Statistics
| Branch: | Tag: | Revision:

root / NEWS @ b0e8ed3f

History | View | Annotate | Download (119.1 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
- Queries for nodes, instances, groups, backups and networks are now
19
  exclusively done via the luxi daemon. Legacy python code was removed,
20
  as well as the --enable-split-queries configuration option.
21
- Orphan volumes errors are demoted to warnings and no longer affect the exit
22
  code of ``gnt-cluster verify``.
23

    
24
New features
25
~~~~~~~~~~~~
26

    
27
- Instance moves, backups and imports can now use compression to transfer the
28
  instance data.
29
- Node groups can be configured to use an SSH port different than the
30
  default 22.
31

    
32
New dependencies
33
~~~~~~~~~~~~~~~~
34
The following new dependencies have been added:
35

    
36
For Haskell:
37

    
38
- ``zlib`` library (http://hackage.haskell.org/package/base64-bytestring)
39

    
40
- ``base64-bytestring`` library (http://hackage.haskell.org/package/zlib),
41
  at least version 1.0.0.0
42

    
43

    
44
Version 2.10.0 alpha1
45
---------------------
46

    
47
*(unreleased)*
48

    
49
Incompatible/important changes
50
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
51

    
52
- Adding disks with 'gnt-instance modify' now waits for the disks to sync per
53
  default. Specify --no-wait-for-sync to override this behavior.
54
- The Ganeti python code now adheres to a private-module layout. In particular,
55
  the module 'ganeti' is no longer in the python search path.
56
- On instance allocation, the iallocator now considers non-LVM storage
57
  properly. In particular, actual file storage space information is used
58
  when allocating space for a file/sharedfile instance.
59
- When disabling disk templates cluster-wide, the cluster now first
60
  checks whether there are instances still using those templates.
61
- 'gnt-node list-storage' now also reports storage information about
62
  file-based storage types.
63

    
64
New features
65
~~~~~~~~~~~~
66

    
67
- KVM hypervisors can now access RBD storage directly without having to
68
  go through a block device.
69
- A new command 'gnt-cluster upgrade' was added that automates the upgrade
70
  procedure between two Ganeti versions that are both 2.10 or higher.
71
- The move-instance command can now change disk templates when moving
72
  instances, and does not require any node placement options to be
73
  specified if the destination cluster has a default iallocator.
74
- Users can now change the soundhw and cpuid settings for XEN hypervisors.
75
- Hail and hbal now have the (optional) capability of accessing average CPU
76
  load information through the monitoring deamon, and to use it to dynamically
77
  adapt the allocation of instances.
78
- Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
79
  so that disk and NIC modifications take effect without the need of actual
80
  reboot. There are a couple of constrains currently for this feature:
81

    
82
   - only KVM hypervisor (versions >= 1.0) supports it,
83
   - one can not (yet) hotplug a disk using userspace access mode for RBD
84
   - in case of a downgrade instances should suffer a reboot in order to
85
     be migratable (due to core change of runtime files)
86

    
87
Misc changes
88
~~~~~~~~~~~~
89

    
90
- A new test framework for logical units was introduced and the test
91
  coverage for logical units was improved significantly.
92
- Opcodes are entirely generated from Haskell using the tool 'hs2py' and
93
  the module 'src/Ganeti/OpCodes.hs'.
94
- Constants are also generated from Haskell using the tool
95
  'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
96
  exception of socket related constants, which require changing the
97
  cluster configuration file, and HVS related constants, because they
98
  are part of a port of instance queries to Haskell.  As a result, these
99
  changes will be part of the next release of Ganeti.
100

    
101
New dependencies
102
~~~~~~~~~~~~~~~~
103

    
104
The following new dependencies have been added/updated.
105

    
106
Python
107

    
108
- The version requirements for ``python-mock`` have increased to at least
109
  version 1.0.1. It is still used for testing only.
110

    
111

    
112
Version 2.9.1
113
-------------
114

    
115
*(Released Wed, 13 Nov 2013)*
116

    
117
- fix bug, that kept nodes offline when readding
118
- when verifying DRBD versions, ignore unavailable nodes
119
- fix bug that made the console unavailable on kvm in split-user
120
  setup (issue 608)
121
- DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
122

    
123

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

    
127
*(Released Tue, 5 Nov 2013)*
128

    
129
Incompatible/important changes
130
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
131

    
132
- hroller now also plans for capacity to move non-redundant instances off
133
  any node to be rebooted; the old behavior of completely ignoring any
134
  non-redundant instances can be restored by adding the --ignore-non-redundant
135
  option.
136
- The cluster option '--no-lvm-storage' was removed in favor of the new option
137
  '--enabled-disk-templates'.
138
- On instance creation, disk templates no longer need to be specified
139
  with '-t'. The default disk template will be taken from the list of
140
  enabled disk templates.
141
- The monitoring daemon is now running as root, in order to be able to collect
142
  information only available to root (such as the state of Xen instances).
143
- The ConfD client is now IPv6 compatible.
144
- File and shared file storage is no longer dis/enabled at configure time,
145
  but using the option '--enabled-disk-templates' at cluster initialization and
146
  modification.
147
- The default directories for file and shared file storage are not anymore
148
  specified at configure time, but taken from the cluster's configuration.
149
  They can be set at cluster initialization and modification with
150
  '--file-storage-dir' and '--shared-file-storage-dir'.
151
- Cluster verification now includes stricter checks regarding the
152
  default file and shared file storage directories. It now checks that
153
  the directories are explicitely allowed in the 'file-storage-paths' file and
154
  that the directories exist on all nodes.
155
- The list of allowed disk templates in the instance policy and the list
156
  of cluster-wide enabled disk templates is now checked for consistency
157
  on cluster or group modification. On cluster initialization, the ipolicy
158
  disk templates are ensured to be a subset of the cluster-wide enabled
159
  disk templates.
160

    
161
New features
162
~~~~~~~~~~~~
163

    
164
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
165
  the correct command syntax. It is possible to use different DRBD versions
166
  on different nodes as long as they are compatible to each other. This
167
  enables rolling upgrades of DRBD with no downtime. As permanent operation
168
  of different DRBD versions within a node group is discouraged,
169
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
170
- New "inst-status-xen" data collector for the monitoring daemon, providing
171
  information about the state of the xen instances on the nodes.
172
- New "lv" data collector for the monitoring daemon, collecting data about the
173
  logical volumes on the nodes, and pairing them with the name of the instances
174
  they belong to.
175
- New "diskstats" data collector, collecting the data from /proc/diskstats and
176
  presenting them over the monitoring daemon interface.
177
- The ConfD client is now IPv6 compatible.
178

    
179
New dependencies
180
~~~~~~~~~~~~~~~~
181
The following new dependencies have been added.
182

    
183
Python
184

    
185
- ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
186
  for the unit tests (and only used for testing).
187

    
188
Haskell
189

    
190
- ``hslogger`` (http://software.complete.org/hslogger) is now always
191
  required, even if confd is not enabled.
192

    
193
Since 2.9.0 rc3
194
~~~~~~~~~~~~~~~
195

    
196
- Correctly start/stop luxid during gnt-cluster master-failover (inherited
197
  from stable-2.8)
198
- Improved error messsages (inherited from stable-2.8)
199

    
200

    
201
Version 2.9.0 rc3
202
-----------------
203

    
204
*(Released Tue, 15 Oct 2013)*
205

    
206
The third release candidate in the 2.9 series. Since 2.9.0 rc2:
207

    
208
- in implicit configuration upgrade, match ipolicy with enabled disk templates
209
- improved harep documentation (inherited from stable-2.8)
210

    
211

    
212
Version 2.9.0 rc2
213
-----------------
214

    
215
*(Released Wed, 9 Oct 2013)*
216

    
217
The second release candidate in the 2.9 series. Since 2.9.0 rc1:
218

    
219
- Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
220
  at least one DRBD instance.
221
- Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
222
  downgrading.
223

    
224

    
225
Version 2.9.0 rc1
226
-----------------
227

    
228
*(Released Tue, 1 Oct 2013)*
229

    
230
The first release candidate in the 2.9 series. Since 2.9.0 beta1:
231

    
232
- various bug fixes
233
- update of the documentation, in particular installation instructions
234
- merging of LD_* constants into DT_* constants
235
- python style changes to be compatible with newer versions of pylint
236

    
237

    
238
Version 2.9.0 beta1
239
-------------------
240

    
241
*(Released Thu, 29 Aug 2013)*
242

    
243
This was the first beta release of the 2.9 series. All important changes
244
are listed in the latest 2.9 entry.
245

    
246

    
247
Version 2.8.2
248
-------------
249

    
250
*(Released Thu, 07 Nov 2013)*
251

    
252
- DRBD: ensure peers are UpToDate for dual-primary
253
- Improve error message for replace-disks
254
- More dependency checks at configure time
255
- Placate warnings on ganeti.outils_unittest.py
256

    
257

    
258
Version 2.8.1
259
-------------
260

    
261
*(Released Thu, 17 Oct 2013)*
262

    
263
- Correctly start/stop luxid during gnt-cluster master-failover
264
- Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
265
- Fix path for the job queue serial file
266
- Improved harep man page
267
- Minor documentation improvements
268

    
269

    
270
Version 2.8.0
271
-------------
272

    
273
*(Released Mon, 30 Sep 2013)*
274

    
275
Incompatible/important changes
276
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
277

    
278
- Instance policy can contain multiple instance specs, as described in
279
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
280
  <design-partitioned>`. As a consequence, it's not possible to partially change
281
  or override instance specs. Bounding specs (min and max) can be specified as a
282
  whole using the new option ``--ipolicy-bounds-specs``, while standard
283
  specs use the new option ``--ipolicy-std-specs``.
284
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
285
  gnt-instance is a valid YAML object.
286
- hail now honors network restrictions when allocating nodes. This led to an
287
  update of the IAllocator protocol. See the IAllocator documentation for
288
  details.
289
- confd now only answers static configuration request over the network. luxid
290
  was extracted, listens on the local LUXI socket and responds to live queries.
291
  This allows finer grained permissions if using separate users.
292

    
293
New features
294
~~~~~~~~~~~~
295

    
296
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
297
  to always require authentication, ``--require-authentication``. It can
298
  be specified in ``$sysconfdir/default/ganeti``.
299
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
300
  be used to manage the disk templates to be used by instances in the cluster.
301
  Initially, it will be set to a list that includes plain, drbd, if they were
302
  enabled by specifying a volume group name, and file and sharedfile, if those
303
  were enabled at configure time. Additionally, it will include all disk
304
  templates that are currently used by instances. The order of disk templates
305
  will be based on Ganeti's history of supporting them. In the future, the
306
  first entry of the list will be used as a default disk template on instance
307
  creation.
308
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
309
  configuration back to the previous stable version.
310
- Disk templates in group ipolicy can be restored to the default value.
311
- Initial support for diskless instances and virtual clusters in QA.
312
- More QA and unit tests for instance policies.
313
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
314
  describing why the opcode itself was executed.
315
- The monitoring daemon is now available. It allows users to query the cluster
316
  for obtaining information about the status of the system. The daemon is only
317
  responsible for providing the information over the network: the actual data
318
  gathering is performed by data collectors (currently, only the DRBD status
319
  collector is available).
320
- In order to help developers work on Ganeti, a new script
321
  (``devel/build_chroot``) is provided, for building a chroot that contains all
322
  the required development libraries and tools for compiling Ganeti on a Debian
323
  Squeeze system.
324
- A new tool, ``harep``, for performing self-repair and recreation of instances
325
  in Ganeti has been added.
326
- Split queries are enabled for tags, network, exports, cluster info, groups,
327
  jobs, nodes.
328
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
329
  It prints the command line to set the current policies, to ease
330
  changing them.
331
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
332
  devices for KVM virtio-net interfaces will get created with VNET_HDR
333
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
334
  virtio-net interfaces, which prevents host kernel tainting and log
335
  flooding, when dealing with broken or malicious virtio-net drivers.
336
  It's set to true by default.
337
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
338
  failures.
339
- Support 'viridian' parameter in Xen HVM
340
- Support DSA SSH keys in bootstrap
341
- To simplify the work of packaging frameworks that want to add the needed users
342
  and groups in a split-user setup themselves, at build time three files in
343
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
344
  the groups to be generated, the ``users`` file contains, one per line, the
345
  users to be generated, optionally followed by their primary group, where
346
  important. The ``groupmemberships`` file contains, one per line, additional
347
  user-group membership relations that need to be established. The syntax of
348
  these files will remain stable in all future versions.
349

    
350

    
351
New dependencies
352
~~~~~~~~~~~~~~~~
353
The following new dependencies have been added:
354

    
355
For Haskell:
356
- The ``curl`` library is not optional anymore for compiling the Haskell code.
357
- ``snap-server`` library (if monitoring is enabled).
358

    
359
For Python:
360
- The minimum Python version needed to run Ganeti is now 2.6.
361
- ``yaml`` library (only for running the QA).
362

    
363
Since 2.8.0 rc3
364
~~~~~~~~~~~~~~~
365
- Perform proper cleanup on termination of Haskell daemons
366
- Fix corner-case in handling of remaining retry time
367

    
368

    
369
Version 2.8.0 rc3
370
-----------------
371

    
372
*(Released Tue, 17 Sep 2013)*
373

    
374
- To simplify the work of packaging frameworks that want to add the needed users
375
  and groups in a split-user setup themselves, at build time three files in
376
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
377
  the groups to be generated, the ``users`` file contains, one per line, the
378
  users to be generated, optionally followed by their primary group, where
379
  important. The ``groupmemberships`` file contains, one per line, additional
380
  user-group membership relations that need to be established. The syntax of
381
  these files will remain stable in all future versions.
382
- Add a default to file-driver when unspecified over RAPI (Issue 571)
383
- Mark the DSA host pubkey as optional, and remove it during config downgrade
384
  (Issue 560)
385
- Some documentation fixes
386

    
387

    
388
Version 2.8.0 rc2
389
-----------------
390

    
391
*(Released Tue, 27 Aug 2013)*
392

    
393
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
394

    
395
- Support 'viridian' parameter in Xen HVM (Issue 233)
396
- Include VCS version in ``gnt-cluster version``
397
- Support DSA SSH keys in bootstrap (Issue 338)
398
- Fix batch creation of instances
399
- Use FQDN to check master node status (Issue 551)
400
- Make the DRBD collector more failure-resilient
401

    
402

    
403
Version 2.8.0 rc1
404
-----------------
405

    
406
*(Released Fri, 2 Aug 2013)*
407

    
408
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
409

    
410
- Fix upgrading/downgrading from 2.7
411
- Increase maximum RAPI message size
412
- Documentation updates
413
- Split ``confd`` between ``luxid`` and ``confd``
414
- Merge 2.7 series up to the 2.7.1 release
415
- Allow the ``modify_etc_hosts`` option to be changed
416
- Add better debugging for ``luxid`` queries
417
- Expose bulk parameter for GetJobs in RAPI client
418
- Expose missing ``network`` fields in RAPI
419
- Add some ``cluster verify`` tests
420
- Some unittest fixes
421
- Fix a malfunction in ``hspace``'s tiered allocation
422
- Fix query compatibility between haskell and python implementations
423
- Add the ``vnet_hdr`` HV parameter for KVM
424
- Add ``--cleanup`` to instance failover
425
- Change the connected groups format in ``gnt-network info`` output; it
426
  was previously displayed as a raw list by mistake. (Merged from 2.7)
427

    
428

    
429
Version 2.8.0 beta1
430
-------------------
431

    
432
*(Released Mon, 24 Jun 2013)*
433

    
434
This was the first beta release of the 2.8 series. All important changes
435
are listed in the latest 2.8 entry.
436

    
437

    
438
Version 2.7.2
439
-------------
440

    
441
*(Released Thu, 26 Sep 2013)*
442

    
443
- Change the connected groups format in ``gnt-network info`` output; it
444
  was previously displayed as a raw list by mistake
445
- Check disk template in right dict when copying
446
- Support multi-instance allocs without iallocator
447
- Fix some errors in the documentation
448
- Fix formatting of tuple in an error message
449

    
450

    
451
Version 2.7.1
452
-------------
453

    
454
*(Released Thu, 25 Jul 2013)*
455

    
456
- Add logrotate functionality in daemon-util
457
- Add logrotate example file
458
- Add missing fields to network queries over rapi
459
- Fix network object timestamps
460
- Add support for querying network timestamps
461
- Fix a typo in the example crontab
462
- Fix a documentation typo
463

    
464

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

    
468
*(Released Thu, 04 Jul 2013)*
469

    
470
Incompatible/important changes
471
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
472

    
473
- Instance policies for disk size were documented to be on a per-disk
474
  basis, but hail applied them to the sum of all disks. This has been
475
  fixed.
476
- ``hbal`` will now exit with status 0 if, during job execution over
477
  LUXI, early exit has been requested and all jobs are successful;
478
  before, exit status 1 was used, which cannot be differentiated from
479
  "job error" case
480
- Compatibility with newer versions of rbd has been fixed
481
- ``gnt-instance batch-create`` has been changed to use the bulk create
482
  opcode from Ganeti. This lead to incompatible changes in the format of
483
  the JSON file. It's now not a custom dict anymore but a dict
484
  compatible with the ``OpInstanceCreate`` opcode.
485
- Parent directories for file storage need to be listed in
486
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
487
  write the file automatically based on old configuration values, but it
488
  can not distribute it across all nodes and the file contents should be
489
  verified. Use ``gnt-cluster copyfile
490
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
491
  upgraded. The reason for requiring this list of paths now is that
492
  before it would have been possible to inject new paths via RPC,
493
  allowing files to be created in arbitrary locations. The RPC protocol
494
  is protected using SSL/X.509 certificates, but as a design principle
495
  Ganeti does not permit arbitrary paths to be passed.
496
- The parsing of the variants file for OSes (see
497
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
498
  lines and comment lines (starting with ``#``) are ignored for better
499
  readability.
500
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
501
  longer available. ``gnt-node add`` now invokes a new tool on the
502
  destination node, named ``prepare-node-join``, to configure the SSH
503
  daemon. Paramiko is no longer necessary to configure nodes' SSH
504
  daemons via ``gnt-node add``.
505
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
506
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
507
  the flag is not reset after a master failover.
508
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
509
  is recommended. 2.6 will be mandatory from the 2.8 series.
510

    
511

    
512
New features
513
~~~~~~~~~~~~
514

    
515
- New network management functionality to support automatic allocation
516
  of IP addresses and managing of network parameters. See
517
  :manpage:`gnt-network(8)` for more details.
518
- New external storage backend, to allow managing arbitrary storage
519
  systems external to the cluster. See
520
  :manpage:`ganeti-extstorage-interface(7)`.
521
- New ``exclusive-storage`` node parameter added, restricted to
522
  nodegroup level. When it's set to true, physical disks are assigned in
523
  an exclusive fashion to instances, as documented in :doc:`Partitioned
524
  Ganeti <design-partitioned>`.  Currently, only instances using the
525
  ``plain`` disk template are supported.
526
- The KVM hypervisor has been updated with many new hypervisor
527
  parameters, including a generic one for passing arbitrary command line
528
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
529
  compatible up to qemu 1.4.
530
- A new tool, called ``mon-collector``, is the stand-alone executor of
531
  the data collectors for a monitoring system. As of this version, it
532
  just includes the DRBD data collector, that can be executed by calling
533
  ``mon-collector`` using the ``drbd`` parameter. See
534
  :manpage:`mon-collector(7)`.
535
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
536
  for RAPI users. It allows granting permissions to query for
537
  information to a specific user without giving
538
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
539
- A new tool named ``node-cleanup`` has been added. It cleans remains of
540
  a cluster from a machine by stopping all daemons, removing
541
  certificates and ssconf files. Unless the ``--no-backup`` option is
542
  given, copies of the certificates are made.
543
- Instance creations now support the use of opportunistic locking,
544
  potentially speeding up the (parallel) creation of multiple instances.
545
  This feature is currently only available via the :doc:`RAPI
546
  <rapi>` interface and when an instance allocator is used. If the
547
  ``opportunistic_locking`` parameter is set the opcode will try to
548
  acquire as many locks as possible, but will not wait for any locks
549
  held by other opcodes. If not enough resources can be found to
550
  allocate the instance, the temporary error code
551
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
552
  retried thereafter, with or without opportunistic locking.
553
- New experimental linux-ha resource scripts.
554
- Restricted-commands support: ganeti can now be asked (via command line
555
  or rapi) to perform commands on a node. These are passed via ganeti
556
  RPC rather than ssh. This functionality is restricted to commands
557
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
558
  reasons. The file is not copied automatically.
559

    
560

    
561
Misc changes
562
~~~~~~~~~~~~
563

    
564
- Diskless instances are now externally mirrored (Issue 237). This for
565
  now has only been tested in conjunction with explicit target nodes for
566
  migration/failover.
567
- Queries not needing locks or RPC access to the node can now be
568
  performed by the confd daemon, making them independent from jobs, and
569
  thus faster to execute. This is selectable at configure time.
570
- The functionality for allocating multiple instances at once has been
571
  overhauled and is now also available through :doc:`RAPI <rapi>`.
572

    
573
There are no significant changes from version 2.7.0~rc3.
574

    
575

    
576
Version 2.7.0 rc3
577
-----------------
578

    
579
*(Released Tue, 25 Jun 2013)*
580

    
581
- Fix permissions on the confd query socket (Issue 477)
582
- Fix permissions on the job archive dir (Issue 498)
583
- Fix handling of an internal exception in replace-disks (Issue 472)
584
- Fix gnt-node info handling of shortened names (Issue 497)
585
- Fix gnt-instance grow-disk when wiping is enabled
586
- Documentation improvements, and support for newer pandoc
587
- Fix hspace honoring ipolicy for disks (Issue 484)
588
- Improve handling of the ``kvm_extra`` HV parameter
589

    
590

    
591
Version 2.7.0 rc2
592
-----------------
593

    
594
*(Released Fri, 24 May 2013)*
595

    
596
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
597
  symlink.
598
- Disks added through ``gnt-instance modify`` or created through
599
  ``gnt-instance recreate-disks`` are wiped, if the
600
  ``prealloc_wipe_disks`` flag is set.
601
- If wiping newly created disks fails, the disks are removed. Also,
602
  partial failures in creating disks through ``gnt-instance modify``
603
  triggers a cleanup of the partially-created disks.
604
- Removing the master IP address doesn't fail if the address has been
605
  already removed.
606
- Fix ownership of the OS log dir
607
- Workaround missing SO_PEERCRED constant (Issue 191)
608

    
609

    
610
Version 2.7.0 rc1
611
-----------------
612

    
613
*(Released Fri, 3 May 2013)*
614

    
615
This was the first release candidate of the 2.7 series. Since beta3:
616

    
617
- Fix kvm compatibility with qemu 1.4 (Issue 389)
618
- Documentation updates (admin guide, upgrade notes, install
619
  instructions) (Issue 372)
620
- Fix gnt-group list nodes and instances count (Issue 436)
621
- Fix compilation without non-mandatory libraries (Issue 441)
622
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
623
- Make confd logging more verbose at INFO level (Issue 435)
624
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
625
- Fix failure path for instance storage type conversion (Issue 229)
626
- Update htools text backend documentation
627
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
628
- Disable inter-cluster instance move for file-based instances, because
629
  it is dependant on instance export, which is not supported for
630
  file-based instances. (Issue 414)
631
- Fix gnt-job crashes on non-ascii characters (Issue 427)
632
- Fix volume group checks on non-vm-capable nodes (Issue 432)
633

    
634

    
635
Version 2.7.0 beta3
636
-------------------
637

    
638
*(Released Mon, 22 Apr 2013)*
639

    
640
This was the third beta release of the 2.7 series. Since beta2:
641

    
642
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
643
- Fix data loss on wrong usage of ``gnt-instance move``
644
- Properly export errors in confd-based job queries
645
- Add ``users-setup`` tool
646
- Fix iallocator protocol to report 0 as a disk size for diskless
647
  instances. This avoids hail breaking when a diskless instance is
648
  present.
649
- Fix job queue directory permission problem that made confd job queries
650
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
651
  for access to archived jobs (Issue 406).
652
- Limit the sizes of networks supported by ``gnt-network`` to something
653
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
654
- Fix bugs in instance disk template conversion
655
- Fix GHC 7 compatibility
656
- Fix ``burnin`` install path (Issue 426).
657
- Allow very small disk grows (Issue 347).
658
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
659
  that noded doesn't import masterd code (Issue 419).
660
- Make sure the default metavg at cluster init is the same as the vg, if
661
  unspecified (Issue 358).
662
- Fix cleanup of partially created disks (part of Issue 416)
663

    
664

    
665
Version 2.7.0 beta2
666
-------------------
667

    
668
*(Released Tue, 2 Apr 2013)*
669

    
670
This was the second beta release of the 2.7 series. Since beta1:
671

    
672
- Networks no longer have a "type" slot, since this information was
673
  unused in Ganeti: instead of it tags should be used.
674
- The rapi client now has a ``target_node`` option to MigrateInstance.
675
- Fix early exit return code for hbal (Issue 386).
676
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
677
- Fix ``rbd showmapped`` output parsing (Issue 312).
678
- Networks are now referenced indexed by UUID, rather than name. This
679
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
680
  use.
681
- The OS environment now includes network information.
682
- Deleting of a network is now disallowed if any instance nic is using
683
  it, to prevent dangling references.
684
- External storage is now documented in man pages.
685
- The exclusive_storage flag can now only be set at nodegroup level.
686
- Hbal can now submit an explicit priority with its jobs.
687
- Many network related locking fixes.
688
- Bump up the required pylint version to 0.25.1.
689
- Fix the ``no_remember`` option in RAPI client.
690
- Many ipolicy related tests, qa, and fixes.
691
- Many documentation improvements and fixes.
692
- Fix building with ``--disable-file-storage``.
693
- Fix ``-q`` option in htools, which was broken if passed more than
694
  once.
695
- Some haskell/python interaction improvements and fixes.
696
- Fix iallocator in case of missing LVM storage.
697
- Fix confd config load in case of ``--no-lvm-storage``.
698
- The confd/query functionality is now mentioned in the security
699
  documentation.
700

    
701

    
702
Version 2.7.0 beta1
703
-------------------
704

    
705
*(Released Wed, 6 Feb 2013)*
706

    
707
This was the first beta release of the 2.7 series. All important changes
708
are listed in the latest 2.7 entry.
709

    
710

    
711
Version 2.6.2
712
-------------
713

    
714
*(Released Fri, 21 Dec 2012)*
715

    
716
Important behaviour change: hbal won't rebalance anymore instances which
717
have the ``auto_balance`` attribute set to false. This was the intention
718
all along, but until now it only skipped those from the N+1 memory
719
reservation (DRBD-specific).
720

    
721
A significant number of bug fixes in this release:
722

    
723
- Fixed disk adoption interaction with ipolicy checks.
724
- Fixed networking issues when instances are started, stopped or
725
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
726
- Fixed the warning in cluster verify for shared storage instances not
727
  being redundant.
728
- Fixed removal of storage directory on shared file storage (issue 262).
729
- Fixed validation of LVM volume group name in OpClusterSetParams
730
  (``gnt-cluster modify``) (issue 285).
731
- Fixed runtime memory increases (``gnt-instance modify -m``).
732
- Fixed live migration under Xen's ``xl`` mode.
733
- Fixed ``gnt-instance console`` with ``xl``.
734
- Fixed building with newer Haskell compiler/libraries.
735
- Fixed PID file writing in Haskell daemons (confd); this prevents
736
  restart issues if confd was launched manually (outside of
737
  ``daemon-util``) while another copy of it was running
738
- Fixed a type error when doing live migrations with KVM (issue 297) and
739
  the error messages for failing migrations have been improved.
740
- Fixed opcode validation for the out-of-band commands (``gnt-node
741
  power``).
742
- Fixed a type error when unsetting OS hypervisor parameters (issue
743
  311); now it's possible to unset all OS-specific hypervisor
744
  parameters.
745
- Fixed the ``dry-run`` mode for many operations: verification of
746
  results was over-zealous but didn't take into account the ``dry-run``
747
  operation, resulting in "wrong" failures.
748
- Fixed bash completion in ``gnt-job list`` when the job queue has
749
  hundreds of entries; especially with older ``bash`` versions, this
750
  results in significant CPU usage.
751

    
752
And lastly, a few other improvements have been made:
753

    
754
- Added option to force master-failover without voting (issue 282).
755
- Clarified error message on lock conflict (issue 287).
756
- Logging of newly submitted jobs has been improved (issue 290).
757
- Hostname checks have been made uniform between instance rename and
758
  create (issue 291).
759
- The ``--submit`` option is now supported by ``gnt-debug delay``.
760
- Shutting down the master daemon by sending SIGTERM now stops it from
761
  processing jobs waiting for locks; instead, those jobs will be started
762
  once again after the master daemon is started the next time (issue
763
  296).
764
- Support for Xen's ``xl`` program has been improved (besides the fixes
765
  above).
766
- Reduced logging noise in the Haskell confd daemon (only show one log
767
  entry for each config reload, instead of two).
768
- Several man page updates and typo fixes.
769

    
770

    
771
Version 2.6.1
772
-------------
773

    
774
*(Released Fri, 12 Oct 2012)*
775

    
776
A small bugfix release. Among the bugs fixed:
777

    
778
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
779
  made the command unusable.
780
- Commands that issue many jobs don't fail anymore just because some jobs
781
  take so long that other jobs are archived.
782
- Failures during ``gnt-instance reinstall`` are reflected by the exit
783
  status.
784
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
785
  DRBD is enabled.
786
- When ``always_failover`` is set, ``--allow-failover`` is not required
787
  in migrate commands anymore.
788
- ``bash_completion`` works even if extglob is disabled.
789
- Fixed bug with locks that made failover for RDB-based instances fail.
790
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
791
  a random node instead of one based on the allocator metric.
792
- Support for newer versions of pylint and pep8.
793
- Hail doesn't fail anymore when trying to add an instance of type
794
  ``file``, ``sharedfile`` or ``rbd``.
795
- Added new Makefile target to rebuild the whole distribution, so that
796
  all files are included.
797

    
798

    
799
Version 2.6.0
800
-------------
801

    
802
*(Released Fri, 27 Jul 2012)*
803

    
804

    
805
.. attention:: The ``LUXI`` protocol has been made more consistent
806
   regarding its handling of command arguments. This, however, leads to
807
   incompatibility issues with previous versions. Please ensure that you
808
   restart Ganeti daemons soon after the upgrade, otherwise most
809
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
810
   pausing/resuming the watcher, cancelling and archiving jobs, querying
811
   the cluster configuration) will fail.
812

    
813

    
814
New features
815
~~~~~~~~~~~~
816

    
817
Instance run status
818
+++++++++++++++++++
819

    
820
The current ``admin_up`` field, which used to denote whether an instance
821
should be running or not, has been removed. Instead, ``admin_state`` is
822
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
823

    
824
The rational behind this is that an instance being “down” can have
825
different meanings:
826

    
827
- it could be down during a reboot
828
- it could be temporarily be down for a reinstall
829
- or it could be down because it is deprecated and kept just for its
830
  disk
831

    
832
The previous Boolean state was making it difficult to do capacity
833
calculations: should Ganeti reserve memory for a down instance? Now, the
834
tri-state field makes it clear:
835

    
836
- in ``up`` and ``down`` state, all resources are reserved for the
837
  instance, and it can be at any time brought up if it is down
838
- in ``offline`` state, only disk space is reserved for it, but not
839
  memory or CPUs
840

    
841
The field can have an extra use: since the transition between ``up`` and
842
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
843
transition between ``offline`` and ``down`` is done via ``gnt-instance
844
modify``, it is possible to given different rights to users. For
845
example, owners of an instance could be allowed to start/stop it, but
846
not transition it out of the offline state.
847

    
848
Instance policies and specs
849
+++++++++++++++++++++++++++
850

    
851
In previous Ganeti versions, an instance creation request was not
852
limited on the minimum size and on the maximum size just by the cluster
853
resources. As such, any policy could be implemented only in third-party
854
clients (RAPI clients, or shell wrappers over ``gnt-*``
855
tools). Furthermore, calculating cluster capacity via ``hspace`` again
856
required external input with regards to instance sizes.
857

    
858
In order to improve these workflows and to allow for example better
859
per-node group differentiation, we introduced instance specs, which
860
allow declaring:
861

    
862
- minimum instance disk size, disk count, memory size, cpu count
863
- maximum values for the above metrics
864
- and “standard” values (used in ``hspace`` to calculate the standard
865
  sized instances)
866

    
867
The minimum/maximum values can be also customised at node-group level,
868
for example allowing more powerful hardware to support bigger instance
869
memory sizes.
870

    
871
Beside the instance specs, there are a few other settings belonging to
872
the instance policy framework. It is possible now to customise, per
873
cluster and node-group:
874

    
875
- the list of allowed disk templates
876
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
877
- the maximum ratio of instance to spindles (see below for more
878
  information) for local storage
879

    
880
All these together should allow all tools that talk to Ganeti to know
881
what are the ranges of allowed values for instances and the
882
over-subscription that is allowed.
883

    
884
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
885
instance configuration, and the physical CPU configuration from the
886
node. For the spindle ratios however, we didn't track before these
887
values, so new parameters have been added:
888

    
889
- a new node parameter ``spindle_count``, defaults to 1, customisable at
890
  node group or node level
891
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
892

    
893
Note that spindles in this context doesn't need to mean actual
894
mechanical hard-drives; it's just a relative number for both the node
895
I/O capacity and instance I/O consumption.
896

    
897
Instance migration behaviour
898
++++++++++++++++++++++++++++
899

    
900
While live-migration is in general desirable over failover, it is
901
possible that for some workloads it is actually worse, due to the
902
variable time of the “suspend” phase during live migration.
903

    
904
To allow the tools to work consistently over such instances (without
905
having to hard-code instance names), a new backend parameter
906
``always_failover`` has been added to control the migration/failover
907
behaviour. When set to True, all migration requests for an instance will
908
instead fall-back to failover.
909

    
910
Instance memory ballooning
911
++++++++++++++++++++++++++
912

    
913
Initial support for memory ballooning has been added. The memory for an
914
instance is no longer fixed (backend parameter ``memory``), but instead
915
can vary between minimum and maximum values (backend parameters
916
``minmem`` and ``maxmem``). Currently we only change an instance's
917
memory when:
918

    
919
- live migrating or failing over and instance and the target node
920
  doesn't have enough memory
921
- user requests changing the memory via ``gnt-instance modify
922
  --runtime-memory``
923

    
924
Instance CPU pinning
925
++++++++++++++++++++
926

    
927
In order to control the use of specific CPUs by instance, support for
928
controlling CPU pinning has been added for the Xen, HVM and LXC
929
hypervisors. This is controlled by a new hypervisor parameter
930
``cpu_mask``; details about possible values for this are in the
931
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
932
VCPU-to-CPU mapping) form will work well only when all nodes in your
933
cluster have the same amount of CPUs.
934

    
935
Disk parameters
936
+++++++++++++++
937

    
938
Another area in which Ganeti was not customisable were the parameters
939
used for storage configuration, e.g. how many stripes to use for LVM,
940
DRBD resync configuration, etc.
941

    
942
To improve this area, we've added disks parameters, which are
943
customisable at cluster and node group level, and which allow to
944
specify various parameters for disks (DRBD has the most parameters
945
currently), for example:
946

    
947
- DRBD resync algorithm and parameters (e.g. speed)
948
- the default VG for meta-data volumes for DRBD
949
- number of stripes for LVM (plain disk template)
950
- the RBD pool
951

    
952
These parameters can be modified via ``gnt-cluster modify -D …`` and
953
``gnt-group modify -D …``, and are used at either instance creation (in
954
case of LVM stripes, for example) or at disk “activation” time
955
(e.g. resync speed).
956

    
957
Rados block device support
958
++++++++++++++++++++++++++
959

    
960
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
961
denoted by the ``rbd`` disk template type. This is considered
962
experimental, feedback is welcome. For details on configuring it, see
963
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
964

    
965
Master IP setup
966
+++++++++++++++
967

    
968
The existing master IP functionality works well only in simple setups (a
969
single network shared by all nodes); however, if nodes belong to
970
different networks, then the ``/32`` setup and lack of routing
971
information is not enough.
972

    
973
To allow the master IP to function well in more complex cases, the
974
system was reworked as follows:
975

    
976
- a master IP netmask setting has been added
977
- the master IP activation/turn-down code was moved from the node daemon
978
  to a separate script
979
- whether to run the Ganeti-supplied master IP script or a user-supplied
980
  on is a ``gnt-cluster init`` setting
981

    
982
Details about the location of the standard and custom setup scripts are
983
in the man page :manpage:`gnt-cluster(8)`; for information about the
984
setup script protocol, look at the Ganeti-supplied script.
985

    
986
SPICE support
987
+++++++++++++
988

    
989
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
990
improved.
991

    
992
It is now possible to use TLS-protected connections, and when renewing
993
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
994
it is now possible to specify spice or spice CA certificates. Also, it
995
is possible to configure a password for SPICE sessions via the
996
hypervisor parameter ``spice_password_file``.
997

    
998
There are also new parameters to control the compression and streaming
999
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1000
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1001
for the spice parameters.
1002

    
1003
Lastly, it is now possible to see the SPICE connection information via
1004
``gnt-instance console``.
1005

    
1006
OVF converter
1007
+++++++++++++
1008

    
1009
A new tool (``tools/ovfconverter``) has been added that supports
1010
conversion between Ganeti and the `Open Virtualization Format
1011
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1012
from).
1013

    
1014
This relies on the ``qemu-img`` tool to convert the disk formats, so the
1015
actual compatibility with other virtualization solutions depends on it.
1016

    
1017
Confd daemon changes
1018
++++++++++++++++++++
1019

    
1020
The configuration query daemon (``ganeti-confd``) is now optional, and
1021
has been rewritten in Haskell; whether to use the daemon at all, use the
1022
Python (default) or the Haskell version is selectable at configure time
1023
via the ``--enable-confd`` parameter, which can take one of the
1024
``haskell``, ``python`` or ``no`` values. If not used, disabling the
1025
daemon will result in a smaller footprint; for larger systems, we
1026
welcome feedback on the Haskell version which might become the default
1027
in future versions.
1028

    
1029
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1030
daemon running. The Python version doesn't implement the new call.
1031

    
1032

    
1033
User interface changes
1034
~~~~~~~~~~~~~~~~~~~~~~
1035

    
1036
We have replaced the ``--disks`` option of ``gnt-instance
1037
replace-disks`` with a more flexible ``--disk`` option, which allows
1038
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1039
disk size and mode can be changed upon recreation (via ``gnt-instance
1040
recreate-disks``, which accepts the same ``--disk`` option).
1041

    
1042
As many people are used to a ``show`` command, we have added that as an
1043
alias to ``info`` on all ``gnt-*`` commands.
1044

    
1045
The ``gnt-instance grow-disk`` command has a new mode in which it can
1046
accept the target size of the disk, instead of the delta; this can be
1047
more safe since two runs in absolute mode will be idempotent, and
1048
sometimes it's also easier to specify the desired size directly.
1049

    
1050
Also the handling of instances with regard to offline secondaries has
1051
been improved. Instance operations should not fail because one of it's
1052
secondary nodes is offline, even though it's safe to proceed.
1053

    
1054
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1055
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1056
minors to instance name.
1057

    
1058
API changes
1059
~~~~~~~~~~~
1060

    
1061
RAPI coverage has improved, with (for example) new resources for
1062
recreate-disks, node power-cycle, etc.
1063

    
1064
Compatibility
1065
~~~~~~~~~~~~~
1066

    
1067
There is partial support for ``xl`` in the Xen hypervisor; feedback is
1068
welcome.
1069

    
1070
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1071
whether to still support Python 2.4 or move to Python 2.6 as minimum
1072
required version.
1073

    
1074
Support for Fedora has been slightly improved; the provided example
1075
init.d script should work better on it and the INSTALL file should
1076
document the needed dependencies.
1077

    
1078
Internal changes
1079
~~~~~~~~~~~~~~~~
1080

    
1081
The deprecated ``QueryLocks`` LUXI request has been removed. Use
1082
``Query(what=QR_LOCK, ...)`` instead.
1083

    
1084
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1085
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1086
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1087
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1088
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1089

    
1090
RAPI client: ``CertificateError`` now derives from
1091
``GanetiApiError``. This should make it more easy to handle Ganeti
1092
errors.
1093

    
1094
Deprecation warnings due to PyCrypto/paramiko import in
1095
``tools/setup-ssh`` have been silenced, as usually they are safe; please
1096
make sure to run an up-to-date paramiko version, if you use this tool.
1097

    
1098
The QA scripts now depend on Python 2.5 or above (the main code base
1099
still works with Python 2.4).
1100

    
1101
The configuration file (``config.data``) is now written without
1102
indentation for performance reasons; if you want to edit it, it can be
1103
re-formatted via ``tools/fmtjson``.
1104

    
1105
A number of bugs has been fixed in the cluster merge tool.
1106

    
1107
``x509`` certification verification (used in import-export) has been
1108
changed to allow the same clock skew as permitted by the cluster
1109
verification. This will remove some rare but hard to diagnose errors in
1110
import-export.
1111

    
1112

    
1113
Version 2.6.0 rc4
1114
-----------------
1115

    
1116
*(Released Thu, 19 Jul 2012)*
1117

    
1118
Very few changes from rc4 to the final release, only bugfixes:
1119

    
1120
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
1121
  instance, fix CDROM booting for KVM instances)
1122
- fixed node group modification of node parameters
1123
- fixed issue in LUClusterVerifyGroup with multi-group clusters
1124
- fixed generation of bash completion to ensure a stable ordering
1125
- fixed a few typos
1126

    
1127

    
1128
Version 2.6.0 rc3
1129
-----------------
1130

    
1131
*(Released Fri, 13 Jul 2012)*
1132

    
1133
Third release candidate for 2.6. The following changes were done from
1134
rc3 to rc4:
1135

    
1136
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1137
- Fixed an inconsistency in the LUXI protocol with the provided
1138
  arguments (NOT backwards compatible)
1139
- Fixed a bug with node groups ipolicy where ``min`` was greater than
1140
  the cluster ``std`` value
1141
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1142
  easier instance debugging on nodes (requires ``hconfd`` to work)
1143

    
1144

    
1145
Version 2.6.0 rc2
1146
-----------------
1147

    
1148
*(Released Tue, 03 Jul 2012)*
1149

    
1150
Second release candidate for 2.6. The following changes were done from
1151
rc2 to rc3:
1152

    
1153
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1154
  master candidates
1155
- Fixed a RAPI regression on missing beparams/memory
1156
- Fixed redistribution of files on offline nodes
1157
- Added possibility to run activate-disks even though secondaries are
1158
  offline. With this change it relaxes also the strictness on some other
1159
  commands which use activate disks internally:
1160
  * ``gnt-instance start|reboot|rename|backup|export``
1161
- Made it possible to remove safely an instance if its secondaries are
1162
  offline
1163
- Made it possible to reinstall even though secondaries are offline
1164

    
1165

    
1166
Version 2.6.0 rc1
1167
-----------------
1168

    
1169
*(Released Mon, 25 Jun 2012)*
1170

    
1171
First release candidate for 2.6. The following changes were done from
1172
rc1 to rc2:
1173

    
1174
- Fixed bugs with disk parameters and ``rbd`` templates as well as
1175
  ``instance_os_add``
1176
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1177
  behaviour. It supports now the modify operation
1178
- ``hcheck`` implemented to analyze cluster health and possibility of
1179
  improving health by rebalance
1180
- ``hbal`` has been improved in dealing with split instances
1181

    
1182

    
1183
Version 2.6.0 beta2
1184
-------------------
1185

    
1186
*(Released Mon, 11 Jun 2012)*
1187

    
1188
Second beta release of 2.6. The following changes were done from beta2
1189
to rc1:
1190

    
1191
- Fixed ``daemon-util`` with non-root user models
1192
- Fixed creation of plain instances with ``--no-wait-for-sync``
1193
- Fix wrong iv_names when running ``cfgupgrade``
1194
- Export more information in RAPI group queries
1195
- Fixed bug when changing instance network interfaces
1196
- Extended burnin to do NIC changes
1197
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1198
- Changed default for DRBD barriers
1199
- Fixed DRBD error reporting for syncer rate
1200
- Verify the options on disk parameters
1201

    
1202
And of course various fixes to documentation and improved unittests and
1203
QA.
1204

    
1205

    
1206
Version 2.6.0 beta1
1207
-------------------
1208

    
1209
*(Released Wed, 23 May 2012)*
1210

    
1211
First beta release of 2.6. The following changes were done from beta1 to
1212
beta2:
1213

    
1214
- integrated patch for distributions without ``start-stop-daemon``
1215
- adapted example init.d script to work on Fedora
1216
- fixed log handling in Haskell daemons
1217
- adapted checks in the watcher for pycurl linked against libnss
1218
- add partial support for ``xl`` instead of ``xm`` for Xen
1219
- fixed a type issue in cluster verification
1220
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1221
  clusters)
1222

    
1223
Plus integrated fixes from the 2.5 branch:
1224

    
1225
- fixed ``kvm-ifup`` to use ``/bin/bash``
1226
- fixed parallel build failures
1227
- KVM live migration when using a custom keymap
1228

    
1229

    
1230
Version 2.5.2
1231
-------------
1232

    
1233
*(Released Tue, 24 Jul 2012)*
1234

    
1235
A small bugfix release, with no new features:
1236

    
1237
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1238
  different default shell (e.g. Debian, Ubuntu)
1239
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1240
  243 and Debian bug #650664)
1241
- fixed compatibility with KVM versions that don't support multiple boot
1242
  devices (fixes Issue 230 and Debian bug #624256)
1243

    
1244
Additionally, a few fixes were done to the build system (fixed parallel
1245
build failures) and to the unittests (fixed race condition in test for
1246
FileID functions, and the default enable/disable mode for QA test is now
1247
customisable).
1248

    
1249

    
1250
Version 2.5.1
1251
-------------
1252

    
1253
*(Released Fri, 11 May 2012)*
1254

    
1255
A small bugfix release.
1256

    
1257
The main issues solved are on the topic of compatibility with newer LVM
1258
releases:
1259

    
1260
- fixed parsing of ``lv_attr`` field
1261
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1262
  the ``--force`` flag is needed
1263

    
1264
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1265
to accept kernel 3.x too (was hardcoded to 2.6.*).
1266

    
1267
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1268
of hook results and that also made display of other errors suboptimal
1269
was fixed; the code behaves now like 2.4 and earlier.
1270

    
1271
Another change in 2.5, the cleanup of the OS scripts environment, is too
1272
aggressive: it removed even the ``PATH`` variable, which requires the OS
1273
scripts to *always* need to export it. Since this is a bit too strict,
1274
we now export a minimal PATH, the same that we export for hooks.
1275

    
1276
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1277
integrated into this release.
1278

    
1279
Finally, a few other miscellaneous changes were done (no new features,
1280
just small improvements):
1281

    
1282
- Fix ``gnt-group --help`` display
1283
- Fix hardcoded Xen kernel path
1284
- Fix grow-disk handling of invalid units
1285
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1286
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1287

    
1288

    
1289
Version 2.5.0
1290
-------------
1291

    
1292
*(Released Thu, 12 Apr 2012)*
1293

    
1294
Incompatible/important changes and bugfixes
1295
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1296

    
1297
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1298
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1299
  to match the underlying LUXI interface.
1300
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1301
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1302
  not maintain backwards-compatibility as the underlying operation
1303
  changed in an incompatible way. The RAPI client can talk to old
1304
  servers, but it needs to be told so as the return value changed.
1305
- When creating file-based instances via RAPI, the ``file_driver``
1306
  parameter no longer defaults to ``loop`` and must be specified.
1307
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1308
  ``link`` instead.
1309
- Support for the undocumented and deprecated RAPI instance creation
1310
  request format version 0 has been dropped. Use version 1, supported
1311
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1312
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1313
  documentation <install>`.
1314
- The "cluster-verify" hooks are now executed per group by the
1315
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1316
  if you just run ``gnt-cluster verify``, which generates one opcode per
1317
  group.
1318
- The environment as passed to the OS scripts is cleared, and thus no
1319
  environment variables defined in the node daemon's environment will be
1320
  inherited by the scripts.
1321
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1322
  deprecated.
1323
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1324
  support operations involving multiple node groups.
1325
- Offline nodes are ignored when failing over an instance.
1326
- Support for KVM version 1.0, which changed the version reporting format
1327
  from 3 to 2 digits.
1328
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1329
  removal.
1330
- ``Makefile`` is now compatible with Automake 1.11.2
1331
- Includes all bugfixes made in the 2.4 series
1332

    
1333
New features
1334
~~~~~~~~~~~~
1335

    
1336
- The ganeti-htools project has been merged into the ganeti-core source
1337
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1338
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1339
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1340
  GPT-style partition tables (requires `parted
1341
  <http://www.gnu.org/s/parted/>`_).
1342
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1343
- Allowed adding tags on instance creation.
1344
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1345
- Allow instances to be started in a paused state, enabling the user to
1346
  see the complete console output on boot using the console.
1347
- Added new hypervisor flag to control default reboot behaviour
1348
  (``reboot_behavior``).
1349
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1350
- Improved out-of-band management support:
1351

    
1352
  - Added ``gnt-node health`` command reporting the health status of
1353
    nodes.
1354
  - Added ``gnt-node power`` command to manage power status of nodes.
1355
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1356

    
1357
- Instance migration can fall back to failover if instance is not
1358
  running.
1359
- Filters can be used when listing nodes, instances, groups and locks;
1360
  see :manpage:`ganeti(7)` manpage.
1361
- Added post-execution status as variables to :doc:`hooks <hooks>`
1362
  environment.
1363
- Instance tags are exported/imported together with the instance.
1364
- When given an explicit job ID, ``gnt-job info`` will work for archived
1365
  jobs.
1366
- Jobs can define dependencies on other jobs (not yet supported via
1367
  RAPI or command line, but used by internal commands and usable via
1368
  LUXI).
1369

    
1370
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1371
    dependencies.
1372

    
1373
- Instance failover is now available as a RAPI resource
1374
  (``/2/instances/[instance_name]/failover``).
1375
- ``gnt-instance info`` defaults to static information if primary node
1376
  is offline.
1377
- Opcodes have a new ``comment`` attribute.
1378
- Added basic SPICE support to KVM hypervisor.
1379
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1380

    
1381
Node group improvements
1382
~~~~~~~~~~~~~~~~~~~~~~~
1383

    
1384
- ``gnt-cluster verify`` has been modified to check groups separately,
1385
  thereby improving performance.
1386
- Node group support has been added to ``gnt-cluster verify-disks``,
1387
  which now operates per node group.
1388
- Watcher has been changed to work better with node groups.
1389

    
1390
  - One process and state file per node group.
1391
  - Slow watcher in one group doesn't block other group's watcher.
1392

    
1393
- Added new command, ``gnt-group evacuate``, to move all instances in a
1394
  node group to other groups.
1395
- Added ``gnt-instance change-group`` to move an instance to another
1396
  node group.
1397
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1398
  per-group operations.
1399
- Node groups can be tagged.
1400
- Some operations switch from an exclusive to a shared lock as soon as
1401
  possible.
1402
- Instance's primary and secondary nodes' groups are now available as
1403
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1404
  and ``snodes.group.uuid``).
1405

    
1406
Misc
1407
~~~~
1408

    
1409
- Numerous updates to documentation and manpages.
1410

    
1411
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1412
    descriptions.
1413
  - Some opcode/job results are now also documented, see :doc:`RAPI
1414
    <rapi>`.
1415

    
1416
- A lockset's internal lock is now also visible in lock monitor.
1417
- Log messages from job queue workers now contain information about the
1418
  opcode they're processing.
1419
- ``gnt-instance console`` no longer requires the instance lock.
1420
- A short delay when waiting for job changes reduces the number of LUXI
1421
  requests significantly.
1422
- DRBD metadata volumes are overwritten with zeros during disk creation.
1423
- Out-of-band commands no longer acquire the cluster lock in exclusive
1424
  mode.
1425
- ``devel/upload`` now uses correct permissions for directories.
1426

    
1427

    
1428
Version 2.5.0 rc6
1429
-----------------
1430

    
1431
*(Released Fri, 23 Mar 2012)*
1432

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

    
1435

    
1436
Version 2.5.0 rc5
1437
-----------------
1438

    
1439
*(Released Mon, 9 Jan 2012)*
1440

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

    
1443

    
1444
Version 2.5.0 rc4
1445
-----------------
1446

    
1447
*(Released Thu, 27 Oct 2011)*
1448

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

    
1451

    
1452
Version 2.5.0 rc3
1453
-----------------
1454

    
1455
*(Released Wed, 26 Oct 2011)*
1456

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

    
1459

    
1460
Version 2.5.0 rc2
1461
-----------------
1462

    
1463
*(Released Tue, 18 Oct 2011)*
1464

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

    
1467

    
1468
Version 2.5.0 rc1
1469
-----------------
1470

    
1471
*(Released Tue, 4 Oct 2011)*
1472

    
1473
This was the first release candidate of the 2.5 series.
1474

    
1475

    
1476
Version 2.5.0 beta3
1477
-------------------
1478

    
1479
*(Released Wed, 31 Aug 2011)*
1480

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

    
1483

    
1484
Version 2.5.0 beta2
1485
-------------------
1486

    
1487
*(Released Mon, 22 Aug 2011)*
1488

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

    
1491

    
1492
Version 2.5.0 beta1
1493
-------------------
1494

    
1495
*(Released Fri, 12 Aug 2011)*
1496

    
1497
This was the first beta release of the 2.5 series.
1498

    
1499

    
1500
Version 2.4.5
1501
-------------
1502

    
1503
*(Released Thu, 27 Oct 2011)*
1504

    
1505
- Fixed bug when parsing command line parameter values ending in
1506
  backslash
1507
- Fixed assertion error after unclean master shutdown
1508
- Disable HTTP client pool for RPC, significantly reducing memory usage
1509
  of master daemon
1510
- Fixed queue archive creation with wrong permissions
1511

    
1512

    
1513
Version 2.4.4
1514
-------------
1515

    
1516
*(Released Tue, 23 Aug 2011)*
1517

    
1518
Small bug-fixes:
1519

    
1520
- Fixed documentation for importing with ``--src-dir`` option
1521
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1522
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1523

    
1524

    
1525
Version 2.4.3
1526
-------------
1527

    
1528
*(Released Fri, 5 Aug 2011)*
1529

    
1530
Many bug-fixes and a few small features:
1531

    
1532
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1533
  issues when you tried to add an instance with two MAC addresses in one
1534
  request
1535
- KVM: fixed per-instance stored UID value
1536
- KVM: configure bridged NICs at migration start
1537
- KVM: Fix a bug where instance will not start with never KVM versions
1538
  (>= 0.14)
1539
- Added OS search path to ``gnt-cluster info``
1540
- Fixed an issue with ``file_storage_dir`` where you were forced to
1541
  provide an absolute path, but the documentation states it is a
1542
  relative path, the documentation was right
1543
- Added a new parameter to instance stop/start called ``--no-remember``
1544
  that will make the state change to not be remembered
1545
- Implemented ``no_remember`` at RAPI level
1546
- Improved the documentation
1547
- Node evacuation: don't call IAllocator if node is already empty
1548
- Fixed bug in DRBD8 replace disks on current nodes
1549
- Fixed bug in recreate-disks for DRBD instances
1550
- Moved assertion checking locks in ``gnt-instance replace-disks``
1551
  causing it to abort with not owning the right locks for some situation
1552
- Job queue: Fixed potential race condition when cancelling queued jobs
1553
- Fixed off-by-one bug in job serial generation
1554
- ``gnt-node volumes``: Fix instance names
1555
- Fixed aliases in bash completion
1556
- Fixed a bug in reopening log files after being sent a SIGHUP
1557
- Added a flag to burnin to allow specifying VCPU count
1558
- Bugfixes to non-root Ganeti configuration
1559

    
1560

    
1561
Version 2.4.2
1562
-------------
1563

    
1564
*(Released Thu, 12 May 2011)*
1565

    
1566
Many bug-fixes and a few new small features:
1567

    
1568
- Fixed a bug related to log opening failures
1569
- Fixed a bug in instance listing with orphan instances
1570
- Fixed a bug which prevented resetting the cluster-level node parameter
1571
  ``oob_program`` to the default
1572
- Many fixes related to the ``cluster-merge`` tool
1573
- Fixed a race condition in the lock monitor, which caused failures
1574
  during (at least) creation of many instances in parallel
1575
- Improved output for gnt-job info
1576
- Removed the quiet flag on some ssh calls which prevented debugging
1577
  failures
1578
- Improved the N+1 failure messages in cluster verify by actually
1579
  showing the memory values (needed and available)
1580
- Increased lock attempt timeouts so that when executing long operations
1581
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1582
  too early and thus prevent the use of the 'fair' mechanism
1583
- Changed instance query data (``gnt-instance info``) to not acquire
1584
  locks unless needed, thus allowing its use on locked instance if only
1585
  static information is asked for
1586
- Improved behaviour with filesystems that do not support rename on an
1587
  opened file
1588
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1589
  kept locks on all nodes during the wipe, which is unneeded
1590
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1591
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1592
  10GiB) which caused the wipe to fail right at the end in some cases
1593
- Fixed master IP activation when doing master failover with no-voting
1594
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1595
  the master node itself
1596
- Fixed potential data-loss in under disk full conditions, where Ganeti
1597
  wouldn't check correctly the return code and would consider
1598
  partially-written files 'correct'
1599
- Fixed bug related to multiple VGs and DRBD disk replacing
1600
- Added new disk parameter ``metavg`` that allows placement of the meta
1601
  device for DRBD in a different volume group
1602
- Fixed error handling in the node daemon when the system libc doesn't
1603
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1604
- Fixed lock release during replace-disks, which kept cluster-wide locks
1605
  when doing disk replaces with an iallocator script
1606
- Added check for missing bridges in cluster verify
1607
- Handle EPIPE errors while writing to the terminal better, so that
1608
  piping the output to e.g. ``less`` doesn't cause a backtrace
1609
- Fixed rare case where a ^C during Luxi calls could have been
1610
  interpreted as server errors, instead of simply terminating
1611
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1612
  assign-nodes``)
1613
- Added a few more parameters to the KVM hypervisor, allowing a second
1614
  CDROM, custom disk type for CDROMs and a floppy image
1615
- Removed redundant message in instance rename when the name is given
1616
  already as a FQDN
1617
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1618
  disks on new nodes, allowing recreation when the original instance
1619
  nodes are completely gone
1620
- Added option when converting disk templates to DRBD to skip waiting
1621
  for the resync, in order to make the instance available sooner
1622
- Added two new variables to the OS scripts environment (containing the
1623
  instance's nodes)
1624
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1625
  to allow use of ``pvgrub`` as bootloader
1626
- Changed the instance memory modifications to only check out-of-memory
1627
  conditions on memory increases, and turned the secondary node warnings
1628
  into errors (they can still be overridden via ``--force``)
1629
- Fixed the handling of a corner case when the Python installation gets
1630
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1631
  to execute a command that doesn't exist
1632
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1633
  node of the instance returned failures during instance shutdown; this
1634
  adds the option ``--ignore-consistency`` to gnt-instance move
1635

    
1636
And as usual, various improvements to the error messages, documentation
1637
and man pages.
1638

    
1639

    
1640
Version 2.4.1
1641
-------------
1642

    
1643
*(Released Wed, 09 Mar 2011)*
1644

    
1645
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1646
the RAPI users file if run twice (even with ``--dry-run``).
1647

    
1648
The release fixes that bug (nothing else changed).
1649

    
1650

    
1651
Version 2.4.0
1652
-------------
1653

    
1654
*(Released Mon, 07 Mar 2011)*
1655

    
1656
Final 2.4.0 release. Just a few small fixes:
1657

    
1658
- Fixed RAPI node evacuate
1659
- Fixed the kvm-ifup script
1660
- Fixed internal error handling for special job cases
1661
- Updated man page to specify the escaping feature for options
1662

    
1663

    
1664
Version 2.4.0 rc3
1665
-----------------
1666

    
1667
*(Released Mon, 28 Feb 2011)*
1668

    
1669
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1670
that this feature wiped the disks of the wrong instance, leading to loss
1671
of data.
1672

    
1673
Other changes:
1674

    
1675
- Fixed title of query field containing instance name
1676
- Expanded the glossary in the documentation
1677
- Fixed one unittest (internal issue)
1678

    
1679

    
1680
Version 2.4.0 rc2
1681
-----------------
1682

    
1683
*(Released Mon, 21 Feb 2011)*
1684

    
1685
A number of bug fixes plus just a couple functionality changes.
1686

    
1687
On the user-visible side, the ``gnt-* list`` command output has changed
1688
with respect to "special" field states. The current rc1 style of display
1689
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1690
the default output mode special fields states are displayed as follows:
1691

    
1692
- Offline resource: ``*``
1693
- Unavailable/not applicable: ``-``
1694
- Data missing (RPC failure): ``?``
1695
- Unknown field: ``??``
1696

    
1697
Another user-visible change is the addition of ``--force-join`` to
1698
``gnt-node add``.
1699

    
1700
As for bug fixes:
1701

    
1702
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1703
- Fixed regression in RAPI/instance reinstall where all parameters were
1704
  required (instead of optional)
1705
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1706
- Fixed iallocator usage (offline nodes were not considered offline)
1707
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1708
- Fixed hypervisor and OS parameter validation with respect to
1709
  non-vm_capable nodes
1710
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1711
  cosmetic)
1712
- Fixed ``tools/listrunner`` with respect to agent-based usage
1713

    
1714

    
1715
Version 2.4.0 rc1
1716
-----------------
1717

    
1718
*(Released Fri,  4 Feb 2011)*
1719

    
1720
Many changes and fixes since the beta1 release. While there were some
1721
internal changes, the code has been mostly stabilised for the RC
1722
release.
1723

    
1724
Note: the dumb allocator was removed in this release, as it was not kept
1725
up-to-date with the IAllocator protocol changes. It is recommended to
1726
use the ``hail`` command from the ganeti-htools package.
1727

    
1728
Note: the 2.4 and up versions of Ganeti are not compatible with the
1729
0.2.x branch of ganeti-htools. You need to upgrade to
1730
ganeti-htools-0.3.0 (or later).
1731

    
1732
Regressions fixed from 2.3
1733
~~~~~~~~~~~~~~~~~~~~~~~~~~
1734

    
1735
- Fixed the ``gnt-cluster verify-disks`` command
1736
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1737
  serially on nodes)
1738
- Fixed disk adoption breakage
1739
- Fixed wrong headers in instance listing for field aliases
1740

    
1741
Other bugs fixed
1742
~~~~~~~~~~~~~~~~
1743

    
1744
- Fixed corner case in KVM handling of NICs
1745
- Fixed many cases of wrong handling of non-vm_capable nodes
1746
- Fixed a bug where a missing instance symlink was not possible to
1747
  recreate with any ``gnt-*`` command (now ``gnt-instance
1748
  activate-disks`` does it)
1749
- Fixed the volume group name as reported by ``gnt-cluster
1750
  verify-disks``
1751
- Increased timeouts for the import-export code, hopefully leading to
1752
  fewer aborts due network or instance timeouts
1753
- Fixed bug in ``gnt-node list-storage``
1754
- Fixed bug where not all daemons were started on cluster
1755
  initialisation, but only at the first watcher run
1756
- Fixed many bugs in the OOB implementation
1757
- Fixed watcher behaviour in presence of instances with offline
1758
  secondaries
1759
- Fixed instance list output for instances running on the wrong node
1760
- a few fixes to the cluster-merge tool, but it still cannot merge
1761
  multi-node groups (currently it is not recommended to use this tool)
1762

    
1763

    
1764
Improvements
1765
~~~~~~~~~~~~
1766

    
1767
- Improved network configuration for the KVM hypervisor
1768
- Added e1000 as a supported NIC for Xen-HVM
1769
- Improved the lvmstrap tool to also be able to use partitions, as
1770
  opposed to full disks
1771
- Improved speed of disk wiping (the cluster parameter
1772
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1773
  of instance creations
1774
- Added documentation for the OS parameters
1775
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1776
  hypervisor is not responding
1777
- Added display of blacklisted and hidden OS information in
1778
  ``gnt-cluster info``
1779
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1780
  NIC and node parameters, which might create problems with currently
1781
  invalid (but undetected) configuration files, but prevents validation
1782
  failures when unrelated parameters are modified
1783
- Changed cluster initialisation to wait for the master daemon to become
1784
  available
1785
- Expanded the RAPI interface:
1786

    
1787
  - Added config redistribution resource
1788
  - Added activation/deactivation of instance disks
1789
  - Added export of console information
1790

    
1791
- Implemented log file reopening on SIGHUP, which allows using
1792
  logrotate(8) for the Ganeti log files
1793
- Added a basic OOB helper script as an example
1794

    
1795

    
1796
Version 2.4.0 beta1
1797
-------------------
1798

    
1799
*(Released Fri, 14 Jan 2011)*
1800

    
1801
User-visible
1802
~~~~~~~~~~~~
1803

    
1804
- Fixed timezone issues when formatting timestamps
1805
- Added support for node groups, available via ``gnt-group`` and other
1806
  commands
1807
- Added out-of-band framework and management, see :doc:`design
1808
  document <design-oob>`
1809
- Removed support for roman numbers from ``gnt-node list`` and
1810
  ``gnt-instance list``.
1811
- Allowed modification of master network interface via ``gnt-cluster
1812
  modify --master-netdev``
1813
- Accept offline secondaries while shutting down instance disks
1814
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1815
- Added support for multiple LVM volume groups
1816
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1817
  requested
1818
- Added commands to list available fields:
1819

    
1820
  - ``gnt-node list-fields``
1821
  - ``gnt-group list-fields``
1822
  - ``gnt-instance list-fields``
1823

    
1824
- Updated documentation and man pages
1825

    
1826
Integration
1827
~~~~~~~~~~~
1828

    
1829
- Moved ``rapi_users`` file into separate directory, now named
1830
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1831
  symlink
1832
- Added new tool for running commands on many machines,
1833
  ``tools/ganeti-listrunner``
1834
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1835
  improving the ``gnt-instance console`` output
1836
- Allowed customisation of disk index separator at ``configure`` time
1837
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1838
- Added support for non-partitioned md disks in ``lvmstrap``
1839
- Added script to gracefully power off KVM instances
1840
- Split ``utils`` module into smaller parts
1841
- Changed query operations to return more detailed information, e.g.
1842
  whether an information is unavailable due to an offline node. To use
1843
  this new functionality, the LUXI call ``Query`` must be used. Field
1844
  information is now stored by the master daemon and can be retrieved
1845
  using ``QueryFields``. Instances, nodes and groups can also be queried
1846
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1847
  exposed via RAPI). The following commands make use of this
1848
  infrastructure change:
1849

    
1850
  - ``gnt-group list``
1851
  - ``gnt-group list-fields``
1852
  - ``gnt-node list``
1853
  - ``gnt-node list-fields``
1854
  - ``gnt-instance list``
1855
  - ``gnt-instance list-fields``
1856
  - ``gnt-debug locks``
1857

    
1858
Remote API
1859
~~~~~~~~~~
1860

    
1861
- New RAPI resources (see :doc:`rapi`):
1862

    
1863
  - ``/2/modify``
1864
  - ``/2/groups``
1865
  - ``/2/groups/[group_name]``
1866
  - ``/2/groups/[group_name]/assign-nodes``
1867
  - ``/2/groups/[group_name]/modify``
1868
  - ``/2/groups/[group_name]/rename``
1869
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1870

    
1871
- RAPI changes:
1872

    
1873
  - Implemented ``no_install`` for instance creation
1874
  - Implemented OS parameters for instance reinstallation, allowing
1875
    use of special settings on reinstallation (e.g. for preserving data)
1876

    
1877
Misc
1878
~~~~
1879

    
1880
- Added IPv6 support in import/export
1881
- Pause DRBD synchronization while wiping disks on instance creation
1882
- Updated unittests and QA scripts
1883
- Improved network parameters passed to KVM
1884
- Converted man pages from docbook to reStructuredText
1885

    
1886

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

    
1890
*(Released Mon, 20 Dec 2010)*
1891

    
1892
Released version 2.3.1~rc1 without any changes.
1893

    
1894

    
1895
Version 2.3.1 rc1
1896
-----------------
1897

    
1898
*(Released Wed, 1 Dec 2010)*
1899

    
1900
- impexpd: Disable OpenSSL compression in socat if possible (backport
1901
  from master, commit e90739d625b, see :doc:`installation guide
1902
  <install-quick>` for details)
1903
- Changed unittest coverage report to exclude test scripts
1904
- Added script to check version format
1905

    
1906

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

    
1910
*(Released Wed, 1 Dec 2010)*
1911

    
1912
Released version 2.3.0~rc1 without any changes.
1913

    
1914

    
1915
Version 2.3.0 rc1
1916
-----------------
1917

    
1918
*(Released Fri, 19 Nov 2010)*
1919

    
1920
A number of bugfixes and documentation updates:
1921

    
1922
- Update ganeti-os-interface documentation
1923
- Fixed a bug related to duplicate MACs or similar items which should be
1924
  unique
1925
- Fix breakage in OS state modify
1926
- Reinstall instance: disallow offline secondaries (fixes bug related to
1927
  OS changing but reinstall failing)
1928
- plus all the other fixes between 2.2.1 and 2.2.2
1929

    
1930

    
1931
Version 2.3.0 rc0
1932
-----------------
1933

    
1934
*(Released Tue, 2 Nov 2010)*
1935

    
1936
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1937
- Fixed master failover race with watcher
1938
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1939
  configuration
1940
- Accept previously stopped instance for export with instance removal
1941
- Simplify and extend the environment variables for instance OS scripts
1942
- Added new node flags, ``master_capable`` and ``vm_capable``
1943
- Added optional instance disk wiping prior during allocation. This is a
1944
  cluster-wide option and can be set/modified using
1945
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1946
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1947
  :doc:`install-quick`
1948
- Added a new watcher option (``--ignore-pause``)
1949
- Added option to ignore offline node on instance start/stop
1950
  (``--ignore-offline``)
1951
- Allow overriding OS parameters with ``gnt-instance reinstall``
1952
- Added ability to change node's secondary IP address using ``gnt-node
1953
  modify``
1954
- Implemented privilege separation for all daemons except
1955
  ``ganeti-noded``, see ``configure`` options
1956
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1957
  verify``
1958
- Implemented job priorities (see ``ganeti(7)`` manpage)
1959
- Ignore failures while shutting down instances during failover from
1960
  offline node
1961
- Exit daemon's bootstrap process only once daemon is ready
1962
- Export more information via ``LUInstanceQuery``/remote API
1963
- Improved documentation, QA and unittests
1964
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1965
  restart if the file was created or changed
1966
- Added LUXI protocol version sent with each request and response,
1967
  allowing detection of server/client mismatches
1968
- Moved the Python scripts among gnt-* and ganeti-* into modules
1969
- Moved all code related to setting up SSH to an external script,
1970
  ``setup-ssh``
1971
- Infrastructure changes for node group support in future versions
1972

    
1973

    
1974
Version 2.2.2
1975
-------------
1976

    
1977
*(Released Fri, 19 Nov 2010)*
1978

    
1979
A few small bugs fixed, and some improvements to the build system:
1980

    
1981
- Fix documentation regarding conversion to drbd
1982
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1983
  -B``)
1984
- Fix error handling in node modify with multiple changes
1985
- Allow remote imports without checked names
1986

    
1987

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

    
1991
*(Released Tue, 19 Oct 2010)*
1992

    
1993
- Disable SSL session ID cache in RPC client
1994

    
1995

    
1996
Version 2.2.1 rc1
1997
-----------------
1998

    
1999
*(Released Thu, 14 Oct 2010)*
2000

    
2001
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
2002
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2003
  with GnuTLS
2004
- Fix problems with interaction between Curl and Python's HTTP server,
2005
  resulting in increased speed in many RPC calls
2006
- Improve our release script to prevent breakage with older aclocal and
2007
  Python 2.6
2008

    
2009

    
2010
Version 2.2.1 rc0
2011
-----------------
2012

    
2013
*(Released Thu, 7 Oct 2010)*
2014

    
2015
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2016
  value retrieved from master
2017
- Added support for blacklisted or hidden OS definitions
2018
- Added simple lock monitor (accessible via (``gnt-debug locks``)
2019
- Added support for -mem-path in KVM hypervisor abstraction layer
2020
- Allow overriding instance parameters in tool for inter-cluster
2021
  instance moves (``tools/move-instance``)
2022
- Improved opcode summaries (e.g. in ``gnt-job list``)
2023
- Improve consistency of OS listing by sorting it
2024
- Documentation updates
2025

    
2026

    
2027
Version 2.2.0.1
2028
---------------
2029

    
2030
*(Released Fri, 8 Oct 2010)*
2031

    
2032
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
2033

    
2034

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

    
2038
*(Released Mon, 4 Oct 2010)*
2039

    
2040
- Fixed regression in ``gnt-instance rename``
2041

    
2042

    
2043
Version 2.2.0 rc2
2044
-----------------
2045

    
2046
*(Released Wed, 22 Sep 2010)*
2047

    
2048
- Fixed OS_VARIANT variable for OS scripts
2049
- Fixed cluster tag operations via RAPI
2050
- Made ``setup-ssh`` exit with non-zero code if an error occurred
2051
- Disabled RAPI CA checks in watcher
2052

    
2053

    
2054
Version 2.2.0 rc1
2055
-----------------
2056

    
2057
*(Released Mon, 23 Aug 2010)*
2058

    
2059
- Support DRBD versions of the format "a.b.c.d"
2060
- Updated manpages
2061
- Re-introduce support for usage from multiple threads in RAPI client
2062
- Instance renames and modify via RAPI
2063
- Work around race condition between processing and archival in job
2064
  queue
2065
- Mark opcodes following failed one as failed, too
2066
- Job field ``lock_status`` was removed due to difficulties making it
2067
  work with the changed job queue in Ganeti 2.2; a better way to monitor
2068
  locks is expected for a later 2.2.x release
2069
- Fixed dry-run behaviour with many commands
2070
- Support ``ssh-agent`` again when adding nodes
2071
- Many additional bugfixes
2072

    
2073

    
2074
Version 2.2.0 rc0
2075
-----------------
2076

    
2077
*(Released Fri, 30 Jul 2010)*
2078

    
2079
Important change: the internal RPC mechanism between Ganeti nodes has
2080
changed from using a home-grown http library (based on the Python base
2081
libraries) to use the PycURL library. This requires that PycURL is
2082
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2083
against GnuTLS by default. cURL's support for GnuTLS had known issues
2084
before cURL 7.21.0 and we recommend using the latest cURL release or
2085
linking against OpenSSL. Most other distributions already link PycURL
2086
and cURL against OpenSSL. The command::
2087

    
2088
  python -c 'import pycurl; print pycurl.version'
2089

    
2090
can be used to determine the libraries PycURL and cURL are linked
2091
against.
2092

    
2093
Other significant changes:
2094

    
2095
- Rewrote much of the internals of the job queue, in order to achieve
2096
  better parallelism; this decouples job query operations from the job
2097
  processing, and it should allow much nicer behaviour of the master
2098
  daemon under load, and it also has uncovered some long-standing bugs
2099
  related to the job serialisation (now fixed)
2100
- Added a default iallocator setting to the cluster parameters,
2101
  eliminating the need to always pass nodes or an iallocator for
2102
  operations that require selection of new node(s)
2103
- Added experimental support for the LXC virtualization method
2104
- Added support for OS parameters, which allows the installation of
2105
  instances to pass parameter to OS scripts in order to customise the
2106
  instance
2107
- Added a hypervisor parameter controlling the migration type (live or
2108
  non-live), since hypervisors have various levels of reliability; this
2109
  has renamed the 'live' parameter to 'mode'
2110
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
2111
  logical volumes, meaning that cluster verify will ignore them and not
2112
  flag their presence as errors
2113
- The watcher will now reset the error count for failed instances after
2114
  8 hours, thus allowing self-healing if the problem that caused the
2115
  instances to be down/fail to start has cleared in the meantime
2116
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2117
  check for, and warn, if the drbd module parameter ``usermode_helper``
2118
  is not consistent with the cluster-wide setting; this is needed to
2119
  make diagnose easier of failed drbd creations
2120
- Started adding base IPv6 support, but this is not yet
2121
  enabled/available for use
2122
- Rename operations (cluster, instance) will now return the new name,
2123
  which is especially useful if a short name was passed in
2124
- Added support for instance migration in RAPI
2125
- Added a tool to pre-configure nodes for the SSH setup, before joining
2126
  them to the cluster; this will allow in the future a simplified model
2127
  for node joining (but not yet fully enabled in 2.2); this needs the
2128
  paramiko python library
2129
- Fixed handling of name-resolving errors
2130
- Fixed consistency of job results on the error path
2131
- Fixed master-failover race condition when executed multiple times in
2132
  sequence
2133
- Fixed many bugs related to the job queue (mostly introduced during the
2134
  2.2 development cycle, so not all are impacting 2.1)
2135
- Fixed instance migration with missing disk symlinks
2136
- Fixed handling of unknown jobs in ``gnt-job archive``
2137
- And many other small fixes/improvements
2138

    
2139
Internal changes:
2140

    
2141
- Enhanced both the unittest and the QA coverage
2142
- Switched the opcode validation to a generic model, and extended the
2143
  validation to all opcode parameters
2144
- Changed more parts of the code that write shell scripts to use the
2145
  same class for this
2146
- Switched the master daemon to use the asyncore library for the Luxi
2147
  server endpoint
2148

    
2149

    
2150
Version 2.2.0 beta0
2151
-------------------
2152

    
2153
*(Released Thu, 17 Jun 2010)*
2154

    
2155
- Added tool (``move-instance``) and infrastructure to move instances
2156
  between separate clusters (see :doc:`separate documentation
2157
  <move-instance>` and :doc:`design document <design-2.2>`)
2158
- Added per-request RPC timeout
2159
- RAPI now requires a Content-Type header for requests with a body (e.g.
2160
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2161
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
2162
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2163
  reachable
2164
- Implemented initial support for running Ganeti daemons as separate
2165
  users, see configure-time flags ``--with-user-prefix`` and
2166
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2167
  time)
2168
- Instances can be removed after export (``gnt-backup export
2169
  --remove-instance``)
2170
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2171
  key (instead of 1024 bit)
2172
- Added new cluster configuration file for cluster domain secret
2173
- Import/export now use SSL instead of SSH
2174
- Added support for showing estimated time when exporting an instance,
2175
  see the ``ganeti-os-interface(7)`` manpage and look for
2176
  ``EXP_SIZE_FD``
2177

    
2178

    
2179
Version 2.1.8
2180
-------------
2181

    
2182
*(Released Tue, 16 Nov 2010)*
2183

    
2184
Some more bugfixes. Unless critical bugs occur, this will be the last
2185
2.1 release:
2186

    
2187
- Fix case of MAC special-values
2188
- Fix mac checker regex
2189
- backend: Fix typo causing "out of range" error
2190
- Add missing --units in gnt-instance list man page
2191

    
2192

    
2193
Version 2.1.7
2194
-------------
2195

    
2196
*(Released Tue, 24 Aug 2010)*
2197

    
2198
Bugfixes only:
2199
  - Don't ignore secondary node silently on non-mirrored disk templates
2200
    (issue 113)
2201
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2202
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2203
  - Properly document the usb_mouse parameter
2204
  - Fix path in ganeti-rapi(8) (issue 116)
2205
  - Adjust error message when the ganeti user's .ssh directory is
2206
    missing
2207
  - Add same-node-check when changing the disk template to drbd
2208

    
2209

    
2210
Version 2.1.6
2211
-------------
2212

    
2213
*(Released Fri, 16 Jul 2010)*
2214

    
2215
Bugfixes only:
2216
  - Add an option to only select some reboot types during qa/burnin.
2217
    (on some hypervisors consequent reboots are not supported)
2218
  - Fix infrequent race condition in master failover. Sometimes the old
2219
    master ip address would be still detected as up for a short time
2220
    after it was removed, causing failover to fail.
2221
  - Decrease mlockall warnings when the ctypes module is missing. On
2222
    Python 2.4 we support running even if no ctypes module is installed,
2223
    but we were too verbose about this issue.
2224
  - Fix building on old distributions, on which man doesn't have a
2225
    --warnings option.
2226
  - Fix RAPI not to ignore the MAC address on instance creation
2227
  - Implement the old instance creation format in the RAPI client.
2228

    
2229

    
2230
Version 2.1.5
2231
-------------
2232

    
2233
*(Released Thu, 01 Jul 2010)*
2234

    
2235
A small bugfix release:
2236
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2237
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2238
    a non-existing option
2239
  - Fix instance create: the --force-variant option was ignored
2240
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2241
  - Fix modify node storage with non-FQDN arguments
2242
  - Fix RAPI client to authenticate under Python 2.6 when used
2243
    for more than 5 requests needing authentication
2244
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2245
    when converting a non-shutdown drbd instance to plain
2246

    
2247

    
2248
Version 2.1.4
2249
-------------
2250

    
2251
*(Released Fri, 18 Jun 2010)*
2252

    
2253
A small bugfix release:
2254

    
2255
  - Fix live migration of KVM instances started with older Ganeti
2256
    versions which had fewer hypervisor parameters
2257
  - Fix gnt-instance grow-disk on down instances
2258
  - Fix an error-reporting bug during instance migration
2259
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2260
    silently ignoring broken ones
2261
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2262
    users
2263
  - Fix bug triggered by different API version os-es on different nodes
2264
  - Fix a bug in instance startup with custom hvparams: OS level
2265
    parameters would fail to be applied.
2266
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2267
    make it work completely well with OpenSSL)
2268
  - Fix handling of errors when resolving names from DNS
2269

    
2270

    
2271
Version 2.1.3
2272
-------------
2273

    
2274
*(Released Thu, 3 Jun 2010)*
2275

    
2276
A medium sized development cycle. Some new features, and some
2277
fixes/small improvements/cleanups.
2278

    
2279
Significant features
2280
~~~~~~~~~~~~~~~~~~~~
2281

    
2282
The node deamon now tries to mlock itself into memory, unless the
2283
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2284
its logs, and falls back to console logging. This allows emergency
2285
features such as ``gnt-node powercycle`` to work even in the event of a
2286
broken node disk (tested offlining the disk hosting the node's
2287
filesystem and dropping its memory caches; don't try this at home)
2288

    
2289
KVM: add vhost-net acceleration support. It can be tested with a new
2290
enough version of the kernel and of qemu-kvm.
2291

    
2292
KVM: Add instance chrooting feature. If you use privilege dropping for
2293
your VMs you can also now force them to chroot to an empty directory,
2294
before starting the emulated guest.
2295

    
2296
KVM: Add maximum migration bandwith and maximum downtime tweaking
2297
support (requires a new-enough version of qemu-kvm).
2298

    
2299
Cluster verify will now warn if the master node doesn't have the master
2300
ip configured on it.
2301

    
2302
Add a new (incompatible) instance creation request format to RAPI which
2303
supports all parameters (previously only a subset was supported, and it
2304
wasn't possible to extend the old format to accomodate all the new
2305
features. The old format is still supported, and a client can check for
2306
this feature, before using it, by checking for its presence in the
2307
``features`` RAPI resource.
2308

    
2309
Now with ancient latin support. Try it passing the ``--roman`` option to
2310
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2311
(requires the python-roman module to be installed, in order to work).
2312

    
2313
Other changes
2314
~~~~~~~~~~~~~
2315

    
2316
As usual many internal code refactorings, documentation updates, and
2317
such. Among others:
2318

    
2319
  - Lots of improvements and cleanups to the experimental Remote API
2320
    (RAPI) client library.
2321
  - A new unit test suite for the core daemon libraries.
2322
  - A fix to creating missing directories makes sure the umask is not
2323
    applied anymore. This enforces the same directory permissions
2324
    everywhere.
2325
  - Better handling terminating daemons with ctrl+c (used when running
2326
    them in debugging mode).
2327
  - Fix a race condition in live migrating a KVM instance, when stat()
2328
    on the old proc status file returned EINVAL, which is an unexpected
2329
    value.
2330
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2331
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2332

    
2333

    
2334
Version 2.1.2.1
2335
---------------
2336

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

    
2339
Fix a bug which prevented untagged KVM instances from starting.
2340

    
2341

    
2342
Version 2.1.2
2343
-------------
2344

    
2345
*(Released Fri, 7 May 2010)*
2346

    
2347
Another release with a long development cycle, during which many
2348
different features were added.
2349

    
2350
Significant features
2351
~~~~~~~~~~~~~~~~~~~~
2352

    
2353
The KVM hypervisor now can run the individual instances as non-root, to
2354
reduce the impact of a VM being hijacked due to bugs in the
2355
hypervisor. It is possible to run all instances as a single (non-root)
2356
user, to manually specify a user for each instance, or to dynamically
2357
allocate a user out of a cluster-wide pool to each instance, with the
2358
guarantee that no two instances will run under the same user ID on any
2359
given node.
2360

    
2361
An experimental RAPI client library, that can be used standalone
2362
(without the other Ganeti libraries), is provided in the source tree as
2363
``lib/rapi/client.py``. Note this client might change its interface in
2364
the future, as we iterate on its capabilities.
2365

    
2366
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2367
replace the cluster's certificates and crypto keys. This might help in
2368
case they have been compromised, or have simply expired.
2369

    
2370
A new disk option for instance creation has been added that allows one
2371
to "adopt" currently existing logical volumes, with data
2372
preservation. This should allow easier migration to Ganeti from
2373
unmanaged (or managed via other software) instances.
2374

    
2375
Another disk improvement is the possibility to convert between redundant
2376
(DRBD) and plain (LVM) disk configuration for an instance. This should
2377
allow better scalability (starting with one node and growing the
2378
cluster, or shrinking a two-node cluster to one node).
2379

    
2380
A new feature that could help with automated node failovers has been
2381
implemented: if a node sees itself as offline (by querying the master
2382
candidates), it will try to shutdown (hard) all instances and any active
2383
DRBD devices. This reduces the risk of duplicate instances if an
2384
external script automatically failovers the instances on such nodes. To
2385
enable this, the cluster parameter ``maintain_node_health`` should be
2386
enabled; in the future this option (per the name) will enable other
2387
automatic maintenance features.
2388

    
2389
Instance export/import now will reuse the original instance
2390
specifications for all parameters; that means exporting an instance,
2391
deleting it and the importing it back should give an almost identical
2392
instance. Note that the default import behaviour has changed from
2393
before, where it created only one NIC; now it recreates the original
2394
number of NICs.
2395

    
2396
Cluster verify has added a few new checks: SSL certificates validity,
2397
/etc/hosts consistency across the cluster, etc.
2398

    
2399
Other changes
2400
~~~~~~~~~~~~~
2401

    
2402
As usual, many internal changes were done, documentation fixes,
2403
etc. Among others:
2404

    
2405
- Fixed cluster initialization with disabled cluster storage (regression
2406
  introduced in 2.1.1)
2407
- File-based storage supports growing the disks
2408
- Fixed behaviour of node role changes
2409
- Fixed cluster verify for some corner cases, plus a general rewrite of
2410
  cluster verify to allow future extension with more checks
2411
- Fixed log spamming by watcher and node daemon (regression introduced
2412
  in 2.1.1)
2413
- Fixed possible validation issues when changing the list of enabled
2414
  hypervisors
2415
- Fixed cleanup of /etc/hosts during node removal
2416
- Fixed RAPI response for invalid methods
2417
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2418
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2419
  from ide disks, etc.)
2420
- Allow OS changes without re-installation (to record a changed OS
2421
  outside of Ganeti, or to allow OS renames)
2422
- Allow instance creation without OS installation (useful for example if
2423
  the OS will be installed manually, or restored from a backup not in
2424
  Ganeti format)
2425
- Implemented option to make cluster ``copyfile`` use the replication
2426
  network
2427
- Added list of enabled hypervisors to ssconf (possibly useful for
2428
  external scripts)
2429
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2430
  1.2 clusters
2431
- A partial form of node re-IP is possible via node readd, which now
2432
  allows changed node primary IP
2433
- Command line utilities now show an informational message if the job is
2434
  waiting for a lock
2435
- The logs of the master daemon now show the PID/UID/GID of the
2436
  connected client
2437

    
2438

    
2439
Version 2.1.1
2440
-------------
2441

    
2442
*(Released Fri, 12 Mar 2010)*
2443

    
2444
During the 2.1.0 long release candidate cycle, a lot of improvements and
2445
changes have accumulated with were released later as 2.1.1.
2446

    
2447
Major changes
2448
~~~~~~~~~~~~~
2449

    
2450
The node evacuate command (``gnt-node evacuate``) was significantly
2451
rewritten, and as such the IAllocator protocol was changed - a new
2452
request type has been added. This unfortunate change during a stable
2453
series is designed to improve performance of node evacuations; on
2454
clusters with more than about five nodes and which are well-balanced,
2455
evacuation should proceed in parallel for all instances of the node
2456
being evacuated. As such, any existing IAllocator scripts need to be
2457
updated, otherwise the above command will fail due to the unknown
2458
request. The provided "dumb" allocator has not been updated; but the
2459
ganeti-htools package supports the new protocol since version 0.2.4.
2460

    
2461
Another important change is increased validation of node and instance
2462
names. This might create problems in special cases, if invalid host
2463
names are being used.
2464

    
2465
Also, a new layer of hypervisor parameters has been added, that sits at
2466
OS level between the cluster defaults and the instance ones. This allows
2467
customisation of virtualization parameters depending on the installed
2468
OS. For example instances with OS 'X' may have a different KVM kernel
2469
(or any other parameter) than the cluster defaults. This is intended to
2470
help managing a multiple OSes on the same cluster, without manual
2471
modification of each instance's parameters.
2472

    
2473
A tool for merging clusters, ``cluster-merge``, has been added in the
2474
tools sub-directory.
2475

    
2476
Bug fixes
2477
~~~~~~~~~
2478

    
2479
- Improved the int/float conversions that should make the code more
2480
  robust in face of errors from the node daemons
2481
- Fixed the remove node code in case of internal configuration errors
2482
- Fixed the node daemon behaviour in face of inconsistent queue
2483
  directory (e.g. read-only file-system where we can't open the files
2484
  read-write, etc.)
2485
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2486
  now it either aborts cleanly or, if given the new "auto_promote"
2487
  parameter, will automatically promote other nodes as needed
2488
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2489
  completely prevent Ganeti from working
2490
- Fixed bug for instance export when not all disks were successfully
2491
  exported
2492
- Fixed behaviour of node add when the new node is slow in starting up
2493
  the node daemon
2494
- Fixed handling of signals in the LUXI client, which should improve
2495
  behaviour of command-line scripts
2496
- Added checks for invalid node/instance names in the configuration (now
2497
  flagged during cluster verify)
2498
- Fixed watcher behaviour for disk activation errors
2499
- Fixed two potentially endless loops in http library, which led to the
2500
  RAPI daemon hanging and consuming 100% CPU in some cases
2501
- Fixed bug in RAPI daemon related to hashed passwords
2502
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2503
  instances
2504
- Enhanced compatibility with non-Debian OSes, but not using absolute
2505
  path in some commands and allowing customisation of the ssh
2506
  configuration directory
2507
- Fixed possible future issue with new Python versions by abiding to the
2508
  proper use of ``__slots__`` attribute on classes
2509
- Added checks that should prevent directory traversal attacks
2510
- Many documentation fixes based on feedback from users
2511

    
2512
New features
2513
~~~~~~~~~~~~
2514

    
2515
- Added an "early_release" more for instance replace disks and node
2516
  evacuate, where we release locks earlier and thus allow higher
2517
  parallelism within the cluster
2518
- Added watcher hooks, intended to allow the watcher to restart other
2519
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2520
  course for any other purpose
2521
- Added a compile-time disable for DRBD barriers, to increase
2522
  performance if the administrator trusts the power supply or the
2523
  storage system to not lose writes
2524
- Added the option of using syslog for logging instead of, or in
2525
  addition to, Ganeti's own log files
2526
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2527
  can indeed boot from a paravirtual NIC
2528
- Added a generic debug level for many operations; while this is not
2529
  used widely yet, it allows one to pass the debug value all the way to
2530
  the OS scripts
2531
- Enhanced the hooks environment for instance moves (failovers,
2532
  migrations) where the primary/secondary nodes changed during the
2533
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2534
- Enhanced data validations for many user-supplied values; one important
2535
  item is the restrictions imposed on instance and node names, which
2536
  might reject some (invalid) host names
2537
- Add a configure-time option to disable file-based storage, if it's not
2538
  needed; this allows greater security separation between the master
2539
  node and the other nodes from the point of view of the inter-node RPC
2540
  protocol
2541
- Added user notification in interactive tools if job is waiting in the
2542
  job queue or trying to acquire locks
2543
- Added log messages when a job is waiting for locks
2544
- Added filtering by node tags in instance operations which admit
2545
  multiple instances (start, stop, reboot, reinstall)
2546
- Added a new tool for cluster mergers, ``cluster-merge``
2547
- Parameters from command line which are of the form ``a=b,c=d`` can now
2548
  use backslash escapes to pass in values which contain commas,
2549
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2550
  ``b,c``
2551
- For KVM, the instance name is the first parameter passed to KVM, so
2552
  that it's more visible in the process list
2553

    
2554

    
2555
Version 2.1.0
2556
-------------
2557

    
2558
*(Released Tue, 2 Mar 2010)*
2559

    
2560
Ganeti 2.1 brings many improvements with it. Major changes:
2561

    
2562
- Added infrastructure to ease automated disk repairs
2563
- Added new daemon to export configuration data in a cheaper way than
2564
  using the remote API
2565
- Instance NICs can now be routed instead of being associated with a
2566
  networking bridge
2567
- Improved job locking logic to reduce impact of jobs acquiring multiple
2568
  locks waiting for other long-running jobs
2569

    
2570
In-depth implementation details can be found in the Ganeti 2.1 design
2571
document.
2572

    
2573
Details
2574
~~~~~~~
2575

    
2576
- Added chroot hypervisor
2577
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2578
  ``device_model``)
2579
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2580
  ``bootloader_path`` and ``bootloader_args``)
2581
- Added the ``use_localtime`` option for the xen-hvm and kvm
2582
  hypervisors, and the default value for this has changed to false (in
2583
  2.0 xen-hvm always enabled it)
2584
- Added luxi call to submit multiple jobs in one go
2585
- Added cluster initialization option to not modify ``/etc/hosts``
2586
  file on nodes
2587
- Added network interface parameters
2588
- Added dry run mode to some LUs
2589
- Added RAPI resources:
2590

    
2591
  - ``/2/instances/[instance_name]/info``
2592
  - ``/2/instances/[instance_name]/replace-disks``
2593
  - ``/2/nodes/[node_name]/evacuate``
2594
  - ``/2/nodes/[node_name]/migrate``
2595
  - ``/2/nodes/[node_name]/role``
2596
  - ``/2/nodes/[node_name]/storage``
2597
  - ``/2/nodes/[node_name]/storage/modify``
2598
  - ``/2/nodes/[node_name]/storage/repair``
2599

    
2600
- Added OpCodes to evacuate or migrate all instances on a node
2601
- Added new command to list storage elements on nodes (``gnt-node
2602
  list-storage``) and modify them (``gnt-node modify-storage``)
2603
- Added new ssconf files with master candidate IP address
2604
  (``ssconf_master_candidates_ips``), node primary IP address
2605
  (``ssconf_node_primary_ips``) and node secondary IP address
2606
  (``ssconf_node_secondary_ips``)
2607
- Added ``ganeti-confd`` and a client library to query the Ganeti
2608
  configuration via UDP
2609
- Added ability to run hooks after cluster initialization and before
2610
  cluster destruction
2611
- Added automatic mode for disk replace (``gnt-instance replace-disks
2612
  --auto``)
2613
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2614
  after catastrophic data-loss
2615
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2616
  groups
2617
- Added ``gnt-instance move`` command to move instances
2618
- Added ``gnt-cluster watcher`` command to control watcher
2619
- Added ``gnt-node powercycle`` command to powercycle nodes
2620
- Added new job status field ``lock_status``
2621
- Added parseable error codes to cluster verification (``gnt-cluster
2622
  verify --error-codes``) and made output less verbose (use
2623
  ``--verbose`` to restore previous behaviour)
2624
- Added UUIDs to the main config entities (cluster, nodes, instances)
2625
- Added support for OS variants
2626
- Added support for hashed passwords in the Ganeti remote API users file
2627
  (``rapi_users``)
2628
- Added option to specify maximum timeout on instance shutdown
2629
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2630
- Added new helper script to start and stop Ganeti daemons
2631
  (``daemon-util``), with the intent to reduce the work necessary to
2632
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2633
  from one place
2634
- Added more unittests
2635
- Fixed critical bug in ganeti-masterd startup
2636
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2637
  now customisable at the cluster level for both the KVM and Xen
2638
  hypervisors using the new ``migration_port`` parameter
2639
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2640
  reinstalling an instance
2641
- Allowed ``@`` in tag names
2642
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2643
- Many documentation updates
2644
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2645
- ``gnt-instance reinstall`` can now reinstall multiple instances
2646
- Updated many command line parameters
2647
- Introduced new OS API version 15
2648
- No longer support a default hypervisor
2649
- Treat virtual LVs as inexistent
2650
- Improved job locking logic to reduce lock contention
2651
- Match instance and node names case insensitively
2652
- Reimplemented bash completion script to be more complete
2653
- Improved burnin
2654

    
2655

    
2656
Version 2.0.6
2657
-------------
2658

    
2659
*(Released Thu, 4 Feb 2010)*
2660

    
2661
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2662
- Fix a string formatting bug
2663
- Improve safety of the code in some error paths
2664
- Improve data validation in the master of values returned from nodes
2665

    
2666

    
2667
Version 2.0.5
2668
-------------
2669

    
2670
*(Released Thu, 17 Dec 2009)*
2671

    
2672
- Fix security issue due to missing validation of iallocator names; this
2673
  allows local and remote execution of arbitrary executables
2674
- Fix failure of gnt-node list during instance removal
2675
- Ship the RAPI documentation in the archive
2676

    
2677

    
2678
Version 2.0.4
2679
-------------
2680

    
2681
*(Released Wed, 30 Sep 2009)*
2682

    
2683
- Fixed many wrong messages
2684
- Fixed a few bugs related to the locking library
2685
- Fixed MAC checking at instance creation time
2686
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2687
- Fixed a few issues related to signal handling in both daemons and
2688
  scripts
2689
- Fixed the example startup script provided
2690
- Fixed insserv dependencies in the example startup script (patch from
2691
  Debian)
2692
- Fixed handling of drained nodes in the iallocator framework
2693
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2694
  #528618)
2695
- Fixed error related to invalid job IDs in job polling
2696
- Fixed job/opcode persistence on unclean master shutdown
2697
- Fixed handling of partial job processing after unclean master
2698
  shutdown
2699
- Fixed error reporting from LUs, previously all errors were converted
2700
  into execution errors
2701
- Fixed error reporting from burnin
2702
- Decreased significantly the memory usage of the job queue
2703
- Optimised slightly multi-job submission
2704
- Optimised slightly opcode loading
2705
- Backported the multi-job submit framework from the development
2706
  branch; multi-instance start and stop should be faster
2707
- Added script to clean archived jobs after 21 days; this will reduce
2708
  the size of the queue directory
2709
- Added some extra checks in disk size tracking
2710
- Added an example ethers hook script
2711
- Added a cluster parameter that prevents Ganeti from modifying of
2712
  /etc/hosts
2713
- Added more node information to RAPI responses
2714
- Added a ``gnt-job watch`` command that allows following the ouput of a
2715
  job
2716
- Added a bind-address option to ganeti-rapi
2717
- Added more checks to the configuration verify
2718
- Enhanced the burnin script such that some operations can be retried
2719
  automatically
2720
- Converted instance reinstall to multi-instance model
2721

    
2722

    
2723
Version 2.0.3
2724
-------------
2725

    
2726
*(Released Fri, 7 Aug 2009)*
2727

    
2728
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2729
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2730
  instances have mismatched disk sizes in the configuration
2731
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2732
  any configuration mismatches for disk sizes
2733
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2734
  failover to work on two-node clusters
2735
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2736
  unusable
2737
- Fixed detection of OS script errors in ``gnt-backup export``
2738
- Fixed exit code of ``gnt-backup export``
2739

    
2740

    
2741
Version 2.0.2
2742
-------------
2743

    
2744
*(Released Fri, 17 Jul 2009)*
2745

    
2746
- Added experimental support for stripped logical volumes; this should
2747
  enhance performance but comes with a higher complexity in the block
2748
  device handling; stripping is only enabled when passing
2749
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2750
  affected even in the non-stripped mode
2751
- Improved resiliency against transient failures at the end of DRBD
2752
  resyncs, and in general of DRBD resync checks
2753
- Fixed a couple of issues with exports and snapshot errors
2754
- Fixed a couple of issues in instance listing
2755
- Added display of the disk size in ``gnt-instance info``
2756
- Fixed checking for valid OSes in instance creation
2757
- Fixed handling of the "vcpus" parameter in instance listing and in
2758
  general of invalid parameters
2759
- Fixed http server library, and thus RAPI, to handle invalid
2760
  username/password combinations correctly; this means that now they
2761
  report unauthorized for queries too, not only for modifications,
2762
  allowing earlier detect of configuration problems
2763
- Added a new "role" node list field, equivalent to the master/master
2764
  candidate/drained/offline flags combinations
2765
- Fixed cluster modify and changes of candidate pool size
2766
- Fixed cluster verify error messages for wrong files on regular nodes
2767
- Fixed a couple of issues with node demotion from master candidate role
2768
- Fixed node readd issues
2769
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2770
- Added a new ``--no-voting`` option for masterfailover to fix failover
2771
  on two-nodes clusters when the former master node is unreachable
2772
- Added instance reinstall over RAPI
2773

    
2774

    
2775
Version 2.0.1
2776
-------------
2777

    
2778
*(Released Tue, 16 Jun 2009)*
2779

    
2780
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2781
  allow re-adding the start in single-user option (regression from 1.2)
2782
- the watcher writes the instance status to a file, to allow monitoring
2783
  to report the instance status (from the master) based on cached
2784
  results of the watcher's queries; while this can get stale if the
2785
  watcher is being locked due to other work on the cluster, this is
2786
  still an improvement
2787
- the watcher now also restarts the node daemon and the rapi daemon if
2788
  they died
2789
- fixed the watcher to handle full and drained queue cases
2790
- hooks export more instance data in the environment, which helps if
2791
  hook scripts need to take action based on the instance's properties
2792
  (no longer need to query back into ganeti)
2793
- instance failovers when the instance is stopped do not check for free
2794
  RAM, so that failing over a stopped instance is possible in low memory
2795
  situations
2796
- rapi uses queries for tags instead of jobs (for less job traffic), and
2797
  for cluster tags it won't talk to masterd at all but read them from
2798
  ssconf
2799
- a couple of error handling fixes in RAPI
2800
- drbd handling: improved the error handling of inconsistent disks after
2801
  resync to reduce the frequency of "there are some degraded disks for
2802
  this instance" messages
2803
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2804
  error handling path called a wrong function name)
2805

    
2806

    
2807
Version 2.0.0
2808
-------------
2809

    
2810
*(Released Wed, 27 May 2009)*
2811

    
2812
- no changes from rc5
2813

    
2814

    
2815
Version 2.0 rc5
2816
---------------
2817

    
2818
*(Released Wed, 20 May 2009)*
2819

    
2820
- fix a couple of bugs (validation, argument checks)
2821
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2822
- some small improvements to RAPI and IAllocator
2823
- make watcher automatically start the master daemon if down
2824

    
2825

    
2826
Version 2.0 rc4
2827
---------------
2828

    
2829
*(Released Mon, 27 Apr 2009)*
2830

    
2831
- change the OS list to not require locks; this helps with big clusters
2832
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2833
  volume group is broken
2834
- ``gnt-instance info``, without any arguments, doesn't run for all
2835
  instances anymore; either pass ``--all`` or pass the desired
2836
  instances; this helps against mistakes on big clusters where listing
2837
  the information for all instances takes a long time
2838
- miscellaneous doc and man pages fixes
2839

    
2840

    
2841
Version 2.0 rc3
2842
---------------
2843

    
2844
*(Released Wed, 8 Apr 2009)*
2845

    
2846
- Change the internal locking model of some ``gnt-node`` commands, in
2847
  order to reduce contention (and blocking of master daemon) when
2848
  batching many creation/reinstall jobs
2849
- Fixes to Xen soft reboot
2850
- No longer build documentation at build time, instead distribute it in
2851
  the archive, in order to reduce the need for the whole docbook/rst
2852
  toolchains
2853

    
2854

    
2855
Version 2.0 rc2
2856
---------------
2857

    
2858
*(Released Fri, 27 Mar 2009)*
2859

    
2860
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2861
- Fix watcher startup sequence, improves the behaviour of busy clusters
2862
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2863
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2864
  bind address and other places
2865
- Some documentation fixes and updates
2866

    
2867

    
2868
Version 2.0 rc1
2869
---------------
2870

    
2871
*(Released Mon, 2 Mar 2009)*
2872

    
2873
- More documentation updates, now all docs should be more-or-less
2874
  up-to-date
2875
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2876
  etc.)
2877
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2878
  and KVM)
2879
- Fix an issue related to $libdir/run/ganeti and cluster creation
2880

    
2881

    
2882
Version 2.0 beta2
2883
-----------------
2884

    
2885
*(Released Thu, 19 Feb 2009)*
2886

    
2887
- Xen PVM and KVM have switched the default value for the instance root
2888
  disk to the first partition on the first drive, instead of the whole
2889
  drive; this means that the OS installation scripts must be changed
2890
  accordingly
2891
- Man pages have been updated
2892
- RAPI has been switched by default to HTTPS, and the exported functions
2893
  should all work correctly
2894
- RAPI v1 has been removed
2895
- Many improvements to the KVM hypervisor
2896
- Block device errors are now better reported
2897
- Many other bugfixes and small improvements
2898

    
2899

    
2900
Version 2.0 beta1
2901
-----------------
2902

    
2903
*(Released Mon, 26 Jan 2009)*
2904

    
2905
- Version 2 is a general rewrite of the code and therefore the
2906
  differences are too many to list, see the design document for 2.0 in
2907
  the ``doc/`` subdirectory for more details
2908
- In this beta version there is not yet a migration path from 1.2 (there
2909
  will be one in the final 2.0 release)
2910
- A few significant changes are:
2911

    
2912
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2913
    various ``gnt-*`` commands are just front-ends to it
2914
  - all the commands are entered into, and executed from a job queue,
2915
    see the ``gnt-job(8)`` manpage
2916
  - the RAPI daemon supports read-write operations, secured by basic
2917
    HTTP authentication on top of HTTPS
2918
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2919
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2920
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2921
  - DRBD devices are using statically allocated minor numbers, which
2922
    will be assigned to existing instances during the migration process
2923
  - there is support for both Xen PVM and Xen HVM instances running on
2924
    the same cluster
2925
  - KVM virtualization is supported too
2926
  - file-based storage has been implemented, which means that it is
2927
    possible to run the cluster without LVM and DRBD storage, for
2928
    example using a shared filesystem exported from shared storage (and
2929
    still have live migration)
2930

    
2931

    
2932
Version 1.2.7
2933
-------------
2934

    
2935
*(Released Tue, 13 Jan 2009)*
2936

    
2937
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2938
- Reuse the old instance mac address by default on instance import, if
2939
  the instance name is the same.
2940
- Handle situations in which the node info rpc returns incomplete
2941
  results (issue 46)
2942
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2943
- Improved version of batcher:
2944

    
2945
  - state file support
2946
  - instance mac address support
2947
  - support for HVM clusters/instances
2948

    
2949
- Add an option to show the number of cpu sockets and nodes in
2950
  ``gnt-node list``
2951
- Support OSes that handle more than one version of the OS api (but do
2952
  not change the current API in any other way)
2953
- Fix ``gnt-node migrate``
2954
- ``gnt-debug`` man page
2955
- Fixes various more typos and small issues
2956
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2957

    
2958

    
2959
Version 1.2.6
2960
-------------
2961

    
2962
*(Released Wed, 24 Sep 2008)*
2963

    
2964
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2965
  type of disk exported to fully virtualized instances.
2966
- provide access to the serial console of HVM instances
2967
- instance auto_balance flag, set by default. If turned off it will
2968
  avoid warnings on cluster verify if there is not enough memory to fail
2969
  over an instance. in the future it will prevent automatically failing
2970
  it over when we will support that.
2971
- batcher tool for instance creation, see ``tools/README.batcher``
2972
- ``gnt-instance reinstall --select-os`` to interactively select a new
2973
  operating system when reinstalling an instance.
2974
- when changing the memory amount on instance modify a check has been
2975
  added that the instance will be able to start. also warnings are
2976
  emitted if the instance will not be able to fail over, if auto_balance
2977
  is true.
2978
- documentation fixes
2979
- sync fields between ``gnt-instance list/modify/add/import``
2980
- fix a race condition in drbd when the sync speed was set after giving
2981
  the device a remote peer.
2982

    
2983

    
2984
Version 1.2.5
2985
-------------
2986

    
2987
*(Released Tue, 22 Jul 2008)*
2988

    
2989
- note: the allowed size and number of tags per object were reduced
2990
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2991
- fixed twisted 8.x compatibility
2992
- fixed ``gnt-instance replace-disks`` with iallocator
2993
- add TCP keepalives on twisted connections to detect restarted nodes
2994
- disk increase support, see ``gnt-instance grow-disk``
2995
- implement bulk node/instance query for RAPI
2996
- add tags in node/instance listing (optional)
2997
- experimental migration (and live migration) support, read the man page
2998
  for ``gnt-instance migrate``
2999
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3000
  has some small improvements in handling its state file
3001

    
3002

    
3003
Version 1.2.4
3004
-------------
3005

    
3006
*(Released Fri, 13 Jun 2008)*
3007

    
3008
- Experimental readonly, REST-based remote API implementation;
3009
  automatically started on master node, TCP port 5080, if enabled by
3010
  ``--enable-rapi`` parameter to configure script.
3011
- Instance allocator support. Add and import instance accept a
3012
  ``--iallocator`` parameter, and call that instance allocator to decide
3013
  which node to use for the instance. The iallocator document describes
3014
  what's expected from an allocator script.
3015
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3016
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3017
  node is lost there is still enough memory to fail over the instances
3018
  that reside on it.
3019
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3020
  ``gnt-cluster verify``, to check for site-specific compliance. All the
3021
  hooks will run, and their output, if any, will be displayed. Any
3022
  failing hook will make the verification return an error value.
3023
- ``gnt-cluster verify`` now checks that its peers are reachable on the
3024
  primary and secondary interfaces
3025
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3026
  that is still declared as part of the cluster and has failed.
3027
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
3028
  specifying output fields, that just adds the chosen fields to the
3029
  default ones.
3030
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
3031
  export from the filesystem.
3032
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3033
  have been added. Using them you can enable/disable acpi and pae
3034
  support, and specify a path for a cd image to be exported to the
3035
  instance. These parameters as the name suggest only work on HVM
3036
  clusters.
3037
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3038
  PAE support will be set to the previously hardcoded values, but the
3039
  (previously hardcoded) path to the CDROM ISO image will be unset and
3040
  if required, needs to be set manually with ``gnt-instance modify``
3041
  after the upgrade.
3042
- The address to which an instance's VNC console is bound is now
3043
  selectable per-instance, rather than being cluster wide. Of course
3044
  this only applies to instances controlled via VNC, so currently just
3045
  applies to HVM clusters.
3046

    
3047

    
3048
Version 1.2.3
3049
-------------
3050

    
3051
*(Released Mon, 18 Feb 2008)*
3052

    
3053
- more tweaks to the disk activation code (especially helpful for DRBD)
3054
- change the default ``gnt-instance list`` output format, now there is
3055
  one combined status field (see the manpage for the exact values this
3056
  field will have)
3057
- some more fixes for the mac export to hooks change
3058
- make Ganeti not break with DRBD 8.2.x (which changed the version
3059
  format in ``/proc/drbd``) (issue 24)
3060
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3061
  template, allowing migration from DRBD0.7+MD to DRBD8
3062

    
3063

    
3064
Version 1.2.2
3065
-------------
3066

    
3067
*(Released Wed, 30 Jan 2008)*
3068

    
3069
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3070
  support (issue 23)
3071
- add command aliases infrastructure and a few aliases
3072
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
3073
  man pages and the ``--help`` option of ``gnt-node
3074
  list``/``gnt-instance list``
3075
- fix ``gnt-backup list`` with down nodes (issue 21)
3076
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3077
- fix the dist archive and add a check for including svn/git files in
3078
  the future
3079
- some developer-related changes: improve the burnin and the QA suite,
3080
  add an upload script for testing during development
3081

    
3082

    
3083
Version 1.2.1
3084
-------------
3085

    
3086
*(Released Wed, 16 Jan 2008)*
3087

    
3088
- experimental HVM support, read the install document, section
3089
  "Initializing the cluster"
3090
- allow for the PVM hypervisor per-instance kernel and initrd paths
3091
- add a new command ``gnt-cluster verify-disks`` which uses a new
3092
  algorithm to improve the reconnection of the DRBD pairs if the device
3093
  on the secondary node has gone away
3094
- make logical volume code auto-activate LVs at disk activation time
3095
- slightly improve the speed of activating disks
3096
- allow specification of the MAC address at instance creation time, and
3097
  changing it later via ``gnt-instance modify``
3098
- fix handling of external commands that generate lots of output on
3099
  stderr
3100
- update documentation with regard to minimum version of DRBD8 supported
3101

    
3102

    
3103
Version 1.2.0
3104
-------------
3105

    
3106
*(Released Tue, 4 Dec 2007)*
3107

    
3108
- Log the ``xm create`` output to the node daemon log on failure (to
3109
  help diagnosing the error)
3110
- In debug mode, log all external commands output if failed to the logs
3111
- Change parsing of lvm commands to ignore stderr
3112

    
3113

    
3114
Version 1.2 beta3
3115
-----------------
3116

    
3117
*(Released Wed, 28 Nov 2007)*
3118

    
3119
- Another round of updates to the DRBD 8 code to deal with more failures
3120
  in the replace secondary node operation
3121
- Some more logging of failures in disk operations (lvm, drbd)
3122
- A few documentation updates
3123
- QA updates
3124

    
3125

    
3126
Version 1.2 beta2
3127
-----------------
3128

    
3129
*(Released Tue, 13 Nov 2007)*
3130

    
3131
- Change configuration file format from Python's Pickle to JSON.
3132
  Upgrading is possible using the cfgupgrade utility.
3133
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3134
  faster replace disks and is more stable (DRBD 8 has many improvements
3135
  compared to DRBD 0.7)
3136
- Added command line tags support (see man pages for ``gnt-instance``,
3137
  ``gnt-node``, ``gnt-cluster``)
3138
- Added instance rename support
3139
- Added multi-instance startup/shutdown
3140
- Added cluster rename support
3141
- Added ``gnt-node evacuate`` to simplify some node operations
3142
- Added instance reboot operation that can speedup reboot as compared to
3143
  stop and start
3144
- Soften the requirement that hostnames are in FQDN format
3145
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
3146
  reboots
3147
- Removed dependency on debian's patched fping that uses the
3148
  non-standard ``-S`` option
3149
- Now the OS definitions are searched for in multiple, configurable
3150
  paths (easier for distros to package)
3151
- Some changes to the hooks infrastructure (especially the new
3152
  post-configuration update hook)
3153
- Other small bugfixes
3154

    
3155
.. vim: set textwidth=72 syntax=rst :
3156
.. Local Variables:
3157
.. mode: rst
3158
.. fill-column: 72
3159
.. End: