Statistics
| Branch: | Tag: | Revision:

root / NEWS @ dd4a4aa9

History | View | Annotate | Download (120 kB)

1
News
2
====
3

    
4

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

    
8
*(unreleased)*
9

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

    
13
- ``gnt-node list`` no longer shows disk space information for shared file
14
  disk templates because it is not a node attribute. (For example, if you have
15
  both the file and shared file disk templates enabled, ``gnt-node list`` now
16
  only shows information about the file disk template.)
17
- The shared file disk template is now in the new 'sharedfile' storage type.
18
  As a result, ``gnt-node list-storage -t file`` now only shows information
19
  about the file disk template and you may use ``gnt-node list-storage -t
20
  sharedfile`` to query storage information for the shared file disk template.
21
- Over luxi, syntactially incorrect queries are now rejected as a whole;
22
  before, a 'SumbmitManyJobs' request was partially executed, if the outer
23
  structure of the request was syntactically correct. As the luxi protocol
24
  is internal (external applications are expected to use RAPI), the impact
25
  of this incompatible change should be limited.
26
- Queries for nodes, instances, groups, backups and networks are now
27
  exclusively done via the luxi daemon. Legacy python code was removed,
28
  as well as the --enable-split-queries configuration option.
29
- Orphan volumes errors are demoted to warnings and no longer affect the exit
30
  code of ``gnt-cluster verify``.
31

    
32
New features
33
~~~~~~~~~~~~
34

    
35
- Instance moves, backups and imports can now use compression to transfer the
36
  instance data.
37
- Node groups can be configured to use an SSH port different than the
38
  default 22.
39
- Added experimental support for Gluster distributed file storage as the
40
  ``gluster`` disk template under the new ``sharedfile`` storage type through
41
  automatic management of per-node FUSE mount points. You can configure the
42
  mount point location at ``gnt-cluster init`` time by using the new
43
  ``--gluster-storage-dir`` switch.
44

    
45
New dependencies
46
~~~~~~~~~~~~~~~~
47
The following new dependencies have been added:
48

    
49
For Haskell:
50

    
51
- ``zlib`` library (http://hackage.haskell.org/package/base64-bytestring)
52

    
53
- ``base64-bytestring`` library (http://hackage.haskell.org/package/zlib),
54
  at least version 1.0.0.0
55

    
56

    
57
Version 2.10.0 alpha1
58
---------------------
59

    
60
*(unreleased)*
61

    
62
Incompatible/important changes
63
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
64

    
65
- Adding disks with 'gnt-instance modify' now waits for the disks to sync per
66
  default. Specify --no-wait-for-sync to override this behavior.
67
- The Ganeti python code now adheres to a private-module layout. In particular,
68
  the module 'ganeti' is no longer in the python search path.
69
- On instance allocation, the iallocator now considers non-LVM storage
70
  properly. In particular, actual file storage space information is used
71
  when allocating space for a file/sharedfile instance.
72
- When disabling disk templates cluster-wide, the cluster now first
73
  checks whether there are instances still using those templates.
74
- 'gnt-node list-storage' now also reports storage information about
75
  file-based storage types.
76

    
77
New features
78
~~~~~~~~~~~~
79

    
80
- KVM hypervisors can now access RBD storage directly without having to
81
  go through a block device.
82
- A new command 'gnt-cluster upgrade' was added that automates the upgrade
83
  procedure between two Ganeti versions that are both 2.10 or higher.
84
- The move-instance command can now change disk templates when moving
85
  instances, and does not require any node placement options to be
86
  specified if the destination cluster has a default iallocator.
87
- Users can now change the soundhw and cpuid settings for XEN hypervisors.
88
- Hail and hbal now have the (optional) capability of accessing average CPU
89
  load information through the monitoring deamon, and to use it to dynamically
90
  adapt the allocation of instances.
91
- Hotplug support. Introduce new option '--hotplug' to ``gnt-instance modify``
92
  so that disk and NIC modifications take effect without the need of actual
93
  reboot. There are a couple of constrains currently for this feature:
94

    
95
   - only KVM hypervisor (versions >= 1.0) supports it,
96
   - one can not (yet) hotplug a disk using userspace access mode for RBD
97
   - in case of a downgrade instances should suffer a reboot in order to
98
     be migratable (due to core change of runtime files)
99

    
100
Misc changes
101
~~~~~~~~~~~~
102

    
103
- A new test framework for logical units was introduced and the test
104
  coverage for logical units was improved significantly.
105
- Opcodes are entirely generated from Haskell using the tool 'hs2py' and
106
  the module 'src/Ganeti/OpCodes.hs'.
107
- Constants are also generated from Haskell using the tool
108
  'hs2py-constants' and the module 'src/Ganeti/Constants.hs', with the
109
  exception of socket related constants, which require changing the
110
  cluster configuration file, and HVS related constants, because they
111
  are part of a port of instance queries to Haskell.  As a result, these
112
  changes will be part of the next release of Ganeti.
113

    
114
New dependencies
115
~~~~~~~~~~~~~~~~
116

    
117
The following new dependencies have been added/updated.
118

    
119
Python
120

    
121
- The version requirements for ``python-mock`` have increased to at least
122
  version 1.0.1. It is still used for testing only.
123

    
124

    
125
Version 2.9.1
126
-------------
127

    
128
*(Released Wed, 13 Nov 2013)*
129

    
130
- fix bug, that kept nodes offline when readding
131
- when verifying DRBD versions, ignore unavailable nodes
132
- fix bug that made the console unavailable on kvm in split-user
133
  setup (issue 608)
134
- DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
135

    
136

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

    
140
*(Released Tue, 5 Nov 2013)*
141

    
142
Incompatible/important changes
143
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
144

    
145
- hroller now also plans for capacity to move non-redundant instances off
146
  any node to be rebooted; the old behavior of completely ignoring any
147
  non-redundant instances can be restored by adding the --ignore-non-redundant
148
  option.
149
- The cluster option '--no-lvm-storage' was removed in favor of the new option
150
  '--enabled-disk-templates'.
151
- On instance creation, disk templates no longer need to be specified
152
  with '-t'. The default disk template will be taken from the list of
153
  enabled disk templates.
154
- The monitoring daemon is now running as root, in order to be able to collect
155
  information only available to root (such as the state of Xen instances).
156
- The ConfD client is now IPv6 compatible.
157
- File and shared file storage is no longer dis/enabled at configure time,
158
  but using the option '--enabled-disk-templates' at cluster initialization and
159
  modification.
160
- The default directories for file and shared file storage are not anymore
161
  specified at configure time, but taken from the cluster's configuration.
162
  They can be set at cluster initialization and modification with
163
  '--file-storage-dir' and '--shared-file-storage-dir'.
164
- Cluster verification now includes stricter checks regarding the
165
  default file and shared file storage directories. It now checks that
166
  the directories are explicitely allowed in the 'file-storage-paths' file and
167
  that the directories exist on all nodes.
168
- The list of allowed disk templates in the instance policy and the list
169
  of cluster-wide enabled disk templates is now checked for consistency
170
  on cluster or group modification. On cluster initialization, the ipolicy
171
  disk templates are ensured to be a subset of the cluster-wide enabled
172
  disk templates.
173

    
174
New features
175
~~~~~~~~~~~~
176

    
177
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
178
  the correct command syntax. It is possible to use different DRBD versions
179
  on different nodes as long as they are compatible to each other. This
180
  enables rolling upgrades of DRBD with no downtime. As permanent operation
181
  of different DRBD versions within a node group is discouraged,
182
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
183
- New "inst-status-xen" data collector for the monitoring daemon, providing
184
  information about the state of the xen instances on the nodes.
185
- New "lv" data collector for the monitoring daemon, collecting data about the
186
  logical volumes on the nodes, and pairing them with the name of the instances
187
  they belong to.
188
- New "diskstats" data collector, collecting the data from /proc/diskstats and
189
  presenting them over the monitoring daemon interface.
190
- The ConfD client is now IPv6 compatible.
191

    
192
New dependencies
193
~~~~~~~~~~~~~~~~
194
The following new dependencies have been added.
195

    
196
Python
197

    
198
- ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
199
  for the unit tests (and only used for testing).
200

    
201
Haskell
202

    
203
- ``hslogger`` (http://software.complete.org/hslogger) is now always
204
  required, even if confd is not enabled.
205

    
206
Since 2.9.0 rc3
207
~~~~~~~~~~~~~~~
208

    
209
- Correctly start/stop luxid during gnt-cluster master-failover (inherited
210
  from stable-2.8)
211
- Improved error messsages (inherited from stable-2.8)
212

    
213

    
214
Version 2.9.0 rc3
215
-----------------
216

    
217
*(Released Tue, 15 Oct 2013)*
218

    
219
The third release candidate in the 2.9 series. Since 2.9.0 rc2:
220

    
221
- in implicit configuration upgrade, match ipolicy with enabled disk templates
222
- improved harep documentation (inherited from stable-2.8)
223

    
224

    
225
Version 2.9.0 rc2
226
-----------------
227

    
228
*(Released Wed, 9 Oct 2013)*
229

    
230
The second release candidate in the 2.9 series. Since 2.9.0 rc1:
231

    
232
- Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
233
  at least one DRBD instance.
234
- Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
235
  downgrading.
236

    
237

    
238
Version 2.9.0 rc1
239
-----------------
240

    
241
*(Released Tue, 1 Oct 2013)*
242

    
243
The first release candidate in the 2.9 series. Since 2.9.0 beta1:
244

    
245
- various bug fixes
246
- update of the documentation, in particular installation instructions
247
- merging of LD_* constants into DT_* constants
248
- python style changes to be compatible with newer versions of pylint
249

    
250

    
251
Version 2.9.0 beta1
252
-------------------
253

    
254
*(Released Thu, 29 Aug 2013)*
255

    
256
This was the first beta release of the 2.9 series. All important changes
257
are listed in the latest 2.9 entry.
258

    
259

    
260
Version 2.8.2
261
-------------
262

    
263
*(Released Thu, 07 Nov 2013)*
264

    
265
- DRBD: ensure peers are UpToDate for dual-primary
266
- Improve error message for replace-disks
267
- More dependency checks at configure time
268
- Placate warnings on ganeti.outils_unittest.py
269

    
270

    
271
Version 2.8.1
272
-------------
273

    
274
*(Released Thu, 17 Oct 2013)*
275

    
276
- Correctly start/stop luxid during gnt-cluster master-failover
277
- Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
278
- Fix path for the job queue serial file
279
- Improved harep man page
280
- Minor documentation improvements
281

    
282

    
283
Version 2.8.0
284
-------------
285

    
286
*(Released Mon, 30 Sep 2013)*
287

    
288
Incompatible/important changes
289
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
290

    
291
- Instance policy can contain multiple instance specs, as described in
292
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
293
  <design-partitioned>`. As a consequence, it's not possible to partially change
294
  or override instance specs. Bounding specs (min and max) can be specified as a
295
  whole using the new option ``--ipolicy-bounds-specs``, while standard
296
  specs use the new option ``--ipolicy-std-specs``.
297
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
298
  gnt-instance is a valid YAML object.
299
- hail now honors network restrictions when allocating nodes. This led to an
300
  update of the IAllocator protocol. See the IAllocator documentation for
301
  details.
302
- confd now only answers static configuration request over the network. luxid
303
  was extracted, listens on the local LUXI socket and responds to live queries.
304
  This allows finer grained permissions if using separate users.
305

    
306
New features
307
~~~~~~~~~~~~
308

    
309
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
310
  to always require authentication, ``--require-authentication``. It can
311
  be specified in ``$sysconfdir/default/ganeti``.
312
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
313
  be used to manage the disk templates to be used by instances in the cluster.
314
  Initially, it will be set to a list that includes plain, drbd, if they were
315
  enabled by specifying a volume group name, and file and sharedfile, if those
316
  were enabled at configure time. Additionally, it will include all disk
317
  templates that are currently used by instances. The order of disk templates
318
  will be based on Ganeti's history of supporting them. In the future, the
319
  first entry of the list will be used as a default disk template on instance
320
  creation.
321
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
322
  configuration back to the previous stable version.
323
- Disk templates in group ipolicy can be restored to the default value.
324
- Initial support for diskless instances and virtual clusters in QA.
325
- More QA and unit tests for instance policies.
326
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
327
  describing why the opcode itself was executed.
328
- The monitoring daemon is now available. It allows users to query the cluster
329
  for obtaining information about the status of the system. The daemon is only
330
  responsible for providing the information over the network: the actual data
331
  gathering is performed by data collectors (currently, only the DRBD status
332
  collector is available).
333
- In order to help developers work on Ganeti, a new script
334
  (``devel/build_chroot``) is provided, for building a chroot that contains all
335
  the required development libraries and tools for compiling Ganeti on a Debian
336
  Squeeze system.
337
- A new tool, ``harep``, for performing self-repair and recreation of instances
338
  in Ganeti has been added.
339
- Split queries are enabled for tags, network, exports, cluster info, groups,
340
  jobs, nodes.
341
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
342
  It prints the command line to set the current policies, to ease
343
  changing them.
344
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
345
  devices for KVM virtio-net interfaces will get created with VNET_HDR
346
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
347
  virtio-net interfaces, which prevents host kernel tainting and log
348
  flooding, when dealing with broken or malicious virtio-net drivers.
349
  It's set to true by default.
350
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
351
  failures.
352
- Support 'viridian' parameter in Xen HVM
353
- Support DSA SSH keys in bootstrap
354
- To simplify the work of packaging frameworks that want to add the needed users
355
  and groups in a split-user setup themselves, at build time three files in
356
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
357
  the groups to be generated, the ``users`` file contains, one per line, the
358
  users to be generated, optionally followed by their primary group, where
359
  important. The ``groupmemberships`` file contains, one per line, additional
360
  user-group membership relations that need to be established. The syntax of
361
  these files will remain stable in all future versions.
362

    
363

    
364
New dependencies
365
~~~~~~~~~~~~~~~~
366
The following new dependencies have been added:
367

    
368
For Haskell:
369
- The ``curl`` library is not optional anymore for compiling the Haskell code.
370
- ``snap-server`` library (if monitoring is enabled).
371

    
372
For Python:
373
- The minimum Python version needed to run Ganeti is now 2.6.
374
- ``yaml`` library (only for running the QA).
375

    
376
Since 2.8.0 rc3
377
~~~~~~~~~~~~~~~
378
- Perform proper cleanup on termination of Haskell daemons
379
- Fix corner-case in handling of remaining retry time
380

    
381

    
382
Version 2.8.0 rc3
383
-----------------
384

    
385
*(Released Tue, 17 Sep 2013)*
386

    
387
- To simplify the work of packaging frameworks that want to add the needed users
388
  and groups in a split-user setup themselves, at build time three files in
389
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
390
  the groups to be generated, the ``users`` file contains, one per line, the
391
  users to be generated, optionally followed by their primary group, where
392
  important. The ``groupmemberships`` file contains, one per line, additional
393
  user-group membership relations that need to be established. The syntax of
394
  these files will remain stable in all future versions.
395
- Add a default to file-driver when unspecified over RAPI (Issue 571)
396
- Mark the DSA host pubkey as optional, and remove it during config downgrade
397
  (Issue 560)
398
- Some documentation fixes
399

    
400

    
401
Version 2.8.0 rc2
402
-----------------
403

    
404
*(Released Tue, 27 Aug 2013)*
405

    
406
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
407

    
408
- Support 'viridian' parameter in Xen HVM (Issue 233)
409
- Include VCS version in ``gnt-cluster version``
410
- Support DSA SSH keys in bootstrap (Issue 338)
411
- Fix batch creation of instances
412
- Use FQDN to check master node status (Issue 551)
413
- Make the DRBD collector more failure-resilient
414

    
415

    
416
Version 2.8.0 rc1
417
-----------------
418

    
419
*(Released Fri, 2 Aug 2013)*
420

    
421
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
422

    
423
- Fix upgrading/downgrading from 2.7
424
- Increase maximum RAPI message size
425
- Documentation updates
426
- Split ``confd`` between ``luxid`` and ``confd``
427
- Merge 2.7 series up to the 2.7.1 release
428
- Allow the ``modify_etc_hosts`` option to be changed
429
- Add better debugging for ``luxid`` queries
430
- Expose bulk parameter for GetJobs in RAPI client
431
- Expose missing ``network`` fields in RAPI
432
- Add some ``cluster verify`` tests
433
- Some unittest fixes
434
- Fix a malfunction in ``hspace``'s tiered allocation
435
- Fix query compatibility between haskell and python implementations
436
- Add the ``vnet_hdr`` HV parameter for KVM
437
- Add ``--cleanup`` to instance failover
438
- Change the connected groups format in ``gnt-network info`` output; it
439
  was previously displayed as a raw list by mistake. (Merged from 2.7)
440

    
441

    
442
Version 2.8.0 beta1
443
-------------------
444

    
445
*(Released Mon, 24 Jun 2013)*
446

    
447
This was the first beta release of the 2.8 series. All important changes
448
are listed in the latest 2.8 entry.
449

    
450

    
451
Version 2.7.2
452
-------------
453

    
454
*(Released Thu, 26 Sep 2013)*
455

    
456
- Change the connected groups format in ``gnt-network info`` output; it
457
  was previously displayed as a raw list by mistake
458
- Check disk template in right dict when copying
459
- Support multi-instance allocs without iallocator
460
- Fix some errors in the documentation
461
- Fix formatting of tuple in an error message
462

    
463

    
464
Version 2.7.1
465
-------------
466

    
467
*(Released Thu, 25 Jul 2013)*
468

    
469
- Add logrotate functionality in daemon-util
470
- Add logrotate example file
471
- Add missing fields to network queries over rapi
472
- Fix network object timestamps
473
- Add support for querying network timestamps
474
- Fix a typo in the example crontab
475
- Fix a documentation typo
476

    
477

    
478
Version 2.7.0
479
-------------
480

    
481
*(Released Thu, 04 Jul 2013)*
482

    
483
Incompatible/important changes
484
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
485

    
486
- Instance policies for disk size were documented to be on a per-disk
487
  basis, but hail applied them to the sum of all disks. This has been
488
  fixed.
489
- ``hbal`` will now exit with status 0 if, during job execution over
490
  LUXI, early exit has been requested and all jobs are successful;
491
  before, exit status 1 was used, which cannot be differentiated from
492
  "job error" case
493
- Compatibility with newer versions of rbd has been fixed
494
- ``gnt-instance batch-create`` has been changed to use the bulk create
495
  opcode from Ganeti. This lead to incompatible changes in the format of
496
  the JSON file. It's now not a custom dict anymore but a dict
497
  compatible with the ``OpInstanceCreate`` opcode.
498
- Parent directories for file storage need to be listed in
499
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
500
  write the file automatically based on old configuration values, but it
501
  can not distribute it across all nodes and the file contents should be
502
  verified. Use ``gnt-cluster copyfile
503
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
504
  upgraded. The reason for requiring this list of paths now is that
505
  before it would have been possible to inject new paths via RPC,
506
  allowing files to be created in arbitrary locations. The RPC protocol
507
  is protected using SSL/X.509 certificates, but as a design principle
508
  Ganeti does not permit arbitrary paths to be passed.
509
- The parsing of the variants file for OSes (see
510
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
511
  lines and comment lines (starting with ``#``) are ignored for better
512
  readability.
513
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
514
  longer available. ``gnt-node add`` now invokes a new tool on the
515
  destination node, named ``prepare-node-join``, to configure the SSH
516
  daemon. Paramiko is no longer necessary to configure nodes' SSH
517
  daemons via ``gnt-node add``.
518
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
519
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
520
  the flag is not reset after a master failover.
521
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
522
  is recommended. 2.6 will be mandatory from the 2.8 series.
523

    
524

    
525
New features
526
~~~~~~~~~~~~
527

    
528
- New network management functionality to support automatic allocation
529
  of IP addresses and managing of network parameters. See
530
  :manpage:`gnt-network(8)` for more details.
531
- New external storage backend, to allow managing arbitrary storage
532
  systems external to the cluster. See
533
  :manpage:`ganeti-extstorage-interface(7)`.
534
- New ``exclusive-storage`` node parameter added, restricted to
535
  nodegroup level. When it's set to true, physical disks are assigned in
536
  an exclusive fashion to instances, as documented in :doc:`Partitioned
537
  Ganeti <design-partitioned>`.  Currently, only instances using the
538
  ``plain`` disk template are supported.
539
- The KVM hypervisor has been updated with many new hypervisor
540
  parameters, including a generic one for passing arbitrary command line
541
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
542
  compatible up to qemu 1.4.
543
- A new tool, called ``mon-collector``, is the stand-alone executor of
544
  the data collectors for a monitoring system. As of this version, it
545
  just includes the DRBD data collector, that can be executed by calling
546
  ``mon-collector`` using the ``drbd`` parameter. See
547
  :manpage:`mon-collector(7)`.
548
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
549
  for RAPI users. It allows granting permissions to query for
550
  information to a specific user without giving
551
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
552
- A new tool named ``node-cleanup`` has been added. It cleans remains of
553
  a cluster from a machine by stopping all daemons, removing
554
  certificates and ssconf files. Unless the ``--no-backup`` option is
555
  given, copies of the certificates are made.
556
- Instance creations now support the use of opportunistic locking,
557
  potentially speeding up the (parallel) creation of multiple instances.
558
  This feature is currently only available via the :doc:`RAPI
559
  <rapi>` interface and when an instance allocator is used. If the
560
  ``opportunistic_locking`` parameter is set the opcode will try to
561
  acquire as many locks as possible, but will not wait for any locks
562
  held by other opcodes. If not enough resources can be found to
563
  allocate the instance, the temporary error code
564
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
565
  retried thereafter, with or without opportunistic locking.
566
- New experimental linux-ha resource scripts.
567
- Restricted-commands support: ganeti can now be asked (via command line
568
  or rapi) to perform commands on a node. These are passed via ganeti
569
  RPC rather than ssh. This functionality is restricted to commands
570
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
571
  reasons. The file is not copied automatically.
572

    
573

    
574
Misc changes
575
~~~~~~~~~~~~
576

    
577
- Diskless instances are now externally mirrored (Issue 237). This for
578
  now has only been tested in conjunction with explicit target nodes for
579
  migration/failover.
580
- Queries not needing locks or RPC access to the node can now be
581
  performed by the confd daemon, making them independent from jobs, and
582
  thus faster to execute. This is selectable at configure time.
583
- The functionality for allocating multiple instances at once has been
584
  overhauled and is now also available through :doc:`RAPI <rapi>`.
585

    
586
There are no significant changes from version 2.7.0~rc3.
587

    
588

    
589
Version 2.7.0 rc3
590
-----------------
591

    
592
*(Released Tue, 25 Jun 2013)*
593

    
594
- Fix permissions on the confd query socket (Issue 477)
595
- Fix permissions on the job archive dir (Issue 498)
596
- Fix handling of an internal exception in replace-disks (Issue 472)
597
- Fix gnt-node info handling of shortened names (Issue 497)
598
- Fix gnt-instance grow-disk when wiping is enabled
599
- Documentation improvements, and support for newer pandoc
600
- Fix hspace honoring ipolicy for disks (Issue 484)
601
- Improve handling of the ``kvm_extra`` HV parameter
602

    
603

    
604
Version 2.7.0 rc2
605
-----------------
606

    
607
*(Released Fri, 24 May 2013)*
608

    
609
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
610
  symlink.
611
- Disks added through ``gnt-instance modify`` or created through
612
  ``gnt-instance recreate-disks`` are wiped, if the
613
  ``prealloc_wipe_disks`` flag is set.
614
- If wiping newly created disks fails, the disks are removed. Also,
615
  partial failures in creating disks through ``gnt-instance modify``
616
  triggers a cleanup of the partially-created disks.
617
- Removing the master IP address doesn't fail if the address has been
618
  already removed.
619
- Fix ownership of the OS log dir
620
- Workaround missing SO_PEERCRED constant (Issue 191)
621

    
622

    
623
Version 2.7.0 rc1
624
-----------------
625

    
626
*(Released Fri, 3 May 2013)*
627

    
628
This was the first release candidate of the 2.7 series. Since beta3:
629

    
630
- Fix kvm compatibility with qemu 1.4 (Issue 389)
631
- Documentation updates (admin guide, upgrade notes, install
632
  instructions) (Issue 372)
633
- Fix gnt-group list nodes and instances count (Issue 436)
634
- Fix compilation without non-mandatory libraries (Issue 441)
635
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
636
- Make confd logging more verbose at INFO level (Issue 435)
637
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
638
- Fix failure path for instance storage type conversion (Issue 229)
639
- Update htools text backend documentation
640
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
641
- Disable inter-cluster instance move for file-based instances, because
642
  it is dependant on instance export, which is not supported for
643
  file-based instances. (Issue 414)
644
- Fix gnt-job crashes on non-ascii characters (Issue 427)
645
- Fix volume group checks on non-vm-capable nodes (Issue 432)
646

    
647

    
648
Version 2.7.0 beta3
649
-------------------
650

    
651
*(Released Mon, 22 Apr 2013)*
652

    
653
This was the third beta release of the 2.7 series. Since beta2:
654

    
655
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
656
- Fix data loss on wrong usage of ``gnt-instance move``
657
- Properly export errors in confd-based job queries
658
- Add ``users-setup`` tool
659
- Fix iallocator protocol to report 0 as a disk size for diskless
660
  instances. This avoids hail breaking when a diskless instance is
661
  present.
662
- Fix job queue directory permission problem that made confd job queries
663
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
664
  for access to archived jobs (Issue 406).
665
- Limit the sizes of networks supported by ``gnt-network`` to something
666
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
667
- Fix bugs in instance disk template conversion
668
- Fix GHC 7 compatibility
669
- Fix ``burnin`` install path (Issue 426).
670
- Allow very small disk grows (Issue 347).
671
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
672
  that noded doesn't import masterd code (Issue 419).
673
- Make sure the default metavg at cluster init is the same as the vg, if
674
  unspecified (Issue 358).
675
- Fix cleanup of partially created disks (part of Issue 416)
676

    
677

    
678
Version 2.7.0 beta2
679
-------------------
680

    
681
*(Released Tue, 2 Apr 2013)*
682

    
683
This was the second beta release of the 2.7 series. Since beta1:
684

    
685
- Networks no longer have a "type" slot, since this information was
686
  unused in Ganeti: instead of it tags should be used.
687
- The rapi client now has a ``target_node`` option to MigrateInstance.
688
- Fix early exit return code for hbal (Issue 386).
689
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
690
- Fix ``rbd showmapped`` output parsing (Issue 312).
691
- Networks are now referenced indexed by UUID, rather than name. This
692
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
693
  use.
694
- The OS environment now includes network information.
695
- Deleting of a network is now disallowed if any instance nic is using
696
  it, to prevent dangling references.
697
- External storage is now documented in man pages.
698
- The exclusive_storage flag can now only be set at nodegroup level.
699
- Hbal can now submit an explicit priority with its jobs.
700
- Many network related locking fixes.
701
- Bump up the required pylint version to 0.25.1.
702
- Fix the ``no_remember`` option in RAPI client.
703
- Many ipolicy related tests, qa, and fixes.
704
- Many documentation improvements and fixes.
705
- Fix building with ``--disable-file-storage``.
706
- Fix ``-q`` option in htools, which was broken if passed more than
707
  once.
708
- Some haskell/python interaction improvements and fixes.
709
- Fix iallocator in case of missing LVM storage.
710
- Fix confd config load in case of ``--no-lvm-storage``.
711
- The confd/query functionality is now mentioned in the security
712
  documentation.
713

    
714

    
715
Version 2.7.0 beta1
716
-------------------
717

    
718
*(Released Wed, 6 Feb 2013)*
719

    
720
This was the first beta release of the 2.7 series. All important changes
721
are listed in the latest 2.7 entry.
722

    
723

    
724
Version 2.6.2
725
-------------
726

    
727
*(Released Fri, 21 Dec 2012)*
728

    
729
Important behaviour change: hbal won't rebalance anymore instances which
730
have the ``auto_balance`` attribute set to false. This was the intention
731
all along, but until now it only skipped those from the N+1 memory
732
reservation (DRBD-specific).
733

    
734
A significant number of bug fixes in this release:
735

    
736
- Fixed disk adoption interaction with ipolicy checks.
737
- Fixed networking issues when instances are started, stopped or
738
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
739
- Fixed the warning in cluster verify for shared storage instances not
740
  being redundant.
741
- Fixed removal of storage directory on shared file storage (issue 262).
742
- Fixed validation of LVM volume group name in OpClusterSetParams
743
  (``gnt-cluster modify``) (issue 285).
744
- Fixed runtime memory increases (``gnt-instance modify -m``).
745
- Fixed live migration under Xen's ``xl`` mode.
746
- Fixed ``gnt-instance console`` with ``xl``.
747
- Fixed building with newer Haskell compiler/libraries.
748
- Fixed PID file writing in Haskell daemons (confd); this prevents
749
  restart issues if confd was launched manually (outside of
750
  ``daemon-util``) while another copy of it was running
751
- Fixed a type error when doing live migrations with KVM (issue 297) and
752
  the error messages for failing migrations have been improved.
753
- Fixed opcode validation for the out-of-band commands (``gnt-node
754
  power``).
755
- Fixed a type error when unsetting OS hypervisor parameters (issue
756
  311); now it's possible to unset all OS-specific hypervisor
757
  parameters.
758
- Fixed the ``dry-run`` mode for many operations: verification of
759
  results was over-zealous but didn't take into account the ``dry-run``
760
  operation, resulting in "wrong" failures.
761
- Fixed bash completion in ``gnt-job list`` when the job queue has
762
  hundreds of entries; especially with older ``bash`` versions, this
763
  results in significant CPU usage.
764

    
765
And lastly, a few other improvements have been made:
766

    
767
- Added option to force master-failover without voting (issue 282).
768
- Clarified error message on lock conflict (issue 287).
769
- Logging of newly submitted jobs has been improved (issue 290).
770
- Hostname checks have been made uniform between instance rename and
771
  create (issue 291).
772
- The ``--submit`` option is now supported by ``gnt-debug delay``.
773
- Shutting down the master daemon by sending SIGTERM now stops it from
774
  processing jobs waiting for locks; instead, those jobs will be started
775
  once again after the master daemon is started the next time (issue
776
  296).
777
- Support for Xen's ``xl`` program has been improved (besides the fixes
778
  above).
779
- Reduced logging noise in the Haskell confd daemon (only show one log
780
  entry for each config reload, instead of two).
781
- Several man page updates and typo fixes.
782

    
783

    
784
Version 2.6.1
785
-------------
786

    
787
*(Released Fri, 12 Oct 2012)*
788

    
789
A small bugfix release. Among the bugs fixed:
790

    
791
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
792
  made the command unusable.
793
- Commands that issue many jobs don't fail anymore just because some jobs
794
  take so long that other jobs are archived.
795
- Failures during ``gnt-instance reinstall`` are reflected by the exit
796
  status.
797
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
798
  DRBD is enabled.
799
- When ``always_failover`` is set, ``--allow-failover`` is not required
800
  in migrate commands anymore.
801
- ``bash_completion`` works even if extglob is disabled.
802
- Fixed bug with locks that made failover for RDB-based instances fail.
803
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
804
  a random node instead of one based on the allocator metric.
805
- Support for newer versions of pylint and pep8.
806
- Hail doesn't fail anymore when trying to add an instance of type
807
  ``file``, ``sharedfile`` or ``rbd``.
808
- Added new Makefile target to rebuild the whole distribution, so that
809
  all files are included.
810

    
811

    
812
Version 2.6.0
813
-------------
814

    
815
*(Released Fri, 27 Jul 2012)*
816

    
817

    
818
.. attention:: The ``LUXI`` protocol has been made more consistent
819
   regarding its handling of command arguments. This, however, leads to
820
   incompatibility issues with previous versions. Please ensure that you
821
   restart Ganeti daemons soon after the upgrade, otherwise most
822
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
823
   pausing/resuming the watcher, cancelling and archiving jobs, querying
824
   the cluster configuration) will fail.
825

    
826

    
827
New features
828
~~~~~~~~~~~~
829

    
830
Instance run status
831
+++++++++++++++++++
832

    
833
The current ``admin_up`` field, which used to denote whether an instance
834
should be running or not, has been removed. Instead, ``admin_state`` is
835
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
836

    
837
The rational behind this is that an instance being “down” can have
838
different meanings:
839

    
840
- it could be down during a reboot
841
- it could be temporarily be down for a reinstall
842
- or it could be down because it is deprecated and kept just for its
843
  disk
844

    
845
The previous Boolean state was making it difficult to do capacity
846
calculations: should Ganeti reserve memory for a down instance? Now, the
847
tri-state field makes it clear:
848

    
849
- in ``up`` and ``down`` state, all resources are reserved for the
850
  instance, and it can be at any time brought up if it is down
851
- in ``offline`` state, only disk space is reserved for it, but not
852
  memory or CPUs
853

    
854
The field can have an extra use: since the transition between ``up`` and
855
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
856
transition between ``offline`` and ``down`` is done via ``gnt-instance
857
modify``, it is possible to given different rights to users. For
858
example, owners of an instance could be allowed to start/stop it, but
859
not transition it out of the offline state.
860

    
861
Instance policies and specs
862
+++++++++++++++++++++++++++
863

    
864
In previous Ganeti versions, an instance creation request was not
865
limited on the minimum size and on the maximum size just by the cluster
866
resources. As such, any policy could be implemented only in third-party
867
clients (RAPI clients, or shell wrappers over ``gnt-*``
868
tools). Furthermore, calculating cluster capacity via ``hspace`` again
869
required external input with regards to instance sizes.
870

    
871
In order to improve these workflows and to allow for example better
872
per-node group differentiation, we introduced instance specs, which
873
allow declaring:
874

    
875
- minimum instance disk size, disk count, memory size, cpu count
876
- maximum values for the above metrics
877
- and “standard” values (used in ``hspace`` to calculate the standard
878
  sized instances)
879

    
880
The minimum/maximum values can be also customised at node-group level,
881
for example allowing more powerful hardware to support bigger instance
882
memory sizes.
883

    
884
Beside the instance specs, there are a few other settings belonging to
885
the instance policy framework. It is possible now to customise, per
886
cluster and node-group:
887

    
888
- the list of allowed disk templates
889
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
890
- the maximum ratio of instance to spindles (see below for more
891
  information) for local storage
892

    
893
All these together should allow all tools that talk to Ganeti to know
894
what are the ranges of allowed values for instances and the
895
over-subscription that is allowed.
896

    
897
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
898
instance configuration, and the physical CPU configuration from the
899
node. For the spindle ratios however, we didn't track before these
900
values, so new parameters have been added:
901

    
902
- a new node parameter ``spindle_count``, defaults to 1, customisable at
903
  node group or node level
904
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
905

    
906
Note that spindles in this context doesn't need to mean actual
907
mechanical hard-drives; it's just a relative number for both the node
908
I/O capacity and instance I/O consumption.
909

    
910
Instance migration behaviour
911
++++++++++++++++++++++++++++
912

    
913
While live-migration is in general desirable over failover, it is
914
possible that for some workloads it is actually worse, due to the
915
variable time of the “suspend” phase during live migration.
916

    
917
To allow the tools to work consistently over such instances (without
918
having to hard-code instance names), a new backend parameter
919
``always_failover`` has been added to control the migration/failover
920
behaviour. When set to True, all migration requests for an instance will
921
instead fall-back to failover.
922

    
923
Instance memory ballooning
924
++++++++++++++++++++++++++
925

    
926
Initial support for memory ballooning has been added. The memory for an
927
instance is no longer fixed (backend parameter ``memory``), but instead
928
can vary between minimum and maximum values (backend parameters
929
``minmem`` and ``maxmem``). Currently we only change an instance's
930
memory when:
931

    
932
- live migrating or failing over and instance and the target node
933
  doesn't have enough memory
934
- user requests changing the memory via ``gnt-instance modify
935
  --runtime-memory``
936

    
937
Instance CPU pinning
938
++++++++++++++++++++
939

    
940
In order to control the use of specific CPUs by instance, support for
941
controlling CPU pinning has been added for the Xen, HVM and LXC
942
hypervisors. This is controlled by a new hypervisor parameter
943
``cpu_mask``; details about possible values for this are in the
944
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
945
VCPU-to-CPU mapping) form will work well only when all nodes in your
946
cluster have the same amount of CPUs.
947

    
948
Disk parameters
949
+++++++++++++++
950

    
951
Another area in which Ganeti was not customisable were the parameters
952
used for storage configuration, e.g. how many stripes to use for LVM,
953
DRBD resync configuration, etc.
954

    
955
To improve this area, we've added disks parameters, which are
956
customisable at cluster and node group level, and which allow to
957
specify various parameters for disks (DRBD has the most parameters
958
currently), for example:
959

    
960
- DRBD resync algorithm and parameters (e.g. speed)
961
- the default VG for meta-data volumes for DRBD
962
- number of stripes for LVM (plain disk template)
963
- the RBD pool
964

    
965
These parameters can be modified via ``gnt-cluster modify -D …`` and
966
``gnt-group modify -D …``, and are used at either instance creation (in
967
case of LVM stripes, for example) or at disk “activation” time
968
(e.g. resync speed).
969

    
970
Rados block device support
971
++++++++++++++++++++++++++
972

    
973
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
974
denoted by the ``rbd`` disk template type. This is considered
975
experimental, feedback is welcome. For details on configuring it, see
976
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
977

    
978
Master IP setup
979
+++++++++++++++
980

    
981
The existing master IP functionality works well only in simple setups (a
982
single network shared by all nodes); however, if nodes belong to
983
different networks, then the ``/32`` setup and lack of routing
984
information is not enough.
985

    
986
To allow the master IP to function well in more complex cases, the
987
system was reworked as follows:
988

    
989
- a master IP netmask setting has been added
990
- the master IP activation/turn-down code was moved from the node daemon
991
  to a separate script
992
- whether to run the Ganeti-supplied master IP script or a user-supplied
993
  on is a ``gnt-cluster init`` setting
994

    
995
Details about the location of the standard and custom setup scripts are
996
in the man page :manpage:`gnt-cluster(8)`; for information about the
997
setup script protocol, look at the Ganeti-supplied script.
998

    
999
SPICE support
1000
+++++++++++++
1001

    
1002
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
1003
improved.
1004

    
1005
It is now possible to use TLS-protected connections, and when renewing
1006
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
1007
it is now possible to specify spice or spice CA certificates. Also, it
1008
is possible to configure a password for SPICE sessions via the
1009
hypervisor parameter ``spice_password_file``.
1010

    
1011
There are also new parameters to control the compression and streaming
1012
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1013
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1014
for the spice parameters.
1015

    
1016
Lastly, it is now possible to see the SPICE connection information via
1017
``gnt-instance console``.
1018

    
1019
OVF converter
1020
+++++++++++++
1021

    
1022
A new tool (``tools/ovfconverter``) has been added that supports
1023
conversion between Ganeti and the `Open Virtualization Format
1024
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1025
from).
1026

    
1027
This relies on the ``qemu-img`` tool to convert the disk formats, so the
1028
actual compatibility with other virtualization solutions depends on it.
1029

    
1030
Confd daemon changes
1031
++++++++++++++++++++
1032

    
1033
The configuration query daemon (``ganeti-confd``) is now optional, and
1034
has been rewritten in Haskell; whether to use the daemon at all, use the
1035
Python (default) or the Haskell version is selectable at configure time
1036
via the ``--enable-confd`` parameter, which can take one of the
1037
``haskell``, ``python`` or ``no`` values. If not used, disabling the
1038
daemon will result in a smaller footprint; for larger systems, we
1039
welcome feedback on the Haskell version which might become the default
1040
in future versions.
1041

    
1042
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1043
daemon running. The Python version doesn't implement the new call.
1044

    
1045

    
1046
User interface changes
1047
~~~~~~~~~~~~~~~~~~~~~~
1048

    
1049
We have replaced the ``--disks`` option of ``gnt-instance
1050
replace-disks`` with a more flexible ``--disk`` option, which allows
1051
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1052
disk size and mode can be changed upon recreation (via ``gnt-instance
1053
recreate-disks``, which accepts the same ``--disk`` option).
1054

    
1055
As many people are used to a ``show`` command, we have added that as an
1056
alias to ``info`` on all ``gnt-*`` commands.
1057

    
1058
The ``gnt-instance grow-disk`` command has a new mode in which it can
1059
accept the target size of the disk, instead of the delta; this can be
1060
more safe since two runs in absolute mode will be idempotent, and
1061
sometimes it's also easier to specify the desired size directly.
1062

    
1063
Also the handling of instances with regard to offline secondaries has
1064
been improved. Instance operations should not fail because one of it's
1065
secondary nodes is offline, even though it's safe to proceed.
1066

    
1067
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1068
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1069
minors to instance name.
1070

    
1071
API changes
1072
~~~~~~~~~~~
1073

    
1074
RAPI coverage has improved, with (for example) new resources for
1075
recreate-disks, node power-cycle, etc.
1076

    
1077
Compatibility
1078
~~~~~~~~~~~~~
1079

    
1080
There is partial support for ``xl`` in the Xen hypervisor; feedback is
1081
welcome.
1082

    
1083
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1084
whether to still support Python 2.4 or move to Python 2.6 as minimum
1085
required version.
1086

    
1087
Support for Fedora has been slightly improved; the provided example
1088
init.d script should work better on it and the INSTALL file should
1089
document the needed dependencies.
1090

    
1091
Internal changes
1092
~~~~~~~~~~~~~~~~
1093

    
1094
The deprecated ``QueryLocks`` LUXI request has been removed. Use
1095
``Query(what=QR_LOCK, ...)`` instead.
1096

    
1097
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1098
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1099
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1100
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1101
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1102

    
1103
RAPI client: ``CertificateError`` now derives from
1104
``GanetiApiError``. This should make it more easy to handle Ganeti
1105
errors.
1106

    
1107
Deprecation warnings due to PyCrypto/paramiko import in
1108
``tools/setup-ssh`` have been silenced, as usually they are safe; please
1109
make sure to run an up-to-date paramiko version, if you use this tool.
1110

    
1111
The QA scripts now depend on Python 2.5 or above (the main code base
1112
still works with Python 2.4).
1113

    
1114
The configuration file (``config.data``) is now written without
1115
indentation for performance reasons; if you want to edit it, it can be
1116
re-formatted via ``tools/fmtjson``.
1117

    
1118
A number of bugs has been fixed in the cluster merge tool.
1119

    
1120
``x509`` certification verification (used in import-export) has been
1121
changed to allow the same clock skew as permitted by the cluster
1122
verification. This will remove some rare but hard to diagnose errors in
1123
import-export.
1124

    
1125

    
1126
Version 2.6.0 rc4
1127
-----------------
1128

    
1129
*(Released Thu, 19 Jul 2012)*
1130

    
1131
Very few changes from rc4 to the final release, only bugfixes:
1132

    
1133
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
1134
  instance, fix CDROM booting for KVM instances)
1135
- fixed node group modification of node parameters
1136
- fixed issue in LUClusterVerifyGroup with multi-group clusters
1137
- fixed generation of bash completion to ensure a stable ordering
1138
- fixed a few typos
1139

    
1140

    
1141
Version 2.6.0 rc3
1142
-----------------
1143

    
1144
*(Released Fri, 13 Jul 2012)*
1145

    
1146
Third release candidate for 2.6. The following changes were done from
1147
rc3 to rc4:
1148

    
1149
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1150
- Fixed an inconsistency in the LUXI protocol with the provided
1151
  arguments (NOT backwards compatible)
1152
- Fixed a bug with node groups ipolicy where ``min`` was greater than
1153
  the cluster ``std`` value
1154
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1155
  easier instance debugging on nodes (requires ``hconfd`` to work)
1156

    
1157

    
1158
Version 2.6.0 rc2
1159
-----------------
1160

    
1161
*(Released Tue, 03 Jul 2012)*
1162

    
1163
Second release candidate for 2.6. The following changes were done from
1164
rc2 to rc3:
1165

    
1166
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1167
  master candidates
1168
- Fixed a RAPI regression on missing beparams/memory
1169
- Fixed redistribution of files on offline nodes
1170
- Added possibility to run activate-disks even though secondaries are
1171
  offline. With this change it relaxes also the strictness on some other
1172
  commands which use activate disks internally:
1173
  * ``gnt-instance start|reboot|rename|backup|export``
1174
- Made it possible to remove safely an instance if its secondaries are
1175
  offline
1176
- Made it possible to reinstall even though secondaries are offline
1177

    
1178

    
1179
Version 2.6.0 rc1
1180
-----------------
1181

    
1182
*(Released Mon, 25 Jun 2012)*
1183

    
1184
First release candidate for 2.6. The following changes were done from
1185
rc1 to rc2:
1186

    
1187
- Fixed bugs with disk parameters and ``rbd`` templates as well as
1188
  ``instance_os_add``
1189
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1190
  behaviour. It supports now the modify operation
1191
- ``hcheck`` implemented to analyze cluster health and possibility of
1192
  improving health by rebalance
1193
- ``hbal`` has been improved in dealing with split instances
1194

    
1195

    
1196
Version 2.6.0 beta2
1197
-------------------
1198

    
1199
*(Released Mon, 11 Jun 2012)*
1200

    
1201
Second beta release of 2.6. The following changes were done from beta2
1202
to rc1:
1203

    
1204
- Fixed ``daemon-util`` with non-root user models
1205
- Fixed creation of plain instances with ``--no-wait-for-sync``
1206
- Fix wrong iv_names when running ``cfgupgrade``
1207
- Export more information in RAPI group queries
1208
- Fixed bug when changing instance network interfaces
1209
- Extended burnin to do NIC changes
1210
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1211
- Changed default for DRBD barriers
1212
- Fixed DRBD error reporting for syncer rate
1213
- Verify the options on disk parameters
1214

    
1215
And of course various fixes to documentation and improved unittests and
1216
QA.
1217

    
1218

    
1219
Version 2.6.0 beta1
1220
-------------------
1221

    
1222
*(Released Wed, 23 May 2012)*
1223

    
1224
First beta release of 2.6. The following changes were done from beta1 to
1225
beta2:
1226

    
1227
- integrated patch for distributions without ``start-stop-daemon``
1228
- adapted example init.d script to work on Fedora
1229
- fixed log handling in Haskell daemons
1230
- adapted checks in the watcher for pycurl linked against libnss
1231
- add partial support for ``xl`` instead of ``xm`` for Xen
1232
- fixed a type issue in cluster verification
1233
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1234
  clusters)
1235

    
1236
Plus integrated fixes from the 2.5 branch:
1237

    
1238
- fixed ``kvm-ifup`` to use ``/bin/bash``
1239
- fixed parallel build failures
1240
- KVM live migration when using a custom keymap
1241

    
1242

    
1243
Version 2.5.2
1244
-------------
1245

    
1246
*(Released Tue, 24 Jul 2012)*
1247

    
1248
A small bugfix release, with no new features:
1249

    
1250
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1251
  different default shell (e.g. Debian, Ubuntu)
1252
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1253
  243 and Debian bug #650664)
1254
- fixed compatibility with KVM versions that don't support multiple boot
1255
  devices (fixes Issue 230 and Debian bug #624256)
1256

    
1257
Additionally, a few fixes were done to the build system (fixed parallel
1258
build failures) and to the unittests (fixed race condition in test for
1259
FileID functions, and the default enable/disable mode for QA test is now
1260
customisable).
1261

    
1262

    
1263
Version 2.5.1
1264
-------------
1265

    
1266
*(Released Fri, 11 May 2012)*
1267

    
1268
A small bugfix release.
1269

    
1270
The main issues solved are on the topic of compatibility with newer LVM
1271
releases:
1272

    
1273
- fixed parsing of ``lv_attr`` field
1274
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1275
  the ``--force`` flag is needed
1276

    
1277
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1278
to accept kernel 3.x too (was hardcoded to 2.6.*).
1279

    
1280
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1281
of hook results and that also made display of other errors suboptimal
1282
was fixed; the code behaves now like 2.4 and earlier.
1283

    
1284
Another change in 2.5, the cleanup of the OS scripts environment, is too
1285
aggressive: it removed even the ``PATH`` variable, which requires the OS
1286
scripts to *always* need to export it. Since this is a bit too strict,
1287
we now export a minimal PATH, the same that we export for hooks.
1288

    
1289
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1290
integrated into this release.
1291

    
1292
Finally, a few other miscellaneous changes were done (no new features,
1293
just small improvements):
1294

    
1295
- Fix ``gnt-group --help`` display
1296
- Fix hardcoded Xen kernel path
1297
- Fix grow-disk handling of invalid units
1298
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1299
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1300

    
1301

    
1302
Version 2.5.0
1303
-------------
1304

    
1305
*(Released Thu, 12 Apr 2012)*
1306

    
1307
Incompatible/important changes and bugfixes
1308
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1309

    
1310
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1311
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1312
  to match the underlying LUXI interface.
1313
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1314
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1315
  not maintain backwards-compatibility as the underlying operation
1316
  changed in an incompatible way. The RAPI client can talk to old
1317
  servers, but it needs to be told so as the return value changed.
1318
- When creating file-based instances via RAPI, the ``file_driver``
1319
  parameter no longer defaults to ``loop`` and must be specified.
1320
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1321
  ``link`` instead.
1322
- Support for the undocumented and deprecated RAPI instance creation
1323
  request format version 0 has been dropped. Use version 1, supported
1324
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1325
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1326
  documentation <install>`.
1327
- The "cluster-verify" hooks are now executed per group by the
1328
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1329
  if you just run ``gnt-cluster verify``, which generates one opcode per
1330
  group.
1331
- The environment as passed to the OS scripts is cleared, and thus no
1332
  environment variables defined in the node daemon's environment will be
1333
  inherited by the scripts.
1334
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1335
  deprecated.
1336
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1337
  support operations involving multiple node groups.
1338
- Offline nodes are ignored when failing over an instance.
1339
- Support for KVM version 1.0, which changed the version reporting format
1340
  from 3 to 2 digits.
1341
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1342
  removal.
1343
- ``Makefile`` is now compatible with Automake 1.11.2
1344
- Includes all bugfixes made in the 2.4 series
1345

    
1346
New features
1347
~~~~~~~~~~~~
1348

    
1349
- The ganeti-htools project has been merged into the ganeti-core source
1350
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1351
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1352
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1353
  GPT-style partition tables (requires `parted
1354
  <http://www.gnu.org/s/parted/>`_).
1355
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1356
- Allowed adding tags on instance creation.
1357
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1358
- Allow instances to be started in a paused state, enabling the user to
1359
  see the complete console output on boot using the console.
1360
- Added new hypervisor flag to control default reboot behaviour
1361
  (``reboot_behavior``).
1362
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1363
- Improved out-of-band management support:
1364

    
1365
  - Added ``gnt-node health`` command reporting the health status of
1366
    nodes.
1367
  - Added ``gnt-node power`` command to manage power status of nodes.
1368
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1369

    
1370
- Instance migration can fall back to failover if instance is not
1371
  running.
1372
- Filters can be used when listing nodes, instances, groups and locks;
1373
  see :manpage:`ganeti(7)` manpage.
1374
- Added post-execution status as variables to :doc:`hooks <hooks>`
1375
  environment.
1376
- Instance tags are exported/imported together with the instance.
1377
- When given an explicit job ID, ``gnt-job info`` will work for archived
1378
  jobs.
1379
- Jobs can define dependencies on other jobs (not yet supported via
1380
  RAPI or command line, but used by internal commands and usable via
1381
  LUXI).
1382

    
1383
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1384
    dependencies.
1385

    
1386
- Instance failover is now available as a RAPI resource
1387
  (``/2/instances/[instance_name]/failover``).
1388
- ``gnt-instance info`` defaults to static information if primary node
1389
  is offline.
1390
- Opcodes have a new ``comment`` attribute.
1391
- Added basic SPICE support to KVM hypervisor.
1392
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1393

    
1394
Node group improvements
1395
~~~~~~~~~~~~~~~~~~~~~~~
1396

    
1397
- ``gnt-cluster verify`` has been modified to check groups separately,
1398
  thereby improving performance.
1399
- Node group support has been added to ``gnt-cluster verify-disks``,
1400
  which now operates per node group.
1401
- Watcher has been changed to work better with node groups.
1402

    
1403
  - One process and state file per node group.
1404
  - Slow watcher in one group doesn't block other group's watcher.
1405

    
1406
- Added new command, ``gnt-group evacuate``, to move all instances in a
1407
  node group to other groups.
1408
- Added ``gnt-instance change-group`` to move an instance to another
1409
  node group.
1410
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1411
  per-group operations.
1412
- Node groups can be tagged.
1413
- Some operations switch from an exclusive to a shared lock as soon as
1414
  possible.
1415
- Instance's primary and secondary nodes' groups are now available as
1416
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1417
  and ``snodes.group.uuid``).
1418

    
1419
Misc
1420
~~~~
1421

    
1422
- Numerous updates to documentation and manpages.
1423

    
1424
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1425
    descriptions.
1426
  - Some opcode/job results are now also documented, see :doc:`RAPI
1427
    <rapi>`.
1428

    
1429
- A lockset's internal lock is now also visible in lock monitor.
1430
- Log messages from job queue workers now contain information about the
1431
  opcode they're processing.
1432
- ``gnt-instance console`` no longer requires the instance lock.
1433
- A short delay when waiting for job changes reduces the number of LUXI
1434
  requests significantly.
1435
- DRBD metadata volumes are overwritten with zeros during disk creation.
1436
- Out-of-band commands no longer acquire the cluster lock in exclusive
1437
  mode.
1438
- ``devel/upload`` now uses correct permissions for directories.
1439

    
1440

    
1441
Version 2.5.0 rc6
1442
-----------------
1443

    
1444
*(Released Fri, 23 Mar 2012)*
1445

    
1446
This was the sixth release candidate of the 2.5 series.
1447

    
1448

    
1449
Version 2.5.0 rc5
1450
-----------------
1451

    
1452
*(Released Mon, 9 Jan 2012)*
1453

    
1454
This was the fifth release candidate of the 2.5 series.
1455

    
1456

    
1457
Version 2.5.0 rc4
1458
-----------------
1459

    
1460
*(Released Thu, 27 Oct 2011)*
1461

    
1462
This was the fourth release candidate of the 2.5 series.
1463

    
1464

    
1465
Version 2.5.0 rc3
1466
-----------------
1467

    
1468
*(Released Wed, 26 Oct 2011)*
1469

    
1470
This was the third release candidate of the 2.5 series.
1471

    
1472

    
1473
Version 2.5.0 rc2
1474
-----------------
1475

    
1476
*(Released Tue, 18 Oct 2011)*
1477

    
1478
This was the second release candidate of the 2.5 series.
1479

    
1480

    
1481
Version 2.5.0 rc1
1482
-----------------
1483

    
1484
*(Released Tue, 4 Oct 2011)*
1485

    
1486
This was the first release candidate of the 2.5 series.
1487

    
1488

    
1489
Version 2.5.0 beta3
1490
-------------------
1491

    
1492
*(Released Wed, 31 Aug 2011)*
1493

    
1494
This was the third beta release of the 2.5 series.
1495

    
1496

    
1497
Version 2.5.0 beta2
1498
-------------------
1499

    
1500
*(Released Mon, 22 Aug 2011)*
1501

    
1502
This was the second beta release of the 2.5 series.
1503

    
1504

    
1505
Version 2.5.0 beta1
1506
-------------------
1507

    
1508
*(Released Fri, 12 Aug 2011)*
1509

    
1510
This was the first beta release of the 2.5 series.
1511

    
1512

    
1513
Version 2.4.5
1514
-------------
1515

    
1516
*(Released Thu, 27 Oct 2011)*
1517

    
1518
- Fixed bug when parsing command line parameter values ending in
1519
  backslash
1520
- Fixed assertion error after unclean master shutdown
1521
- Disable HTTP client pool for RPC, significantly reducing memory usage
1522
  of master daemon
1523
- Fixed queue archive creation with wrong permissions
1524

    
1525

    
1526
Version 2.4.4
1527
-------------
1528

    
1529
*(Released Tue, 23 Aug 2011)*
1530

    
1531
Small bug-fixes:
1532

    
1533
- Fixed documentation for importing with ``--src-dir`` option
1534
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1535
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1536

    
1537

    
1538
Version 2.4.3
1539
-------------
1540

    
1541
*(Released Fri, 5 Aug 2011)*
1542

    
1543
Many bug-fixes and a few small features:
1544

    
1545
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1546
  issues when you tried to add an instance with two MAC addresses in one
1547
  request
1548
- KVM: fixed per-instance stored UID value
1549
- KVM: configure bridged NICs at migration start
1550
- KVM: Fix a bug where instance will not start with never KVM versions
1551
  (>= 0.14)
1552
- Added OS search path to ``gnt-cluster info``
1553
- Fixed an issue with ``file_storage_dir`` where you were forced to
1554
  provide an absolute path, but the documentation states it is a
1555
  relative path, the documentation was right
1556
- Added a new parameter to instance stop/start called ``--no-remember``
1557
  that will make the state change to not be remembered
1558
- Implemented ``no_remember`` at RAPI level
1559
- Improved the documentation
1560
- Node evacuation: don't call IAllocator if node is already empty
1561
- Fixed bug in DRBD8 replace disks on current nodes
1562
- Fixed bug in recreate-disks for DRBD instances
1563
- Moved assertion checking locks in ``gnt-instance replace-disks``
1564
  causing it to abort with not owning the right locks for some situation
1565
- Job queue: Fixed potential race condition when cancelling queued jobs
1566
- Fixed off-by-one bug in job serial generation
1567
- ``gnt-node volumes``: Fix instance names
1568
- Fixed aliases in bash completion
1569
- Fixed a bug in reopening log files after being sent a SIGHUP
1570
- Added a flag to burnin to allow specifying VCPU count
1571
- Bugfixes to non-root Ganeti configuration
1572

    
1573

    
1574
Version 2.4.2
1575
-------------
1576

    
1577
*(Released Thu, 12 May 2011)*
1578

    
1579
Many bug-fixes and a few new small features:
1580

    
1581
- Fixed a bug related to log opening failures
1582
- Fixed a bug in instance listing with orphan instances
1583
- Fixed a bug which prevented resetting the cluster-level node parameter
1584
  ``oob_program`` to the default
1585
- Many fixes related to the ``cluster-merge`` tool
1586
- Fixed a race condition in the lock monitor, which caused failures
1587
  during (at least) creation of many instances in parallel
1588
- Improved output for gnt-job info
1589
- Removed the quiet flag on some ssh calls which prevented debugging
1590
  failures
1591
- Improved the N+1 failure messages in cluster verify by actually
1592
  showing the memory values (needed and available)
1593
- Increased lock attempt timeouts so that when executing long operations
1594
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1595
  too early and thus prevent the use of the 'fair' mechanism
1596
- Changed instance query data (``gnt-instance info``) to not acquire
1597
  locks unless needed, thus allowing its use on locked instance if only
1598
  static information is asked for
1599
- Improved behaviour with filesystems that do not support rename on an
1600
  opened file
1601
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1602
  kept locks on all nodes during the wipe, which is unneeded
1603
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1604
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1605
  10GiB) which caused the wipe to fail right at the end in some cases
1606
- Fixed master IP activation when doing master failover with no-voting
1607
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1608
  the master node itself
1609
- Fixed potential data-loss in under disk full conditions, where Ganeti
1610
  wouldn't check correctly the return code and would consider
1611
  partially-written files 'correct'
1612
- Fixed bug related to multiple VGs and DRBD disk replacing
1613
- Added new disk parameter ``metavg`` that allows placement of the meta
1614
  device for DRBD in a different volume group
1615
- Fixed error handling in the node daemon when the system libc doesn't
1616
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1617
- Fixed lock release during replace-disks, which kept cluster-wide locks
1618
  when doing disk replaces with an iallocator script
1619
- Added check for missing bridges in cluster verify
1620
- Handle EPIPE errors while writing to the terminal better, so that
1621
  piping the output to e.g. ``less`` doesn't cause a backtrace
1622
- Fixed rare case where a ^C during Luxi calls could have been
1623
  interpreted as server errors, instead of simply terminating
1624
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1625
  assign-nodes``)
1626
- Added a few more parameters to the KVM hypervisor, allowing a second
1627
  CDROM, custom disk type for CDROMs and a floppy image
1628
- Removed redundant message in instance rename when the name is given
1629
  already as a FQDN
1630
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1631
  disks on new nodes, allowing recreation when the original instance
1632
  nodes are completely gone
1633
- Added option when converting disk templates to DRBD to skip waiting
1634
  for the resync, in order to make the instance available sooner
1635
- Added two new variables to the OS scripts environment (containing the
1636
  instance's nodes)
1637
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1638
  to allow use of ``pvgrub`` as bootloader
1639
- Changed the instance memory modifications to only check out-of-memory
1640
  conditions on memory increases, and turned the secondary node warnings
1641
  into errors (they can still be overridden via ``--force``)
1642
- Fixed the handling of a corner case when the Python installation gets
1643
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1644
  to execute a command that doesn't exist
1645
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1646
  node of the instance returned failures during instance shutdown; this
1647
  adds the option ``--ignore-consistency`` to gnt-instance move
1648

    
1649
And as usual, various improvements to the error messages, documentation
1650
and man pages.
1651

    
1652

    
1653
Version 2.4.1
1654
-------------
1655

    
1656
*(Released Wed, 09 Mar 2011)*
1657

    
1658
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1659
the RAPI users file if run twice (even with ``--dry-run``).
1660

    
1661
The release fixes that bug (nothing else changed).
1662

    
1663

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

    
1667
*(Released Mon, 07 Mar 2011)*
1668

    
1669
Final 2.4.0 release. Just a few small fixes:
1670

    
1671
- Fixed RAPI node evacuate
1672
- Fixed the kvm-ifup script
1673
- Fixed internal error handling for special job cases
1674
- Updated man page to specify the escaping feature for options
1675

    
1676

    
1677
Version 2.4.0 rc3
1678
-----------------
1679

    
1680
*(Released Mon, 28 Feb 2011)*
1681

    
1682
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1683
that this feature wiped the disks of the wrong instance, leading to loss
1684
of data.
1685

    
1686
Other changes:
1687

    
1688
- Fixed title of query field containing instance name
1689
- Expanded the glossary in the documentation
1690
- Fixed one unittest (internal issue)
1691

    
1692

    
1693
Version 2.4.0 rc2
1694
-----------------
1695

    
1696
*(Released Mon, 21 Feb 2011)*
1697

    
1698
A number of bug fixes plus just a couple functionality changes.
1699

    
1700
On the user-visible side, the ``gnt-* list`` command output has changed
1701
with respect to "special" field states. The current rc1 style of display
1702
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1703
the default output mode special fields states are displayed as follows:
1704

    
1705
- Offline resource: ``*``
1706
- Unavailable/not applicable: ``-``
1707
- Data missing (RPC failure): ``?``
1708
- Unknown field: ``??``
1709

    
1710
Another user-visible change is the addition of ``--force-join`` to
1711
``gnt-node add``.
1712

    
1713
As for bug fixes:
1714

    
1715
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1716
- Fixed regression in RAPI/instance reinstall where all parameters were
1717
  required (instead of optional)
1718
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1719
- Fixed iallocator usage (offline nodes were not considered offline)
1720
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1721
- Fixed hypervisor and OS parameter validation with respect to
1722
  non-vm_capable nodes
1723
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1724
  cosmetic)
1725
- Fixed ``tools/listrunner`` with respect to agent-based usage
1726

    
1727

    
1728
Version 2.4.0 rc1
1729
-----------------
1730

    
1731
*(Released Fri,  4 Feb 2011)*
1732

    
1733
Many changes and fixes since the beta1 release. While there were some
1734
internal changes, the code has been mostly stabilised for the RC
1735
release.
1736

    
1737
Note: the dumb allocator was removed in this release, as it was not kept
1738
up-to-date with the IAllocator protocol changes. It is recommended to
1739
use the ``hail`` command from the ganeti-htools package.
1740

    
1741
Note: the 2.4 and up versions of Ganeti are not compatible with the
1742
0.2.x branch of ganeti-htools. You need to upgrade to
1743
ganeti-htools-0.3.0 (or later).
1744

    
1745
Regressions fixed from 2.3
1746
~~~~~~~~~~~~~~~~~~~~~~~~~~
1747

    
1748
- Fixed the ``gnt-cluster verify-disks`` command
1749
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1750
  serially on nodes)
1751
- Fixed disk adoption breakage
1752
- Fixed wrong headers in instance listing for field aliases
1753

    
1754
Other bugs fixed
1755
~~~~~~~~~~~~~~~~
1756

    
1757
- Fixed corner case in KVM handling of NICs
1758
- Fixed many cases of wrong handling of non-vm_capable nodes
1759
- Fixed a bug where a missing instance symlink was not possible to
1760
  recreate with any ``gnt-*`` command (now ``gnt-instance
1761
  activate-disks`` does it)
1762
- Fixed the volume group name as reported by ``gnt-cluster
1763
  verify-disks``
1764
- Increased timeouts for the import-export code, hopefully leading to
1765
  fewer aborts due network or instance timeouts
1766
- Fixed bug in ``gnt-node list-storage``
1767
- Fixed bug where not all daemons were started on cluster
1768
  initialisation, but only at the first watcher run
1769
- Fixed many bugs in the OOB implementation
1770
- Fixed watcher behaviour in presence of instances with offline
1771
  secondaries
1772
- Fixed instance list output for instances running on the wrong node
1773
- a few fixes to the cluster-merge tool, but it still cannot merge
1774
  multi-node groups (currently it is not recommended to use this tool)
1775

    
1776

    
1777
Improvements
1778
~~~~~~~~~~~~
1779

    
1780
- Improved network configuration for the KVM hypervisor
1781
- Added e1000 as a supported NIC for Xen-HVM
1782
- Improved the lvmstrap tool to also be able to use partitions, as
1783
  opposed to full disks
1784
- Improved speed of disk wiping (the cluster parameter
1785
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1786
  of instance creations
1787
- Added documentation for the OS parameters
1788
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1789
  hypervisor is not responding
1790
- Added display of blacklisted and hidden OS information in
1791
  ``gnt-cluster info``
1792
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1793
  NIC and node parameters, which might create problems with currently
1794
  invalid (but undetected) configuration files, but prevents validation
1795
  failures when unrelated parameters are modified
1796
- Changed cluster initialisation to wait for the master daemon to become
1797
  available
1798
- Expanded the RAPI interface:
1799

    
1800
  - Added config redistribution resource
1801
  - Added activation/deactivation of instance disks
1802
  - Added export of console information
1803

    
1804
- Implemented log file reopening on SIGHUP, which allows using
1805
  logrotate(8) for the Ganeti log files
1806
- Added a basic OOB helper script as an example
1807

    
1808

    
1809
Version 2.4.0 beta1
1810
-------------------
1811

    
1812
*(Released Fri, 14 Jan 2011)*
1813

    
1814
User-visible
1815
~~~~~~~~~~~~
1816

    
1817
- Fixed timezone issues when formatting timestamps
1818
- Added support for node groups, available via ``gnt-group`` and other
1819
  commands
1820
- Added out-of-band framework and management, see :doc:`design
1821
  document <design-oob>`
1822
- Removed support for roman numbers from ``gnt-node list`` and
1823
  ``gnt-instance list``.
1824
- Allowed modification of master network interface via ``gnt-cluster
1825
  modify --master-netdev``
1826
- Accept offline secondaries while shutting down instance disks
1827
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1828
- Added support for multiple LVM volume groups
1829
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1830
  requested
1831
- Added commands to list available fields:
1832

    
1833
  - ``gnt-node list-fields``
1834
  - ``gnt-group list-fields``
1835
  - ``gnt-instance list-fields``
1836

    
1837
- Updated documentation and man pages
1838

    
1839
Integration
1840
~~~~~~~~~~~
1841

    
1842
- Moved ``rapi_users`` file into separate directory, now named
1843
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1844
  symlink
1845
- Added new tool for running commands on many machines,
1846
  ``tools/ganeti-listrunner``
1847
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1848
  improving the ``gnt-instance console`` output
1849
- Allowed customisation of disk index separator at ``configure`` time
1850
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1851
- Added support for non-partitioned md disks in ``lvmstrap``
1852
- Added script to gracefully power off KVM instances
1853
- Split ``utils`` module into smaller parts
1854
- Changed query operations to return more detailed information, e.g.
1855
  whether an information is unavailable due to an offline node. To use
1856
  this new functionality, the LUXI call ``Query`` must be used. Field
1857
  information is now stored by the master daemon and can be retrieved
1858
  using ``QueryFields``. Instances, nodes and groups can also be queried
1859
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1860
  exposed via RAPI). The following commands make use of this
1861
  infrastructure change:
1862

    
1863
  - ``gnt-group list``
1864
  - ``gnt-group list-fields``
1865
  - ``gnt-node list``
1866
  - ``gnt-node list-fields``
1867
  - ``gnt-instance list``
1868
  - ``gnt-instance list-fields``
1869
  - ``gnt-debug locks``
1870

    
1871
Remote API
1872
~~~~~~~~~~
1873

    
1874
- New RAPI resources (see :doc:`rapi`):
1875

    
1876
  - ``/2/modify``
1877
  - ``/2/groups``
1878
  - ``/2/groups/[group_name]``
1879
  - ``/2/groups/[group_name]/assign-nodes``
1880
  - ``/2/groups/[group_name]/modify``
1881
  - ``/2/groups/[group_name]/rename``
1882
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1883

    
1884
- RAPI changes:
1885

    
1886
  - Implemented ``no_install`` for instance creation
1887
  - Implemented OS parameters for instance reinstallation, allowing
1888
    use of special settings on reinstallation (e.g. for preserving data)
1889

    
1890
Misc
1891
~~~~
1892

    
1893
- Added IPv6 support in import/export
1894
- Pause DRBD synchronization while wiping disks on instance creation
1895
- Updated unittests and QA scripts
1896
- Improved network parameters passed to KVM
1897
- Converted man pages from docbook to reStructuredText
1898

    
1899

    
1900
Version 2.3.1
1901
-------------
1902

    
1903
*(Released Mon, 20 Dec 2010)*
1904

    
1905
Released version 2.3.1~rc1 without any changes.
1906

    
1907

    
1908
Version 2.3.1 rc1
1909
-----------------
1910

    
1911
*(Released Wed, 1 Dec 2010)*
1912

    
1913
- impexpd: Disable OpenSSL compression in socat if possible (backport
1914
  from master, commit e90739d625b, see :doc:`installation guide
1915
  <install-quick>` for details)
1916
- Changed unittest coverage report to exclude test scripts
1917
- Added script to check version format
1918

    
1919

    
1920
Version 2.3.0
1921
-------------
1922

    
1923
*(Released Wed, 1 Dec 2010)*
1924

    
1925
Released version 2.3.0~rc1 without any changes.
1926

    
1927

    
1928
Version 2.3.0 rc1
1929
-----------------
1930

    
1931
*(Released Fri, 19 Nov 2010)*
1932

    
1933
A number of bugfixes and documentation updates:
1934

    
1935
- Update ganeti-os-interface documentation
1936
- Fixed a bug related to duplicate MACs or similar items which should be
1937
  unique
1938
- Fix breakage in OS state modify
1939
- Reinstall instance: disallow offline secondaries (fixes bug related to
1940
  OS changing but reinstall failing)
1941
- plus all the other fixes between 2.2.1 and 2.2.2
1942

    
1943

    
1944
Version 2.3.0 rc0
1945
-----------------
1946

    
1947
*(Released Tue, 2 Nov 2010)*
1948

    
1949
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1950
- Fixed master failover race with watcher
1951
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1952
  configuration
1953
- Accept previously stopped instance for export with instance removal
1954
- Simplify and extend the environment variables for instance OS scripts
1955
- Added new node flags, ``master_capable`` and ``vm_capable``
1956
- Added optional instance disk wiping prior during allocation. This is a
1957
  cluster-wide option and can be set/modified using
1958
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1959
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1960
  :doc:`install-quick`
1961
- Added a new watcher option (``--ignore-pause``)
1962
- Added option to ignore offline node on instance start/stop
1963
  (``--ignore-offline``)
1964
- Allow overriding OS parameters with ``gnt-instance reinstall``
1965
- Added ability to change node's secondary IP address using ``gnt-node
1966
  modify``
1967
- Implemented privilege separation for all daemons except
1968
  ``ganeti-noded``, see ``configure`` options
1969
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1970
  verify``
1971
- Implemented job priorities (see ``ganeti(7)`` manpage)
1972
- Ignore failures while shutting down instances during failover from
1973
  offline node
1974
- Exit daemon's bootstrap process only once daemon is ready
1975
- Export more information via ``LUInstanceQuery``/remote API
1976
- Improved documentation, QA and unittests
1977
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1978
  restart if the file was created or changed
1979
- Added LUXI protocol version sent with each request and response,
1980
  allowing detection of server/client mismatches
1981
- Moved the Python scripts among gnt-* and ganeti-* into modules
1982
- Moved all code related to setting up SSH to an external script,
1983
  ``setup-ssh``
1984
- Infrastructure changes for node group support in future versions
1985

    
1986

    
1987
Version 2.2.2
1988
-------------
1989

    
1990
*(Released Fri, 19 Nov 2010)*
1991

    
1992
A few small bugs fixed, and some improvements to the build system:
1993

    
1994
- Fix documentation regarding conversion to drbd
1995
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1996
  -B``)
1997
- Fix error handling in node modify with multiple changes
1998
- Allow remote imports without checked names
1999

    
2000

    
2001
Version 2.2.1
2002
-------------
2003

    
2004
*(Released Tue, 19 Oct 2010)*
2005

    
2006
- Disable SSL session ID cache in RPC client
2007

    
2008

    
2009
Version 2.2.1 rc1
2010
-----------------
2011

    
2012
*(Released Thu, 14 Oct 2010)*
2013

    
2014
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
2015
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2016
  with GnuTLS
2017
- Fix problems with interaction between Curl and Python's HTTP server,
2018
  resulting in increased speed in many RPC calls
2019
- Improve our release script to prevent breakage with older aclocal and
2020
  Python 2.6
2021

    
2022

    
2023
Version 2.2.1 rc0
2024
-----------------
2025

    
2026
*(Released Thu, 7 Oct 2010)*
2027

    
2028
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2029
  value retrieved from master
2030
- Added support for blacklisted or hidden OS definitions
2031
- Added simple lock monitor (accessible via (``gnt-debug locks``)
2032
- Added support for -mem-path in KVM hypervisor abstraction layer
2033
- Allow overriding instance parameters in tool for inter-cluster
2034
  instance moves (``tools/move-instance``)
2035
- Improved opcode summaries (e.g. in ``gnt-job list``)
2036
- Improve consistency of OS listing by sorting it
2037
- Documentation updates
2038

    
2039

    
2040
Version 2.2.0.1
2041
---------------
2042

    
2043
*(Released Fri, 8 Oct 2010)*
2044

    
2045
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
2046

    
2047

    
2048
Version 2.2.0
2049
-------------
2050

    
2051
*(Released Mon, 4 Oct 2010)*
2052

    
2053
- Fixed regression in ``gnt-instance rename``
2054

    
2055

    
2056
Version 2.2.0 rc2
2057
-----------------
2058

    
2059
*(Released Wed, 22 Sep 2010)*
2060

    
2061
- Fixed OS_VARIANT variable for OS scripts
2062
- Fixed cluster tag operations via RAPI
2063
- Made ``setup-ssh`` exit with non-zero code if an error occurred
2064
- Disabled RAPI CA checks in watcher
2065

    
2066

    
2067
Version 2.2.0 rc1
2068
-----------------
2069

    
2070
*(Released Mon, 23 Aug 2010)*
2071

    
2072
- Support DRBD versions of the format "a.b.c.d"
2073
- Updated manpages
2074
- Re-introduce support for usage from multiple threads in RAPI client
2075
- Instance renames and modify via RAPI
2076
- Work around race condition between processing and archival in job
2077
  queue
2078
- Mark opcodes following failed one as failed, too
2079
- Job field ``lock_status`` was removed due to difficulties making it
2080
  work with the changed job queue in Ganeti 2.2; a better way to monitor
2081
  locks is expected for a later 2.2.x release
2082
- Fixed dry-run behaviour with many commands
2083
- Support ``ssh-agent`` again when adding nodes
2084
- Many additional bugfixes
2085

    
2086

    
2087
Version 2.2.0 rc0
2088
-----------------
2089

    
2090
*(Released Fri, 30 Jul 2010)*
2091

    
2092
Important change: the internal RPC mechanism between Ganeti nodes has
2093
changed from using a home-grown http library (based on the Python base
2094
libraries) to use the PycURL library. This requires that PycURL is
2095
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2096
against GnuTLS by default. cURL's support for GnuTLS had known issues
2097
before cURL 7.21.0 and we recommend using the latest cURL release or
2098
linking against OpenSSL. Most other distributions already link PycURL
2099
and cURL against OpenSSL. The command::
2100

    
2101
  python -c 'import pycurl; print pycurl.version'
2102

    
2103
can be used to determine the libraries PycURL and cURL are linked
2104
against.
2105

    
2106
Other significant changes:
2107

    
2108
- Rewrote much of the internals of the job queue, in order to achieve
2109
  better parallelism; this decouples job query operations from the job
2110
  processing, and it should allow much nicer behaviour of the master
2111
  daemon under load, and it also has uncovered some long-standing bugs
2112
  related to the job serialisation (now fixed)
2113
- Added a default iallocator setting to the cluster parameters,
2114
  eliminating the need to always pass nodes or an iallocator for
2115
  operations that require selection of new node(s)
2116
- Added experimental support for the LXC virtualization method
2117
- Added support for OS parameters, which allows the installation of
2118
  instances to pass parameter to OS scripts in order to customise the
2119
  instance
2120
- Added a hypervisor parameter controlling the migration type (live or
2121
  non-live), since hypervisors have various levels of reliability; this
2122
  has renamed the 'live' parameter to 'mode'
2123
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
2124
  logical volumes, meaning that cluster verify will ignore them and not
2125
  flag their presence as errors
2126
- The watcher will now reset the error count for failed instances after
2127
  8 hours, thus allowing self-healing if the problem that caused the
2128
  instances to be down/fail to start has cleared in the meantime
2129
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2130
  check for, and warn, if the drbd module parameter ``usermode_helper``
2131
  is not consistent with the cluster-wide setting; this is needed to
2132
  make diagnose easier of failed drbd creations
2133
- Started adding base IPv6 support, but this is not yet
2134
  enabled/available for use
2135
- Rename operations (cluster, instance) will now return the new name,
2136
  which is especially useful if a short name was passed in
2137
- Added support for instance migration in RAPI
2138
- Added a tool to pre-configure nodes for the SSH setup, before joining
2139
  them to the cluster; this will allow in the future a simplified model
2140
  for node joining (but not yet fully enabled in 2.2); this needs the
2141
  paramiko python library
2142
- Fixed handling of name-resolving errors
2143
- Fixed consistency of job results on the error path
2144
- Fixed master-failover race condition when executed multiple times in
2145
  sequence
2146
- Fixed many bugs related to the job queue (mostly introduced during the
2147
  2.2 development cycle, so not all are impacting 2.1)
2148
- Fixed instance migration with missing disk symlinks
2149
- Fixed handling of unknown jobs in ``gnt-job archive``
2150
- And many other small fixes/improvements
2151

    
2152
Internal changes:
2153

    
2154
- Enhanced both the unittest and the QA coverage
2155
- Switched the opcode validation to a generic model, and extended the
2156
  validation to all opcode parameters
2157
- Changed more parts of the code that write shell scripts to use the
2158
  same class for this
2159
- Switched the master daemon to use the asyncore library for the Luxi
2160
  server endpoint
2161

    
2162

    
2163
Version 2.2.0 beta0
2164
-------------------
2165

    
2166
*(Released Thu, 17 Jun 2010)*
2167

    
2168
- Added tool (``move-instance``) and infrastructure to move instances
2169
  between separate clusters (see :doc:`separate documentation
2170
  <move-instance>` and :doc:`design document <design-2.2>`)
2171
- Added per-request RPC timeout
2172
- RAPI now requires a Content-Type header for requests with a body (e.g.
2173
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2174
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
2175
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2176
  reachable
2177
- Implemented initial support for running Ganeti daemons as separate
2178
  users, see configure-time flags ``--with-user-prefix`` and
2179
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2180
  time)
2181
- Instances can be removed after export (``gnt-backup export
2182
  --remove-instance``)
2183
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2184
  key (instead of 1024 bit)
2185
- Added new cluster configuration file for cluster domain secret
2186
- Import/export now use SSL instead of SSH
2187
- Added support for showing estimated time when exporting an instance,
2188
  see the ``ganeti-os-interface(7)`` manpage and look for
2189
  ``EXP_SIZE_FD``
2190

    
2191

    
2192
Version 2.1.8
2193
-------------
2194

    
2195
*(Released Tue, 16 Nov 2010)*
2196

    
2197
Some more bugfixes. Unless critical bugs occur, this will be the last
2198
2.1 release:
2199

    
2200
- Fix case of MAC special-values
2201
- Fix mac checker regex
2202
- backend: Fix typo causing "out of range" error
2203
- Add missing --units in gnt-instance list man page
2204

    
2205

    
2206
Version 2.1.7
2207
-------------
2208

    
2209
*(Released Tue, 24 Aug 2010)*
2210

    
2211
Bugfixes only:
2212
  - Don't ignore secondary node silently on non-mirrored disk templates
2213
    (issue 113)
2214
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2215
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2216
  - Properly document the usb_mouse parameter
2217
  - Fix path in ganeti-rapi(8) (issue 116)
2218
  - Adjust error message when the ganeti user's .ssh directory is
2219
    missing
2220
  - Add same-node-check when changing the disk template to drbd
2221

    
2222

    
2223
Version 2.1.6
2224
-------------
2225

    
2226
*(Released Fri, 16 Jul 2010)*
2227

    
2228
Bugfixes only:
2229
  - Add an option to only select some reboot types during qa/burnin.
2230
    (on some hypervisors consequent reboots are not supported)
2231
  - Fix infrequent race condition in master failover. Sometimes the old
2232
    master ip address would be still detected as up for a short time
2233
    after it was removed, causing failover to fail.
2234
  - Decrease mlockall warnings when the ctypes module is missing. On
2235
    Python 2.4 we support running even if no ctypes module is installed,
2236
    but we were too verbose about this issue.
2237
  - Fix building on old distributions, on which man doesn't have a
2238
    --warnings option.
2239
  - Fix RAPI not to ignore the MAC address on instance creation
2240
  - Implement the old instance creation format in the RAPI client.
2241

    
2242

    
2243
Version 2.1.5
2244
-------------
2245

    
2246
*(Released Thu, 01 Jul 2010)*
2247

    
2248
A small bugfix release:
2249
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2250
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2251
    a non-existing option
2252
  - Fix instance create: the --force-variant option was ignored
2253
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2254
  - Fix modify node storage with non-FQDN arguments
2255
  - Fix RAPI client to authenticate under Python 2.6 when used
2256
    for more than 5 requests needing authentication
2257
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2258
    when converting a non-shutdown drbd instance to plain
2259

    
2260

    
2261
Version 2.1.4
2262
-------------
2263

    
2264
*(Released Fri, 18 Jun 2010)*
2265

    
2266
A small bugfix release:
2267

    
2268
  - Fix live migration of KVM instances started with older Ganeti
2269
    versions which had fewer hypervisor parameters
2270
  - Fix gnt-instance grow-disk on down instances
2271
  - Fix an error-reporting bug during instance migration
2272
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2273
    silently ignoring broken ones
2274
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2275
    users
2276
  - Fix bug triggered by different API version os-es on different nodes
2277
  - Fix a bug in instance startup with custom hvparams: OS level
2278
    parameters would fail to be applied.
2279
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2280
    make it work completely well with OpenSSL)
2281
  - Fix handling of errors when resolving names from DNS
2282

    
2283

    
2284
Version 2.1.3
2285
-------------
2286

    
2287
*(Released Thu, 3 Jun 2010)*
2288

    
2289
A medium sized development cycle. Some new features, and some
2290
fixes/small improvements/cleanups.
2291

    
2292
Significant features
2293
~~~~~~~~~~~~~~~~~~~~
2294

    
2295
The node deamon now tries to mlock itself into memory, unless the
2296
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2297
its logs, and falls back to console logging. This allows emergency
2298
features such as ``gnt-node powercycle`` to work even in the event of a
2299
broken node disk (tested offlining the disk hosting the node's
2300
filesystem and dropping its memory caches; don't try this at home)
2301

    
2302
KVM: add vhost-net acceleration support. It can be tested with a new
2303
enough version of the kernel and of qemu-kvm.
2304

    
2305
KVM: Add instance chrooting feature. If you use privilege dropping for
2306
your VMs you can also now force them to chroot to an empty directory,
2307
before starting the emulated guest.
2308

    
2309
KVM: Add maximum migration bandwith and maximum downtime tweaking
2310
support (requires a new-enough version of qemu-kvm).
2311

    
2312
Cluster verify will now warn if the master node doesn't have the master
2313
ip configured on it.
2314

    
2315
Add a new (incompatible) instance creation request format to RAPI which
2316
supports all parameters (previously only a subset was supported, and it
2317
wasn't possible to extend the old format to accomodate all the new
2318
features. The old format is still supported, and a client can check for
2319
this feature, before using it, by checking for its presence in the
2320
``features`` RAPI resource.
2321

    
2322
Now with ancient latin support. Try it passing the ``--roman`` option to
2323
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2324
(requires the python-roman module to be installed, in order to work).
2325

    
2326
Other changes
2327
~~~~~~~~~~~~~
2328

    
2329
As usual many internal code refactorings, documentation updates, and
2330
such. Among others:
2331

    
2332
  - Lots of improvements and cleanups to the experimental Remote API
2333
    (RAPI) client library.
2334
  - A new unit test suite for the core daemon libraries.
2335
  - A fix to creating missing directories makes sure the umask is not
2336
    applied anymore. This enforces the same directory permissions
2337
    everywhere.
2338
  - Better handling terminating daemons with ctrl+c (used when running
2339
    them in debugging mode).
2340
  - Fix a race condition in live migrating a KVM instance, when stat()
2341
    on the old proc status file returned EINVAL, which is an unexpected
2342
    value.
2343
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2344
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2345

    
2346

    
2347
Version 2.1.2.1
2348
---------------
2349

    
2350
*(Released Fri, 7 May 2010)*
2351

    
2352
Fix a bug which prevented untagged KVM instances from starting.
2353

    
2354

    
2355
Version 2.1.2
2356
-------------
2357

    
2358
*(Released Fri, 7 May 2010)*
2359

    
2360
Another release with a long development cycle, during which many
2361
different features were added.
2362

    
2363
Significant features
2364
~~~~~~~~~~~~~~~~~~~~
2365

    
2366
The KVM hypervisor now can run the individual instances as non-root, to
2367
reduce the impact of a VM being hijacked due to bugs in the
2368
hypervisor. It is possible to run all instances as a single (non-root)
2369
user, to manually specify a user for each instance, or to dynamically
2370
allocate a user out of a cluster-wide pool to each instance, with the
2371
guarantee that no two instances will run under the same user ID on any
2372
given node.
2373

    
2374
An experimental RAPI client library, that can be used standalone
2375
(without the other Ganeti libraries), is provided in the source tree as
2376
``lib/rapi/client.py``. Note this client might change its interface in
2377
the future, as we iterate on its capabilities.
2378

    
2379
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2380
replace the cluster's certificates and crypto keys. This might help in
2381
case they have been compromised, or have simply expired.
2382

    
2383
A new disk option for instance creation has been added that allows one
2384
to "adopt" currently existing logical volumes, with data
2385
preservation. This should allow easier migration to Ganeti from
2386
unmanaged (or managed via other software) instances.
2387

    
2388
Another disk improvement is the possibility to convert between redundant
2389
(DRBD) and plain (LVM) disk configuration for an instance. This should
2390
allow better scalability (starting with one node and growing the
2391
cluster, or shrinking a two-node cluster to one node).
2392

    
2393
A new feature that could help with automated node failovers has been
2394
implemented: if a node sees itself as offline (by querying the master
2395
candidates), it will try to shutdown (hard) all instances and any active
2396
DRBD devices. This reduces the risk of duplicate instances if an
2397
external script automatically failovers the instances on such nodes. To
2398
enable this, the cluster parameter ``maintain_node_health`` should be
2399
enabled; in the future this option (per the name) will enable other
2400
automatic maintenance features.
2401

    
2402
Instance export/import now will reuse the original instance
2403
specifications for all parameters; that means exporting an instance,
2404
deleting it and the importing it back should give an almost identical
2405
instance. Note that the default import behaviour has changed from
2406
before, where it created only one NIC; now it recreates the original
2407
number of NICs.
2408

    
2409
Cluster verify has added a few new checks: SSL certificates validity,
2410
/etc/hosts consistency across the cluster, etc.
2411

    
2412
Other changes
2413
~~~~~~~~~~~~~
2414

    
2415
As usual, many internal changes were done, documentation fixes,
2416
etc. Among others:
2417

    
2418
- Fixed cluster initialization with disabled cluster storage (regression
2419
  introduced in 2.1.1)
2420
- File-based storage supports growing the disks
2421
- Fixed behaviour of node role changes
2422
- Fixed cluster verify for some corner cases, plus a general rewrite of
2423
  cluster verify to allow future extension with more checks
2424
- Fixed log spamming by watcher and node daemon (regression introduced
2425
  in 2.1.1)
2426
- Fixed possible validation issues when changing the list of enabled
2427
  hypervisors
2428
- Fixed cleanup of /etc/hosts during node removal
2429
- Fixed RAPI response for invalid methods
2430
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2431
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2432
  from ide disks, etc.)
2433
- Allow OS changes without re-installation (to record a changed OS
2434
  outside of Ganeti, or to allow OS renames)
2435
- Allow instance creation without OS installation (useful for example if
2436
  the OS will be installed manually, or restored from a backup not in
2437
  Ganeti format)
2438
- Implemented option to make cluster ``copyfile`` use the replication
2439
  network
2440
- Added list of enabled hypervisors to ssconf (possibly useful for
2441
  external scripts)
2442
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2443
  1.2 clusters
2444
- A partial form of node re-IP is possible via node readd, which now
2445
  allows changed node primary IP
2446
- Command line utilities now show an informational message if the job is
2447
  waiting for a lock
2448
- The logs of the master daemon now show the PID/UID/GID of the
2449
  connected client
2450

    
2451

    
2452
Version 2.1.1
2453
-------------
2454

    
2455
*(Released Fri, 12 Mar 2010)*
2456

    
2457
During the 2.1.0 long release candidate cycle, a lot of improvements and
2458
changes have accumulated with were released later as 2.1.1.
2459

    
2460
Major changes
2461
~~~~~~~~~~~~~
2462

    
2463
The node evacuate command (``gnt-node evacuate``) was significantly
2464
rewritten, and as such the IAllocator protocol was changed - a new
2465
request type has been added. This unfortunate change during a stable
2466
series is designed to improve performance of node evacuations; on
2467
clusters with more than about five nodes and which are well-balanced,
2468
evacuation should proceed in parallel for all instances of the node
2469
being evacuated. As such, any existing IAllocator scripts need to be
2470
updated, otherwise the above command will fail due to the unknown
2471
request. The provided "dumb" allocator has not been updated; but the
2472
ganeti-htools package supports the new protocol since version 0.2.4.
2473

    
2474
Another important change is increased validation of node and instance
2475
names. This might create problems in special cases, if invalid host
2476
names are being used.
2477

    
2478
Also, a new layer of hypervisor parameters has been added, that sits at
2479
OS level between the cluster defaults and the instance ones. This allows
2480
customisation of virtualization parameters depending on the installed
2481
OS. For example instances with OS 'X' may have a different KVM kernel
2482
(or any other parameter) than the cluster defaults. This is intended to
2483
help managing a multiple OSes on the same cluster, without manual
2484
modification of each instance's parameters.
2485

    
2486
A tool for merging clusters, ``cluster-merge``, has been added in the
2487
tools sub-directory.
2488

    
2489
Bug fixes
2490
~~~~~~~~~
2491

    
2492
- Improved the int/float conversions that should make the code more
2493
  robust in face of errors from the node daemons
2494
- Fixed the remove node code in case of internal configuration errors
2495
- Fixed the node daemon behaviour in face of inconsistent queue
2496
  directory (e.g. read-only file-system where we can't open the files
2497
  read-write, etc.)
2498
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2499
  now it either aborts cleanly or, if given the new "auto_promote"
2500
  parameter, will automatically promote other nodes as needed
2501
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2502
  completely prevent Ganeti from working
2503
- Fixed bug for instance export when not all disks were successfully
2504
  exported
2505
- Fixed behaviour of node add when the new node is slow in starting up
2506
  the node daemon
2507
- Fixed handling of signals in the LUXI client, which should improve
2508
  behaviour of command-line scripts
2509
- Added checks for invalid node/instance names in the configuration (now
2510
  flagged during cluster verify)
2511
- Fixed watcher behaviour for disk activation errors
2512
- Fixed two potentially endless loops in http library, which led to the
2513
  RAPI daemon hanging and consuming 100% CPU in some cases
2514
- Fixed bug in RAPI daemon related to hashed passwords
2515
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2516
  instances
2517
- Enhanced compatibility with non-Debian OSes, but not using absolute
2518
  path in some commands and allowing customisation of the ssh
2519
  configuration directory
2520
- Fixed possible future issue with new Python versions by abiding to the
2521
  proper use of ``__slots__`` attribute on classes
2522
- Added checks that should prevent directory traversal attacks
2523
- Many documentation fixes based on feedback from users
2524

    
2525
New features
2526
~~~~~~~~~~~~
2527

    
2528
- Added an "early_release" more for instance replace disks and node
2529
  evacuate, where we release locks earlier and thus allow higher
2530
  parallelism within the cluster
2531
- Added watcher hooks, intended to allow the watcher to restart other
2532
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2533
  course for any other purpose
2534
- Added a compile-time disable for DRBD barriers, to increase
2535
  performance if the administrator trusts the power supply or the
2536
  storage system to not lose writes
2537
- Added the option of using syslog for logging instead of, or in
2538
  addition to, Ganeti's own log files
2539
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2540
  can indeed boot from a paravirtual NIC
2541
- Added a generic debug level for many operations; while this is not
2542
  used widely yet, it allows one to pass the debug value all the way to
2543
  the OS scripts
2544
- Enhanced the hooks environment for instance moves (failovers,
2545
  migrations) where the primary/secondary nodes changed during the
2546
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2547
- Enhanced data validations for many user-supplied values; one important
2548
  item is the restrictions imposed on instance and node names, which
2549
  might reject some (invalid) host names
2550
- Add a configure-time option to disable file-based storage, if it's not
2551
  needed; this allows greater security separation between the master
2552
  node and the other nodes from the point of view of the inter-node RPC
2553
  protocol
2554
- Added user notification in interactive tools if job is waiting in the
2555
  job queue or trying to acquire locks
2556
- Added log messages when a job is waiting for locks
2557
- Added filtering by node tags in instance operations which admit
2558
  multiple instances (start, stop, reboot, reinstall)
2559
- Added a new tool for cluster mergers, ``cluster-merge``
2560
- Parameters from command line which are of the form ``a=b,c=d`` can now
2561
  use backslash escapes to pass in values which contain commas,
2562
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2563
  ``b,c``
2564
- For KVM, the instance name is the first parameter passed to KVM, so
2565
  that it's more visible in the process list
2566

    
2567

    
2568
Version 2.1.0
2569
-------------
2570

    
2571
*(Released Tue, 2 Mar 2010)*
2572

    
2573
Ganeti 2.1 brings many improvements with it. Major changes:
2574

    
2575
- Added infrastructure to ease automated disk repairs
2576
- Added new daemon to export configuration data in a cheaper way than
2577
  using the remote API
2578
- Instance NICs can now be routed instead of being associated with a
2579
  networking bridge
2580
- Improved job locking logic to reduce impact of jobs acquiring multiple
2581
  locks waiting for other long-running jobs
2582

    
2583
In-depth implementation details can be found in the Ganeti 2.1 design
2584
document.
2585

    
2586
Details
2587
~~~~~~~
2588

    
2589
- Added chroot hypervisor
2590
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2591
  ``device_model``)
2592
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2593
  ``bootloader_path`` and ``bootloader_args``)
2594
- Added the ``use_localtime`` option for the xen-hvm and kvm
2595
  hypervisors, and the default value for this has changed to false (in
2596
  2.0 xen-hvm always enabled it)
2597
- Added luxi call to submit multiple jobs in one go
2598
- Added cluster initialization option to not modify ``/etc/hosts``
2599
  file on nodes
2600
- Added network interface parameters
2601
- Added dry run mode to some LUs
2602
- Added RAPI resources:
2603

    
2604
  - ``/2/instances/[instance_name]/info``
2605
  - ``/2/instances/[instance_name]/replace-disks``
2606
  - ``/2/nodes/[node_name]/evacuate``
2607
  - ``/2/nodes/[node_name]/migrate``
2608
  - ``/2/nodes/[node_name]/role``
2609
  - ``/2/nodes/[node_name]/storage``
2610
  - ``/2/nodes/[node_name]/storage/modify``
2611
  - ``/2/nodes/[node_name]/storage/repair``
2612

    
2613
- Added OpCodes to evacuate or migrate all instances on a node
2614
- Added new command to list storage elements on nodes (``gnt-node
2615
  list-storage``) and modify them (``gnt-node modify-storage``)
2616
- Added new ssconf files with master candidate IP address
2617
  (``ssconf_master_candidates_ips``), node primary IP address
2618
  (``ssconf_node_primary_ips``) and node secondary IP address
2619
  (``ssconf_node_secondary_ips``)
2620
- Added ``ganeti-confd`` and a client library to query the Ganeti
2621
  configuration via UDP
2622
- Added ability to run hooks after cluster initialization and before
2623
  cluster destruction
2624
- Added automatic mode for disk replace (``gnt-instance replace-disks
2625
  --auto``)
2626
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2627
  after catastrophic data-loss
2628
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2629
  groups
2630
- Added ``gnt-instance move`` command to move instances
2631
- Added ``gnt-cluster watcher`` command to control watcher
2632
- Added ``gnt-node powercycle`` command to powercycle nodes
2633
- Added new job status field ``lock_status``
2634
- Added parseable error codes to cluster verification (``gnt-cluster
2635
  verify --error-codes``) and made output less verbose (use
2636
  ``--verbose`` to restore previous behaviour)
2637
- Added UUIDs to the main config entities (cluster, nodes, instances)
2638
- Added support for OS variants
2639
- Added support for hashed passwords in the Ganeti remote API users file
2640
  (``rapi_users``)
2641
- Added option to specify maximum timeout on instance shutdown
2642
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2643
- Added new helper script to start and stop Ganeti daemons
2644
  (``daemon-util``), with the intent to reduce the work necessary to
2645
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2646
  from one place
2647
- Added more unittests
2648
- Fixed critical bug in ganeti-masterd startup
2649
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2650
  now customisable at the cluster level for both the KVM and Xen
2651
  hypervisors using the new ``migration_port`` parameter
2652
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2653
  reinstalling an instance
2654
- Allowed ``@`` in tag names
2655
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2656
- Many documentation updates
2657
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2658
- ``gnt-instance reinstall`` can now reinstall multiple instances
2659
- Updated many command line parameters
2660
- Introduced new OS API version 15
2661
- No longer support a default hypervisor
2662
- Treat virtual LVs as inexistent
2663
- Improved job locking logic to reduce lock contention
2664
- Match instance and node names case insensitively
2665
- Reimplemented bash completion script to be more complete
2666
- Improved burnin
2667

    
2668

    
2669
Version 2.0.6
2670
-------------
2671

    
2672
*(Released Thu, 4 Feb 2010)*
2673

    
2674
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2675
- Fix a string formatting bug
2676
- Improve safety of the code in some error paths
2677
- Improve data validation in the master of values returned from nodes
2678

    
2679

    
2680
Version 2.0.5
2681
-------------
2682

    
2683
*(Released Thu, 17 Dec 2009)*
2684

    
2685
- Fix security issue due to missing validation of iallocator names; this
2686
  allows local and remote execution of arbitrary executables
2687
- Fix failure of gnt-node list during instance removal
2688
- Ship the RAPI documentation in the archive
2689

    
2690

    
2691
Version 2.0.4
2692
-------------
2693

    
2694
*(Released Wed, 30 Sep 2009)*
2695

    
2696
- Fixed many wrong messages
2697
- Fixed a few bugs related to the locking library
2698
- Fixed MAC checking at instance creation time
2699
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2700
- Fixed a few issues related to signal handling in both daemons and
2701
  scripts
2702
- Fixed the example startup script provided
2703
- Fixed insserv dependencies in the example startup script (patch from
2704
  Debian)
2705
- Fixed handling of drained nodes in the iallocator framework
2706
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2707
  #528618)
2708
- Fixed error related to invalid job IDs in job polling
2709
- Fixed job/opcode persistence on unclean master shutdown
2710
- Fixed handling of partial job processing after unclean master
2711
  shutdown
2712
- Fixed error reporting from LUs, previously all errors were converted
2713
  into execution errors
2714
- Fixed error reporting from burnin
2715
- Decreased significantly the memory usage of the job queue
2716
- Optimised slightly multi-job submission
2717
- Optimised slightly opcode loading
2718
- Backported the multi-job submit framework from the development
2719
  branch; multi-instance start and stop should be faster
2720
- Added script to clean archived jobs after 21 days; this will reduce
2721
  the size of the queue directory
2722
- Added some extra checks in disk size tracking
2723
- Added an example ethers hook script
2724
- Added a cluster parameter that prevents Ganeti from modifying of
2725
  /etc/hosts
2726
- Added more node information to RAPI responses
2727
- Added a ``gnt-job watch`` command that allows following the ouput of a
2728
  job
2729
- Added a bind-address option to ganeti-rapi
2730
- Added more checks to the configuration verify
2731
- Enhanced the burnin script such that some operations can be retried
2732
  automatically
2733
- Converted instance reinstall to multi-instance model
2734

    
2735

    
2736
Version 2.0.3
2737
-------------
2738

    
2739
*(Released Fri, 7 Aug 2009)*
2740

    
2741
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2742
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2743
  instances have mismatched disk sizes in the configuration
2744
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2745
  any configuration mismatches for disk sizes
2746
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2747
  failover to work on two-node clusters
2748
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2749
  unusable
2750
- Fixed detection of OS script errors in ``gnt-backup export``
2751
- Fixed exit code of ``gnt-backup export``
2752

    
2753

    
2754
Version 2.0.2
2755
-------------
2756

    
2757
*(Released Fri, 17 Jul 2009)*
2758

    
2759
- Added experimental support for stripped logical volumes; this should
2760
  enhance performance but comes with a higher complexity in the block
2761
  device handling; stripping is only enabled when passing
2762
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2763
  affected even in the non-stripped mode
2764
- Improved resiliency against transient failures at the end of DRBD
2765
  resyncs, and in general of DRBD resync checks
2766
- Fixed a couple of issues with exports and snapshot errors
2767
- Fixed a couple of issues in instance listing
2768
- Added display of the disk size in ``gnt-instance info``
2769
- Fixed checking for valid OSes in instance creation
2770
- Fixed handling of the "vcpus" parameter in instance listing and in
2771
  general of invalid parameters
2772
- Fixed http server library, and thus RAPI, to handle invalid
2773
  username/password combinations correctly; this means that now they
2774
  report unauthorized for queries too, not only for modifications,
2775
  allowing earlier detect of configuration problems
2776
- Added a new "role" node list field, equivalent to the master/master
2777
  candidate/drained/offline flags combinations
2778
- Fixed cluster modify and changes of candidate pool size
2779
- Fixed cluster verify error messages for wrong files on regular nodes
2780
- Fixed a couple of issues with node demotion from master candidate role
2781
- Fixed node readd issues
2782
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2783
- Added a new ``--no-voting`` option for masterfailover to fix failover
2784
  on two-nodes clusters when the former master node is unreachable
2785
- Added instance reinstall over RAPI
2786

    
2787

    
2788
Version 2.0.1
2789
-------------
2790

    
2791
*(Released Tue, 16 Jun 2009)*
2792

    
2793
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2794
  allow re-adding the start in single-user option (regression from 1.2)
2795
- the watcher writes the instance status to a file, to allow monitoring
2796
  to report the instance status (from the master) based on cached
2797
  results of the watcher's queries; while this can get stale if the
2798
  watcher is being locked due to other work on the cluster, this is
2799
  still an improvement
2800
- the watcher now also restarts the node daemon and the rapi daemon if
2801
  they died
2802
- fixed the watcher to handle full and drained queue cases
2803
- hooks export more instance data in the environment, which helps if
2804
  hook scripts need to take action based on the instance's properties
2805
  (no longer need to query back into ganeti)
2806
- instance failovers when the instance is stopped do not check for free
2807
  RAM, so that failing over a stopped instance is possible in low memory
2808
  situations
2809
- rapi uses queries for tags instead of jobs (for less job traffic), and
2810
  for cluster tags it won't talk to masterd at all but read them from
2811
  ssconf
2812
- a couple of error handling fixes in RAPI
2813
- drbd handling: improved the error handling of inconsistent disks after
2814
  resync to reduce the frequency of "there are some degraded disks for
2815
  this instance" messages
2816
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2817
  error handling path called a wrong function name)
2818

    
2819

    
2820
Version 2.0.0
2821
-------------
2822

    
2823
*(Released Wed, 27 May 2009)*
2824

    
2825
- no changes from rc5
2826

    
2827

    
2828
Version 2.0 rc5
2829
---------------
2830

    
2831
*(Released Wed, 20 May 2009)*
2832

    
2833
- fix a couple of bugs (validation, argument checks)
2834
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2835
- some small improvements to RAPI and IAllocator
2836
- make watcher automatically start the master daemon if down
2837

    
2838

    
2839
Version 2.0 rc4
2840
---------------
2841

    
2842
*(Released Mon, 27 Apr 2009)*
2843

    
2844
- change the OS list to not require locks; this helps with big clusters
2845
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2846
  volume group is broken
2847
- ``gnt-instance info``, without any arguments, doesn't run for all
2848
  instances anymore; either pass ``--all`` or pass the desired
2849
  instances; this helps against mistakes on big clusters where listing
2850
  the information for all instances takes a long time
2851
- miscellaneous doc and man pages fixes
2852

    
2853

    
2854
Version 2.0 rc3
2855
---------------
2856

    
2857
*(Released Wed, 8 Apr 2009)*
2858

    
2859
- Change the internal locking model of some ``gnt-node`` commands, in
2860
  order to reduce contention (and blocking of master daemon) when
2861
  batching many creation/reinstall jobs
2862
- Fixes to Xen soft reboot
2863
- No longer build documentation at build time, instead distribute it in
2864
  the archive, in order to reduce the need for the whole docbook/rst
2865
  toolchains
2866

    
2867

    
2868
Version 2.0 rc2
2869
---------------
2870

    
2871
*(Released Fri, 27 Mar 2009)*
2872

    
2873
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2874
- Fix watcher startup sequence, improves the behaviour of busy clusters
2875
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2876
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2877
  bind address and other places
2878
- Some documentation fixes and updates
2879

    
2880

    
2881
Version 2.0 rc1
2882
---------------
2883

    
2884
*(Released Mon, 2 Mar 2009)*
2885

    
2886
- More documentation updates, now all docs should be more-or-less
2887
  up-to-date
2888
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2889
  etc.)
2890
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2891
  and KVM)
2892
- Fix an issue related to $libdir/run/ganeti and cluster creation
2893

    
2894

    
2895
Version 2.0 beta2
2896
-----------------
2897

    
2898
*(Released Thu, 19 Feb 2009)*
2899

    
2900
- Xen PVM and KVM have switched the default value for the instance root
2901
  disk to the first partition on the first drive, instead of the whole
2902
  drive; this means that the OS installation scripts must be changed
2903
  accordingly
2904
- Man pages have been updated
2905
- RAPI has been switched by default to HTTPS, and the exported functions
2906
  should all work correctly
2907
- RAPI v1 has been removed
2908
- Many improvements to the KVM hypervisor
2909
- Block device errors are now better reported
2910
- Many other bugfixes and small improvements
2911

    
2912

    
2913
Version 2.0 beta1
2914
-----------------
2915

    
2916
*(Released Mon, 26 Jan 2009)*
2917

    
2918
- Version 2 is a general rewrite of the code and therefore the
2919
  differences are too many to list, see the design document for 2.0 in
2920
  the ``doc/`` subdirectory for more details
2921
- In this beta version there is not yet a migration path from 1.2 (there
2922
  will be one in the final 2.0 release)
2923
- A few significant changes are:
2924

    
2925
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2926
    various ``gnt-*`` commands are just front-ends to it
2927
  - all the commands are entered into, and executed from a job queue,
2928
    see the ``gnt-job(8)`` manpage
2929
  - the RAPI daemon supports read-write operations, secured by basic
2930
    HTTP authentication on top of HTTPS
2931
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2932
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2933
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2934
  - DRBD devices are using statically allocated minor numbers, which
2935
    will be assigned to existing instances during the migration process
2936
  - there is support for both Xen PVM and Xen HVM instances running on
2937
    the same cluster
2938
  - KVM virtualization is supported too
2939
  - file-based storage has been implemented, which means that it is
2940
    possible to run the cluster without LVM and DRBD storage, for
2941
    example using a shared filesystem exported from shared storage (and
2942
    still have live migration)
2943

    
2944

    
2945
Version 1.2.7
2946
-------------
2947

    
2948
*(Released Tue, 13 Jan 2009)*
2949

    
2950
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2951
- Reuse the old instance mac address by default on instance import, if
2952
  the instance name is the same.
2953
- Handle situations in which the node info rpc returns incomplete
2954
  results (issue 46)
2955
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2956
- Improved version of batcher:
2957

    
2958
  - state file support
2959
  - instance mac address support
2960
  - support for HVM clusters/instances
2961

    
2962
- Add an option to show the number of cpu sockets and nodes in
2963
  ``gnt-node list``
2964
- Support OSes that handle more than one version of the OS api (but do
2965
  not change the current API in any other way)
2966
- Fix ``gnt-node migrate``
2967
- ``gnt-debug`` man page
2968
- Fixes various more typos and small issues
2969
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2970

    
2971

    
2972
Version 1.2.6
2973
-------------
2974

    
2975
*(Released Wed, 24 Sep 2008)*
2976

    
2977
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2978
  type of disk exported to fully virtualized instances.
2979
- provide access to the serial console of HVM instances
2980
- instance auto_balance flag, set by default. If turned off it will
2981
  avoid warnings on cluster verify if there is not enough memory to fail
2982
  over an instance. in the future it will prevent automatically failing
2983
  it over when we will support that.
2984
- batcher tool for instance creation, see ``tools/README.batcher``
2985
- ``gnt-instance reinstall --select-os`` to interactively select a new
2986
  operating system when reinstalling an instance.
2987
- when changing the memory amount on instance modify a check has been
2988
  added that the instance will be able to start. also warnings are
2989
  emitted if the instance will not be able to fail over, if auto_balance
2990
  is true.
2991
- documentation fixes
2992
- sync fields between ``gnt-instance list/modify/add/import``
2993
- fix a race condition in drbd when the sync speed was set after giving
2994
  the device a remote peer.
2995

    
2996

    
2997
Version 1.2.5
2998
-------------
2999

    
3000
*(Released Tue, 22 Jul 2008)*
3001

    
3002
- note: the allowed size and number of tags per object were reduced
3003
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
3004
- fixed twisted 8.x compatibility
3005
- fixed ``gnt-instance replace-disks`` with iallocator
3006
- add TCP keepalives on twisted connections to detect restarted nodes
3007
- disk increase support, see ``gnt-instance grow-disk``
3008
- implement bulk node/instance query for RAPI
3009
- add tags in node/instance listing (optional)
3010
- experimental migration (and live migration) support, read the man page
3011
  for ``gnt-instance migrate``
3012
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3013
  has some small improvements in handling its state file
3014

    
3015

    
3016
Version 1.2.4
3017
-------------
3018

    
3019
*(Released Fri, 13 Jun 2008)*
3020

    
3021
- Experimental readonly, REST-based remote API implementation;
3022
  automatically started on master node, TCP port 5080, if enabled by
3023
  ``--enable-rapi`` parameter to configure script.
3024
- Instance allocator support. Add and import instance accept a
3025
  ``--iallocator`` parameter, and call that instance allocator to decide
3026
  which node to use for the instance. The iallocator document describes
3027
  what's expected from an allocator script.
3028
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3029
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3030
  node is lost there is still enough memory to fail over the instances
3031
  that reside on it.
3032
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3033
  ``gnt-cluster verify``, to check for site-specific compliance. All the
3034
  hooks will run, and their output, if any, will be displayed. Any
3035
  failing hook will make the verification return an error value.
3036
- ``gnt-cluster verify`` now checks that its peers are reachable on the
3037
  primary and secondary interfaces
3038
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3039
  that is still declared as part of the cluster and has failed.
3040
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
3041
  specifying output fields, that just adds the chosen fields to the
3042
  default ones.
3043
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
3044
  export from the filesystem.
3045
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3046
  have been added. Using them you can enable/disable acpi and pae
3047
  support, and specify a path for a cd image to be exported to the
3048
  instance. These parameters as the name suggest only work on HVM
3049
  clusters.
3050
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3051
  PAE support will be set to the previously hardcoded values, but the
3052
  (previously hardcoded) path to the CDROM ISO image will be unset and
3053
  if required, needs to be set manually with ``gnt-instance modify``
3054
  after the upgrade.
3055
- The address to which an instance's VNC console is bound is now
3056
  selectable per-instance, rather than being cluster wide. Of course
3057
  this only applies to instances controlled via VNC, so currently just
3058
  applies to HVM clusters.
3059

    
3060

    
3061
Version 1.2.3
3062
-------------
3063

    
3064
*(Released Mon, 18 Feb 2008)*
3065

    
3066
- more tweaks to the disk activation code (especially helpful for DRBD)
3067
- change the default ``gnt-instance list`` output format, now there is
3068
  one combined status field (see the manpage for the exact values this
3069
  field will have)
3070
- some more fixes for the mac export to hooks change
3071
- make Ganeti not break with DRBD 8.2.x (which changed the version
3072
  format in ``/proc/drbd``) (issue 24)
3073
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3074
  template, allowing migration from DRBD0.7+MD to DRBD8
3075

    
3076

    
3077
Version 1.2.2
3078
-------------
3079

    
3080
*(Released Wed, 30 Jan 2008)*
3081

    
3082
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3083
  support (issue 23)
3084
- add command aliases infrastructure and a few aliases
3085
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
3086
  man pages and the ``--help`` option of ``gnt-node
3087
  list``/``gnt-instance list``
3088
- fix ``gnt-backup list`` with down nodes (issue 21)
3089
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3090
- fix the dist archive and add a check for including svn/git files in
3091
  the future
3092
- some developer-related changes: improve the burnin and the QA suite,
3093
  add an upload script for testing during development
3094

    
3095

    
3096
Version 1.2.1
3097
-------------
3098

    
3099
*(Released Wed, 16 Jan 2008)*
3100

    
3101
- experimental HVM support, read the install document, section
3102
  "Initializing the cluster"
3103
- allow for the PVM hypervisor per-instance kernel and initrd paths
3104
- add a new command ``gnt-cluster verify-disks`` which uses a new
3105
  algorithm to improve the reconnection of the DRBD pairs if the device
3106
  on the secondary node has gone away
3107
- make logical volume code auto-activate LVs at disk activation time
3108
- slightly improve the speed of activating disks
3109
- allow specification of the MAC address at instance creation time, and
3110
  changing it later via ``gnt-instance modify``
3111
- fix handling of external commands that generate lots of output on
3112
  stderr
3113
- update documentation with regard to minimum version of DRBD8 supported
3114

    
3115

    
3116
Version 1.2.0
3117
-------------
3118

    
3119
*(Released Tue, 4 Dec 2007)*
3120

    
3121
- Log the ``xm create`` output to the node daemon log on failure (to
3122
  help diagnosing the error)
3123
- In debug mode, log all external commands output if failed to the logs
3124
- Change parsing of lvm commands to ignore stderr
3125

    
3126

    
3127
Version 1.2 beta3
3128
-----------------
3129

    
3130
*(Released Wed, 28 Nov 2007)*
3131

    
3132
- Another round of updates to the DRBD 8 code to deal with more failures
3133
  in the replace secondary node operation
3134
- Some more logging of failures in disk operations (lvm, drbd)
3135
- A few documentation updates
3136
- QA updates
3137

    
3138

    
3139
Version 1.2 beta2
3140
-----------------
3141

    
3142
*(Released Tue, 13 Nov 2007)*
3143

    
3144
- Change configuration file format from Python's Pickle to JSON.
3145
  Upgrading is possible using the cfgupgrade utility.
3146
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3147
  faster replace disks and is more stable (DRBD 8 has many improvements
3148
  compared to DRBD 0.7)
3149
- Added command line tags support (see man pages for ``gnt-instance``,
3150
  ``gnt-node``, ``gnt-cluster``)
3151
- Added instance rename support
3152
- Added multi-instance startup/shutdown
3153
- Added cluster rename support
3154
- Added ``gnt-node evacuate`` to simplify some node operations
3155
- Added instance reboot operation that can speedup reboot as compared to
3156
  stop and start
3157
- Soften the requirement that hostnames are in FQDN format
3158
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
3159
  reboots
3160
- Removed dependency on debian's patched fping that uses the
3161
  non-standard ``-S`` option
3162
- Now the OS definitions are searched for in multiple, configurable
3163
  paths (easier for distros to package)
3164
- Some changes to the hooks infrastructure (especially the new
3165
  post-configuration update hook)
3166
- Other small bugfixes
3167

    
3168
.. vim: set textwidth=72 syntax=rst :
3169
.. Local Variables:
3170
.. mode: rst
3171
.. fill-column: 72
3172
.. End: