Statistics
| Branch: | Tag: | Revision:

root / NEWS @ acd59944

History | View | Annotate | Download (119.1 kB)

1
News
2
====
3

    
4

    
5
Version 2.9.4
6
-------------
7

    
8
*(unreleased)*
9

    
10
- Fix the RAPI instances-multi-alloc call
11
- assign unique filenames to file-based disks
12

    
13

    
14
Version 2.9.3
15
-------------
16

    
17
*(Released Mon, 27 Jan 2014)*
18

    
19
- Ensure that all the hypervisors exist in the config file (Issue 640)
20
- Correctly recognise the role as master node (Issue 687)
21
- configure: allow detection of Sphinx 1.2+ (Issue 502)
22
- gnt-instance now honors the KVM path correctly (Issue 691)
23

    
24
Inherited from the 2.8 branch:
25

    
26
- Change the list separator for the usb_devices parameter from comma to space.
27
  Commas could not work because they are already the hypervisor option
28
  separator (Issue 649)
29
- Add support for blktap2 file-driver (Issue 638)
30
- Add network tag definitions to the haskell codebase (Issue 641)
31
- Fix RAPI network tag handling
32
- Add the network tags to the tags searched by gnt-cluster search-tags
33
- Fix caching bug preventing jobs from being cancelled
34
- Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
35

    
36

    
37
Version 2.9.2
38
-------------
39

    
40
*(Released Fri, 13 Dec 2013)*
41

    
42
- use custom KVM path if set for version checking
43
- SingleNotifyPipeCondition: don't share pollers
44

    
45
Inherited from the 2.8 branch:
46

    
47
- Fixed Luxi daemon socket permissions after master-failover
48
- Improve IP version detection code directly checking for colons rather than
49
  passing the family from the cluster object
50
- Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
51
  opportunistically anymore (Issue 622)
52
- Allow link local IPv6 gateways (Issue 624)
53
- Fix error printing (Issue 616)
54
- Fix a bug in InstanceSetParams concerning names: in case no name is passed in
55
  disk modifications, keep the old one. If name=none then set disk name to
56
  None.
57
- Update build_chroot script to work with the latest hackage packages
58
- Add a packet number limit to "fping" in master-ip-setup (Issue 630)
59
- Fix evacuation out of drained node (Issue 615)
60
- Add default file_driver if missing (Issue 571)
61
- Fix job error message after unclean master shutdown (Issue 618)
62
- Lock group(s) when creating instances (Issue 621)
63
- SetDiskID() before accepting an instance (Issue 633)
64
- Allow the ext template disks to receive arbitrary parameters, both at creation
65
  time and while being modified
66
- Xen handle domain shutdown (future proofing cherry-pick)
67
- Refactor reading live data in htools (future proofing cherry-pick)
68

    
69

    
70
Version 2.9.1
71
-------------
72

    
73
*(Released Wed, 13 Nov 2013)*
74

    
75
- fix bug, that kept nodes offline when readding
76
- when verifying DRBD versions, ignore unavailable nodes
77
- fix bug that made the console unavailable on kvm in split-user
78
  setup (issue 608)
79
- DRBD: ensure peers are UpToDate for dual-primary (inherited 2.8.2)
80

    
81

    
82
Version 2.9.0
83
-------------
84

    
85
*(Released Tue, 5 Nov 2013)*
86

    
87
Incompatible/important changes
88
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
89

    
90
- hroller now also plans for capacity to move non-redundant instances off
91
  any node to be rebooted; the old behavior of completely ignoring any
92
  non-redundant instances can be restored by adding the --ignore-non-redundant
93
  option.
94
- The cluster option '--no-lvm-storage' was removed in favor of the new option
95
  '--enabled-disk-templates'.
96
- On instance creation, disk templates no longer need to be specified
97
  with '-t'. The default disk template will be taken from the list of
98
  enabled disk templates.
99
- The monitoring daemon is now running as root, in order to be able to collect
100
  information only available to root (such as the state of Xen instances).
101
- The ConfD client is now IPv6 compatible.
102
- File and shared file storage is no longer dis/enabled at configure time,
103
  but using the option '--enabled-disk-templates' at cluster initialization and
104
  modification.
105
- The default directories for file and shared file storage are not anymore
106
  specified at configure time, but taken from the cluster's configuration.
107
  They can be set at cluster initialization and modification with
108
  '--file-storage-dir' and '--shared-file-storage-dir'.
109
- Cluster verification now includes stricter checks regarding the
110
  default file and shared file storage directories. It now checks that
111
  the directories are explicitely allowed in the 'file-storage-paths' file and
112
  that the directories exist on all nodes.
113
- The list of allowed disk templates in the instance policy and the list
114
  of cluster-wide enabled disk templates is now checked for consistency
115
  on cluster or group modification. On cluster initialization, the ipolicy
116
  disk templates are ensured to be a subset of the cluster-wide enabled
117
  disk templates.
118

    
119
New features
120
~~~~~~~~~~~~
121

    
122
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
123
  the correct command syntax. It is possible to use different DRBD versions
124
  on different nodes as long as they are compatible to each other. This
125
  enables rolling upgrades of DRBD with no downtime. As permanent operation
126
  of different DRBD versions within a node group is discouraged,
127
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
128
- New "inst-status-xen" data collector for the monitoring daemon, providing
129
  information about the state of the xen instances on the nodes.
130
- New "lv" data collector for the monitoring daemon, collecting data about the
131
  logical volumes on the nodes, and pairing them with the name of the instances
132
  they belong to.
133
- New "diskstats" data collector, collecting the data from /proc/diskstats and
134
  presenting them over the monitoring daemon interface.
135
- The ConfD client is now IPv6 compatible.
136

    
137
New dependencies
138
~~~~~~~~~~~~~~~~
139
The following new dependencies have been added.
140

    
141
Python
142

    
143
- ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
144
  for the unit tests (and only used for testing).
145

    
146
Haskell
147

    
148
- ``hslogger`` (http://software.complete.org/hslogger) is now always
149
  required, even if confd is not enabled.
150

    
151
Since 2.9.0 rc3
152
~~~~~~~~~~~~~~~
153

    
154
- Correctly start/stop luxid during gnt-cluster master-failover (inherited
155
  from stable-2.8)
156
- Improved error messsages (inherited from stable-2.8)
157

    
158

    
159
Version 2.9.0 rc3
160
-----------------
161

    
162
*(Released Tue, 15 Oct 2013)*
163

    
164
The third release candidate in the 2.9 series. Since 2.9.0 rc2:
165

    
166
- in implicit configuration upgrade, match ipolicy with enabled disk templates
167
- improved harep documentation (inherited from stable-2.8)
168

    
169

    
170
Version 2.9.0 rc2
171
-----------------
172

    
173
*(Released Wed, 9 Oct 2013)*
174

    
175
The second release candidate in the 2.9 series. Since 2.9.0 rc1:
176

    
177
- Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
178
  at least one DRBD instance.
179
- Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
180
  downgrading.
181

    
182

    
183
Version 2.9.0 rc1
184
-----------------
185

    
186
*(Released Tue, 1 Oct 2013)*
187

    
188
The first release candidate in the 2.9 series. Since 2.9.0 beta1:
189

    
190
- various bug fixes
191
- update of the documentation, in particular installation instructions
192
- merging of LD_* constants into DT_* constants
193
- python style changes to be compatible with newer versions of pylint
194

    
195

    
196
Version 2.9.0 beta1
197
-------------------
198

    
199
*(Released Thu, 29 Aug 2013)*
200

    
201
This was the first beta release of the 2.9 series. All important changes
202
are listed in the latest 2.9 entry.
203

    
204

    
205
Version 2.8.4
206
-------------
207

    
208
*(Released Thu, 23 Jan 2014)*
209

    
210
- Change the list separator for the usb_devices parameter from comma to space.
211
  Commas could not work because they are already the hypervisor option
212
  separator (Issue 649)
213
- Add support for blktap2 file-driver (Issue 638)
214
- Add network tag definitions to the haskell codebase (Issue 641)
215
- Fix RAPI network tag handling
216
- Add the network tags to the tags searched by gnt-cluster search-tags
217
- Fix caching bug preventing jobs from being cancelled
218
- Start-master/stop-master was always failing if ConfD was disabled. (Issue 685)
219

    
220

    
221
Version 2.8.3
222
-------------
223

    
224
*(Released Thu, 12 Dec 2013)*
225

    
226
- Fixed Luxi daemon socket permissions after master-failover
227
- Improve IP version detection code directly checking for colons rather than
228
  passing the family from the cluster object
229
- Fix NODE/NODE_RES locking in LUInstanceCreate by not acquiring NODE_RES locks
230
  opportunistically anymore (Issue 622)
231
- Allow link local IPv6 gateways (Issue 624)
232
- Fix error printing (Issue 616)
233
- Fix a bug in InstanceSetParams concerning names: in case no name is passed in
234
  disk modifications, keep the old one. If name=none then set disk name to
235
  None.
236
- Update build_chroot script to work with the latest hackage packages
237
- Add a packet number limit to "fping" in master-ip-setup (Issue 630)
238
- Fix evacuation out of drained node (Issue 615)
239
- Add default file_driver if missing (Issue 571)
240
- Fix job error message after unclean master shutdown (Issue 618)
241
- Lock group(s) when creating instances (Issue 621)
242
- SetDiskID() before accepting an instance (Issue 633)
243
- Allow the ext template disks to receive arbitrary parameters, both at creation
244
  time and while being modified
245
- Xen handle domain shutdown (future proofing cherry-pick)
246
- Refactor reading live data in htools (future proofing cherry-pick)
247

    
248

    
249
Version 2.8.2
250
-------------
251

    
252
*(Released Thu, 07 Nov 2013)*
253

    
254
- DRBD: ensure peers are UpToDate for dual-primary
255
- Improve error message for replace-disks
256
- More dependency checks at configure time
257
- Placate warnings on ganeti.outils_unittest.py
258

    
259

    
260
Version 2.8.1
261
-------------
262

    
263
*(Released Thu, 17 Oct 2013)*
264

    
265
- Correctly start/stop luxid during gnt-cluster master-failover
266
- Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
267
- Fix path for the job queue serial file
268
- Improved harep man page
269
- Minor documentation improvements
270

    
271

    
272
Version 2.8.0
273
-------------
274

    
275
*(Released Mon, 30 Sep 2013)*
276

    
277
Incompatible/important changes
278
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
279

    
280
- Instance policy can contain multiple instance specs, as described in
281
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
282
  <design-partitioned>`. As a consequence, it's not possible to partially change
283
  or override instance specs. Bounding specs (min and max) can be specified as a
284
  whole using the new option ``--ipolicy-bounds-specs``, while standard
285
  specs use the new option ``--ipolicy-std-specs``.
286
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
287
  gnt-instance is a valid YAML object.
288
- hail now honors network restrictions when allocating nodes. This led to an
289
  update of the IAllocator protocol. See the IAllocator documentation for
290
  details.
291
- confd now only answers static configuration request over the network. luxid
292
  was extracted, listens on the local LUXI socket and responds to live queries.
293
  This allows finer grained permissions if using separate users.
294

    
295
New features
296
~~~~~~~~~~~~
297

    
298
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
299
  to always require authentication, ``--require-authentication``. It can
300
  be specified in ``$sysconfdir/default/ganeti``.
301
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
302
  be used to manage the disk templates to be used by instances in the cluster.
303
  Initially, it will be set to a list that includes plain, drbd, if they were
304
  enabled by specifying a volume group name, and file and sharedfile, if those
305
  were enabled at configure time. Additionally, it will include all disk
306
  templates that are currently used by instances. The order of disk templates
307
  will be based on Ganeti's history of supporting them. In the future, the
308
  first entry of the list will be used as a default disk template on instance
309
  creation.
310
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
311
  configuration back to the previous stable version.
312
- Disk templates in group ipolicy can be restored to the default value.
313
- Initial support for diskless instances and virtual clusters in QA.
314
- More QA and unit tests for instance policies.
315
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
316
  describing why the opcode itself was executed.
317
- The monitoring daemon is now available. It allows users to query the cluster
318
  for obtaining information about the status of the system. The daemon is only
319
  responsible for providing the information over the network: the actual data
320
  gathering is performed by data collectors (currently, only the DRBD status
321
  collector is available).
322
- In order to help developers work on Ganeti, a new script
323
  (``devel/build_chroot``) is provided, for building a chroot that contains all
324
  the required development libraries and tools for compiling Ganeti on a Debian
325
  Squeeze system.
326
- A new tool, ``harep``, for performing self-repair and recreation of instances
327
  in Ganeti has been added.
328
- Split queries are enabled for tags, network, exports, cluster info, groups,
329
  jobs, nodes.
330
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
331
  It prints the command line to set the current policies, to ease
332
  changing them.
333
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
334
  devices for KVM virtio-net interfaces will get created with VNET_HDR
335
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
336
  virtio-net interfaces, which prevents host kernel tainting and log
337
  flooding, when dealing with broken or malicious virtio-net drivers.
338
  It's set to true by default.
339
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
340
  failures.
341
- Support 'viridian' parameter in Xen HVM
342
- Support DSA SSH keys in bootstrap
343
- To simplify the work of packaging frameworks that want to add the needed users
344
  and groups in a split-user setup themselves, at build time three files in
345
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
346
  the groups to be generated, the ``users`` file contains, one per line, the
347
  users to be generated, optionally followed by their primary group, where
348
  important. The ``groupmemberships`` file contains, one per line, additional
349
  user-group membership relations that need to be established. The syntax of
350
  these files will remain stable in all future versions.
351

    
352

    
353
New dependencies
354
~~~~~~~~~~~~~~~~
355
The following new dependencies have been added:
356

    
357
For Haskell:
358
- The ``curl`` library is not optional anymore for compiling the Haskell code.
359
- ``snap-server`` library (if monitoring is enabled).
360

    
361
For Python:
362
- The minimum Python version needed to run Ganeti is now 2.6.
363
- ``yaml`` library (only for running the QA).
364

    
365
Since 2.8.0 rc3
366
~~~~~~~~~~~~~~~
367
- Perform proper cleanup on termination of Haskell daemons
368
- Fix corner-case in handling of remaining retry time
369

    
370

    
371
Version 2.8.0 rc3
372
-----------------
373

    
374
*(Released Tue, 17 Sep 2013)*
375

    
376
- To simplify the work of packaging frameworks that want to add the needed users
377
  and groups in a split-user setup themselves, at build time three files in
378
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
379
  the groups to be generated, the ``users`` file contains, one per line, the
380
  users to be generated, optionally followed by their primary group, where
381
  important. The ``groupmemberships`` file contains, one per line, additional
382
  user-group membership relations that need to be established. The syntax of
383
  these files will remain stable in all future versions.
384
- Add a default to file-driver when unspecified over RAPI (Issue 571)
385
- Mark the DSA host pubkey as optional, and remove it during config downgrade
386
  (Issue 560)
387
- Some documentation fixes
388

    
389

    
390
Version 2.8.0 rc2
391
-----------------
392

    
393
*(Released Tue, 27 Aug 2013)*
394

    
395
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
396

    
397
- Support 'viridian' parameter in Xen HVM (Issue 233)
398
- Include VCS version in ``gnt-cluster version``
399
- Support DSA SSH keys in bootstrap (Issue 338)
400
- Fix batch creation of instances
401
- Use FQDN to check master node status (Issue 551)
402
- Make the DRBD collector more failure-resilient
403

    
404

    
405
Version 2.8.0 rc1
406
-----------------
407

    
408
*(Released Fri, 2 Aug 2013)*
409

    
410
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
411

    
412
- Fix upgrading/downgrading from 2.7
413
- Increase maximum RAPI message size
414
- Documentation updates
415
- Split ``confd`` between ``luxid`` and ``confd``
416
- Merge 2.7 series up to the 2.7.1 release
417
- Allow the ``modify_etc_hosts`` option to be changed
418
- Add better debugging for ``luxid`` queries
419
- Expose bulk parameter for GetJobs in RAPI client
420
- Expose missing ``network`` fields in RAPI
421
- Add some ``cluster verify`` tests
422
- Some unittest fixes
423
- Fix a malfunction in ``hspace``'s tiered allocation
424
- Fix query compatibility between haskell and python implementations
425
- Add the ``vnet_hdr`` HV parameter for KVM
426
- Add ``--cleanup`` to instance failover
427
- Change the connected groups format in ``gnt-network info`` output; it
428
  was previously displayed as a raw list by mistake. (Merged from 2.7)
429

    
430

    
431
Version 2.8.0 beta1
432
-------------------
433

    
434
*(Released Mon, 24 Jun 2013)*
435

    
436
This was the first beta release of the 2.8 series. All important changes
437
are listed in the latest 2.8 entry.
438

    
439

    
440
Version 2.7.2
441
-------------
442

    
443
*(Released Thu, 26 Sep 2013)*
444

    
445
- Change the connected groups format in ``gnt-network info`` output; it
446
  was previously displayed as a raw list by mistake
447
- Check disk template in right dict when copying
448
- Support multi-instance allocs without iallocator
449
- Fix some errors in the documentation
450
- Fix formatting of tuple in an error message
451

    
452

    
453
Version 2.7.1
454
-------------
455

    
456
*(Released Thu, 25 Jul 2013)*
457

    
458
- Add logrotate functionality in daemon-util
459
- Add logrotate example file
460
- Add missing fields to network queries over rapi
461
- Fix network object timestamps
462
- Add support for querying network timestamps
463
- Fix a typo in the example crontab
464
- Fix a documentation typo
465

    
466

    
467
Version 2.7.0
468
-------------
469

    
470
*(Released Thu, 04 Jul 2013)*
471

    
472
Incompatible/important changes
473
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
474

    
475
- Instance policies for disk size were documented to be on a per-disk
476
  basis, but hail applied them to the sum of all disks. This has been
477
  fixed.
478
- ``hbal`` will now exit with status 0 if, during job execution over
479
  LUXI, early exit has been requested and all jobs are successful;
480
  before, exit status 1 was used, which cannot be differentiated from
481
  "job error" case
482
- Compatibility with newer versions of rbd has been fixed
483
- ``gnt-instance batch-create`` has been changed to use the bulk create
484
  opcode from Ganeti. This lead to incompatible changes in the format of
485
  the JSON file. It's now not a custom dict anymore but a dict
486
  compatible with the ``OpInstanceCreate`` opcode.
487
- Parent directories for file storage need to be listed in
488
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
489
  write the file automatically based on old configuration values, but it
490
  can not distribute it across all nodes and the file contents should be
491
  verified. Use ``gnt-cluster copyfile
492
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
493
  upgraded. The reason for requiring this list of paths now is that
494
  before it would have been possible to inject new paths via RPC,
495
  allowing files to be created in arbitrary locations. The RPC protocol
496
  is protected using SSL/X.509 certificates, but as a design principle
497
  Ganeti does not permit arbitrary paths to be passed.
498
- The parsing of the variants file for OSes (see
499
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
500
  lines and comment lines (starting with ``#``) are ignored for better
501
  readability.
502
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
503
  longer available. ``gnt-node add`` now invokes a new tool on the
504
  destination node, named ``prepare-node-join``, to configure the SSH
505
  daemon. Paramiko is no longer necessary to configure nodes' SSH
506
  daemons via ``gnt-node add``.
507
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
508
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
509
  the flag is not reset after a master failover.
510
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
511
  is recommended. 2.6 will be mandatory from the 2.8 series.
512

    
513

    
514
New features
515
~~~~~~~~~~~~
516

    
517
- New network management functionality to support automatic allocation
518
  of IP addresses and managing of network parameters. See
519
  :manpage:`gnt-network(8)` for more details.
520
- New external storage backend, to allow managing arbitrary storage
521
  systems external to the cluster. See
522
  :manpage:`ganeti-extstorage-interface(7)`.
523
- New ``exclusive-storage`` node parameter added, restricted to
524
  nodegroup level. When it's set to true, physical disks are assigned in
525
  an exclusive fashion to instances, as documented in :doc:`Partitioned
526
  Ganeti <design-partitioned>`.  Currently, only instances using the
527
  ``plain`` disk template are supported.
528
- The KVM hypervisor has been updated with many new hypervisor
529
  parameters, including a generic one for passing arbitrary command line
530
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
531
  compatible up to qemu 1.4.
532
- A new tool, called ``mon-collector``, is the stand-alone executor of
533
  the data collectors for a monitoring system. As of this version, it
534
  just includes the DRBD data collector, that can be executed by calling
535
  ``mon-collector`` using the ``drbd`` parameter. See
536
  :manpage:`mon-collector(7)`.
537
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
538
  for RAPI users. It allows granting permissions to query for
539
  information to a specific user without giving
540
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
541
- A new tool named ``node-cleanup`` has been added. It cleans remains of
542
  a cluster from a machine by stopping all daemons, removing
543
  certificates and ssconf files. Unless the ``--no-backup`` option is
544
  given, copies of the certificates are made.
545
- Instance creations now support the use of opportunistic locking,
546
  potentially speeding up the (parallel) creation of multiple instances.
547
  This feature is currently only available via the :doc:`RAPI
548
  <rapi>` interface and when an instance allocator is used. If the
549
  ``opportunistic_locking`` parameter is set the opcode will try to
550
  acquire as many locks as possible, but will not wait for any locks
551
  held by other opcodes. If not enough resources can be found to
552
  allocate the instance, the temporary error code
553
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
554
  retried thereafter, with or without opportunistic locking.
555
- New experimental linux-ha resource scripts.
556
- Restricted-commands support: ganeti can now be asked (via command line
557
  or rapi) to perform commands on a node. These are passed via ganeti
558
  RPC rather than ssh. This functionality is restricted to commands
559
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
560
  reasons. The file is not copied automatically.
561

    
562

    
563
Misc changes
564
~~~~~~~~~~~~
565

    
566
- Diskless instances are now externally mirrored (Issue 237). This for
567
  now has only been tested in conjunction with explicit target nodes for
568
  migration/failover.
569
- Queries not needing locks or RPC access to the node can now be
570
  performed by the confd daemon, making them independent from jobs, and
571
  thus faster to execute. This is selectable at configure time.
572
- The functionality for allocating multiple instances at once has been
573
  overhauled and is now also available through :doc:`RAPI <rapi>`.
574

    
575
There are no significant changes from version 2.7.0~rc3.
576

    
577

    
578
Version 2.7.0 rc3
579
-----------------
580

    
581
*(Released Tue, 25 Jun 2013)*
582

    
583
- Fix permissions on the confd query socket (Issue 477)
584
- Fix permissions on the job archive dir (Issue 498)
585
- Fix handling of an internal exception in replace-disks (Issue 472)
586
- Fix gnt-node info handling of shortened names (Issue 497)
587
- Fix gnt-instance grow-disk when wiping is enabled
588
- Documentation improvements, and support for newer pandoc
589
- Fix hspace honoring ipolicy for disks (Issue 484)
590
- Improve handling of the ``kvm_extra`` HV parameter
591

    
592

    
593
Version 2.7.0 rc2
594
-----------------
595

    
596
*(Released Fri, 24 May 2013)*
597

    
598
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
599
  symlink.
600
- Disks added through ``gnt-instance modify`` or created through
601
  ``gnt-instance recreate-disks`` are wiped, if the
602
  ``prealloc_wipe_disks`` flag is set.
603
- If wiping newly created disks fails, the disks are removed. Also,
604
  partial failures in creating disks through ``gnt-instance modify``
605
  triggers a cleanup of the partially-created disks.
606
- Removing the master IP address doesn't fail if the address has been
607
  already removed.
608
- Fix ownership of the OS log dir
609
- Workaround missing SO_PEERCRED constant (Issue 191)
610

    
611

    
612
Version 2.7.0 rc1
613
-----------------
614

    
615
*(Released Fri, 3 May 2013)*
616

    
617
This was the first release candidate of the 2.7 series. Since beta3:
618

    
619
- Fix kvm compatibility with qemu 1.4 (Issue 389)
620
- Documentation updates (admin guide, upgrade notes, install
621
  instructions) (Issue 372)
622
- Fix gnt-group list nodes and instances count (Issue 436)
623
- Fix compilation without non-mandatory libraries (Issue 441)
624
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
625
- Make confd logging more verbose at INFO level (Issue 435)
626
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
627
- Fix failure path for instance storage type conversion (Issue 229)
628
- Update htools text backend documentation
629
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
630
- Disable inter-cluster instance move for file-based instances, because
631
  it is dependant on instance export, which is not supported for
632
  file-based instances. (Issue 414)
633
- Fix gnt-job crashes on non-ascii characters (Issue 427)
634
- Fix volume group checks on non-vm-capable nodes (Issue 432)
635

    
636

    
637
Version 2.7.0 beta3
638
-------------------
639

    
640
*(Released Mon, 22 Apr 2013)*
641

    
642
This was the third beta release of the 2.7 series. Since beta2:
643

    
644
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
645
- Fix data loss on wrong usage of ``gnt-instance move``
646
- Properly export errors in confd-based job queries
647
- Add ``users-setup`` tool
648
- Fix iallocator protocol to report 0 as a disk size for diskless
649
  instances. This avoids hail breaking when a diskless instance is
650
  present.
651
- Fix job queue directory permission problem that made confd job queries
652
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
653
  for access to archived jobs (Issue 406).
654
- Limit the sizes of networks supported by ``gnt-network`` to something
655
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
656
- Fix bugs in instance disk template conversion
657
- Fix GHC 7 compatibility
658
- Fix ``burnin`` install path (Issue 426).
659
- Allow very small disk grows (Issue 347).
660
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
661
  that noded doesn't import masterd code (Issue 419).
662
- Make sure the default metavg at cluster init is the same as the vg, if
663
  unspecified (Issue 358).
664
- Fix cleanup of partially created disks (part of Issue 416)
665

    
666

    
667
Version 2.7.0 beta2
668
-------------------
669

    
670
*(Released Tue, 2 Apr 2013)*
671

    
672
This was the second beta release of the 2.7 series. Since beta1:
673

    
674
- Networks no longer have a "type" slot, since this information was
675
  unused in Ganeti: instead of it tags should be used.
676
- The rapi client now has a ``target_node`` option to MigrateInstance.
677
- Fix early exit return code for hbal (Issue 386).
678
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
679
- Fix ``rbd showmapped`` output parsing (Issue 312).
680
- Networks are now referenced indexed by UUID, rather than name. This
681
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
682
  use.
683
- The OS environment now includes network information.
684
- Deleting of a network is now disallowed if any instance nic is using
685
  it, to prevent dangling references.
686
- External storage is now documented in man pages.
687
- The exclusive_storage flag can now only be set at nodegroup level.
688
- Hbal can now submit an explicit priority with its jobs.
689
- Many network related locking fixes.
690
- Bump up the required pylint version to 0.25.1.
691
- Fix the ``no_remember`` option in RAPI client.
692
- Many ipolicy related tests, qa, and fixes.
693
- Many documentation improvements and fixes.
694
- Fix building with ``--disable-file-storage``.
695
- Fix ``-q`` option in htools, which was broken if passed more than
696
  once.
697
- Some haskell/python interaction improvements and fixes.
698
- Fix iallocator in case of missing LVM storage.
699
- Fix confd config load in case of ``--no-lvm-storage``.
700
- The confd/query functionality is now mentioned in the security
701
  documentation.
702

    
703

    
704
Version 2.7.0 beta1
705
-------------------
706

    
707
*(Released Wed, 6 Feb 2013)*
708

    
709
This was the first beta release of the 2.7 series. All important changes
710
are listed in the latest 2.7 entry.
711

    
712

    
713
Version 2.6.2
714
-------------
715

    
716
*(Released Fri, 21 Dec 2012)*
717

    
718
Important behaviour change: hbal won't rebalance anymore instances which
719
have the ``auto_balance`` attribute set to false. This was the intention
720
all along, but until now it only skipped those from the N+1 memory
721
reservation (DRBD-specific).
722

    
723
A significant number of bug fixes in this release:
724

    
725
- Fixed disk adoption interaction with ipolicy checks.
726
- Fixed networking issues when instances are started, stopped or
727
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
728
- Fixed the warning in cluster verify for shared storage instances not
729
  being redundant.
730
- Fixed removal of storage directory on shared file storage (issue 262).
731
- Fixed validation of LVM volume group name in OpClusterSetParams
732
  (``gnt-cluster modify``) (issue 285).
733
- Fixed runtime memory increases (``gnt-instance modify -m``).
734
- Fixed live migration under Xen's ``xl`` mode.
735
- Fixed ``gnt-instance console`` with ``xl``.
736
- Fixed building with newer Haskell compiler/libraries.
737
- Fixed PID file writing in Haskell daemons (confd); this prevents
738
  restart issues if confd was launched manually (outside of
739
  ``daemon-util``) while another copy of it was running
740
- Fixed a type error when doing live migrations with KVM (issue 297) and
741
  the error messages for failing migrations have been improved.
742
- Fixed opcode validation for the out-of-band commands (``gnt-node
743
  power``).
744
- Fixed a type error when unsetting OS hypervisor parameters (issue
745
  311); now it's possible to unset all OS-specific hypervisor
746
  parameters.
747
- Fixed the ``dry-run`` mode for many operations: verification of
748
  results was over-zealous but didn't take into account the ``dry-run``
749
  operation, resulting in "wrong" failures.
750
- Fixed bash completion in ``gnt-job list`` when the job queue has
751
  hundreds of entries; especially with older ``bash`` versions, this
752
  results in significant CPU usage.
753

    
754
And lastly, a few other improvements have been made:
755

    
756
- Added option to force master-failover without voting (issue 282).
757
- Clarified error message on lock conflict (issue 287).
758
- Logging of newly submitted jobs has been improved (issue 290).
759
- Hostname checks have been made uniform between instance rename and
760
  create (issue 291).
761
- The ``--submit`` option is now supported by ``gnt-debug delay``.
762
- Shutting down the master daemon by sending SIGTERM now stops it from
763
  processing jobs waiting for locks; instead, those jobs will be started
764
  once again after the master daemon is started the next time (issue
765
  296).
766
- Support for Xen's ``xl`` program has been improved (besides the fixes
767
  above).
768
- Reduced logging noise in the Haskell confd daemon (only show one log
769
  entry for each config reload, instead of two).
770
- Several man page updates and typo fixes.
771

    
772

    
773
Version 2.6.1
774
-------------
775

    
776
*(Released Fri, 12 Oct 2012)*
777

    
778
A small bugfix release. Among the bugs fixed:
779

    
780
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
781
  made the command unusable.
782
- Commands that issue many jobs don't fail anymore just because some jobs
783
  take so long that other jobs are archived.
784
- Failures during ``gnt-instance reinstall`` are reflected by the exit
785
  status.
786
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
787
  DRBD is enabled.
788
- When ``always_failover`` is set, ``--allow-failover`` is not required
789
  in migrate commands anymore.
790
- ``bash_completion`` works even if extglob is disabled.
791
- Fixed bug with locks that made failover for RDB-based instances fail.
792
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
793
  a random node instead of one based on the allocator metric.
794
- Support for newer versions of pylint and pep8.
795
- Hail doesn't fail anymore when trying to add an instance of type
796
  ``file``, ``sharedfile`` or ``rbd``.
797
- Added new Makefile target to rebuild the whole distribution, so that
798
  all files are included.
799

    
800

    
801
Version 2.6.0
802
-------------
803

    
804
*(Released Fri, 27 Jul 2012)*
805

    
806

    
807
.. attention:: The ``LUXI`` protocol has been made more consistent
808
   regarding its handling of command arguments. This, however, leads to
809
   incompatibility issues with previous versions. Please ensure that you
810
   restart Ganeti daemons soon after the upgrade, otherwise most
811
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
812
   pausing/resuming the watcher, cancelling and archiving jobs, querying
813
   the cluster configuration) will fail.
814

    
815

    
816
New features
817
~~~~~~~~~~~~
818

    
819
Instance run status
820
+++++++++++++++++++
821

    
822
The current ``admin_up`` field, which used to denote whether an instance
823
should be running or not, has been removed. Instead, ``admin_state`` is
824
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
825

    
826
The rational behind this is that an instance being “down” can have
827
different meanings:
828

    
829
- it could be down during a reboot
830
- it could be temporarily be down for a reinstall
831
- or it could be down because it is deprecated and kept just for its
832
  disk
833

    
834
The previous Boolean state was making it difficult to do capacity
835
calculations: should Ganeti reserve memory for a down instance? Now, the
836
tri-state field makes it clear:
837

    
838
- in ``up`` and ``down`` state, all resources are reserved for the
839
  instance, and it can be at any time brought up if it is down
840
- in ``offline`` state, only disk space is reserved for it, but not
841
  memory or CPUs
842

    
843
The field can have an extra use: since the transition between ``up`` and
844
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
845
transition between ``offline`` and ``down`` is done via ``gnt-instance
846
modify``, it is possible to given different rights to users. For
847
example, owners of an instance could be allowed to start/stop it, but
848
not transition it out of the offline state.
849

    
850
Instance policies and specs
851
+++++++++++++++++++++++++++
852

    
853
In previous Ganeti versions, an instance creation request was not
854
limited on the minimum size and on the maximum size just by the cluster
855
resources. As such, any policy could be implemented only in third-party
856
clients (RAPI clients, or shell wrappers over ``gnt-*``
857
tools). Furthermore, calculating cluster capacity via ``hspace`` again
858
required external input with regards to instance sizes.
859

    
860
In order to improve these workflows and to allow for example better
861
per-node group differentiation, we introduced instance specs, which
862
allow declaring:
863

    
864
- minimum instance disk size, disk count, memory size, cpu count
865
- maximum values for the above metrics
866
- and “standard” values (used in ``hspace`` to calculate the standard
867
  sized instances)
868

    
869
The minimum/maximum values can be also customised at node-group level,
870
for example allowing more powerful hardware to support bigger instance
871
memory sizes.
872

    
873
Beside the instance specs, there are a few other settings belonging to
874
the instance policy framework. It is possible now to customise, per
875
cluster and node-group:
876

    
877
- the list of allowed disk templates
878
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
879
- the maximum ratio of instance to spindles (see below for more
880
  information) for local storage
881

    
882
All these together should allow all tools that talk to Ganeti to know
883
what are the ranges of allowed values for instances and the
884
over-subscription that is allowed.
885

    
886
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
887
instance configuration, and the physical CPU configuration from the
888
node. For the spindle ratios however, we didn't track before these
889
values, so new parameters have been added:
890

    
891
- a new node parameter ``spindle_count``, defaults to 1, customisable at
892
  node group or node level
893
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
894

    
895
Note that spindles in this context doesn't need to mean actual
896
mechanical hard-drives; it's just a relative number for both the node
897
I/O capacity and instance I/O consumption.
898

    
899
Instance migration behaviour
900
++++++++++++++++++++++++++++
901

    
902
While live-migration is in general desirable over failover, it is
903
possible that for some workloads it is actually worse, due to the
904
variable time of the “suspend” phase during live migration.
905

    
906
To allow the tools to work consistently over such instances (without
907
having to hard-code instance names), a new backend parameter
908
``always_failover`` has been added to control the migration/failover
909
behaviour. When set to True, all migration requests for an instance will
910
instead fall-back to failover.
911

    
912
Instance memory ballooning
913
++++++++++++++++++++++++++
914

    
915
Initial support for memory ballooning has been added. The memory for an
916
instance is no longer fixed (backend parameter ``memory``), but instead
917
can vary between minimum and maximum values (backend parameters
918
``minmem`` and ``maxmem``). Currently we only change an instance's
919
memory when:
920

    
921
- live migrating or failing over and instance and the target node
922
  doesn't have enough memory
923
- user requests changing the memory via ``gnt-instance modify
924
  --runtime-memory``
925

    
926
Instance CPU pinning
927
++++++++++++++++++++
928

    
929
In order to control the use of specific CPUs by instance, support for
930
controlling CPU pinning has been added for the Xen, HVM and LXC
931
hypervisors. This is controlled by a new hypervisor parameter
932
``cpu_mask``; details about possible values for this are in the
933
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
934
VCPU-to-CPU mapping) form will work well only when all nodes in your
935
cluster have the same amount of CPUs.
936

    
937
Disk parameters
938
+++++++++++++++
939

    
940
Another area in which Ganeti was not customisable were the parameters
941
used for storage configuration, e.g. how many stripes to use for LVM,
942
DRBD resync configuration, etc.
943

    
944
To improve this area, we've added disks parameters, which are
945
customisable at cluster and node group level, and which allow to
946
specify various parameters for disks (DRBD has the most parameters
947
currently), for example:
948

    
949
- DRBD resync algorithm and parameters (e.g. speed)
950
- the default VG for meta-data volumes for DRBD
951
- number of stripes for LVM (plain disk template)
952
- the RBD pool
953

    
954
These parameters can be modified via ``gnt-cluster modify -D …`` and
955
``gnt-group modify -D …``, and are used at either instance creation (in
956
case of LVM stripes, for example) or at disk “activation” time
957
(e.g. resync speed).
958

    
959
Rados block device support
960
++++++++++++++++++++++++++
961

    
962
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
963
denoted by the ``rbd`` disk template type. This is considered
964
experimental, feedback is welcome. For details on configuring it, see
965
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
966

    
967
Master IP setup
968
+++++++++++++++
969

    
970
The existing master IP functionality works well only in simple setups (a
971
single network shared by all nodes); however, if nodes belong to
972
different networks, then the ``/32`` setup and lack of routing
973
information is not enough.
974

    
975
To allow the master IP to function well in more complex cases, the
976
system was reworked as follows:
977

    
978
- a master IP netmask setting has been added
979
- the master IP activation/turn-down code was moved from the node daemon
980
  to a separate script
981
- whether to run the Ganeti-supplied master IP script or a user-supplied
982
  on is a ``gnt-cluster init`` setting
983

    
984
Details about the location of the standard and custom setup scripts are
985
in the man page :manpage:`gnt-cluster(8)`; for information about the
986
setup script protocol, look at the Ganeti-supplied script.
987

    
988
SPICE support
989
+++++++++++++
990

    
991
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
992
improved.
993

    
994
It is now possible to use TLS-protected connections, and when renewing
995
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
996
it is now possible to specify spice or spice CA certificates. Also, it
997
is possible to configure a password for SPICE sessions via the
998
hypervisor parameter ``spice_password_file``.
999

    
1000
There are also new parameters to control the compression and streaming
1001
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
1002
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
1003
for the spice parameters.
1004

    
1005
Lastly, it is now possible to see the SPICE connection information via
1006
``gnt-instance console``.
1007

    
1008
OVF converter
1009
+++++++++++++
1010

    
1011
A new tool (``tools/ovfconverter``) has been added that supports
1012
conversion between Ganeti and the `Open Virtualization Format
1013
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
1014
from).
1015

    
1016
This relies on the ``qemu-img`` tool to convert the disk formats, so the
1017
actual compatibility with other virtualization solutions depends on it.
1018

    
1019
Confd daemon changes
1020
++++++++++++++++++++
1021

    
1022
The configuration query daemon (``ganeti-confd``) is now optional, and
1023
has been rewritten in Haskell; whether to use the daemon at all, use the
1024
Python (default) or the Haskell version is selectable at configure time
1025
via the ``--enable-confd`` parameter, which can take one of the
1026
``haskell``, ``python`` or ``no`` values. If not used, disabling the
1027
daemon will result in a smaller footprint; for larger systems, we
1028
welcome feedback on the Haskell version which might become the default
1029
in future versions.
1030

    
1031
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
1032
daemon running. The Python version doesn't implement the new call.
1033

    
1034

    
1035
User interface changes
1036
~~~~~~~~~~~~~~~~~~~~~~
1037

    
1038
We have replaced the ``--disks`` option of ``gnt-instance
1039
replace-disks`` with a more flexible ``--disk`` option, which allows
1040
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
1041
disk size and mode can be changed upon recreation (via ``gnt-instance
1042
recreate-disks``, which accepts the same ``--disk`` option).
1043

    
1044
As many people are used to a ``show`` command, we have added that as an
1045
alias to ``info`` on all ``gnt-*`` commands.
1046

    
1047
The ``gnt-instance grow-disk`` command has a new mode in which it can
1048
accept the target size of the disk, instead of the delta; this can be
1049
more safe since two runs in absolute mode will be idempotent, and
1050
sometimes it's also easier to specify the desired size directly.
1051

    
1052
Also the handling of instances with regard to offline secondaries has
1053
been improved. Instance operations should not fail because one of it's
1054
secondary nodes is offline, even though it's safe to proceed.
1055

    
1056
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
1057
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1058
minors to instance name.
1059

    
1060
API changes
1061
~~~~~~~~~~~
1062

    
1063
RAPI coverage has improved, with (for example) new resources for
1064
recreate-disks, node power-cycle, etc.
1065

    
1066
Compatibility
1067
~~~~~~~~~~~~~
1068

    
1069
There is partial support for ``xl`` in the Xen hypervisor; feedback is
1070
welcome.
1071

    
1072
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1073
whether to still support Python 2.4 or move to Python 2.6 as minimum
1074
required version.
1075

    
1076
Support for Fedora has been slightly improved; the provided example
1077
init.d script should work better on it and the INSTALL file should
1078
document the needed dependencies.
1079

    
1080
Internal changes
1081
~~~~~~~~~~~~~~~~
1082

    
1083
The deprecated ``QueryLocks`` LUXI request has been removed. Use
1084
``Query(what=QR_LOCK, ...)`` instead.
1085

    
1086
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1087
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1088
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1089
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1090
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1091

    
1092
RAPI client: ``CertificateError`` now derives from
1093
``GanetiApiError``. This should make it more easy to handle Ganeti
1094
errors.
1095

    
1096
Deprecation warnings due to PyCrypto/paramiko import in
1097
``tools/setup-ssh`` have been silenced, as usually they are safe; please
1098
make sure to run an up-to-date paramiko version, if you use this tool.
1099

    
1100
The QA scripts now depend on Python 2.5 or above (the main code base
1101
still works with Python 2.4).
1102

    
1103
The configuration file (``config.data``) is now written without
1104
indentation for performance reasons; if you want to edit it, it can be
1105
re-formatted via ``tools/fmtjson``.
1106

    
1107
A number of bugs has been fixed in the cluster merge tool.
1108

    
1109
``x509`` certification verification (used in import-export) has been
1110
changed to allow the same clock skew as permitted by the cluster
1111
verification. This will remove some rare but hard to diagnose errors in
1112
import-export.
1113

    
1114

    
1115
Version 2.6.0 rc4
1116
-----------------
1117

    
1118
*(Released Thu, 19 Jul 2012)*
1119

    
1120
Very few changes from rc4 to the final release, only bugfixes:
1121

    
1122
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
1123
  instance, fix CDROM booting for KVM instances)
1124
- fixed node group modification of node parameters
1125
- fixed issue in LUClusterVerifyGroup with multi-group clusters
1126
- fixed generation of bash completion to ensure a stable ordering
1127
- fixed a few typos
1128

    
1129

    
1130
Version 2.6.0 rc3
1131
-----------------
1132

    
1133
*(Released Fri, 13 Jul 2012)*
1134

    
1135
Third release candidate for 2.6. The following changes were done from
1136
rc3 to rc4:
1137

    
1138
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1139
- Fixed an inconsistency in the LUXI protocol with the provided
1140
  arguments (NOT backwards compatible)
1141
- Fixed a bug with node groups ipolicy where ``min`` was greater than
1142
  the cluster ``std`` value
1143
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1144
  easier instance debugging on nodes (requires ``hconfd`` to work)
1145

    
1146

    
1147
Version 2.6.0 rc2
1148
-----------------
1149

    
1150
*(Released Tue, 03 Jul 2012)*
1151

    
1152
Second release candidate for 2.6. The following changes were done from
1153
rc2 to rc3:
1154

    
1155
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1156
  master candidates
1157
- Fixed a RAPI regression on missing beparams/memory
1158
- Fixed redistribution of files on offline nodes
1159
- Added possibility to run activate-disks even though secondaries are
1160
  offline. With this change it relaxes also the strictness on some other
1161
  commands which use activate disks internally:
1162
  * ``gnt-instance start|reboot|rename|backup|export``
1163
- Made it possible to remove safely an instance if its secondaries are
1164
  offline
1165
- Made it possible to reinstall even though secondaries are offline
1166

    
1167

    
1168
Version 2.6.0 rc1
1169
-----------------
1170

    
1171
*(Released Mon, 25 Jun 2012)*
1172

    
1173
First release candidate for 2.6. The following changes were done from
1174
rc1 to rc2:
1175

    
1176
- Fixed bugs with disk parameters and ``rbd`` templates as well as
1177
  ``instance_os_add``
1178
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1179
  behaviour. It supports now the modify operation
1180
- ``hcheck`` implemented to analyze cluster health and possibility of
1181
  improving health by rebalance
1182
- ``hbal`` has been improved in dealing with split instances
1183

    
1184

    
1185
Version 2.6.0 beta2
1186
-------------------
1187

    
1188
*(Released Mon, 11 Jun 2012)*
1189

    
1190
Second beta release of 2.6. The following changes were done from beta2
1191
to rc1:
1192

    
1193
- Fixed ``daemon-util`` with non-root user models
1194
- Fixed creation of plain instances with ``--no-wait-for-sync``
1195
- Fix wrong iv_names when running ``cfgupgrade``
1196
- Export more information in RAPI group queries
1197
- Fixed bug when changing instance network interfaces
1198
- Extended burnin to do NIC changes
1199
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1200
- Changed default for DRBD barriers
1201
- Fixed DRBD error reporting for syncer rate
1202
- Verify the options on disk parameters
1203

    
1204
And of course various fixes to documentation and improved unittests and
1205
QA.
1206

    
1207

    
1208
Version 2.6.0 beta1
1209
-------------------
1210

    
1211
*(Released Wed, 23 May 2012)*
1212

    
1213
First beta release of 2.6. The following changes were done from beta1 to
1214
beta2:
1215

    
1216
- integrated patch for distributions without ``start-stop-daemon``
1217
- adapted example init.d script to work on Fedora
1218
- fixed log handling in Haskell daemons
1219
- adapted checks in the watcher for pycurl linked against libnss
1220
- add partial support for ``xl`` instead of ``xm`` for Xen
1221
- fixed a type issue in cluster verification
1222
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1223
  clusters)
1224

    
1225
Plus integrated fixes from the 2.5 branch:
1226

    
1227
- fixed ``kvm-ifup`` to use ``/bin/bash``
1228
- fixed parallel build failures
1229
- KVM live migration when using a custom keymap
1230

    
1231

    
1232
Version 2.5.2
1233
-------------
1234

    
1235
*(Released Tue, 24 Jul 2012)*
1236

    
1237
A small bugfix release, with no new features:
1238

    
1239
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1240
  different default shell (e.g. Debian, Ubuntu)
1241
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1242
  243 and Debian bug #650664)
1243
- fixed compatibility with KVM versions that don't support multiple boot
1244
  devices (fixes Issue 230 and Debian bug #624256)
1245

    
1246
Additionally, a few fixes were done to the build system (fixed parallel
1247
build failures) and to the unittests (fixed race condition in test for
1248
FileID functions, and the default enable/disable mode for QA test is now
1249
customisable).
1250

    
1251

    
1252
Version 2.5.1
1253
-------------
1254

    
1255
*(Released Fri, 11 May 2012)*
1256

    
1257
A small bugfix release.
1258

    
1259
The main issues solved are on the topic of compatibility with newer LVM
1260
releases:
1261

    
1262
- fixed parsing of ``lv_attr`` field
1263
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1264
  the ``--force`` flag is needed
1265

    
1266
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1267
to accept kernel 3.x too (was hardcoded to 2.6.*).
1268

    
1269
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1270
of hook results and that also made display of other errors suboptimal
1271
was fixed; the code behaves now like 2.4 and earlier.
1272

    
1273
Another change in 2.5, the cleanup of the OS scripts environment, is too
1274
aggressive: it removed even the ``PATH`` variable, which requires the OS
1275
scripts to *always* need to export it. Since this is a bit too strict,
1276
we now export a minimal PATH, the same that we export for hooks.
1277

    
1278
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1279
integrated into this release.
1280

    
1281
Finally, a few other miscellaneous changes were done (no new features,
1282
just small improvements):
1283

    
1284
- Fix ``gnt-group --help`` display
1285
- Fix hardcoded Xen kernel path
1286
- Fix grow-disk handling of invalid units
1287
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1288
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1289

    
1290

    
1291
Version 2.5.0
1292
-------------
1293

    
1294
*(Released Thu, 12 Apr 2012)*
1295

    
1296
Incompatible/important changes and bugfixes
1297
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1298

    
1299
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1300
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1301
  to match the underlying LUXI interface.
1302
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1303
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1304
  not maintain backwards-compatibility as the underlying operation
1305
  changed in an incompatible way. The RAPI client can talk to old
1306
  servers, but it needs to be told so as the return value changed.
1307
- When creating file-based instances via RAPI, the ``file_driver``
1308
  parameter no longer defaults to ``loop`` and must be specified.
1309
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1310
  ``link`` instead.
1311
- Support for the undocumented and deprecated RAPI instance creation
1312
  request format version 0 has been dropped. Use version 1, supported
1313
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1314
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1315
  documentation <install>`.
1316
- The "cluster-verify" hooks are now executed per group by the
1317
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1318
  if you just run ``gnt-cluster verify``, which generates one opcode per
1319
  group.
1320
- The environment as passed to the OS scripts is cleared, and thus no
1321
  environment variables defined in the node daemon's environment will be
1322
  inherited by the scripts.
1323
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1324
  deprecated.
1325
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1326
  support operations involving multiple node groups.
1327
- Offline nodes are ignored when failing over an instance.
1328
- Support for KVM version 1.0, which changed the version reporting format
1329
  from 3 to 2 digits.
1330
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1331
  removal.
1332
- ``Makefile`` is now compatible with Automake 1.11.2
1333
- Includes all bugfixes made in the 2.4 series
1334

    
1335
New features
1336
~~~~~~~~~~~~
1337

    
1338
- The ganeti-htools project has been merged into the ganeti-core source
1339
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1340
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1341
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1342
  GPT-style partition tables (requires `parted
1343
  <http://www.gnu.org/s/parted/>`_).
1344
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1345
- Allowed adding tags on instance creation.
1346
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1347
- Allow instances to be started in a paused state, enabling the user to
1348
  see the complete console output on boot using the console.
1349
- Added new hypervisor flag to control default reboot behaviour
1350
  (``reboot_behavior``).
1351
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1352
- Improved out-of-band management support:
1353

    
1354
  - Added ``gnt-node health`` command reporting the health status of
1355
    nodes.
1356
  - Added ``gnt-node power`` command to manage power status of nodes.
1357
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1358

    
1359
- Instance migration can fall back to failover if instance is not
1360
  running.
1361
- Filters can be used when listing nodes, instances, groups and locks;
1362
  see :manpage:`ganeti(7)` manpage.
1363
- Added post-execution status as variables to :doc:`hooks <hooks>`
1364
  environment.
1365
- Instance tags are exported/imported together with the instance.
1366
- When given an explicit job ID, ``gnt-job info`` will work for archived
1367
  jobs.
1368
- Jobs can define dependencies on other jobs (not yet supported via
1369
  RAPI or command line, but used by internal commands and usable via
1370
  LUXI).
1371

    
1372
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1373
    dependencies.
1374

    
1375
- Instance failover is now available as a RAPI resource
1376
  (``/2/instances/[instance_name]/failover``).
1377
- ``gnt-instance info`` defaults to static information if primary node
1378
  is offline.
1379
- Opcodes have a new ``comment`` attribute.
1380
- Added basic SPICE support to KVM hypervisor.
1381
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1382

    
1383
Node group improvements
1384
~~~~~~~~~~~~~~~~~~~~~~~
1385

    
1386
- ``gnt-cluster verify`` has been modified to check groups separately,
1387
  thereby improving performance.
1388
- Node group support has been added to ``gnt-cluster verify-disks``,
1389
  which now operates per node group.
1390
- Watcher has been changed to work better with node groups.
1391

    
1392
  - One process and state file per node group.
1393
  - Slow watcher in one group doesn't block other group's watcher.
1394

    
1395
- Added new command, ``gnt-group evacuate``, to move all instances in a
1396
  node group to other groups.
1397
- Added ``gnt-instance change-group`` to move an instance to another
1398
  node group.
1399
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1400
  per-group operations.
1401
- Node groups can be tagged.
1402
- Some operations switch from an exclusive to a shared lock as soon as
1403
  possible.
1404
- Instance's primary and secondary nodes' groups are now available as
1405
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1406
  and ``snodes.group.uuid``).
1407

    
1408
Misc
1409
~~~~
1410

    
1411
- Numerous updates to documentation and manpages.
1412

    
1413
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1414
    descriptions.
1415
  - Some opcode/job results are now also documented, see :doc:`RAPI
1416
    <rapi>`.
1417

    
1418
- A lockset's internal lock is now also visible in lock monitor.
1419
- Log messages from job queue workers now contain information about the
1420
  opcode they're processing.
1421
- ``gnt-instance console`` no longer requires the instance lock.
1422
- A short delay when waiting for job changes reduces the number of LUXI
1423
  requests significantly.
1424
- DRBD metadata volumes are overwritten with zeros during disk creation.
1425
- Out-of-band commands no longer acquire the cluster lock in exclusive
1426
  mode.
1427
- ``devel/upload`` now uses correct permissions for directories.
1428

    
1429

    
1430
Version 2.5.0 rc6
1431
-----------------
1432

    
1433
*(Released Fri, 23 Mar 2012)*
1434

    
1435
This was the sixth release candidate of the 2.5 series.
1436

    
1437

    
1438
Version 2.5.0 rc5
1439
-----------------
1440

    
1441
*(Released Mon, 9 Jan 2012)*
1442

    
1443
This was the fifth release candidate of the 2.5 series.
1444

    
1445

    
1446
Version 2.5.0 rc4
1447
-----------------
1448

    
1449
*(Released Thu, 27 Oct 2011)*
1450

    
1451
This was the fourth release candidate of the 2.5 series.
1452

    
1453

    
1454
Version 2.5.0 rc3
1455
-----------------
1456

    
1457
*(Released Wed, 26 Oct 2011)*
1458

    
1459
This was the third release candidate of the 2.5 series.
1460

    
1461

    
1462
Version 2.5.0 rc2
1463
-----------------
1464

    
1465
*(Released Tue, 18 Oct 2011)*
1466

    
1467
This was the second release candidate of the 2.5 series.
1468

    
1469

    
1470
Version 2.5.0 rc1
1471
-----------------
1472

    
1473
*(Released Tue, 4 Oct 2011)*
1474

    
1475
This was the first release candidate of the 2.5 series.
1476

    
1477

    
1478
Version 2.5.0 beta3
1479
-------------------
1480

    
1481
*(Released Wed, 31 Aug 2011)*
1482

    
1483
This was the third beta release of the 2.5 series.
1484

    
1485

    
1486
Version 2.5.0 beta2
1487
-------------------
1488

    
1489
*(Released Mon, 22 Aug 2011)*
1490

    
1491
This was the second beta release of the 2.5 series.
1492

    
1493

    
1494
Version 2.5.0 beta1
1495
-------------------
1496

    
1497
*(Released Fri, 12 Aug 2011)*
1498

    
1499
This was the first beta release of the 2.5 series.
1500

    
1501

    
1502
Version 2.4.5
1503
-------------
1504

    
1505
*(Released Thu, 27 Oct 2011)*
1506

    
1507
- Fixed bug when parsing command line parameter values ending in
1508
  backslash
1509
- Fixed assertion error after unclean master shutdown
1510
- Disable HTTP client pool for RPC, significantly reducing memory usage
1511
  of master daemon
1512
- Fixed queue archive creation with wrong permissions
1513

    
1514

    
1515
Version 2.4.4
1516
-------------
1517

    
1518
*(Released Tue, 23 Aug 2011)*
1519

    
1520
Small bug-fixes:
1521

    
1522
- Fixed documentation for importing with ``--src-dir`` option
1523
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1524
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1525

    
1526

    
1527
Version 2.4.3
1528
-------------
1529

    
1530
*(Released Fri, 5 Aug 2011)*
1531

    
1532
Many bug-fixes and a few small features:
1533

    
1534
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1535
  issues when you tried to add an instance with two MAC addresses in one
1536
  request
1537
- KVM: fixed per-instance stored UID value
1538
- KVM: configure bridged NICs at migration start
1539
- KVM: Fix a bug where instance will not start with never KVM versions
1540
  (>= 0.14)
1541
- Added OS search path to ``gnt-cluster info``
1542
- Fixed an issue with ``file_storage_dir`` where you were forced to
1543
  provide an absolute path, but the documentation states it is a
1544
  relative path, the documentation was right
1545
- Added a new parameter to instance stop/start called ``--no-remember``
1546
  that will make the state change to not be remembered
1547
- Implemented ``no_remember`` at RAPI level
1548
- Improved the documentation
1549
- Node evacuation: don't call IAllocator if node is already empty
1550
- Fixed bug in DRBD8 replace disks on current nodes
1551
- Fixed bug in recreate-disks for DRBD instances
1552
- Moved assertion checking locks in ``gnt-instance replace-disks``
1553
  causing it to abort with not owning the right locks for some situation
1554
- Job queue: Fixed potential race condition when cancelling queued jobs
1555
- Fixed off-by-one bug in job serial generation
1556
- ``gnt-node volumes``: Fix instance names
1557
- Fixed aliases in bash completion
1558
- Fixed a bug in reopening log files after being sent a SIGHUP
1559
- Added a flag to burnin to allow specifying VCPU count
1560
- Bugfixes to non-root Ganeti configuration
1561

    
1562

    
1563
Version 2.4.2
1564
-------------
1565

    
1566
*(Released Thu, 12 May 2011)*
1567

    
1568
Many bug-fixes and a few new small features:
1569

    
1570
- Fixed a bug related to log opening failures
1571
- Fixed a bug in instance listing with orphan instances
1572
- Fixed a bug which prevented resetting the cluster-level node parameter
1573
  ``oob_program`` to the default
1574
- Many fixes related to the ``cluster-merge`` tool
1575
- Fixed a race condition in the lock monitor, which caused failures
1576
  during (at least) creation of many instances in parallel
1577
- Improved output for gnt-job info
1578
- Removed the quiet flag on some ssh calls which prevented debugging
1579
  failures
1580
- Improved the N+1 failure messages in cluster verify by actually
1581
  showing the memory values (needed and available)
1582
- Increased lock attempt timeouts so that when executing long operations
1583
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1584
  too early and thus prevent the use of the 'fair' mechanism
1585
- Changed instance query data (``gnt-instance info``) to not acquire
1586
  locks unless needed, thus allowing its use on locked instance if only
1587
  static information is asked for
1588
- Improved behaviour with filesystems that do not support rename on an
1589
  opened file
1590
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1591
  kept locks on all nodes during the wipe, which is unneeded
1592
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1593
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1594
  10GiB) which caused the wipe to fail right at the end in some cases
1595
- Fixed master IP activation when doing master failover with no-voting
1596
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1597
  the master node itself
1598
- Fixed potential data-loss in under disk full conditions, where Ganeti
1599
  wouldn't check correctly the return code and would consider
1600
  partially-written files 'correct'
1601
- Fixed bug related to multiple VGs and DRBD disk replacing
1602
- Added new disk parameter ``metavg`` that allows placement of the meta
1603
  device for DRBD in a different volume group
1604
- Fixed error handling in the node daemon when the system libc doesn't
1605
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1606
- Fixed lock release during replace-disks, which kept cluster-wide locks
1607
  when doing disk replaces with an iallocator script
1608
- Added check for missing bridges in cluster verify
1609
- Handle EPIPE errors while writing to the terminal better, so that
1610
  piping the output to e.g. ``less`` doesn't cause a backtrace
1611
- Fixed rare case where a ^C during Luxi calls could have been
1612
  interpreted as server errors, instead of simply terminating
1613
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1614
  assign-nodes``)
1615
- Added a few more parameters to the KVM hypervisor, allowing a second
1616
  CDROM, custom disk type for CDROMs and a floppy image
1617
- Removed redundant message in instance rename when the name is given
1618
  already as a FQDN
1619
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1620
  disks on new nodes, allowing recreation when the original instance
1621
  nodes are completely gone
1622
- Added option when converting disk templates to DRBD to skip waiting
1623
  for the resync, in order to make the instance available sooner
1624
- Added two new variables to the OS scripts environment (containing the
1625
  instance's nodes)
1626
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1627
  to allow use of ``pvgrub`` as bootloader
1628
- Changed the instance memory modifications to only check out-of-memory
1629
  conditions on memory increases, and turned the secondary node warnings
1630
  into errors (they can still be overridden via ``--force``)
1631
- Fixed the handling of a corner case when the Python installation gets
1632
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1633
  to execute a command that doesn't exist
1634
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1635
  node of the instance returned failures during instance shutdown; this
1636
  adds the option ``--ignore-consistency`` to gnt-instance move
1637

    
1638
And as usual, various improvements to the error messages, documentation
1639
and man pages.
1640

    
1641

    
1642
Version 2.4.1
1643
-------------
1644

    
1645
*(Released Wed, 09 Mar 2011)*
1646

    
1647
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1648
the RAPI users file if run twice (even with ``--dry-run``).
1649

    
1650
The release fixes that bug (nothing else changed).
1651

    
1652

    
1653
Version 2.4.0
1654
-------------
1655

    
1656
*(Released Mon, 07 Mar 2011)*
1657

    
1658
Final 2.4.0 release. Just a few small fixes:
1659

    
1660
- Fixed RAPI node evacuate
1661
- Fixed the kvm-ifup script
1662
- Fixed internal error handling for special job cases
1663
- Updated man page to specify the escaping feature for options
1664

    
1665

    
1666
Version 2.4.0 rc3
1667
-----------------
1668

    
1669
*(Released Mon, 28 Feb 2011)*
1670

    
1671
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1672
that this feature wiped the disks of the wrong instance, leading to loss
1673
of data.
1674

    
1675
Other changes:
1676

    
1677
- Fixed title of query field containing instance name
1678
- Expanded the glossary in the documentation
1679
- Fixed one unittest (internal issue)
1680

    
1681

    
1682
Version 2.4.0 rc2
1683
-----------------
1684

    
1685
*(Released Mon, 21 Feb 2011)*
1686

    
1687
A number of bug fixes plus just a couple functionality changes.
1688

    
1689
On the user-visible side, the ``gnt-* list`` command output has changed
1690
with respect to "special" field states. The current rc1 style of display
1691
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1692
the default output mode special fields states are displayed as follows:
1693

    
1694
- Offline resource: ``*``
1695
- Unavailable/not applicable: ``-``
1696
- Data missing (RPC failure): ``?``
1697
- Unknown field: ``??``
1698

    
1699
Another user-visible change is the addition of ``--force-join`` to
1700
``gnt-node add``.
1701

    
1702
As for bug fixes:
1703

    
1704
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1705
- Fixed regression in RAPI/instance reinstall where all parameters were
1706
  required (instead of optional)
1707
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1708
- Fixed iallocator usage (offline nodes were not considered offline)
1709
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1710
- Fixed hypervisor and OS parameter validation with respect to
1711
  non-vm_capable nodes
1712
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1713
  cosmetic)
1714
- Fixed ``tools/listrunner`` with respect to agent-based usage
1715

    
1716

    
1717
Version 2.4.0 rc1
1718
-----------------
1719

    
1720
*(Released Fri,  4 Feb 2011)*
1721

    
1722
Many changes and fixes since the beta1 release. While there were some
1723
internal changes, the code has been mostly stabilised for the RC
1724
release.
1725

    
1726
Note: the dumb allocator was removed in this release, as it was not kept
1727
up-to-date with the IAllocator protocol changes. It is recommended to
1728
use the ``hail`` command from the ganeti-htools package.
1729

    
1730
Note: the 2.4 and up versions of Ganeti are not compatible with the
1731
0.2.x branch of ganeti-htools. You need to upgrade to
1732
ganeti-htools-0.3.0 (or later).
1733

    
1734
Regressions fixed from 2.3
1735
~~~~~~~~~~~~~~~~~~~~~~~~~~
1736

    
1737
- Fixed the ``gnt-cluster verify-disks`` command
1738
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1739
  serially on nodes)
1740
- Fixed disk adoption breakage
1741
- Fixed wrong headers in instance listing for field aliases
1742

    
1743
Other bugs fixed
1744
~~~~~~~~~~~~~~~~
1745

    
1746
- Fixed corner case in KVM handling of NICs
1747
- Fixed many cases of wrong handling of non-vm_capable nodes
1748
- Fixed a bug where a missing instance symlink was not possible to
1749
  recreate with any ``gnt-*`` command (now ``gnt-instance
1750
  activate-disks`` does it)
1751
- Fixed the volume group name as reported by ``gnt-cluster
1752
  verify-disks``
1753
- Increased timeouts for the import-export code, hopefully leading to
1754
  fewer aborts due network or instance timeouts
1755
- Fixed bug in ``gnt-node list-storage``
1756
- Fixed bug where not all daemons were started on cluster
1757
  initialisation, but only at the first watcher run
1758
- Fixed many bugs in the OOB implementation
1759
- Fixed watcher behaviour in presence of instances with offline
1760
  secondaries
1761
- Fixed instance list output for instances running on the wrong node
1762
- a few fixes to the cluster-merge tool, but it still cannot merge
1763
  multi-node groups (currently it is not recommended to use this tool)
1764

    
1765

    
1766
Improvements
1767
~~~~~~~~~~~~
1768

    
1769
- Improved network configuration for the KVM hypervisor
1770
- Added e1000 as a supported NIC for Xen-HVM
1771
- Improved the lvmstrap tool to also be able to use partitions, as
1772
  opposed to full disks
1773
- Improved speed of disk wiping (the cluster parameter
1774
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1775
  of instance creations
1776
- Added documentation for the OS parameters
1777
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1778
  hypervisor is not responding
1779
- Added display of blacklisted and hidden OS information in
1780
  ``gnt-cluster info``
1781
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1782
  NIC and node parameters, which might create problems with currently
1783
  invalid (but undetected) configuration files, but prevents validation
1784
  failures when unrelated parameters are modified
1785
- Changed cluster initialisation to wait for the master daemon to become
1786
  available
1787
- Expanded the RAPI interface:
1788

    
1789
  - Added config redistribution resource
1790
  - Added activation/deactivation of instance disks
1791
  - Added export of console information
1792

    
1793
- Implemented log file reopening on SIGHUP, which allows using
1794
  logrotate(8) for the Ganeti log files
1795
- Added a basic OOB helper script as an example
1796

    
1797

    
1798
Version 2.4.0 beta1
1799
-------------------
1800

    
1801
*(Released Fri, 14 Jan 2011)*
1802

    
1803
User-visible
1804
~~~~~~~~~~~~
1805

    
1806
- Fixed timezone issues when formatting timestamps
1807
- Added support for node groups, available via ``gnt-group`` and other
1808
  commands
1809
- Added out-of-band framework and management, see :doc:`design
1810
  document <design-oob>`
1811
- Removed support for roman numbers from ``gnt-node list`` and
1812
  ``gnt-instance list``.
1813
- Allowed modification of master network interface via ``gnt-cluster
1814
  modify --master-netdev``
1815
- Accept offline secondaries while shutting down instance disks
1816
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1817
- Added support for multiple LVM volume groups
1818
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1819
  requested
1820
- Added commands to list available fields:
1821

    
1822
  - ``gnt-node list-fields``
1823
  - ``gnt-group list-fields``
1824
  - ``gnt-instance list-fields``
1825

    
1826
- Updated documentation and man pages
1827

    
1828
Integration
1829
~~~~~~~~~~~
1830

    
1831
- Moved ``rapi_users`` file into separate directory, now named
1832
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1833
  symlink
1834
- Added new tool for running commands on many machines,
1835
  ``tools/ganeti-listrunner``
1836
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1837
  improving the ``gnt-instance console`` output
1838
- Allowed customisation of disk index separator at ``configure`` time
1839
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1840
- Added support for non-partitioned md disks in ``lvmstrap``
1841
- Added script to gracefully power off KVM instances
1842
- Split ``utils`` module into smaller parts
1843
- Changed query operations to return more detailed information, e.g.
1844
  whether an information is unavailable due to an offline node. To use
1845
  this new functionality, the LUXI call ``Query`` must be used. Field
1846
  information is now stored by the master daemon and can be retrieved
1847
  using ``QueryFields``. Instances, nodes and groups can also be queried
1848
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1849
  exposed via RAPI). The following commands make use of this
1850
  infrastructure change:
1851

    
1852
  - ``gnt-group list``
1853
  - ``gnt-group list-fields``
1854
  - ``gnt-node list``
1855
  - ``gnt-node list-fields``
1856
  - ``gnt-instance list``
1857
  - ``gnt-instance list-fields``
1858
  - ``gnt-debug locks``
1859

    
1860
Remote API
1861
~~~~~~~~~~
1862

    
1863
- New RAPI resources (see :doc:`rapi`):
1864

    
1865
  - ``/2/modify``
1866
  - ``/2/groups``
1867
  - ``/2/groups/[group_name]``
1868
  - ``/2/groups/[group_name]/assign-nodes``
1869
  - ``/2/groups/[group_name]/modify``
1870
  - ``/2/groups/[group_name]/rename``
1871
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1872

    
1873
- RAPI changes:
1874

    
1875
  - Implemented ``no_install`` for instance creation
1876
  - Implemented OS parameters for instance reinstallation, allowing
1877
    use of special settings on reinstallation (e.g. for preserving data)
1878

    
1879
Misc
1880
~~~~
1881

    
1882
- Added IPv6 support in import/export
1883
- Pause DRBD synchronization while wiping disks on instance creation
1884
- Updated unittests and QA scripts
1885
- Improved network parameters passed to KVM
1886
- Converted man pages from docbook to reStructuredText
1887

    
1888

    
1889
Version 2.3.1
1890
-------------
1891

    
1892
*(Released Mon, 20 Dec 2010)*
1893

    
1894
Released version 2.3.1~rc1 without any changes.
1895

    
1896

    
1897
Version 2.3.1 rc1
1898
-----------------
1899

    
1900
*(Released Wed, 1 Dec 2010)*
1901

    
1902
- impexpd: Disable OpenSSL compression in socat if possible (backport
1903
  from master, commit e90739d625b, see :doc:`installation guide
1904
  <install-quick>` for details)
1905
- Changed unittest coverage report to exclude test scripts
1906
- Added script to check version format
1907

    
1908

    
1909
Version 2.3.0
1910
-------------
1911

    
1912
*(Released Wed, 1 Dec 2010)*
1913

    
1914
Released version 2.3.0~rc1 without any changes.
1915

    
1916

    
1917
Version 2.3.0 rc1
1918
-----------------
1919

    
1920
*(Released Fri, 19 Nov 2010)*
1921

    
1922
A number of bugfixes and documentation updates:
1923

    
1924
- Update ganeti-os-interface documentation
1925
- Fixed a bug related to duplicate MACs or similar items which should be
1926
  unique
1927
- Fix breakage in OS state modify
1928
- Reinstall instance: disallow offline secondaries (fixes bug related to
1929
  OS changing but reinstall failing)
1930
- plus all the other fixes between 2.2.1 and 2.2.2
1931

    
1932

    
1933
Version 2.3.0 rc0
1934
-----------------
1935

    
1936
*(Released Tue, 2 Nov 2010)*
1937

    
1938
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1939
- Fixed master failover race with watcher
1940
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1941
  configuration
1942
- Accept previously stopped instance for export with instance removal
1943
- Simplify and extend the environment variables for instance OS scripts
1944
- Added new node flags, ``master_capable`` and ``vm_capable``
1945
- Added optional instance disk wiping prior during allocation. This is a
1946
  cluster-wide option and can be set/modified using
1947
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1948
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1949
  :doc:`install-quick`
1950
- Added a new watcher option (``--ignore-pause``)
1951
- Added option to ignore offline node on instance start/stop
1952
  (``--ignore-offline``)
1953
- Allow overriding OS parameters with ``gnt-instance reinstall``
1954
- Added ability to change node's secondary IP address using ``gnt-node
1955
  modify``
1956
- Implemented privilege separation for all daemons except
1957
  ``ganeti-noded``, see ``configure`` options
1958
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1959
  verify``
1960
- Implemented job priorities (see ``ganeti(7)`` manpage)
1961
- Ignore failures while shutting down instances during failover from
1962
  offline node
1963
- Exit daemon's bootstrap process only once daemon is ready
1964
- Export more information via ``LUInstanceQuery``/remote API
1965
- Improved documentation, QA and unittests
1966
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1967
  restart if the file was created or changed
1968
- Added LUXI protocol version sent with each request and response,
1969
  allowing detection of server/client mismatches
1970
- Moved the Python scripts among gnt-* and ganeti-* into modules
1971
- Moved all code related to setting up SSH to an external script,
1972
  ``setup-ssh``
1973
- Infrastructure changes for node group support in future versions
1974

    
1975

    
1976
Version 2.2.2
1977
-------------
1978

    
1979
*(Released Fri, 19 Nov 2010)*
1980

    
1981
A few small bugs fixed, and some improvements to the build system:
1982

    
1983
- Fix documentation regarding conversion to drbd
1984
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1985
  -B``)
1986
- Fix error handling in node modify with multiple changes
1987
- Allow remote imports without checked names
1988

    
1989

    
1990
Version 2.2.1
1991
-------------
1992

    
1993
*(Released Tue, 19 Oct 2010)*
1994

    
1995
- Disable SSL session ID cache in RPC client
1996

    
1997

    
1998
Version 2.2.1 rc1
1999
-----------------
2000

    
2001
*(Released Thu, 14 Oct 2010)*
2002

    
2003
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
2004
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
2005
  with GnuTLS
2006
- Fix problems with interaction between Curl and Python's HTTP server,
2007
  resulting in increased speed in many RPC calls
2008
- Improve our release script to prevent breakage with older aclocal and
2009
  Python 2.6
2010

    
2011

    
2012
Version 2.2.1 rc0
2013
-----------------
2014

    
2015
*(Released Thu, 7 Oct 2010)*
2016

    
2017
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
2018
  value retrieved from master
2019
- Added support for blacklisted or hidden OS definitions
2020
- Added simple lock monitor (accessible via (``gnt-debug locks``)
2021
- Added support for -mem-path in KVM hypervisor abstraction layer
2022
- Allow overriding instance parameters in tool for inter-cluster
2023
  instance moves (``tools/move-instance``)
2024
- Improved opcode summaries (e.g. in ``gnt-job list``)
2025
- Improve consistency of OS listing by sorting it
2026
- Documentation updates
2027

    
2028

    
2029
Version 2.2.0.1
2030
---------------
2031

    
2032
*(Released Fri, 8 Oct 2010)*
2033

    
2034
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
2035

    
2036

    
2037
Version 2.2.0
2038
-------------
2039

    
2040
*(Released Mon, 4 Oct 2010)*
2041

    
2042
- Fixed regression in ``gnt-instance rename``
2043

    
2044

    
2045
Version 2.2.0 rc2
2046
-----------------
2047

    
2048
*(Released Wed, 22 Sep 2010)*
2049

    
2050
- Fixed OS_VARIANT variable for OS scripts
2051
- Fixed cluster tag operations via RAPI
2052
- Made ``setup-ssh`` exit with non-zero code if an error occurred
2053
- Disabled RAPI CA checks in watcher
2054

    
2055

    
2056
Version 2.2.0 rc1
2057
-----------------
2058

    
2059
*(Released Mon, 23 Aug 2010)*
2060

    
2061
- Support DRBD versions of the format "a.b.c.d"
2062
- Updated manpages
2063
- Re-introduce support for usage from multiple threads in RAPI client
2064
- Instance renames and modify via RAPI
2065
- Work around race condition between processing and archival in job
2066
  queue
2067
- Mark opcodes following failed one as failed, too
2068
- Job field ``lock_status`` was removed due to difficulties making it
2069
  work with the changed job queue in Ganeti 2.2; a better way to monitor
2070
  locks is expected for a later 2.2.x release
2071
- Fixed dry-run behaviour with many commands
2072
- Support ``ssh-agent`` again when adding nodes
2073
- Many additional bugfixes
2074

    
2075

    
2076
Version 2.2.0 rc0
2077
-----------------
2078

    
2079
*(Released Fri, 30 Jul 2010)*
2080

    
2081
Important change: the internal RPC mechanism between Ganeti nodes has
2082
changed from using a home-grown http library (based on the Python base
2083
libraries) to use the PycURL library. This requires that PycURL is
2084
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2085
against GnuTLS by default. cURL's support for GnuTLS had known issues
2086
before cURL 7.21.0 and we recommend using the latest cURL release or
2087
linking against OpenSSL. Most other distributions already link PycURL
2088
and cURL against OpenSSL. The command::
2089

    
2090
  python -c 'import pycurl; print pycurl.version'
2091

    
2092
can be used to determine the libraries PycURL and cURL are linked
2093
against.
2094

    
2095
Other significant changes:
2096

    
2097
- Rewrote much of the internals of the job queue, in order to achieve
2098
  better parallelism; this decouples job query operations from the job
2099
  processing, and it should allow much nicer behaviour of the master
2100
  daemon under load, and it also has uncovered some long-standing bugs
2101
  related to the job serialisation (now fixed)
2102
- Added a default iallocator setting to the cluster parameters,
2103
  eliminating the need to always pass nodes or an iallocator for
2104
  operations that require selection of new node(s)
2105
- Added experimental support for the LXC virtualization method
2106
- Added support for OS parameters, which allows the installation of
2107
  instances to pass parameter to OS scripts in order to customise the
2108
  instance
2109
- Added a hypervisor parameter controlling the migration type (live or
2110
  non-live), since hypervisors have various levels of reliability; this
2111
  has renamed the 'live' parameter to 'mode'
2112
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
2113
  logical volumes, meaning that cluster verify will ignore them and not
2114
  flag their presence as errors
2115
- The watcher will now reset the error count for failed instances after
2116
  8 hours, thus allowing self-healing if the problem that caused the
2117
  instances to be down/fail to start has cleared in the meantime
2118
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2119
  check for, and warn, if the drbd module parameter ``usermode_helper``
2120
  is not consistent with the cluster-wide setting; this is needed to
2121
  make diagnose easier of failed drbd creations
2122
- Started adding base IPv6 support, but this is not yet
2123
  enabled/available for use
2124
- Rename operations (cluster, instance) will now return the new name,
2125
  which is especially useful if a short name was passed in
2126
- Added support for instance migration in RAPI
2127
- Added a tool to pre-configure nodes for the SSH setup, before joining
2128
  them to the cluster; this will allow in the future a simplified model
2129
  for node joining (but not yet fully enabled in 2.2); this needs the
2130
  paramiko python library
2131
- Fixed handling of name-resolving errors
2132
- Fixed consistency of job results on the error path
2133
- Fixed master-failover race condition when executed multiple times in
2134
  sequence
2135
- Fixed many bugs related to the job queue (mostly introduced during the
2136
  2.2 development cycle, so not all are impacting 2.1)
2137
- Fixed instance migration with missing disk symlinks
2138
- Fixed handling of unknown jobs in ``gnt-job archive``
2139
- And many other small fixes/improvements
2140

    
2141
Internal changes:
2142

    
2143
- Enhanced both the unittest and the QA coverage
2144
- Switched the opcode validation to a generic model, and extended the
2145
  validation to all opcode parameters
2146
- Changed more parts of the code that write shell scripts to use the
2147
  same class for this
2148
- Switched the master daemon to use the asyncore library for the Luxi
2149
  server endpoint
2150

    
2151

    
2152
Version 2.2.0 beta0
2153
-------------------
2154

    
2155
*(Released Thu, 17 Jun 2010)*
2156

    
2157
- Added tool (``move-instance``) and infrastructure to move instances
2158
  between separate clusters (see :doc:`separate documentation
2159
  <move-instance>` and :doc:`design document <design-2.2>`)
2160
- Added per-request RPC timeout
2161
- RAPI now requires a Content-Type header for requests with a body (e.g.
2162
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2163
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
2164
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2165
  reachable
2166
- Implemented initial support for running Ganeti daemons as separate
2167
  users, see configure-time flags ``--with-user-prefix`` and
2168
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2169
  time)
2170
- Instances can be removed after export (``gnt-backup export
2171
  --remove-instance``)
2172
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2173
  key (instead of 1024 bit)
2174
- Added new cluster configuration file for cluster domain secret
2175
- Import/export now use SSL instead of SSH
2176
- Added support for showing estimated time when exporting an instance,
2177
  see the ``ganeti-os-interface(7)`` manpage and look for
2178
  ``EXP_SIZE_FD``
2179

    
2180

    
2181
Version 2.1.8
2182
-------------
2183

    
2184
*(Released Tue, 16 Nov 2010)*
2185

    
2186
Some more bugfixes. Unless critical bugs occur, this will be the last
2187
2.1 release:
2188

    
2189
- Fix case of MAC special-values
2190
- Fix mac checker regex
2191
- backend: Fix typo causing "out of range" error
2192
- Add missing --units in gnt-instance list man page
2193

    
2194

    
2195
Version 2.1.7
2196
-------------
2197

    
2198
*(Released Tue, 24 Aug 2010)*
2199

    
2200
Bugfixes only:
2201
  - Don't ignore secondary node silently on non-mirrored disk templates
2202
    (issue 113)
2203
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2204
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2205
  - Properly document the usb_mouse parameter
2206
  - Fix path in ganeti-rapi(8) (issue 116)
2207
  - Adjust error message when the ganeti user's .ssh directory is
2208
    missing
2209
  - Add same-node-check when changing the disk template to drbd
2210

    
2211

    
2212
Version 2.1.6
2213
-------------
2214

    
2215
*(Released Fri, 16 Jul 2010)*
2216

    
2217
Bugfixes only:
2218
  - Add an option to only select some reboot types during qa/burnin.
2219
    (on some hypervisors consequent reboots are not supported)
2220
  - Fix infrequent race condition in master failover. Sometimes the old
2221
    master ip address would be still detected as up for a short time
2222
    after it was removed, causing failover to fail.
2223
  - Decrease mlockall warnings when the ctypes module is missing. On
2224
    Python 2.4 we support running even if no ctypes module is installed,
2225
    but we were too verbose about this issue.
2226
  - Fix building on old distributions, on which man doesn't have a
2227
    --warnings option.
2228
  - Fix RAPI not to ignore the MAC address on instance creation
2229
  - Implement the old instance creation format in the RAPI client.
2230

    
2231

    
2232
Version 2.1.5
2233
-------------
2234

    
2235
*(Released Thu, 01 Jul 2010)*
2236

    
2237
A small bugfix release:
2238
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2239
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2240
    a non-existing option
2241
  - Fix instance create: the --force-variant option was ignored
2242
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2243
  - Fix modify node storage with non-FQDN arguments
2244
  - Fix RAPI client to authenticate under Python 2.6 when used
2245
    for more than 5 requests needing authentication
2246
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2247
    when converting a non-shutdown drbd instance to plain
2248

    
2249

    
2250
Version 2.1.4
2251
-------------
2252

    
2253
*(Released Fri, 18 Jun 2010)*
2254

    
2255
A small bugfix release:
2256

    
2257
  - Fix live migration of KVM instances started with older Ganeti
2258
    versions which had fewer hypervisor parameters
2259
  - Fix gnt-instance grow-disk on down instances
2260
  - Fix an error-reporting bug during instance migration
2261
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2262
    silently ignoring broken ones
2263
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2264
    users
2265
  - Fix bug triggered by different API version os-es on different nodes
2266
  - Fix a bug in instance startup with custom hvparams: OS level
2267
    parameters would fail to be applied.
2268
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2269
    make it work completely well with OpenSSL)
2270
  - Fix handling of errors when resolving names from DNS
2271

    
2272

    
2273
Version 2.1.3
2274
-------------
2275

    
2276
*(Released Thu, 3 Jun 2010)*
2277

    
2278
A medium sized development cycle. Some new features, and some
2279
fixes/small improvements/cleanups.
2280

    
2281
Significant features
2282
~~~~~~~~~~~~~~~~~~~~
2283

    
2284
The node deamon now tries to mlock itself into memory, unless the
2285
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2286
its logs, and falls back to console logging. This allows emergency
2287
features such as ``gnt-node powercycle`` to work even in the event of a
2288
broken node disk (tested offlining the disk hosting the node's
2289
filesystem and dropping its memory caches; don't try this at home)
2290

    
2291
KVM: add vhost-net acceleration support. It can be tested with a new
2292
enough version of the kernel and of qemu-kvm.
2293

    
2294
KVM: Add instance chrooting feature. If you use privilege dropping for
2295
your VMs you can also now force them to chroot to an empty directory,
2296
before starting the emulated guest.
2297

    
2298
KVM: Add maximum migration bandwith and maximum downtime tweaking
2299
support (requires a new-enough version of qemu-kvm).
2300

    
2301
Cluster verify will now warn if the master node doesn't have the master
2302
ip configured on it.
2303

    
2304
Add a new (incompatible) instance creation request format to RAPI which
2305
supports all parameters (previously only a subset was supported, and it
2306
wasn't possible to extend the old format to accomodate all the new
2307
features. The old format is still supported, and a client can check for
2308
this feature, before using it, by checking for its presence in the
2309
``features`` RAPI resource.
2310

    
2311
Now with ancient latin support. Try it passing the ``--roman`` option to
2312
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2313
(requires the python-roman module to be installed, in order to work).
2314

    
2315
Other changes
2316
~~~~~~~~~~~~~
2317

    
2318
As usual many internal code refactorings, documentation updates, and
2319
such. Among others:
2320

    
2321
  - Lots of improvements and cleanups to the experimental Remote API
2322
    (RAPI) client library.
2323
  - A new unit test suite for the core daemon libraries.
2324
  - A fix to creating missing directories makes sure the umask is not
2325
    applied anymore. This enforces the same directory permissions
2326
    everywhere.
2327
  - Better handling terminating daemons with ctrl+c (used when running
2328
    them in debugging mode).
2329
  - Fix a race condition in live migrating a KVM instance, when stat()
2330
    on the old proc status file returned EINVAL, which is an unexpected
2331
    value.
2332
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2333
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2334

    
2335

    
2336
Version 2.1.2.1
2337
---------------
2338

    
2339
*(Released Fri, 7 May 2010)*
2340

    
2341
Fix a bug which prevented untagged KVM instances from starting.
2342

    
2343

    
2344
Version 2.1.2
2345
-------------
2346

    
2347
*(Released Fri, 7 May 2010)*
2348

    
2349
Another release with a long development cycle, during which many
2350
different features were added.
2351

    
2352
Significant features
2353
~~~~~~~~~~~~~~~~~~~~
2354

    
2355
The KVM hypervisor now can run the individual instances as non-root, to
2356
reduce the impact of a VM being hijacked due to bugs in the
2357
hypervisor. It is possible to run all instances as a single (non-root)
2358
user, to manually specify a user for each instance, or to dynamically
2359
allocate a user out of a cluster-wide pool to each instance, with the
2360
guarantee that no two instances will run under the same user ID on any
2361
given node.
2362

    
2363
An experimental RAPI client library, that can be used standalone
2364
(without the other Ganeti libraries), is provided in the source tree as
2365
``lib/rapi/client.py``. Note this client might change its interface in
2366
the future, as we iterate on its capabilities.
2367

    
2368
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2369
replace the cluster's certificates and crypto keys. This might help in
2370
case they have been compromised, or have simply expired.
2371

    
2372
A new disk option for instance creation has been added that allows one
2373
to "adopt" currently existing logical volumes, with data
2374
preservation. This should allow easier migration to Ganeti from
2375
unmanaged (or managed via other software) instances.
2376

    
2377
Another disk improvement is the possibility to convert between redundant
2378
(DRBD) and plain (LVM) disk configuration for an instance. This should
2379
allow better scalability (starting with one node and growing the
2380
cluster, or shrinking a two-node cluster to one node).
2381

    
2382
A new feature that could help with automated node failovers has been
2383
implemented: if a node sees itself as offline (by querying the master
2384
candidates), it will try to shutdown (hard) all instances and any active
2385
DRBD devices. This reduces the risk of duplicate instances if an
2386
external script automatically failovers the instances on such nodes. To
2387
enable this, the cluster parameter ``maintain_node_health`` should be
2388
enabled; in the future this option (per the name) will enable other
2389
automatic maintenance features.
2390

    
2391
Instance export/import now will reuse the original instance
2392
specifications for all parameters; that means exporting an instance,
2393
deleting it and the importing it back should give an almost identical
2394
instance. Note that the default import behaviour has changed from
2395
before, where it created only one NIC; now it recreates the original
2396
number of NICs.
2397

    
2398
Cluster verify has added a few new checks: SSL certificates validity,
2399
/etc/hosts consistency across the cluster, etc.
2400

    
2401
Other changes
2402
~~~~~~~~~~~~~
2403

    
2404
As usual, many internal changes were done, documentation fixes,
2405
etc. Among others:
2406

    
2407
- Fixed cluster initialization with disabled cluster storage (regression
2408
  introduced in 2.1.1)
2409
- File-based storage supports growing the disks
2410
- Fixed behaviour of node role changes
2411
- Fixed cluster verify for some corner cases, plus a general rewrite of
2412
  cluster verify to allow future extension with more checks
2413
- Fixed log spamming by watcher and node daemon (regression introduced
2414
  in 2.1.1)
2415
- Fixed possible validation issues when changing the list of enabled
2416
  hypervisors
2417
- Fixed cleanup of /etc/hosts during node removal
2418
- Fixed RAPI response for invalid methods
2419
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2420
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2421
  from ide disks, etc.)
2422
- Allow OS changes without re-installation (to record a changed OS
2423
  outside of Ganeti, or to allow OS renames)
2424
- Allow instance creation without OS installation (useful for example if
2425
  the OS will be installed manually, or restored from a backup not in
2426
  Ganeti format)
2427
- Implemented option to make cluster ``copyfile`` use the replication
2428
  network
2429
- Added list of enabled hypervisors to ssconf (possibly useful for
2430
  external scripts)
2431
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2432
  1.2 clusters
2433
- A partial form of node re-IP is possible via node readd, which now
2434
  allows changed node primary IP
2435
- Command line utilities now show an informational message if the job is
2436
  waiting for a lock
2437
- The logs of the master daemon now show the PID/UID/GID of the
2438
  connected client
2439

    
2440

    
2441
Version 2.1.1
2442
-------------
2443

    
2444
*(Released Fri, 12 Mar 2010)*
2445

    
2446
During the 2.1.0 long release candidate cycle, a lot of improvements and
2447
changes have accumulated with were released later as 2.1.1.
2448

    
2449
Major changes
2450
~~~~~~~~~~~~~
2451

    
2452
The node evacuate command (``gnt-node evacuate``) was significantly
2453
rewritten, and as such the IAllocator protocol was changed - a new
2454
request type has been added. This unfortunate change during a stable
2455
series is designed to improve performance of node evacuations; on
2456
clusters with more than about five nodes and which are well-balanced,
2457
evacuation should proceed in parallel for all instances of the node
2458
being evacuated. As such, any existing IAllocator scripts need to be
2459
updated, otherwise the above command will fail due to the unknown
2460
request. The provided "dumb" allocator has not been updated; but the
2461
ganeti-htools package supports the new protocol since version 0.2.4.
2462

    
2463
Another important change is increased validation of node and instance
2464
names. This might create problems in special cases, if invalid host
2465
names are being used.
2466

    
2467
Also, a new layer of hypervisor parameters has been added, that sits at
2468
OS level between the cluster defaults and the instance ones. This allows
2469
customisation of virtualization parameters depending on the installed
2470
OS. For example instances with OS 'X' may have a different KVM kernel
2471
(or any other parameter) than the cluster defaults. This is intended to
2472
help managing a multiple OSes on the same cluster, without manual
2473
modification of each instance's parameters.
2474

    
2475
A tool for merging clusters, ``cluster-merge``, has been added in the
2476
tools sub-directory.
2477

    
2478
Bug fixes
2479
~~~~~~~~~
2480

    
2481
- Improved the int/float conversions that should make the code more
2482
  robust in face of errors from the node daemons
2483
- Fixed the remove node code in case of internal configuration errors
2484
- Fixed the node daemon behaviour in face of inconsistent queue
2485
  directory (e.g. read-only file-system where we can't open the files
2486
  read-write, etc.)
2487
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2488
  now it either aborts cleanly or, if given the new "auto_promote"
2489
  parameter, will automatically promote other nodes as needed
2490
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2491
  completely prevent Ganeti from working
2492
- Fixed bug for instance export when not all disks were successfully
2493
  exported
2494
- Fixed behaviour of node add when the new node is slow in starting up
2495
  the node daemon
2496
- Fixed handling of signals in the LUXI client, which should improve
2497
  behaviour of command-line scripts
2498
- Added checks for invalid node/instance names in the configuration (now
2499
  flagged during cluster verify)
2500
- Fixed watcher behaviour for disk activation errors
2501
- Fixed two potentially endless loops in http library, which led to the
2502
  RAPI daemon hanging and consuming 100% CPU in some cases
2503
- Fixed bug in RAPI daemon related to hashed passwords
2504
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2505
  instances
2506
- Enhanced compatibility with non-Debian OSes, but not using absolute
2507
  path in some commands and allowing customisation of the ssh
2508
  configuration directory
2509
- Fixed possible future issue with new Python versions by abiding to the
2510
  proper use of ``__slots__`` attribute on classes
2511
- Added checks that should prevent directory traversal attacks
2512
- Many documentation fixes based on feedback from users
2513

    
2514
New features
2515
~~~~~~~~~~~~
2516

    
2517
- Added an "early_release" more for instance replace disks and node
2518
  evacuate, where we release locks earlier and thus allow higher
2519
  parallelism within the cluster
2520
- Added watcher hooks, intended to allow the watcher to restart other
2521
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2522
  course for any other purpose
2523
- Added a compile-time disable for DRBD barriers, to increase
2524
  performance if the administrator trusts the power supply or the
2525
  storage system to not lose writes
2526
- Added the option of using syslog for logging instead of, or in
2527
  addition to, Ganeti's own log files
2528
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2529
  can indeed boot from a paravirtual NIC
2530
- Added a generic debug level for many operations; while this is not
2531
  used widely yet, it allows one to pass the debug value all the way to
2532
  the OS scripts
2533
- Enhanced the hooks environment for instance moves (failovers,
2534
  migrations) where the primary/secondary nodes changed during the
2535
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2536
- Enhanced data validations for many user-supplied values; one important
2537
  item is the restrictions imposed on instance and node names, which
2538
  might reject some (invalid) host names
2539
- Add a configure-time option to disable file-based storage, if it's not
2540
  needed; this allows greater security separation between the master
2541
  node and the other nodes from the point of view of the inter-node RPC
2542
  protocol
2543
- Added user notification in interactive tools if job is waiting in the
2544
  job queue or trying to acquire locks
2545
- Added log messages when a job is waiting for locks
2546
- Added filtering by node tags in instance operations which admit
2547
  multiple instances (start, stop, reboot, reinstall)
2548
- Added a new tool for cluster mergers, ``cluster-merge``
2549
- Parameters from command line which are of the form ``a=b,c=d`` can now
2550
  use backslash escapes to pass in values which contain commas,
2551
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2552
  ``b,c``
2553
- For KVM, the instance name is the first parameter passed to KVM, so
2554
  that it's more visible in the process list
2555

    
2556

    
2557
Version 2.1.0
2558
-------------
2559

    
2560
*(Released Tue, 2 Mar 2010)*
2561

    
2562
Ganeti 2.1 brings many improvements with it. Major changes:
2563

    
2564
- Added infrastructure to ease automated disk repairs
2565
- Added new daemon to export configuration data in a cheaper way than
2566
  using the remote API
2567
- Instance NICs can now be routed instead of being associated with a
2568
  networking bridge
2569
- Improved job locking logic to reduce impact of jobs acquiring multiple
2570
  locks waiting for other long-running jobs
2571

    
2572
In-depth implementation details can be found in the Ganeti 2.1 design
2573
document.
2574

    
2575
Details
2576
~~~~~~~
2577

    
2578
- Added chroot hypervisor
2579
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2580
  ``device_model``)
2581
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2582
  ``bootloader_path`` and ``bootloader_args``)
2583
- Added the ``use_localtime`` option for the xen-hvm and kvm
2584
  hypervisors, and the default value for this has changed to false (in
2585
  2.0 xen-hvm always enabled it)
2586
- Added luxi call to submit multiple jobs in one go
2587
- Added cluster initialization option to not modify ``/etc/hosts``
2588
  file on nodes
2589
- Added network interface parameters
2590
- Added dry run mode to some LUs
2591
- Added RAPI resources:
2592

    
2593
  - ``/2/instances/[instance_name]/info``
2594
  - ``/2/instances/[instance_name]/replace-disks``
2595
  - ``/2/nodes/[node_name]/evacuate``
2596
  - ``/2/nodes/[node_name]/migrate``
2597
  - ``/2/nodes/[node_name]/role``
2598
  - ``/2/nodes/[node_name]/storage``
2599
  - ``/2/nodes/[node_name]/storage/modify``
2600
  - ``/2/nodes/[node_name]/storage/repair``
2601

    
2602
- Added OpCodes to evacuate or migrate all instances on a node
2603
- Added new command to list storage elements on nodes (``gnt-node
2604
  list-storage``) and modify them (``gnt-node modify-storage``)
2605
- Added new ssconf files with master candidate IP address
2606
  (``ssconf_master_candidates_ips``), node primary IP address
2607
  (``ssconf_node_primary_ips``) and node secondary IP address
2608
  (``ssconf_node_secondary_ips``)
2609
- Added ``ganeti-confd`` and a client library to query the Ganeti
2610
  configuration via UDP
2611
- Added ability to run hooks after cluster initialization and before
2612
  cluster destruction
2613
- Added automatic mode for disk replace (``gnt-instance replace-disks
2614
  --auto``)
2615
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2616
  after catastrophic data-loss
2617
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2618
  groups
2619
- Added ``gnt-instance move`` command to move instances
2620
- Added ``gnt-cluster watcher`` command to control watcher
2621
- Added ``gnt-node powercycle`` command to powercycle nodes
2622
- Added new job status field ``lock_status``
2623
- Added parseable error codes to cluster verification (``gnt-cluster
2624
  verify --error-codes``) and made output less verbose (use
2625
  ``--verbose`` to restore previous behaviour)
2626
- Added UUIDs to the main config entities (cluster, nodes, instances)
2627
- Added support for OS variants
2628
- Added support for hashed passwords in the Ganeti remote API users file
2629
  (``rapi_users``)
2630
- Added option to specify maximum timeout on instance shutdown
2631
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2632
- Added new helper script to start and stop Ganeti daemons
2633
  (``daemon-util``), with the intent to reduce the work necessary to
2634
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2635
  from one place
2636
- Added more unittests
2637
- Fixed critical bug in ganeti-masterd startup
2638
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2639
  now customisable at the cluster level for both the KVM and Xen
2640
  hypervisors using the new ``migration_port`` parameter
2641
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2642
  reinstalling an instance
2643
- Allowed ``@`` in tag names
2644
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2645
- Many documentation updates
2646
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2647
- ``gnt-instance reinstall`` can now reinstall multiple instances
2648
- Updated many command line parameters
2649
- Introduced new OS API version 15
2650
- No longer support a default hypervisor
2651
- Treat virtual LVs as inexistent
2652
- Improved job locking logic to reduce lock contention
2653
- Match instance and node names case insensitively
2654
- Reimplemented bash completion script to be more complete
2655
- Improved burnin
2656

    
2657

    
2658
Version 2.0.6
2659
-------------
2660

    
2661
*(Released Thu, 4 Feb 2010)*
2662

    
2663
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2664
- Fix a string formatting bug
2665
- Improve safety of the code in some error paths
2666
- Improve data validation in the master of values returned from nodes
2667

    
2668

    
2669
Version 2.0.5
2670
-------------
2671

    
2672
*(Released Thu, 17 Dec 2009)*
2673

    
2674
- Fix security issue due to missing validation of iallocator names; this
2675
  allows local and remote execution of arbitrary executables
2676
- Fix failure of gnt-node list during instance removal
2677
- Ship the RAPI documentation in the archive
2678

    
2679

    
2680
Version 2.0.4
2681
-------------
2682

    
2683
*(Released Wed, 30 Sep 2009)*
2684

    
2685
- Fixed many wrong messages
2686
- Fixed a few bugs related to the locking library
2687
- Fixed MAC checking at instance creation time
2688
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2689
- Fixed a few issues related to signal handling in both daemons and
2690
  scripts
2691
- Fixed the example startup script provided
2692
- Fixed insserv dependencies in the example startup script (patch from
2693
  Debian)
2694
- Fixed handling of drained nodes in the iallocator framework
2695
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2696
  #528618)
2697
- Fixed error related to invalid job IDs in job polling
2698
- Fixed job/opcode persistence on unclean master shutdown
2699
- Fixed handling of partial job processing after unclean master
2700
  shutdown
2701
- Fixed error reporting from LUs, previously all errors were converted
2702
  into execution errors
2703
- Fixed error reporting from burnin
2704
- Decreased significantly the memory usage of the job queue
2705
- Optimised slightly multi-job submission
2706
- Optimised slightly opcode loading
2707
- Backported the multi-job submit framework from the development
2708
  branch; multi-instance start and stop should be faster
2709
- Added script to clean archived jobs after 21 days; this will reduce
2710
  the size of the queue directory
2711
- Added some extra checks in disk size tracking
2712
- Added an example ethers hook script
2713
- Added a cluster parameter that prevents Ganeti from modifying of
2714
  /etc/hosts
2715
- Added more node information to RAPI responses
2716
- Added a ``gnt-job watch`` command that allows following the ouput of a
2717
  job
2718
- Added a bind-address option to ganeti-rapi
2719
- Added more checks to the configuration verify
2720
- Enhanced the burnin script such that some operations can be retried
2721
  automatically
2722
- Converted instance reinstall to multi-instance model
2723

    
2724

    
2725
Version 2.0.3
2726
-------------
2727

    
2728
*(Released Fri, 7 Aug 2009)*
2729

    
2730
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2731
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2732
  instances have mismatched disk sizes in the configuration
2733
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2734
  any configuration mismatches for disk sizes
2735
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2736
  failover to work on two-node clusters
2737
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2738
  unusable
2739
- Fixed detection of OS script errors in ``gnt-backup export``
2740
- Fixed exit code of ``gnt-backup export``
2741

    
2742

    
2743
Version 2.0.2
2744
-------------
2745

    
2746
*(Released Fri, 17 Jul 2009)*
2747

    
2748
- Added experimental support for stripped logical volumes; this should
2749
  enhance performance but comes with a higher complexity in the block
2750
  device handling; stripping is only enabled when passing
2751
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2752
  affected even in the non-stripped mode
2753
- Improved resiliency against transient failures at the end of DRBD
2754
  resyncs, and in general of DRBD resync checks
2755
- Fixed a couple of issues with exports and snapshot errors
2756
- Fixed a couple of issues in instance listing
2757
- Added display of the disk size in ``gnt-instance info``
2758
- Fixed checking for valid OSes in instance creation
2759
- Fixed handling of the "vcpus" parameter in instance listing and in
2760
  general of invalid parameters
2761
- Fixed http server library, and thus RAPI, to handle invalid
2762
  username/password combinations correctly; this means that now they
2763
  report unauthorized for queries too, not only for modifications,
2764
  allowing earlier detect of configuration problems
2765
- Added a new "role" node list field, equivalent to the master/master
2766
  candidate/drained/offline flags combinations
2767
- Fixed cluster modify and changes of candidate pool size
2768
- Fixed cluster verify error messages for wrong files on regular nodes
2769
- Fixed a couple of issues with node demotion from master candidate role
2770
- Fixed node readd issues
2771
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2772
- Added a new ``--no-voting`` option for masterfailover to fix failover
2773
  on two-nodes clusters when the former master node is unreachable
2774
- Added instance reinstall over RAPI
2775

    
2776

    
2777
Version 2.0.1
2778
-------------
2779

    
2780
*(Released Tue, 16 Jun 2009)*
2781

    
2782
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2783
  allow re-adding the start in single-user option (regression from 1.2)
2784
- the watcher writes the instance status to a file, to allow monitoring
2785
  to report the instance status (from the master) based on cached
2786
  results of the watcher's queries; while this can get stale if the
2787
  watcher is being locked due to other work on the cluster, this is
2788
  still an improvement
2789
- the watcher now also restarts the node daemon and the rapi daemon if
2790
  they died
2791
- fixed the watcher to handle full and drained queue cases
2792
- hooks export more instance data in the environment, which helps if
2793
  hook scripts need to take action based on the instance's properties
2794
  (no longer need to query back into ganeti)
2795
- instance failovers when the instance is stopped do not check for free
2796
  RAM, so that failing over a stopped instance is possible in low memory
2797
  situations
2798
- rapi uses queries for tags instead of jobs (for less job traffic), and
2799
  for cluster tags it won't talk to masterd at all but read them from
2800
  ssconf
2801
- a couple of error handling fixes in RAPI
2802
- drbd handling: improved the error handling of inconsistent disks after
2803
  resync to reduce the frequency of "there are some degraded disks for
2804
  this instance" messages
2805
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2806
  error handling path called a wrong function name)
2807

    
2808

    
2809
Version 2.0.0
2810
-------------
2811

    
2812
*(Released Wed, 27 May 2009)*
2813

    
2814
- no changes from rc5
2815

    
2816

    
2817
Version 2.0 rc5
2818
---------------
2819

    
2820
*(Released Wed, 20 May 2009)*
2821

    
2822
- fix a couple of bugs (validation, argument checks)
2823
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2824
- some small improvements to RAPI and IAllocator
2825
- make watcher automatically start the master daemon if down
2826

    
2827

    
2828
Version 2.0 rc4
2829
---------------
2830

    
2831
*(Released Mon, 27 Apr 2009)*
2832

    
2833
- change the OS list to not require locks; this helps with big clusters
2834
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2835
  volume group is broken
2836
- ``gnt-instance info``, without any arguments, doesn't run for all
2837
  instances anymore; either pass ``--all`` or pass the desired
2838
  instances; this helps against mistakes on big clusters where listing
2839
  the information for all instances takes a long time
2840
- miscellaneous doc and man pages fixes
2841

    
2842

    
2843
Version 2.0 rc3
2844
---------------
2845

    
2846
*(Released Wed, 8 Apr 2009)*
2847

    
2848
- Change the internal locking model of some ``gnt-node`` commands, in
2849
  order to reduce contention (and blocking of master daemon) when
2850
  batching many creation/reinstall jobs
2851
- Fixes to Xen soft reboot
2852
- No longer build documentation at build time, instead distribute it in
2853
  the archive, in order to reduce the need for the whole docbook/rst
2854
  toolchains
2855

    
2856

    
2857
Version 2.0 rc2
2858
---------------
2859

    
2860
*(Released Fri, 27 Mar 2009)*
2861

    
2862
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2863
- Fix watcher startup sequence, improves the behaviour of busy clusters
2864
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2865
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2866
  bind address and other places
2867
- Some documentation fixes and updates
2868

    
2869

    
2870
Version 2.0 rc1
2871
---------------
2872

    
2873
*(Released Mon, 2 Mar 2009)*
2874

    
2875
- More documentation updates, now all docs should be more-or-less
2876
  up-to-date
2877
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2878
  etc.)
2879
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2880
  and KVM)
2881
- Fix an issue related to $libdir/run/ganeti and cluster creation
2882

    
2883

    
2884
Version 2.0 beta2
2885
-----------------
2886

    
2887
*(Released Thu, 19 Feb 2009)*
2888

    
2889
- Xen PVM and KVM have switched the default value for the instance root
2890
  disk to the first partition on the first drive, instead of the whole
2891
  drive; this means that the OS installation scripts must be changed
2892
  accordingly
2893
- Man pages have been updated
2894
- RAPI has been switched by default to HTTPS, and the exported functions
2895
  should all work correctly
2896
- RAPI v1 has been removed
2897
- Many improvements to the KVM hypervisor
2898
- Block device errors are now better reported
2899
- Many other bugfixes and small improvements
2900

    
2901

    
2902
Version 2.0 beta1
2903
-----------------
2904

    
2905
*(Released Mon, 26 Jan 2009)*
2906

    
2907
- Version 2 is a general rewrite of the code and therefore the
2908
  differences are too many to list, see the design document for 2.0 in
2909
  the ``doc/`` subdirectory for more details
2910
- In this beta version there is not yet a migration path from 1.2 (there
2911
  will be one in the final 2.0 release)
2912
- A few significant changes are:
2913

    
2914
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2915
    various ``gnt-*`` commands are just front-ends to it
2916
  - all the commands are entered into, and executed from a job queue,
2917
    see the ``gnt-job(8)`` manpage
2918
  - the RAPI daemon supports read-write operations, secured by basic
2919
    HTTP authentication on top of HTTPS
2920
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2921
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2922
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2923
  - DRBD devices are using statically allocated minor numbers, which
2924
    will be assigned to existing instances during the migration process
2925
  - there is support for both Xen PVM and Xen HVM instances running on
2926
    the same cluster
2927
  - KVM virtualization is supported too
2928
  - file-based storage has been implemented, which means that it is
2929
    possible to run the cluster without LVM and DRBD storage, for
2930
    example using a shared filesystem exported from shared storage (and
2931
    still have live migration)
2932

    
2933

    
2934
Version 1.2.7
2935
-------------
2936

    
2937
*(Released Tue, 13 Jan 2009)*
2938

    
2939
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2940
- Reuse the old instance mac address by default on instance import, if
2941
  the instance name is the same.
2942
- Handle situations in which the node info rpc returns incomplete
2943
  results (issue 46)
2944
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2945
- Improved version of batcher:
2946

    
2947
  - state file support
2948
  - instance mac address support
2949
  - support for HVM clusters/instances
2950

    
2951
- Add an option to show the number of cpu sockets and nodes in
2952
  ``gnt-node list``
2953
- Support OSes that handle more than one version of the OS api (but do
2954
  not change the current API in any other way)
2955
- Fix ``gnt-node migrate``
2956
- ``gnt-debug`` man page
2957
- Fixes various more typos and small issues
2958
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2959

    
2960

    
2961
Version 1.2.6
2962
-------------
2963

    
2964
*(Released Wed, 24 Sep 2008)*
2965

    
2966
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2967
  type of disk exported to fully virtualized instances.
2968
- provide access to the serial console of HVM instances
2969
- instance auto_balance flag, set by default. If turned off it will
2970
  avoid warnings on cluster verify if there is not enough memory to fail
2971
  over an instance. in the future it will prevent automatically failing
2972
  it over when we will support that.
2973
- batcher tool for instance creation, see ``tools/README.batcher``
2974
- ``gnt-instance reinstall --select-os`` to interactively select a new
2975
  operating system when reinstalling an instance.
2976
- when changing the memory amount on instance modify a check has been
2977
  added that the instance will be able to start. also warnings are
2978
  emitted if the instance will not be able to fail over, if auto_balance
2979
  is true.
2980
- documentation fixes
2981
- sync fields between ``gnt-instance list/modify/add/import``
2982
- fix a race condition in drbd when the sync speed was set after giving
2983
  the device a remote peer.
2984

    
2985

    
2986
Version 1.2.5
2987
-------------
2988

    
2989
*(Released Tue, 22 Jul 2008)*
2990

    
2991
- note: the allowed size and number of tags per object were reduced
2992
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2993
- fixed twisted 8.x compatibility
2994
- fixed ``gnt-instance replace-disks`` with iallocator
2995
- add TCP keepalives on twisted connections to detect restarted nodes
2996
- disk increase support, see ``gnt-instance grow-disk``
2997
- implement bulk node/instance query for RAPI
2998
- add tags in node/instance listing (optional)
2999
- experimental migration (and live migration) support, read the man page
3000
  for ``gnt-instance migrate``
3001
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
3002
  has some small improvements in handling its state file
3003

    
3004

    
3005
Version 1.2.4
3006
-------------
3007

    
3008
*(Released Fri, 13 Jun 2008)*
3009

    
3010
- Experimental readonly, REST-based remote API implementation;
3011
  automatically started on master node, TCP port 5080, if enabled by
3012
  ``--enable-rapi`` parameter to configure script.
3013
- Instance allocator support. Add and import instance accept a
3014
  ``--iallocator`` parameter, and call that instance allocator to decide
3015
  which node to use for the instance. The iallocator document describes
3016
  what's expected from an allocator script.
3017
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
3018
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
3019
  node is lost there is still enough memory to fail over the instances
3020
  that reside on it.
3021
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
3022
  ``gnt-cluster verify``, to check for site-specific compliance. All the
3023
  hooks will run, and their output, if any, will be displayed. Any
3024
  failing hook will make the verification return an error value.
3025
- ``gnt-cluster verify`` now checks that its peers are reachable on the
3026
  primary and secondary interfaces
3027
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
3028
  that is still declared as part of the cluster and has failed.
3029
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
3030
  specifying output fields, that just adds the chosen fields to the
3031
  default ones.
3032
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
3033
  export from the filesystem.
3034
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
3035
  have been added. Using them you can enable/disable acpi and pae
3036
  support, and specify a path for a cd image to be exported to the
3037
  instance. These parameters as the name suggest only work on HVM
3038
  clusters.
3039
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
3040
  PAE support will be set to the previously hardcoded values, but the
3041
  (previously hardcoded) path to the CDROM ISO image will be unset and
3042
  if required, needs to be set manually with ``gnt-instance modify``
3043
  after the upgrade.
3044
- The address to which an instance's VNC console is bound is now
3045
  selectable per-instance, rather than being cluster wide. Of course
3046
  this only applies to instances controlled via VNC, so currently just
3047
  applies to HVM clusters.
3048

    
3049

    
3050
Version 1.2.3
3051
-------------
3052

    
3053
*(Released Mon, 18 Feb 2008)*
3054

    
3055
- more tweaks to the disk activation code (especially helpful for DRBD)
3056
- change the default ``gnt-instance list`` output format, now there is
3057
  one combined status field (see the manpage for the exact values this
3058
  field will have)
3059
- some more fixes for the mac export to hooks change
3060
- make Ganeti not break with DRBD 8.2.x (which changed the version
3061
  format in ``/proc/drbd``) (issue 24)
3062
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3063
  template, allowing migration from DRBD0.7+MD to DRBD8
3064

    
3065

    
3066
Version 1.2.2
3067
-------------
3068

    
3069
*(Released Wed, 30 Jan 2008)*
3070

    
3071
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3072
  support (issue 23)
3073
- add command aliases infrastructure and a few aliases
3074
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
3075
  man pages and the ``--help`` option of ``gnt-node
3076
  list``/``gnt-instance list``
3077
- fix ``gnt-backup list`` with down nodes (issue 21)
3078
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3079
- fix the dist archive and add a check for including svn/git files in
3080
  the future
3081
- some developer-related changes: improve the burnin and the QA suite,
3082
  add an upload script for testing during development
3083

    
3084

    
3085
Version 1.2.1
3086
-------------
3087

    
3088
*(Released Wed, 16 Jan 2008)*
3089

    
3090
- experimental HVM support, read the install document, section
3091
  "Initializing the cluster"
3092
- allow for the PVM hypervisor per-instance kernel and initrd paths
3093
- add a new command ``gnt-cluster verify-disks`` which uses a new
3094
  algorithm to improve the reconnection of the DRBD pairs if the device
3095
  on the secondary node has gone away
3096
- make logical volume code auto-activate LVs at disk activation time
3097
- slightly improve the speed of activating disks
3098
- allow specification of the MAC address at instance creation time, and
3099
  changing it later via ``gnt-instance modify``
3100
- fix handling of external commands that generate lots of output on
3101
  stderr
3102
- update documentation with regard to minimum version of DRBD8 supported
3103

    
3104

    
3105
Version 1.2.0
3106
-------------
3107

    
3108
*(Released Tue, 4 Dec 2007)*
3109

    
3110
- Log the ``xm create`` output to the node daemon log on failure (to
3111
  help diagnosing the error)
3112
- In debug mode, log all external commands output if failed to the logs
3113
- Change parsing of lvm commands to ignore stderr
3114

    
3115

    
3116
Version 1.2 beta3
3117
-----------------
3118

    
3119
*(Released Wed, 28 Nov 2007)*
3120

    
3121
- Another round of updates to the DRBD 8 code to deal with more failures
3122
  in the replace secondary node operation
3123
- Some more logging of failures in disk operations (lvm, drbd)
3124
- A few documentation updates
3125
- QA updates
3126

    
3127

    
3128
Version 1.2 beta2
3129
-----------------
3130

    
3131
*(Released Tue, 13 Nov 2007)*
3132

    
3133
- Change configuration file format from Python's Pickle to JSON.
3134
  Upgrading is possible using the cfgupgrade utility.
3135
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3136
  faster replace disks and is more stable (DRBD 8 has many improvements
3137
  compared to DRBD 0.7)
3138
- Added command line tags support (see man pages for ``gnt-instance``,
3139
  ``gnt-node``, ``gnt-cluster``)
3140
- Added instance rename support
3141
- Added multi-instance startup/shutdown
3142
- Added cluster rename support
3143
- Added ``gnt-node evacuate`` to simplify some node operations
3144
- Added instance reboot operation that can speedup reboot as compared to
3145
  stop and start
3146
- Soften the requirement that hostnames are in FQDN format
3147
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
3148
  reboots
3149
- Removed dependency on debian's patched fping that uses the
3150
  non-standard ``-S`` option
3151
- Now the OS definitions are searched for in multiple, configurable
3152
  paths (easier for distros to package)
3153
- Some changes to the hooks infrastructure (especially the new
3154
  post-configuration update hook)
3155
- Other small bugfixes
3156

    
3157
.. vim: set textwidth=72 syntax=rst :
3158
.. Local Variables:
3159
.. mode: rst
3160
.. fill-column: 72
3161
.. End: