
  different default shell (e.g. Debian, Ubuntu)
1231
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1232
  243 and Debian bug #650664)
1233
- fixed compatibility with KVM versions that don't support multiple boot
1234
  devices (fixes Issue 230 and Debian bug #624256)
1235

    
1236
Additionally, a few fixes were done to the build system (fixed parallel
1237
build failures) and to the unittests (fixed race condition in test for
1238
FileID functions, and the default enable/disable mode for QA test is now
1239
customisable).
1240

    
1241

    
1242
Version 2.5.1
1243
-------------
1244

    
1245
*(Released Fri, 11 May 2012)*
1246

    
1247
A small bugfix release.
1248

    
1249
The main issues solved are on the topic of compatibility with newer LVM
1250
releases:
1251

    
1252
- fixed parsing of ``lv_attr`` field
1253
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1254
  the ``--force`` flag is needed
1255

    
1256
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1257
to accept kernel 3.x too (was hardcoded to 2.6.*).
1258

    
1259
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1260
of hook results and that also made display of other errors suboptimal
1261
was fixed; the code behaves now like 2.4 and earlier.
1262

    
1263
Another change in 2.5, the cleanup of the OS scripts environment, is too
1264
aggressive: it removed even the ``PATH`` variable, which requires the OS
1265
scripts to *always* need to export it. Since this is a bit too strict,
1266
we now export a minimal PATH, the same that we export for hooks.
1267

    
1268
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1269
integrated into this release.
1270

    
1271
Finally, a few other miscellaneous changes were done (no new features,
1272
just small improvements):
1273

    
1274
- Fix ``gnt-group --help`` display
1275
- Fix hardcoded Xen kernel path
1276
- Fix grow-disk handling of invalid units
1277
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1278
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1279

    
1280

    
1281
Version 2.5.0
1282
-------------
1283

    
1284
*(Released Thu, 12 Apr 2012)*
1285

    
1286
Incompatible/important changes and bugfixes
1287
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1288

    
1289
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1290
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1291
  to match the underlying LUXI interface.
1292
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1293
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1294
  not maintain backwards-compatibility as the underlying operation
1295
  changed in an incompatible way. The RAPI client can talk to old
1296
  servers, but it needs to be told so as the return value changed.
1297
- When creating file-based instances via RAPI, the ``file_driver``
1298
  parameter no longer defaults to ``loop`` and must be specified.
1299
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1300
  ``link`` instead.
1301
- Support for the undocumented and deprecated RAPI instance creation
1302
  request format version 0 has been dropped. Use version 1, supported
1303
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1304
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1305
  documentation <install>`.
1306
- The "cluster-verify" hooks are now executed per group by the
1307
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1308
  if you just run ``gnt-cluster verify``, which generates one opcode per
1309
  group.
1310
- The environment as passed to the OS scripts is cleared, and thus no
1311
  environment variables defined in the node daemon's environment will be
1312
  inherited by the scripts.
1313
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1314
  deprecated.
1315
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1316
  support operations involving multiple node groups.
1317
- Offline nodes are ignored when failing over an instance.
1318
- Support for KVM version 1.0, which changed the version reporting format
1319
  from 3 to 2 digits.
1320
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1321
  removal.
1322
- ``Makefile`` is now compatible with Automake 1.11.2
1323
- Includes all bugfixes made in the 2.4 series
1324

    
1325
New features
1326
~~~~~~~~~~~~
1327

    
1328
- The ganeti-htools project has been merged into the ganeti-core source
1329
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1330
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1331
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1332
  GPT-style partition tables (requires `parted
1333
  <http://www.gnu.org/s/parted/>`_).
1334
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1335
- Allowed adding tags on instance creation.
1336
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1337
- Allow instances to be started in a paused state, enabling the user to
1338
  see the complete console output on boot using the console.
1339
- Added new hypervisor flag to control default reboot behaviour
1340
  (``reboot_behavior``).
1341
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1342
- Improved out-of-band management support:
1343

    
1344
  - Added ``gnt-node health`` command reporting the health status of
1345
    nodes.
1346
  - Added ``gnt-node power`` command to manage power status of nodes.
1347
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1348

    
1349
- Instance migration can fall back to failover if instance is not
1350
  running.
1351
- Filters can be used when listing nodes, instances, groups and locks;
1352
  see :manpage:`ganeti(7)` manpage.
1353
- Added post-execution status as variables to :doc:`hooks <hooks>`
1354
  environment.
1355
- Instance tags are exported/imported together with the instance.
1356
- When given an explicit job ID, ``gnt-job info`` will work for archived
1357
  jobs.
1358
- Jobs can define dependencies on other jobs (not yet supported via
1359
  RAPI or command line, but used by internal commands and usable via
1360
  LUXI).
1361

    
1362
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1363
    dependencies.
1364

    
1365
- Instance failover is now available as a RAPI resource
1366
  (``/2/instances/[instance_name]/failover``).
1367
- ``gnt-instance info`` defaults to static information if primary node
1368
  is offline.
1369
- Opcodes have a new ``comment`` attribute.
1370
- Added basic SPICE support to KVM hypervisor.
1371
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1372

    
1373
Node group improvements
1374
~~~~~~~~~~~~~~~~~~~~~~~
1375

    
1376
- ``gnt-cluster verify`` has been modified to check groups separately,
1377
  thereby improving performance.
1378
- Node group support has been added to ``gnt-cluster verify-disks``,
1379
  which now operates per node group.
1380
- Watcher has been changed to work better with node groups.
1381

    
1382
  - One process and state file per node group.
1383
  - Slow watcher in one group doesn't block other group's watcher.
1384

    
1385
- Added new command, ``gnt-group evacuate``, to move all instances in a
1386
  node group to other groups.
1387
- Added ``gnt-instance change-group`` to move an instance to another
1388
  node group.
1389
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1390
  per-group operations.
1391
- Node groups can be tagged.
1392
- Some operations switch from an exclusive to a shared lock as soon as
1393
  possible.
1394
- Instance's primary and secondary nodes' groups are now available as
1395
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1396
  and ``snodes.group.uuid``).
1397

    
1398
Misc
1399
~~~~
1400

    
1401
- Numerous updates to documentation and manpages.
1402

    
1403
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1404
    descriptions.
1405
  - Some opcode/job results are now also documented, see :doc:`RAPI
1406
    <rapi>`.
1407

    
1408
- A lockset's internal lock is now also visible in lock monitor.
1409
- Log messages from job queue workers now contain information about the
1410
  opcode they're processing.
1411
- ``gnt-instance console`` no longer requires the instance lock.
1412
- A short delay when waiting for job changes reduces the number of LUXI
1413
  requests significantly.
1414
- DRBD metadata volumes are overwritten with zeros during disk creation.
1415
- Out-of-band commands no longer acquire the cluster lock in exclusive
1416
  mode.
1417
- ``devel/upload`` now uses correct permissions for directories.
1418

    
1419

    
1420
Version 2.5.0 rc6
1421
-----------------
1422

    
1423
*(Released Fri, 23 Mar 2012)*
1424

    
1425
This was the sixth release candidate of the 2.5 series.
1426

    
1427

    
1428
Version 2.5.0 rc5
1429
-----------------
1430

    
1431
*(Released Mon, 9 Jan 2012)*
1432

    
1433
This was the fifth release candidate of the 2.5 series.
1434

    
1435

    
1436
Version 2.5.0 rc4
1437
-----------------
1438

    
1439
*(Released Thu, 27 Oct 2011)*
1440

    
1441
This was the fourth release candidate of the 2.5 series.
1442

    
1443

    
1444
Version 2.5.0 rc3
1445
-----------------
1446

    
1447
*(Released Wed, 26 Oct 2011)*
1448

    
1449
This was the third release candidate of the 2.5 series.
1450

    
1451

    
1452
Version 2.5.0 rc2
1453
-----------------
1454

    
1455
*(Released Tue, 18 Oct 2011)*
1456

    
1457
This was the second release candidate of the 2.5 series.
1458

    
1459

    
1460
Version 2.5.0 rc1
1461
-----------------
1462

    
1463
*(Released Tue, 4 Oct 2011)*
1464

    
1465
This was the first release candidate of the 2.5 series.
1466

    
1467

    
1468
Version 2.5.0 beta3
1469
-------------------
1470

    
1471
*(Released Wed, 31 Aug 2011)*
1472

    
1473
This was the third beta release of the 2.5 series.
1474

    
1475

    
1476
Version 2.5.0 beta2
1477
-------------------
1478

    
1479
*(Released Mon, 22 Aug 2011)*
1480

    
1481
This was the second beta release of the 2.5 series.
1482

    
1483

    
1484
Version 2.5.0 beta1
1485
-------------------
1486

    
1487
*(Released Fri, 12 Aug 2011)*
1488

    
1489
This was the first beta release of the 2.5 series.
1490

    
1491

    
1492
Version 2.4.5
1493
-------------
1494

    
1495
*(Released Thu, 27 Oct 2011)*
1496

    
1497
- Fixed bug when parsing command line parameter values ending in
1498
  backslash
1499
- Fixed assertion error after unclean master shutdown
1500
- Disable HTTP client pool for RPC, significantly reducing memory usage
1501
  of master daemon
1502
- Fixed queue archive creation with wrong permissions
1503

    
1504

    
1505
Version 2.4.4
1506
-------------
1507

    
1508
*(Released Tue, 23 Aug 2011)*
1509

    
1510
Small bug-fixes:
1511

    
1512
- Fixed documentation for importing with ``--src-dir`` option
1513
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1514
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1515

    
1516

    
1517
Version 2.4.3
1518
-------------
1519

    
1520
*(Released Fri, 5 Aug 2011)*
1521

    
1522
Many bug-fixes and a few small features:
1523

    
1524
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1525
  issues when you tried to add an instance with two MAC addresses in one
1526
  request
1527
- KVM: fixed per-instance stored UID value
1528
- KVM: configure bridged NICs at migration start
1529
- KVM: Fix a bug where instance will not start with never KVM versions
1530
  (>= 0.14)
1531
- Added OS search path to ``gnt-cluster info``
1532
- Fixed an issue with ``file_storage_dir`` where you were forced to
1533
  provide an absolute path, but the documentation states it is a
1534
  relative path, the documentation was right
1535
- Added a new parameter to instance stop/start called ``--no-remember``
1536
  that will make the state change to not be remembered
1537
- Implemented ``no_remember`` at RAPI level
1538
- Improved the documentation
1539
- Node evacuation: don't call IAllocator if node is already empty
1540
- Fixed bug in DRBD8 replace disks on current nodes
1541
- Fixed bug in recreate-disks for DRBD instances
1542
- Moved assertion checking locks in ``gnt-instance replace-disks``
1543
  causing it to abort with not owning the right locks for some situation
1544
- Job queue: Fixed potential race condition when cancelling queued jobs
1545
- Fixed off-by-one bug in job serial generation
1546
- ``gnt-node volumes``: Fix instance names
1547
- Fixed aliases in bash completion
1548
- Fixed a bug in reopening log files after being sent a SIGHUP
1549
- Added a flag to burnin to allow specifying VCPU count
1550
- Bugfixes to non-root Ganeti configuration
1551

    
1552

    
1553
Version 2.4.2
1554
-------------
1555

    
1556
*(Released Thu, 12 May 2011)*
1557

    
1558
Many bug-fixes and a few new small features:
1559

    
1560
- Fixed a bug related to log opening failures
1561
- Fixed a bug in instance listing with orphan instances
1562
- Fixed a bug which prevented resetting the cluster-level node parameter
1563
  ``oob_program`` to the default
1564
- Many fixes related to the ``cluster-merge`` tool
1565
- Fixed a race condition in the lock monitor, which caused failures
1566
  during (at least) creation of many instances in parallel
1567
- Improved output for gnt-job info
1568
- Removed the quiet flag on some ssh calls which prevented debugging
1569
  failures
1570
- Improved the N+1 failure messages in cluster verify by actually
1571
  showing the memory values (needed and available)
1572
- Increased lock attempt timeouts so that when executing long operations
1573
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1574
  too early and thus prevent the use of the 'fair' mechanism
1575
- Changed instance query data (``gnt-instance info``) to not acquire
1576
  locks unless needed, thus allowing its use on locked instance if only
1577
  static information is asked for
1578
- Improved behaviour with filesystems that do not support rename on an
1579
  opened file
1580
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1581
  kept locks on all nodes during the wipe, which is unneeded
1582
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1583
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1584
  10GiB) which caused the wipe to fail right at the end in some cases
1585
- Fixed master IP activation when doing master failover with no-voting
1586
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1587
  the master node itself
1588
- Fixed potential data-loss in under disk full conditions, where Ganeti
1589
  wouldn't check correctly the return code and would consider
1590
  partially-written files 'correct'
1591
- Fixed bug related to multiple VGs and DRBD disk replacing
1592
- Added new disk parameter ``metavg`` that allows placement of the meta
1593
  device for DRBD in a different volume group
1594
- Fixed error handling in the node daemon when the system libc doesn't
1595
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1596
- Fixed lock release during replace-disks, which kept cluster-wide locks
1597
  when doing disk replaces with an iallocator script
1598
- Added check for missing bridges in cluster verify
1599
- Handle EPIPE errors while writing to the terminal better, so that
1600
  piping the output to e.g. ``less`` doesn't cause a backtrace
1601
- Fixed rare case where a ^C during Luxi calls could have been
1602
  interpreted as server errors, instead of simply terminating
1603
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1604
  assign-nodes``)
1605
- Added a few more parameters to the KVM hypervisor, allowing a second
1606
  CDROM, custom disk type for CDROMs and a floppy image
1607
- Removed redundant message in instance rename when the name is given
1608
  already as a FQDN
1609
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1610
  disks on new nodes, allowing recreation when the original instance
1611
  nodes are completely gone
1612
- Added option when converting disk templates to DRBD to skip waiting
1613
  for the resync, in order to make the instance available sooner
1614
- Added two new variables to the OS scripts environment (containing the
1615
  instance's nodes)
1616
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1617
  to allow use of ``pvgrub`` as bootloader
1618
- Changed the instance memory modifications to only check out-of-memory
1619
  conditions on memory increases, and turned the secondary node warnings
1620
  into errors (they can still be overridden via ``--force``)
1621
- Fixed the handling of a corner case when the Python installation gets
1622
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1623
  to execute a command that doesn't exist
1624
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1625
  node of the instance returned failures during instance shutdown; this
1626
  adds the option ``--ignore-consistency`` to gnt-instance move
1627

    
1628
And as usual, various improvements to the error messages, documentation
1629
and man pages.
1630

    
1631

    
1632
Version 2.4.1
1633
-------------
1634

    
1635
*(Released Wed, 09 Mar 2011)*
1636

    
1637
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1638
the RAPI users file if run twice (even with ``--dry-run``).
1639

    
1640
The release fixes that bug (nothing else changed).
1641

    
1642

    
1643
Version 2.4.0
1644
-------------
1645

    
1646
*(Released Mon, 07 Mar 2011)*
1647

    
1648
Final 2.4.0 release. Just a few small fixes:
1649

    
1650
- Fixed RAPI node evacuate
1651
- Fixed the kvm-ifup script
1652
- Fixed internal error handling for special job cases
1653
- Updated man page to specify the escaping feature for options
1654

    
1655

    
1656
Version 2.4.0 rc3
1657
-----------------
1658

    
1659
*(Released Mon, 28 Feb 2011)*
1660

    
1661
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1662
that this feature wiped the disks of the wrong instance, leading to loss
1663
of data.
1664

    
1665
Other changes:
1666

    
1667
- Fixed title of query field containing instance name
1668
- Expanded the glossary in the documentation
1669
- Fixed one unittest (internal issue)
1670

    
1671

    
1672
Version 2.4.0 rc2
1673
-----------------
1674

    
1675
*(Released Mon, 21 Feb 2011)*
1676

    
1677
A number of bug fixes plus just a couple functionality changes.
1678

    
1679
On the user-visible side, the ``gnt-* list`` command output has changed
1680
with respect to "special" field states. The current rc1 style of display
1681
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1682
the default output mode special fields states are displayed as follows:
1683

    
1684
- Offline resource: ``*``
1685
- Unavailable/not applicable: ``-``
1686
- Data missing (RPC failure): ``?``
1687
- Unknown field: ``??``
1688

    
1689
Another user-visible change is the addition of ``--force-join`` to
1690
``gnt-node add``.
1691

    
1692
As for bug fixes:
1693

    
1694
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1695
- Fixed regression in RAPI/instance reinstall where all parameters were
1696
  required (instead of optional)
1697
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1698
- Fixed iallocator usage (offline nodes were not considered offline)
1699
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1700
- Fixed hypervisor and OS parameter validation with respect to
1701
  non-vm_capable nodes
1702
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1703
  cosmetic)
1704
- Fixed ``tools/listrunner`` with respect to agent-based usage
1705

    
1706

    
1707
Version 2.4.0 rc1
1708
-----------------
1709

    
1710
*(Released Fri,  4 Feb 2011)*
1711

    
1712
Many changes and fixes since the beta1 release. While there were some
1713
internal changes, the code has been mostly stabilised for the RC
1714
release.
1715

    
1716
Note: the dumb allocator was removed in this release, as it was not kept
1717
up-to-date with the IAllocator protocol changes. It is recommended to
1718
use the ``hail`` command from the ganeti-htools package.
1719

    
1720
Note: the 2.4 and up versions of Ganeti are not compatible with the
1721
0.2.x branch of ganeti-htools. You need to upgrade to
1722
ganeti-htools-0.3.0 (or later).
1723

    
1724
Regressions fixed from 2.3
1725
~~~~~~~~~~~~~~~~~~~~~~~~~~
1726

    
1727
- Fixed the ``gnt-cluster verify-disks`` command
1728
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1729
  serially on nodes)
1730
- Fixed disk adoption breakage
1731
- Fixed wrong headers in instance listing for field aliases
1732

    
1733
Other bugs fixed
1734
~~~~~~~~~~~~~~~~
1735

    
1736
- Fixed corner case in KVM handling of NICs
1737
- Fixed many cases of wrong handling of non-vm_capable nodes
1738
- Fixed a bug where a missing instance symlink was not possible to
1739
  recreate with any ``gnt-*`` command (now ``gnt-instance
1740
  activate-disks`` does it)
1741
- Fixed the volume group name as reported by ``gnt-cluster
1742
  verify-disks``
1743
- Increased timeouts for the import-export code, hopefully leading to
1744
  fewer aborts due network or instance timeouts
1745
- Fixed bug in ``gnt-node list-storage``
1746
- Fixed bug where not all daemons were started on cluster
1747
  initialisation, but only at the first watcher run
1748
- Fixed many bugs in the OOB implementation
1749
- Fixed watcher behaviour in presence of instances with offline
1750
  secondaries
1751
- Fixed instance list output for instances running on the wrong node
1752
- a few fixes to the cluster-merge tool, but it still cannot merge
1753
  multi-node groups (currently it is not recommended to use this tool)
1754

    
1755

    
1756
Improvements
1757
~~~~~~~~~~~~
1758

    
1759
- Improved network configuration for the KVM hypervisor
1760
- Added e1000 as a supported NIC for Xen-HVM
1761
- Improved the lvmstrap tool to also be able to use partitions, as
1762
  opposed to full disks
1763
- Improved speed of disk wiping (the cluster parameter
1764
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1765
  of instance creations
1766
- Added documentation for the OS parameters
1767
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1768
  hypervisor is not responding
1769
- Added display of blacklisted and hidden OS information in
1770
  ``gnt-cluster info``
1771
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1772
  NIC and node parameters, which might create problems with currently
1773
  invalid (but undetected) configuration files, but prevents validation
1774
  failures when unrelated parameters are modified
1775
- Changed cluster initialisation to wait for the master daemon to become
1776
  available
1777
- Expanded the RAPI interface:
1778

    
1779
  - Added config redistribution resource
1780
  - Added activation/deactivation of instance disks
1781
  - Added export of console information
1782

    
1783
- Implemented log file reopening on SIGHUP, which allows using
1784
  logrotate(8) for the Ganeti log files
1785
- Added a basic OOB helper script as an example
1786

    
1787

    
1788
Version 2.4.0 beta1
1789
-------------------
1790

    
1791
*(Released Fri, 14 Jan 2011)*
1792

    
1793
User-visible
1794
~~~~~~~~~~~~
1795

    
1796
- Fixed timezone issues when formatting timestamps
1797
- Added support for node groups, available via ``gnt-group`` and other
1798
  commands
1799
- Added out-of-band framework and management, see :doc:`design
1800
  document <design-oob>`
1801
- Removed support for roman numbers from ``gnt-node list`` and
1802
  ``gnt-instance list``.
1803
- Allowed modification of master network interface via ``gnt-cluster
1804
  modify --master-netdev``
1805
- Accept offline secondaries while shutting down instance disks
1806
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1807
- Added support for multiple LVM volume groups
1808
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1809
  requested
1810
- Added commands to list available fields:
1811

    
1812
  - ``gnt-node list-fields``
1813
  - ``gnt-group list-fields``
1814
  - ``gnt-instance list-fields``
1815

    
1816
- Updated documentation and man pages
1817

    
1818
Integration
1819
~~~~~~~~~~~
1820

    
1821
- Moved ``rapi_users`` file into separate directory, now named
1822
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1823
  symlink
1824
- Added new tool for running commands on many machines,
1825
  ``tools/ganeti-listrunner``
1826
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1827
  improving the ``gnt-instance console`` output
1828
- Allowed customisation of disk index separator at ``configure`` time
1829
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1830
- Added support for non-partitioned md disks in ``lvmstrap``
1831
- Added script to gracefully power off KVM instances
1832
- Split ``utils`` module into smaller parts
1833
- Changed query operations to return more detailed information, e.g.
1834
  whether an information is unavailable due to an offline node. To use
1835
  this new functionality, the LUXI call ``Query`` must be used. Field
1836
  information is now stored by the master daemon and can be retrieved
1837
  using ``QueryFields``. Instances, nodes and groups can also be queried
1838
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1839
  exposed via RAPI). The following commands make use of this
1840
  infrastructure change:
1841

    
1842
  - ``gnt-group list``
1843
  - ``gnt-group list-fields``
1844
  - ``gnt-node list``
1845
  - ``gnt-node list-fields``
1846
  - ``gnt-instance list``
1847
  - ``gnt-instance list-fields``
1848
  - ``gnt-debug locks``
1849

    
1850
Remote API
1851
~~~~~~~~~~
1852

    
1853
- New RAPI resources (see :doc:`rapi`):
1854

    
1855
  - ``/2/modify``
1856
  - ``/2/groups``
1857
  - ``/2/groups/[group_name]``
1858
  - ``/2/groups/[group_name]/assign-nodes``
1859
  - ``/2/groups/[group_name]/modify``
1860
  - ``/2/groups/[group_name]/rename``
1861
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1862

    
1863
- RAPI changes:
1864

    
1865
  - Implemented ``no_install`` for instance creation
1866
  - Implemented OS parameters for instance reinstallation, allowing
1867
    use of special settings on reinstallation (e.g. for preserving data)
1868

    
1869
Misc
1870
~~~~
1871

    
1872
- Added IPv6 support in import/export
1873
- Pause DRBD synchronization while wiping disks on instance creation
1874
- Updated unittests and QA scripts
1875
- Improved network parameters passed to KVM
1876
- Converted man pages from docbook to reStructuredText
1877

    
1878

    
1879
Version 2.3.1
1880
-------------
1881

    
1882
*(Released Mon, 20 Dec 2010)*
1883

    
1884
Released version 2.3.1~rc1 without any changes.
1885

    
1886

    
1887
Version 2.3.1 rc1
1888
-----------------
1889

    
1890
*(Released Wed, 1 Dec 2010)*
1891

    
1892
- impexpd: Disable OpenSSL compression in socat if possible (backport
1893
  from master, commit e90739d625b, see :doc:`installation guide
1894
  <install-quick>` for details)
1895
- Changed unittest coverage report to exclude test scripts
1896
- Added script to check version format
1897

    
1898

    
1899
Version 2.3.0
1900
-------------
1901

    
1902
*(Released Wed, 1 Dec 2010)*
1903

    
1904
Released version 2.3.0~rc1 without any changes.
1905

    
1906

    
1907
Version 2.3.0 rc1
1908
-----------------
1909

    
1910
*(Released Fri, 19 Nov 2010)*
1911

    
1912
A number of bugfixes and documentation updates:
1913

    
1914
- Update ganeti-os-interface documentation
1915
- Fixed a bug related to duplicate MACs or similar items which should be
1916
  unique
1917
- Fix breakage in OS state modify
1918
- Reinstall instance: disallow offline secondaries (fixes bug related to
1919
  OS changing but reinstall failing)
1920
- plus all the other fixes between 2.2.1 and 2.2.2
1921

    
1922

    
1923
Version 2.3.0 rc0
1924
-----------------
1925

    
1926
*(Released Tue, 2 Nov 2010)*
1927

    
1928
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1929
- Fixed master failover race with watcher
1930
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1931
  configuration
1932
- Accept previously stopped instance for export with instance removal
1933
- Simplify and extend the environment variables for instance OS scripts
1934
- Added new node flags, ``master_capable`` and ``vm_capable``
1935
- Added optional instance disk wiping prior during allocation. This is a
1936
  cluster-wide option and can be set/modified using
1937
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1938
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1939
  :doc:`install-quick`
1940
- Added a new watcher option (``--ignore-pause``)
1941
- Added option to ignore offline node on instance start/stop
1942
  (``--ignore-offline``)
1943
- Allow overriding OS parameters with ``gnt-instance reinstall``
1944
- Added ability to change node's secondary IP address using ``gnt-node
1945
  modify``
1946
- Implemented privilege separation for all daemons except
1947
  ``ganeti-noded``, see ``configure`` options
1948
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1949
  verify``
1950
- Implemented job priorities (see ``ganeti(7)`` manpage)
1951
- Ignore failures while shutting down instances during failover from
1952
  offline node
1953
- Exit daemon's bootstrap process only once daemon is ready
1954
- Export more information via ``LUInstanceQuery``/remote API
1955
- Improved documentation, QA and unittests
1956
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1957
  restart if the file was created or changed
1958
- Added LUXI protocol version sent with each request and response,
1959
  allowing detection of server/client mismatches
1960
- Moved the Python scripts among gnt-* and ganeti-* into modules
1961
- Moved all code related to setting up SSH to an external script,
1962
  ``setup-ssh``
1963
- Infrastructure changes for node group support in future versions
1964

    
1965

    
1966
Version 2.2.2
1967
-------------
1968

    
1969
*(Released Fri, 19 Nov 2010)*
1970

    
1971
A few small bugs fixed, and some improvements to the build system:
1972

    
1973
- Fix documentation regarding conversion to drbd
1974
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1975
  -B``)
1976
- Fix error handling in node modify with multiple changes
1977
- Allow remote imports without checked names
1978

    
1979

    
1980
Version 2.2.1
1981
-------------
1982

    
1983
*(Released Tue, 19 Oct 2010)*
1984

    
1985
- Disable SSL session ID cache in RPC client
1986

    
1987

    
1988
Version 2.2.1 rc1
1989
-----------------
1990

    
1991
*(Released Thu, 14 Oct 2010)*
1992

    
1993
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1994
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1995
  with GnuTLS
1996
- Fix problems with interaction between Curl and Python's HTTP server,
1997
  resulting in increased speed in many RPC calls
1998
- Improve our release script to prevent breakage with older aclocal and
1999
  Python 2.6
2000

    
2001

    
2002
Version 2.2.1 rc0
2003
-----------------
2004

    
2005
*(Released Thu, 7 Oct 2010)*
2006

    
2007
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2008
  value retrieved from master
2009
- Added support for blacklisted or hidden OS definitions
2010
- Added simple lock monitor (accessible via (``gnt-debug locks``)
2011
- Added support for -mem-path in KVM hypervisor abstraction layer
2012
- Allow overriding instance parameters in tool for inter-cluster
2013
  instance moves (``tools/move-instance``)
2014
- Improved opcode summaries (e.g. in ``gnt-job list``)
2015
- Improve consistency of OS listing by sorting it
2016
- Documentation updates
2017

    
2018

    
2019
Version 2.2.0.1
2020
---------------
2021

    
2022
*(Released Fri, 8 Oct 2010)*
2023

    
2024
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
2025

    
2026

    
2027
Version 2.2.0
2028
-------------
2029

    
2030
*(Released Mon, 4 Oct 2010)*
2031

    
2032
- Fixed regression in ``gnt-instance rename``
2033

    
2034

    
2035
Version 2.2.0 rc2
2036
-----------------
2037

    
2038
*(Released Wed, 22 Sep 2010)*
2039

    
2040
- Fixed OS_VARIANT variable for OS scripts
2041
- Fixed cluster tag operations via RAPI
2042
- Made ``setup-ssh`` exit with non-zero code if an error occurred
2043
- Disabled RAPI CA checks in watcher
2044

    
2045

    
2046
Version 2.2.0 rc1
2047
-----------------
2048

    
2049
*(Released Mon, 23 Aug 2010)*
2050

    
2051
- Support DRBD versions of the format "a.b.c.d"
2052
- Updated manpages
2053
- Re-introduce support for usage from multiple threads in RAPI client
2054
- Instance renames and modify via RAPI
2055
- Work around race condition between processing and archival in job
2056
  queue
2057
- Mark opcodes following failed one as failed, too
2058
- Job field ``lock_status`` was removed due to difficulties making it
2059
  work with the changed job queue in Ganeti 2.2; a better way to monitor
2060
  locks is expected for a later 2.2.x release
2061
- Fixed dry-run behaviour with many commands
2062
- Support ``ssh-agent`` again when adding nodes
2063
- Many additional bugfixes
2064

    
2065

    
2066
Version 2.2.0 rc0
2067
-----------------
2068

    
2069
*(Released Fri, 30 Jul 2010)*
2070

    
2071
Important change: the internal RPC mechanism between Ganeti nodes has
2072
changed from using a home-grown http library (based on the Python base
2073
libraries) to use the PycURL library. This requires that PycURL is
2074
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2075
against GnuTLS by default. cURL's support for GnuTLS had known issues
2076
before cURL 7.21.0 and we recommend using the latest cURL release or
2077
linking against OpenSSL. Most other distributions already link PycURL
2078
and cURL against OpenSSL. The command::
2079

    
2080
  python -c 'import pycurl; print pycurl.version'
2081

    
2082
can be used to determine the libraries PycURL and cURL are linked
2083
against.
2084

    
2085
Other significant changes:
2086

    
2087
- Rewrote much of the internals of the job queue, in order to achieve
2088
  better parallelism; this decouples job query operations from the job
2089
  processing, and it should allow much nicer behaviour of the master
2090
  daemon under load, and it also has uncovered some long-standing bugs
2091
  related to the job serialisation (now fixed)
2092
- Added a default iallocator setting to the cluster parameters,
2093
  eliminating the need to always pass nodes or an iallocator for
2094
  operations that require selection of new node(s)
2095
- Added experimental support for the LXC virtualization method
2096
- Added support for OS parameters, which allows the installation of
2097
  instances to pass parameter to OS scripts in order to customise the
2098
  instance
2099
- Added a hypervisor parameter controlling the migration type (live or
2100
  non-live), since hypervisors have various levels of reliability; this
2101
  has renamed the 'live' parameter to 'mode'
2102
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
2103
  logical volumes, meaning that cluster verify will ignore them and not
2104
  flag their presence as errors
2105
- The watcher will now reset the error count for failed instances after
2106
  8 hours, thus allowing self-healing if the problem that caused the
2107
  instances to be down/fail to start has cleared in the meantime
2108
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2109
  check for, and warn, if the drbd module parameter ``usermode_helper``
2110
  is not consistent with the cluster-wide setting; this is needed to
2111
  make diagnose easier of failed drbd creations
2112
- Started adding base IPv6 support, but this is not yet
2113
  enabled/available for use
2114
- Rename operations (cluster, instance) will now return the new name,
2115
  which is especially useful if a short name was passed in
2116
- Added support for instance migration in RAPI
2117
- Added a tool to pre-configure nodes for the SSH setup, before joining
2118
  them to the cluster; this will allow in the future a simplified model
2119
  for node joining (but not yet fully enabled in 2.2); this needs the
2120
  paramiko python library
2121
- Fixed handling of name-resolving errors
2122
- Fixed consistency of job results on the error path
2123
- Fixed master-failover race condition when executed multiple times in
2124
  sequence
2125
- Fixed many bugs related to the job queue (mostly introduced during the
2126
  2.2 development cycle, so not all are impacting 2.1)
2127
- Fixed instance migration with missing disk symlinks
2128
- Fixed handling of unknown jobs in ``gnt-job archive``
2129
- And many other small fixes/improvements
2130

    
2131
Internal changes:
2132

    
2133
- Enhanced both the unittest and the QA coverage
2134
- Switched the opcode validation to a generic model, and extended the
2135
  validation to all opcode parameters
2136
- Changed more parts of the code that write shell scripts to use the
2137
  same class for this
2138
- Switched the master daemon to use the asyncore library for the Luxi
2139
  server endpoint
2140

    
2141

    
2142
Version 2.2.0 beta0
2143
-------------------
2144

    
2145
*(Released Thu, 17 Jun 2010)*
2146

    
2147
- Added tool (``move-instance``) and infrastructure to move instances
2148
  between separate clusters (see :doc:`separate documentation
2149
  <move-instance>` and :doc:`design document <design-2.2>`)
2150
- Added per-request RPC timeout
2151
- RAPI now requires a Content-Type header for requests with a body (e.g.
2152
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2153
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
2154
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2155
  reachable
2156
- Implemented initial support for running Ganeti daemons as separate
2157
  users, see configure-time flags ``--with-user-prefix`` and
2158
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2159
  time)
2160
- Instances can be removed after export (``gnt-backup export
2161
  --remove-instance``)
2162
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2163
  key (instead of 1024 bit)
2164
- Added new cluster configuration file for cluster domain secret
2165
- Import/export now use SSL instead of SSH
2166
- Added support for showing estimated time when exporting an instance,
2167
  see the ``ganeti-os-interface(7)`` manpage and look for
2168
  ``EXP_SIZE_FD``
2169

    
2170

    
2171
Version 2.1.8
2172
-------------
2173

    
2174
*(Released Tue, 16 Nov 2010)*
2175

    
2176
Some more bugfixes. Unless critical bugs occur, this will be the last
2177
2.1 release:
2178

    
2179
- Fix case of MAC special-values
2180
- Fix mac checker regex
2181
- backend: Fix typo causing "out of range" error
2182
- Add missing --units in gnt-instance list man page
2183

    
2184

    
2185
Version 2.1.7
2186
-------------
2187

    
2188
*(Released Tue, 24 Aug 2010)*
2189

    
2190
Bugfixes only:
2191
  - Don't ignore secondary node silently on non-mirrored disk templates
2192
    (issue 113)
2193
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2194
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2195
  - Properly document the usb_mouse parameter
2196
  - Fix path in ganeti-rapi(8) (issue 116)
2197
  - Adjust error message when the ganeti user's .ssh directory is
2198
    missing
2199
  - Add same-node-check when changing the disk template to drbd
2200

    
2201

    
2202
Version 2.1.6
2203
-------------
2204

    
2205
*(Released Fri, 16 Jul 2010)*
2206

    
2207
Bugfixes only:
2208
  - Add an option to only select some reboot types during qa/burnin.
2209
    (on some hypervisors consequent reboots are not supported)
2210
  - Fix infrequent race condition in master failover. Sometimes the old
2211
    master ip address would be still detected as up for a short time
2212
    after it was removed, causing failover to fail.
2213
  - Decrease mlockall warnings when the ctypes module is missing. On
2214
    Python 2.4 we support running even if no ctypes module is installed,
2215
    but we were too verbose about this issue.
2216
  - Fix building on old distributions, on which man doesn't have a
2217
    --warnings option.
2218
  - Fix RAPI not to ignore the MAC address on instance creation
2219
  - Implement the old instance creation format in the RAPI client.
2220

    
2221

    
2222
Version 2.1.5
2223
-------------
2224

    
2225
*(Released Thu, 01 Jul 2010)*
2226

    
2227
A small bugfix release:
2228
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2229
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2230
    a non-existing option
2231
  - Fix instance create: the --force-variant option was ignored
2232
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2233
  - Fix modify node storage with non-FQDN arguments
2234
  - Fix RAPI client to authenticate under Python 2.6 when used
2235
    for more than 5 requests needing authentication
2236
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2237
    when converting a non-shutdown drbd instance to plain
2238

    
2239

    
2240
Version 2.1.4
2241
-------------
2242

    
2243
*(Released Fri, 18 Jun 2010)*
2244

    
2245
A small bugfix release:
2246

    
2247
  - Fix live migration of KVM instances started with older Ganeti
2248
    versions which had fewer hypervisor parameters
2249
  - Fix gnt-instance grow-disk on down instances
2250
  - Fix an error-reporting bug during instance migration
2251
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2252
    silently ignoring broken ones
2253
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2254
    users
2255
  - Fix bug triggered by different API version os-es on different nodes
2256
  - Fix a bug in instance startup with custom hvparams: OS level
2257
    parameters would fail to be applied.
2258
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2259
    make it work completely well with OpenSSL)
2260
  - Fix handling of errors when resolving names from DNS
2261

    
2262

    
2263
Version 2.1.3
2264
-------------
2265

    
2266
*(Released Thu, 3 Jun 2010)*
2267

    
2268
A medium sized development cycle. Some new features, and some
2269
fixes/small improvements/cleanups.
2270

    
2271
Significant features
2272
~~~~~~~~~~~~~~~~~~~~
2273

    
2274
The node deamon now tries to mlock itself into memory, unless the
2275
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2276
its logs, and falls back to console logging. This allows emergency
2277
features such as ``gnt-node powercycle`` to work even in the event of a
2278
broken node disk (tested offlining the disk hosting the node's
2279
filesystem and dropping its memory caches; don't try this at home)
2280

    
2281
KVM: add vhost-net acceleration support. It can be tested with a new
2282
enough version of the kernel and of qemu-kvm.
2283

    
2284
KVM: Add instance chrooting feature. If you use privilege dropping for
2285
your VMs you can also now force them to chroot to an empty directory,
2286
before starting the emulated guest.
2287

    
2288
KVM: Add maximum migration bandwith and maximum downtime tweaking
2289
support (requires a new-enough version of qemu-kvm).
2290

    
2291
Cluster verify will now warn if the master node doesn't have the master
2292
ip configured on it.
2293

    
2294
Add a new (incompatible) instance creation request format to RAPI which
2295
supports all parameters (previously only a subset was supported, and it
2296
wasn't possible to extend the old format to accomodate all the new
2297
features. The old format is still supported, and a client can check for
2298
this feature, before using it, by checking for its presence in the
2299
``features`` RAPI resource.
2300

    
2301
Now with ancient latin support. Try it passing the ``--roman`` option to
2302
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2303
(requires the python-roman module to be installed, in order to work).
2304

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

    
2308
As usual many internal code refactorings, documentation updates, and
2309
such. Among others:
2310

    
2311
  - Lots of improvements and cleanups to the experimental Remote API
2312
    (RAPI) client library.
2313
  - A new unit test suite for the core daemon libraries.
2314
  - A fix to creating missing directories makes sure the umask is not
2315
    applied anymore. This enforces the same directory permissions
2316
    everywhere.
2317
  - Better handling terminating daemons with ctrl+c (used when running
2318
    them in debugging mode).
2319
  - Fix a race condition in live migrating a KVM instance, when stat()
2320
    on the old proc status file returned EINVAL, which is an unexpected
2321
    value.
2322
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2323
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2324

    
2325

    
2326
Version 2.1.2.1
2327
---------------
2328

    
2329
*(Released Fri, 7 May 2010)*
2330

    
2331
Fix a bug which prevented untagged KVM instances from starting.
2332

    
2333

    
2334
Version 2.1.2
2335
-------------
2336

    
2337
*(Released Fri, 7 May 2010)*
2338

    
2339
Another release with a long development cycle, during which many
2340
different features were added.
2341

    
2342
Significant features
2343
~~~~~~~~~~~~~~~~~~~~
2344

    
2345
The KVM hypervisor now can run the individual instances as non-root, to
2346
reduce the impact of a VM being hijacked due to bugs in the
2347
hypervisor. It is possible to run all instances as a single (non-root)
2348
user, to manually specify a user for each instance, or to dynamically
2349
allocate a user out of a cluster-wide pool to each instance, with the
2350
guarantee that no two instances will run under the same user ID on any
2351
given node.
2352

    
2353
An experimental RAPI client library, that can be used standalone
2354
(without the other Ganeti libraries), is provided in the source tree as
2355
``lib/rapi/client.py``. Note this client might change its interface in
2356
the future, as we iterate on its capabilities.
2357

    
2358
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2359
replace the cluster's certificates and crypto keys. This might help in
2360
case they have been compromised, or have simply expired.
2361

    
2362
A new disk option for instance creation has been added that allows one
2363
to "adopt" currently existing logical volumes, with data
2364
preservation. This should allow easier migration to Ganeti from
2365
unmanaged (or managed via other software) instances.
2366

    
2367
Another disk improvement is the possibility to convert between redundant
2368
(DRBD) and plain (LVM) disk configuration for an instance. This should
2369
allow better scalability (starting with one node and growing the
2370
cluster, or shrinking a two-node cluster to one node).
2371

    
2372
A new feature that could help with automated node failovers has been
2373
implemented: if a node sees itself as offline (by querying the master
2374
candidates), it will try to shutdown (hard) all instances and any active
2375
DRBD devices. This reduces the risk of duplicate instances if an
2376
external script automatically failovers the instances on such nodes. To
2377
enable this, the cluster parameter ``maintain_node_health`` should be
2378
enabled; in the future this option (per the name) will enable other
2379
automatic maintenance features.
2380

    
2381
Instance export/import now will reuse the original instance
2382
specifications for all parameters; that means exporting an instance,
2383
deleting it and the importing it back should give an almost identical
2384
instance. Note that the default import behaviour has changed from
2385
before, where it created only one NIC; now it recreates the original
2386
number of NICs.
2387

    
2388
Cluster verify has added a few new checks: SSL certificates validity,
2389
/etc/hosts consistency across the cluster, etc.
2390

    
2391
Other changes
2392
~~~~~~~~~~~~~
2393

    
2394
As usual, many internal changes were done, documentation fixes,
2395
etc. Among others:
2396

    
2397
- Fixed cluster initialization with disabled cluster storage (regression
2398
  introduced in 2.1.1)
2399
- File-based storage supports growing the disks
2400
- Fixed behaviour of node role changes
2401
- Fixed cluster verify for some corner cases, plus a general rewrite of
2402
  cluster verify to allow future extension with more checks
2403
- Fixed log spamming by watcher and node daemon (regression introduced
2404
  in 2.1.1)
2405
- Fixed possible validation issues when changing the list of enabled
2406
  hypervisors
2407
- Fixed cleanup of /etc/hosts during node removal
2408
- Fixed RAPI response for invalid methods
2409
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2410
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2411
  from ide disks, etc.)
2412
- Allow OS changes without re-installation (to record a changed OS
2413
  outside of Ganeti, or to allow OS renames)
2414
- Allow instance creation without OS installation (useful for example if
2415
  the OS will be installed manually, or restored from a backup not in
2416
  Ganeti format)
2417
- Implemented option to make cluster ``copyfile`` use the replication
2418
  network
2419
- Added list of enabled hypervisors to ssconf (possibly useful for
2420
  external scripts)
2421
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2422
  1.2 clusters
2423
- A partial form of node re-IP is possible via node readd, which now
2424
  allows changed node primary IP
2425
- Command line utilities now show an informational message if the job is
2426
  waiting for a lock
2427
- The logs of the master daemon now show the PID/UID/GID of the
2428
  connected client
2429

    
2430

    
2431
Version 2.1.1
2432
-------------
2433

    
2434
*(Released Fri, 12 Mar 2010)*
2435

    
2436
During the 2.1.0 long release candidate cycle, a lot of improvements and
2437
changes have accumulated with were released later as 2.1.1.
2438

    
2439
Major changes
2440
~~~~~~~~~~~~~
2441

    
2442
The node evacuate command (``gnt-node evacuate``) was significantly
2443
rewritten, and as such the IAllocator protocol was changed - a new
2444
request type has been added. This unfortunate change during a stable
2445
series is designed to improve performance of node evacuations; on
2446
clusters with more than about five nodes and which are well-balanced,
2447
evacuation should proceed in parallel for all instances of the node
2448
being evacuated. As such, any existing IAllocator scripts need to be
2449
updated, otherwise the above command will fail due to the unknown
2450
request. The provided "dumb" allocator has not been updated; but the
2451
ganeti-htools package supports the new protocol since version 0.2.4.
2452

    
2453
Another important change is increased validation of node and instance
2454
names. This might create problems in special cases, if invalid host
2455
names are being used.
2456

    
2457
Also, a new layer of hypervisor parameters has been added, that sits at
2458
OS level between the cluster defaults and the instance ones. This allows
2459
customisation of virtualization parameters depending on the installed
2460
OS. For example instances with OS 'X' may have a different KVM kernel
2461
(or any other parameter) than the cluster defaults. This is intended to
2462
help managing a multiple OSes on the same cluster, without manual
2463
modification of each instance's parameters.
2464

    
2465
A tool for merging clusters, ``cluster-merge``, has been added in the
2466
tools sub-directory.
2467

    
2468
Bug fixes
2469
~~~~~~~~~
2470

    
2471
- Improved the int/float conversions that should make the code more
2472
  robust in face of errors from the node daemons
2473
- Fixed the remove node code in case of internal configuration errors
2474
- Fixed the node daemon behaviour in face of inconsistent queue
2475
  directory (e.g. read-only file-system where we can't open the files
2476
  read-write, etc.)
2477
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2478
  now it either aborts cleanly or, if given the new "auto_promote"
2479
  parameter, will automatically promote other nodes as needed
2480
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2481
  completely prevent Ganeti from working
2482
- Fixed bug for instance export when not all disks were successfully
2483
  exported
2484
- Fixed behaviour of node add when the new node is slow in starting up
2485
  the node daemon
2486
- Fixed handling of signals in the LUXI client, which should improve
2487
  behaviour of command-line scripts
2488
- Added checks for invalid node/instance names in the configuration (now
2489
  flagged during cluster verify)
2490
- Fixed watcher behaviour for disk activation errors
2491
- Fixed two potentially endless loops in http library, which led to the
2492
  RAPI daemon hanging and consuming 100% CPU in some cases
2493
- Fixed bug in RAPI daemon related to hashed passwords
2494
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2495
  instances
2496
- Enhanced compatibility with non-Debian OSes, but not using absolute
2497
  path in some commands and allowing customisation of the ssh
2498
  configuration directory
2499
- Fixed possible future issue with new Python versions by abiding to the
2500
  proper use of ``__slots__`` attribute on classes
2501
- Added checks that should prevent directory traversal attacks
2502
- Many documentation fixes based on feedback from users
2503

    
2504
New features
2505
~~~~~~~~~~~~
2506

    
2507
- Added an "early_release" more for instance replace disks and node
2508
  evacuate, where we release locks earlier and thus allow higher
2509
  parallelism within the cluster
2510
- Added watcher hooks, intended to allow the watcher to restart other
2511
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2512
  course for any other purpose
2513
- Added a compile-time disable for DRBD barriers, to increase
2514
  performance if the administrator trusts the power supply or the
2515
  storage system to not lose writes
2516
- Added the option of using syslog for logging instead of, or in
2517
  addition to, Ganeti's own log files
2518
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2519
  can indeed boot from a paravirtual NIC
2520
- Added a generic debug level for many operations; while this is not
2521
  used widely yet, it allows one to pass the debug value all the way to
2522
  the OS scripts
2523
- Enhanced the hooks environment for instance moves (failovers,
2524
  migrations) where the primary/secondary nodes changed during the
2525
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2526
- Enhanced data validations for many user-supplied values; one important
2527
  item is the restrictions imposed on instance and node names, which
2528
  might reject some (invalid) host names
2529
- Add a configure-time option to disable file-based storage, if it's not
2530
  needed; this allows greater security separation between the master
2531
  node and the other nodes from the point of view of the inter-node RPC
2532
  protocol
2533
- Added user notification in interactive tools if job is waiting in the
2534
  job queue or trying to acquire locks
2535
- Added log messages when a job is waiting for locks
2536
- Added filtering by node tags in instance operations which admit
2537
  multiple instances (start, stop, reboot, reinstall)
2538
- Added a new tool for cluster mergers, ``cluster-merge``
2539
- Parameters from command line which are of the form ``a=b,c=d`` can now
2540
  use backslash escapes to pass in values which contain commas,
2541
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2542
  ``b,c``
2543
- For KVM, the instance name is the first parameter passed to KVM, so
2544
  that it's more visible in the process list
2545

    
2546

    
2547
Version 2.1.0
2548
-------------
2549

    
2550
*(Released Tue, 2 Mar 2010)*
2551

    
2552
Ganeti 2.1 brings many improvements with it. Major changes:
2553

    
2554
- Added infrastructure to ease automated disk repairs
2555
- Added new daemon to export configuration data in a cheaper way than
2556
  using the remote API
2557
- Instance NICs can now be routed instead of being associated with a
2558
  networking bridge
2559
- Improved job locking logic to reduce impact of jobs acquiring multiple
2560
  locks waiting for other long-running jobs
2561

    
2562
In-depth implementation details can be found in the Ganeti 2.1 design
2563
document.
2564

    
2565
Details
2566
~~~~~~~
2567

    
2568
- Added chroot hypervisor
2569
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2570
  ``device_model``)
2571
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2572
  ``bootloader_path`` and ``bootloader_args``)
2573
- Added the ``use_localtime`` option for the xen-hvm and kvm
2574
  hypervisors, and the default value for this has changed to false (in
2575
  2.0 xen-hvm always enabled it)
2576
- Added luxi call to submit multiple jobs in one go
2577
- Added cluster initialization option to not modify ``/etc/hosts``
2578
  file on nodes
2579
- Added network interface parameters
2580
- Added dry run mode to some LUs
2581
- Added RAPI resources:
2582

    
2583
  - ``/2/instances/[instance_name]/info``
2584
  - ``/2/instances/[instance_name]/replace-disks``
2585
  - ``/2/nodes/[node_name]/evacuate``
2586
  - ``/2/nodes/[node_name]/migrate``
2587
  - ``/2/nodes/[node_name]/role``
2588
  - ``/2/nodes/[node_name]/storage``
2589
  - ``/2/nodes/[node_name]/storage/modify``
2590
  - ``/2/nodes/[node_name]/storage/repair``
2591

    
2592
- Added OpCodes to evacuate or migrate all instances on a node
2593
- Added new command to list storage elements on nodes (``gnt-node
2594
  list-storage``) and modify them (``gnt-node modify-storage``)
2595
- Added new ssconf files with master candidate IP address
2596
  (``ssconf_master_candidates_ips``), node primary IP address
2597
  (``ssconf_node_primary_ips``) and node secondary IP address
2598
  (``ssconf_node_secondary_ips``)
2599
- Added ``ganeti-confd`` and a client library to query the Ganeti
2600
  configuration via UDP
2601
- Added ability to run hooks after cluster initialization and before
2602
  cluster destruction
2603
- Added automatic mode for disk replace (``gnt-instance replace-disks
2604
  --auto``)
2605
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2606
  after catastrophic data-loss
2607
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2608
  groups
2609
- Added ``gnt-instance move`` command to move instances
2610
- Added ``gnt-cluster watcher`` command to control watcher
2611
- Added ``gnt-node powercycle`` command to powercycle nodes
2612
- Added new job status field ``lock_status``
2613
- Added parseable error codes to cluster verification (``gnt-cluster
2614
  verify --error-codes``) and made output less verbose (use
2615
  ``--verbose`` to restore previous behaviour)
2616
- Added UUIDs to the main config entities (cluster, nodes, instances)
2617
- Added support for OS variants
2618
- Added support for hashed passwords in the Ganeti remote API users file
2619
  (``rapi_users``)
2620
- Added option to specify maximum timeout on instance shutdown
2621
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2622
- Added new helper script to start and stop Ganeti daemons
2623
  (``daemon-util``), with the intent to reduce the work necessary to
2624
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2625
  from one place
2626
- Added more unittests
2627
- Fixed critical bug in ganeti-masterd startup
2628
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2629
  now customisable at the cluster level for both the KVM and Xen
2630
  hypervisors using the new ``migration_port`` parameter
2631
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2632
  reinstalling an instance
2633
- Allowed ``@`` in tag names
2634
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2635
- Many documentation updates
2636
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2637
- ``gnt-instance reinstall`` can now reinstall multiple instances
2638
- Updated many command line parameters
2639
- Introduced new OS API version 15
2640
- No longer support a default hypervisor
2641
- Treat virtual LVs as inexistent
2642
- Improved job locking logic to reduce lock contention
2643
- Match instance and node names case insensitively
2644
- Reimplemented bash completion script to be more complete
2645
- Improved burnin
2646

    
2647

    
2648
Version 2.0.6
2649
-------------
2650

    
2651
*(Released Thu, 4 Feb 2010)*
2652

    
2653
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2654
- Fix a string formatting bug
2655
- Improve safety of the code in some error paths
2656
- Improve data validation in the master of values returned from nodes
2657

    
2658

    
2659
Version 2.0.5
2660
-------------
2661

    
2662
*(Released Thu, 17 Dec 2009)*
2663

    
2664
- Fix security issue due to missing validation of iallocator names; this
2665
  allows local and remote execution of arbitrary executables
2666
- Fix failure of gnt-node list during instance removal
2667
- Ship the RAPI documentation in the archive
2668

    
2669

    
2670
Version 2.0.4
2671
-------------
2672

    
2673
*(Released Wed, 30 Sep 2009)*
2674

    
2675
- Fixed many wrong messages
2676
- Fixed a few bugs related to the locking library
2677
- Fixed MAC checking at instance creation time
2678
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2679
- Fixed a few issues related to signal handling in both daemons and
2680
  scripts
2681
- Fixed the example startup script provided
2682
- Fixed insserv dependencies in the example startup script (patch from
2683
  Debian)
2684
- Fixed handling of drained nodes in the iallocator framework
2685
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2686
  #528618)
2687
- Fixed error related to invalid job IDs in job polling
2688
- Fixed job/opcode persistence on unclean master shutdown
2689
- Fixed handling of partial job processing after unclean master
2690
  shutdown
2691
- Fixed error reporting from LUs, previously all errors were converted
2692
  into execution errors
2693
- Fixed error reporting from burnin
2694
- Decreased significantly the memory usage of the job queue
2695
- Optimised slightly multi-job submission
2696
- Optimised slightly opcode loading
2697
- Backported the multi-job submit framework from the development
2698
  branch; multi-instance start and stop should be faster
2699
- Added script to clean archived jobs after 21 days; this will reduce
2700
  the size of the queue directory
2701
- Added some extra checks in disk size tracking
2702
- Added an example ethers hook script
2703
- Added a cluster parameter that prevents Ganeti from modifying of
2704
  /etc/hosts
2705
- Added more node information to RAPI responses
2706
- Added a ``gnt-job watch`` command that allows following the ouput of a
2707
  job
2708
- Added a bind-address option to ganeti-rapi
2709
- Added more checks to the configuration verify
2710
- Enhanced the burnin script such that some operations can be retried
2711
  automatically
2712
- Converted instance reinstall to multi-instance model
2713

    
2714

    
2715
Version 2.0.3
2716
-------------
2717

    
2718
*(Released Fri, 7 Aug 2009)*
2719

    
2720
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2721
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2722
  instances have mismatched disk sizes in the configuration
2723
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2724
  any configuration mismatches for disk sizes
2725
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2726
  failover to work on two-node clusters
2727
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2728
  unusable
2729
- Fixed detection of OS script errors in ``gnt-backup export``
2730
- Fixed exit code of ``gnt-backup export``
2731

    
2732

    
2733
Version 2.0.2
2734
-------------
2735

    
2736
*(Released Fri, 17 Jul 2009)*
2737

    
2738
- Added experimental support for stripped logical volumes; this should
2739
  enhance performance but comes with a higher complexity in the block
2740
  device handling; stripping is only enabled when passing
2741
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2742
  affected even in the non-stripped mode
2743
- Improved resiliency against transient failures at the end of DRBD
2744
  resyncs, and in general of DRBD resync checks
2745
- Fixed a couple of issues with exports and snapshot errors
2746
- Fixed a couple of issues in instance listing
2747
- Added display of the disk size in ``gnt-instance info``
2748
- Fixed checking for valid OSes in instance creation
2749
- Fixed handling of the "vcpus" parameter in instance listing and in
2750
  general of invalid parameters
2751
- Fixed http server library, and thus RAPI, to handle invalid
2752
  username/password combinations correctly; this means that now they
2753
  report unauthorized for queries too, not only for modifications,
2754
  allowing earlier detect of configuration problems
2755
- Added a new "role" node list field, equivalent to the master/master
2756
  candidate/drained/offline flags combinations
2757
- Fixed cluster modify and changes of candidate pool size
2758
- Fixed cluster verify error messages for wrong files on regular nodes
2759
- Fixed a couple of issues with node demotion from master candidate role
2760
- Fixed node readd issues
2761
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2762
- Added a new ``--no-voting`` option for masterfailover to fix failover
2763
  on two-nodes clusters when the former master node is unreachable
2764
- Added instance reinstall over RAPI
2765

    
2766

    
2767
Version 2.0.1
2768
-------------
2769

    
2770
*(Released Tue, 16 Jun 2009)*
2771

    
2772
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2773
  allow re-adding the start in single-user option (regression from 1.2)
2774
- the watcher writes the instance status to a file, to allow monitoring
2775
  to report the instance status (from the master) based on cached
2776
  results of the watcher's queries; while this can get stale if the
2777
  watcher is being locked due to other work on the cluster, this is
2778
  still an improvement
2779
- the watcher now also restarts the node daemon and the rapi daemon if
2780
  they died
2781
- fixed the watcher to handle full and drained queue cases
2782
- hooks export more instance data in the environment, which helps if
2783
  hook scripts need to take action based on the instance's properties
2784
  (no longer need to query back into ganeti)
2785
- instance failovers when the instance is stopped do not check for free
2786
  RAM, so that failing over a stopped instance is possible in low memory
2787
  situations
2788
- rapi uses queries for tags instead of jobs (for less job traffic), and
2789
  for cluster tags it won't talk to masterd at all but read them from
2790
  ssconf
2791
- a couple of error handling fixes in RAPI
2792
- drbd handling: improved the error handling of inconsistent disks after
2793
  resync to reduce the frequency of "there are some degraded disks for
2794
  this instance" messages
2795
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2796
  error handling path called a wrong function name)
2797

    
2798

    
2799
Version 2.0.0
2800
-------------
2801

    
2802
*(Released Wed, 27 May 2009)*
2803

    
2804
- no changes from rc5
2805

    
2806

    
2807
Version 2.0 rc5
2808
---------------
2809

    
2810
*(Released Wed, 20 May 2009)*
2811

    
2812
- fix a couple of bugs (validation, argument checks)
2813
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2814
- some small improvements to RAPI and IAllocator
2815
- make watcher automatically start the master daemon if down
2816

    
2817

    
2818
Version 2.0 rc4
2819
---------------
2820

    
2821
*(Released Mon, 27 Apr 2009)*
2822

    
2823
- change the OS list to not require locks; this helps with big clusters
2824
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2825
  volume group is broken
2826
- ``gnt-instance info``, without any arguments, doesn't run for all
2827
  instances anymore; either pass ``--all`` or pass the desired
2828
  instances; this helps against mistakes on big clusters where listing
2829
  the information for all instances takes a long time
2830
- miscellaneous doc and man pages fixes
2831

    
2832

    
2833
Version 2.0 rc3
2834
---------------
2835

    
2836
*(Released Wed, 8 Apr 2009)*
2837

    
2838
- Change the internal locking model of some ``gnt-node`` commands, in
2839
  order to reduce contention (and blocking of master daemon) when
2840
  batching many creation/reinstall jobs
2841
- Fixes to Xen soft reboot
2842
- No longer build documentation at build time, instead distribute it in
2843
  the archive, in order to reduce the need for the whole docbook/rst
2844
  toolchains
2845

    
2846

    
2847
Version 2.0 rc2
2848
---------------
2849

    
2850
*(Released Fri, 27 Mar 2009)*
2851

    
2852
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2853
- Fix watcher startup sequence, improves the behaviour of busy clusters
2854
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2855
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2856
  bind address and other places
2857
- Some documentation fixes and updates
2858

    
2859

    
2860
Version 2.0 rc1
2861
---------------
2862

    
2863
*(Released Mon, 2 Mar 2009)*
2864

    
2865
- More documentation updates, now all docs should be more-or-less
2866
  up-to-date
2867
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2868
  etc.)
2869
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2870
  and KVM)
2871
- Fix an issue related to $libdir/run/ganeti and cluster creation
2872

    
2873

    
2874
Version 2.0 beta2
2875
-----------------
2876

    
2877
*(Released Thu, 19 Feb 2009)*
2878

    
2879
- Xen PVM and KVM have switched the default value for the instance root
2880
  disk to the first partition on the first drive, instead of the whole
2881
  drive; this means that the OS installation scripts must be changed
2882
  accordingly
2883
- Man pages have been updated
2884
- RAPI has been switched by default to HTTPS, and the exported functions
2885
  should all work correctly
2886
- RAPI v1 has been removed
2887
- Many improvements to the KVM hypervisor
2888
- Block device errors are now better reported
2889
- Many other bugfixes and small improvements
2890

    
2891

    
2892
Version 2.0 beta1
2893
-----------------
2894

    
2895
*(Released Mon, 26 Jan 2009)*
2896

    
2897
- Version 2 is a general rewrite of the code and therefore the
2898
  differences are too many to list, see the design document for 2.0 in
2899
  the ``doc/`` subdirectory for more details
2900
- In this beta version there is not yet a migration path from 1.2 (there
2901
  will be one in the final 2.0 release)
2902
- A few significant changes are:
2903

    
2904
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2905
    various ``gnt-*`` commands are just front-ends to it
2906
  - all the commands are entered into, and executed from a job queue,
2907
    see the ``gnt-job(8)`` manpage
2908
  - the RAPI daemon supports read-write operations, secured by basic
2909
    HTTP authentication on top of HTTPS
2910
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2911
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2912
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2913
  - DRBD devices are using statically allocated minor numbers, which
2914
    will be assigned to existing instances during the migration process
2915
  - there is support for both Xen PVM and Xen HVM instances running on
2916
    the same cluster
2917
  - KVM virtualization is supported too
2918
  - file-based storage has been implemented, which means that it is
2919
    possible to run the cluster without LVM and DRBD storage, for
2920
    example using a shared filesystem exported from shared storage (and
2921
    still have live migration)
2922

    
2923

    
2924
Version 1.2.7
2925
-------------
2926

    
2927
*(Released Tue, 13 Jan 2009)*
2928

    
2929
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2930
- Reuse the old instance mac address by default on instance import, if
2931
  the instance name is the same.
2932
- Handle situations in which the node info rpc returns incomplete
2933
  results (issue 46)
2934
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2935
- Improved version of batcher:
2936

    
2937
  - state file support
2938
  - instance mac address support
2939
  - support for HVM clusters/instances
2940

    
2941
- Add an option to show the number of cpu sockets and nodes in
2942
  ``gnt-node list``
2943
- Support OSes that handle more than one version of the OS api (but do
2944
  not change the current API in any other way)
2945
- Fix ``gnt-node migrate``
2946
- ``gnt-debug`` man page
2947
- Fixes various more typos and small issues
2948
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2949

    
2950

    
2951
Version 1.2.6
2952
-------------
2953

    
2954
*(Released Wed, 24 Sep 2008)*
2955

    
2956
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2957
  type of disk exported to fully virtualized instances.
2958
- provide access to the serial console of HVM instances
2959
- instance auto_balance flag, set by default. If turned off it will
2960
  avoid warnings on cluster verify if there is not enough memory to fail
2961
  over an instance. in the future it will prevent automatically failing
2962
  it over when we will support that.
2963
- batcher tool for instance creation, see ``tools/README.batcher``
2964
- ``gnt-instance reinstall --select-os`` to interactively select a new
2965
  operating system when reinstalling an instance.
2966
- when changing the memory amount on instance modify a check has been
2967
  added that the instance will be able to start. also warnings are
2968
  emitted if the instance will not be able to fail over, if auto_balance
2969
  is true.
2970
- documentation fixes
2971
- sync fields between ``gnt-instance list/modify/add/import``
2972
- fix a race condition in drbd when the sync speed was set after giving
2973
  the device a remote peer.
2974

    
2975

    
2976
Version 1.2.5
2977
-------------
2978

    
2979
*(Released Tue, 22 Jul 2008)*
2980

    
2981
- note: the allowed size and number of tags per object were reduced
2982
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2983
- fixed twisted 8.x compatibility
2984
- fixed ``gnt-instance replace-disks`` with iallocator
2985
- add TCP keepalives on twisted connections to detect restarted nodes
2986
- disk increase support, see ``gnt-instance grow-disk``
2987
- implement bulk node/instance query for RAPI
2988
- add tags in node/instance listing (optional)
2989
- experimental migration (and live migration) support, read the man page
2990
  for ``gnt-instance migrate``
2991
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2992
  has some small improvements in handling its state file
2993

    
2994

    
2995
Version 1.2.4
2996
-------------
2997

    
2998
*(Released Fri, 13 Jun 2008)*
2999

    
3000
- Experimental readonly, REST-based remote API implementation;
3001
  automatically started on master node, TCP port 5080, if enabled by
3002
  ``--enable-rapi`` parameter to configure script.
3003
- Instance allocator support. Add and import instance accept a
3004
  ``--iallocator`` parameter, and call that instance allocator to decide
3005
  which node to use for the instance. The iallocator document describes
3006
  what's expected from an allocator script.
3007
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3008
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3009
  node is lost there is still enough memory to fail over the instances
3010
  that reside on it.
3011
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3012
  ``gnt-cluster verify``, to check for site-specific compliance. All the
3013
  hooks will run, and their output, if any, will be displayed. Any
3014
  failing hook will make the verification return an error value.
3015
- ``gnt-cluster verify`` now checks that its peers are reachable on the
3016
  primary and secondary interfaces
3017
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3018
  that is still declared as part of the cluster and has failed.
3019
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
3020
  specifying output fields, that just adds the chosen fields to the
3021
  default ones.
3022
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
3023
  export from the filesystem.
3024
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3025
  have been added. Using them you can enable/disable acpi and pae
3026
  support, and specify a path for a cd image to be exported to the
3027
  instance. These parameters as the name suggest only work on HVM
3028
  clusters.
3029
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3030
  PAE support will be set to the previously hardcoded values, but the
3031
  (previously hardcoded) path to the CDROM ISO image will be unset and
3032
  if required, needs to be set manually with ``gnt-instance modify``
3033
  after the upgrade.
3034
- The address to which an instance's VNC console is bound is now
3035
  selectable per-instance, rather than being cluster wide. Of course
3036
  this only applies to instances controlled via VNC, so currently just
3037
  applies to HVM clusters.
3038

    
3039

    
3040
Version 1.2.3
3041
-------------
3042

    
3043
*(Released Mon, 18 Feb 2008)*
3044

    
3045
- more tweaks to the disk activation code (especially helpful for DRBD)
3046
- change the default ``gnt-instance list`` output format, now there is
3047
  one combined status field (see the manpage for the exact values this
3048
  field will have)
3049
- some more fixes for the mac export to hooks change
3050
- make Ganeti not break with DRBD 8.2.x (which changed the version
3051
  format in ``/proc/drbd``) (issue 24)
3052
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3053
  template, allowing migration from DRBD0.7+MD to DRBD8
3054

    
3055

    
3056
Version 1.2.2
3057
-------------
3058

    
3059
*(Released Wed, 30 Jan 2008)*
3060

    
3061
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3062
  support (issue 23)
3063
- add command aliases infrastructure and a few aliases
3064
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
3065
  man pages and the ``--help`` option of ``gnt-node
3066
  list``/``gnt-instance list``
3067
- fix ``gnt-backup list`` with down nodes (issue 21)
3068
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3069
- fix the dist archive and add a check for including svn/git files in
3070
  the future
3071
- some developer-related changes: improve the burnin and the QA suite,
3072
  add an upload script for testing during development
3073

    
3074

    
3075
Version 1.2.1
3076
-------------
3077

    
3078
*(Released Wed, 16 Jan 2008)*
3079

    
3080
- experimental HVM support, read the install document, section
3081
  "Initializing the cluster"
3082
- allow for the PVM hypervisor per-instance kernel and initrd paths
3083
- add a new command ``gnt-cluster verify-disks`` which uses a new
3084
  algorithm to improve the reconnection of the DRBD pairs if the device
3085
  on the secondary node has gone away
3086
- make logical volume code auto-activate LVs at disk activation time
3087
- slightly improve the speed of activating disks
3088
- allow specification of the MAC address at instance creation time, and
3089
  changing it later via ``gnt-instance modify``
3090
- fix handling of external commands that generate lots of output on
3091
  stderr
3092
- update documentation with regard to minimum version of DRBD8 supported
3093

    
3094

    
3095
Version 1.2.0
3096
-------------
3097

    
3098
*(Released Tue, 4 Dec 2007)*
3099

    
3100
- Log the ``xm create`` output to the node daemon log on failure (to
3101
  help diagnosing the error)
3102
- In debug mode, log all external commands output if failed to the logs
3103
- Change parsing of lvm commands to ignore stderr
3104

    
3105

    
3106
Version 1.2 beta3
3107
-----------------
3108

    
3109
*(Released Wed, 28 Nov 2007)*
3110

    
3111
- Another round of updates to the DRBD 8 code to deal with more failures
3112
  in the replace secondary node operation
3113
- Some more logging of failures in disk operations (lvm, drbd)
3114
- A few documentation updates
3115
- QA updates
3116

    
3117

    
3118
Version 1.2 beta2
3119
-----------------
3120

    
3121
*(Released Tue, 13 Nov 2007)*
3122

    
3123
- Change configuration file format from Python's Pickle to JSON.
3124
  Upgrading is possible using the cfgupgrade utility.
3125
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3126
  faster replace disks and is more stable (DRBD 8 has many improvements
3127
  compared to DRBD 0.7)
3128
- Added command line tags support (see man pages for ``gnt-instance``,
3129
  ``gnt-node``, ``gnt-cluster``)
3130
- Added instance rename support
3131
- Added multi-instance startup/shutdown
3132
- Added cluster rename support
3133
- Added ``gnt-node evacuate`` to simplify some node operations
3134
- Added instance reboot operation that can speedup reboot as compared to
3135
  stop and start
3136
- Soften the requirement that hostnames are in FQDN format
3137
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
3138
  reboots
3139
- Removed dependency on debian's patched fping that uses the
3140
  non-standard ``-S`` option
3141
- Now the OS definitions are searched for in multiple, configurable
3142
  paths (easier for distros to package)
3143
- Some changes to the hooks infrastructure (especially the new
3144
  post-configuration update hook)
3145
- Other small bugfixes
3146

    
3147
.. vim: set textwidth=72 syntax=rst :
3148
.. Local Variables:
3149
.. mode: rst
3150
.. fill-column: 72
3151
.. End: