Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 7d993147

History | View | Annotate | Download (117 kB)

1
News
2
====
3

    
4

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

    
8
*(unreleased)*
9

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

    
13
New features
14
~~~~~~~~~~~~
15

    
16
- Instance moves, backups and imports can now use compression to transfer the
17
  instance data.
18

    
19

    
20
Version 2.10.0 alpha1
21
---------------------
22

    
23
*(unreleased)*
24

    
25
Incompatible/important changes
26
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
27

    
28
- Adding disks with 'gnt-instance modify' now waits for the disks to sync per
29
  default. Specify --no-wait-for-sync to override this behavior.
30
- The Ganeti python code now adheres to a private-module layout. In particular,
31
  the module 'ganeti' is no longer in the python search path.
32
- On instance allocation, the iallocator now considers non-LVM storage
33
  properly. In particular, actual file storage space information is used
34
  when allocating space for a file/sharedfile instance.
35
- When disabling disk templates cluster-wide, the cluster now first
36
  checks whether there are instances still using those templates.
37
- 'gnt-node list-storage' now also reports storage information about
38
  file-based storage types.
39

    
40
New features
41
~~~~~~~~~~~~
42

    
43
- KVM hypervisors can now access RBD storage directly without having to
44
  go through a block device.
45
- A new command 'gnt-cluster upgrade' was added that automates the upgrade
46
  procedure between two Ganeti versions that are both 2.10 or higher.
47
- The move-instance command can now change disk templates when moving
48
  instances, and does not require any node placement options to be
49
  specified if the destination cluster has a default iallocator.
50
- Users can now change the soundhw and cpuid settings for XEN hypervisors.
51
- Hail and hbal now have the (optional) capability of accessing average CPU
52
  load information through the monitoring deamon, and to use it to dynamically
53
  adapt the allocation of instances.
54

    
55
Misc changes
56
~~~~~~~~~~~~
57

    
58
- A new test framework for logical units was introduced and the test
59
  coverage for logical units was improved significantly.
60
- Opcodes are entirely generated from Haskell using the tool 'hs2py' and
61
  the module 'src/Ganeti/OpCodes.hs'.
62
- Constants are also generated from Haskell using the tool
63
  'hs2py-constants' and the module 'src/Ganeti/HsConstants.hs', with the
64
  exception of socket related constants, which require changing the
65
  cluster configuration file, and HVS related constants, because they
66
  are part of a port of instance queries to Haskell.  As a result, these
67
  changes will be part of the next release of Ganeti.
68

    
69
New dependencies
70
~~~~~~~~~~~~~~~~
71

    
72
The following new dependencies have been added/updated.
73

    
74
Python
75

    
76
- The version requirements for ``python-mock`` have increased to at least
77
  version 1.0.1. It is still used for testing only.
78

    
79

    
80
Version 2.9.0 rc4
81
-----------------
82

    
83
*(unreleased)*
84

    
85
Incompatible/important changes
86
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
87

    
88
- hroller now also plans for capacity to move non-redundant instances off
89
  any node to be rebooted; the old behavior of completely ignoring any
90
  non-redundant instances can be restored by adding the --ignore-non-redundant
91
  option.
92
- The cluster option '--no-lvm-storage' was removed in favor of the new option
93
  '--enabled-disk-templates'.
94
- On instance creation, disk templates no longer need to be specified
95
  with '-t'. The default disk template will be taken from the list of
96
  enabled disk templates.
97
- The monitoring daemon is now running as root, in order to be able to collect
98
  information only available to root (such as the state of Xen instances).
99
- The ConfD client is now IPv6 compatible.
100
- File and shared file storage is no longer dis/enabled at configure time,
101
  but using the option '--enabled-disk-templates' at cluster initialization and
102
  modification.
103
- The default directories for file and shared file storage are not anymore
104
  specified at configure time, but taken from the cluster's configuration.
105
  They can be set at cluster initialization and modification with
106
  '--file-storage-dir' and '--shared-file-storage-dir'.
107
- Cluster verification now includes stricter checks regarding the
108
  default file and shared file storage directories. It now checks that
109
  the directories are explicitely allowed in the 'file-storage-paths' file and
110
  that the directories exist on all nodes.
111
- The list of allowed disk templates in the instance policy and the list
112
  of cluster-wide enabled disk templates is now checked for consistency
113
  on cluster or group modification. On cluster initialization, the ipolicy
114
  disk templates are ensured to be a subset of the cluster-wide enabled
115
  disk templates.
116

    
117
New features
118
~~~~~~~~~~~~
119

    
120
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
121
  the correct command syntax. It is possible to use different DRBD versions
122
  on different nodes as long as they are compatible to each other. This
123
  enables rolling upgrades of DRBD with no downtime. As permanent operation
124
  of different DRBD versions within a node group is discouraged,
125
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
126
- New "inst-status-xen" data collector for the monitoring daemon, providing
127
  information about the state of the xen instances on the nodes.
128
- New "lv" data collector for the monitoring daemon, collecting data about the
129
  logical volumes on the nodes, and pairing them with the name of the instances
130
  they belong to.
131
- New "diskstats" data collector, collecting the data from /proc/diskstats and
132
  presenting them over the monitoring daemon interface.
133
- The ConfD client is now IPv6 compatible.
134

    
135
New dependencies
136
~~~~~~~~~~~~~~~~
137
The following new dependencies have been added.
138

    
139
Python
140

    
141
- ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
142
  for the unit tests (and only used for testing).
143

    
144
Haskell
145

    
146
- ``hslogger`` (http://software.complete.org/hslogger) is now always
147
  required, even if confd is not enabled.
148

    
149
Since 2.9.0 rc3
150
~~~~~~~~~~~~~~~
151

    
152
- Correctly start/stop luxid during gnt-cluster master-failover (inherited
153
  from stable-2.8)
154

    
155

    
156
Version 2.9.0 rc3
157
-----------------
158

    
159
*(Released Tue, 15 Oct 2013)*
160

    
161
The third release candidate in the 2.9 series. Since 2.9.0 rc2:
162

    
163
- in implicit configuration upgrade, match ipolicy with enabled disk templates
164
- improved harep documentation (inherited from stable-2.8)
165

    
166

    
167
Version 2.9.0 rc2
168
-----------------
169

    
170
*(Released Wed, 9 Oct 2013)*
171

    
172
The second release candidate in the 2.9 series. Since 2.9.0 rc1:
173

    
174
- Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
175
  at least one DRBD instance.
176
- Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
177
  downgrading.
178

    
179

    
180
Version 2.9.0 rc1
181
-----------------
182

    
183
*(Released Tue, 1 Oct 2013)*
184

    
185
The first release candidate in the 2.9 series. Since 2.9.0 beta1:
186

    
187
- various bug fixes
188
- update of the documentation, in particular installation instructions
189
- merging of LD_* constants into DT_* constants
190
- python style changes to be compatible with newer versions of pylint
191

    
192

    
193
Version 2.9.0 beta1
194
-------------------
195

    
196
*(Released Thu, 29 Aug 2013)*
197

    
198
This was the first beta release of the 2.9 series. All important changes
199
are listed in the latest 2.9 entry.
200

    
201

    
202
Version 2.8.1
203
-------------
204

    
205
*(Released Thu, 17 Oct 2013)*
206

    
207
- Correctly start/stop luxid during gnt-cluster master-failover
208
- Don't attempt IPv6 ssh in case of IPv4 cluster (Issue 595)
209
- Fix path for the job queue serial file
210
- Improved harep man page
211
- Minor documentation improvements
212

    
213

    
214
Version 2.8.0
215
-------------
216

    
217
*(Released Mon, 30 Sep 2013)*
218

    
219
Incompatible/important changes
220
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
221

    
222
- Instance policy can contain multiple instance specs, as described in
223
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
224
  <design-partitioned>`. As a consequence, it's not possible to partially change
225
  or override instance specs. Bounding specs (min and max) can be specified as a
226
  whole using the new option ``--ipolicy-bounds-specs``, while standard
227
  specs use the new option ``--ipolicy-std-specs``.
228
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
229
  gnt-instance is a valid YAML object.
230
- hail now honors network restrictions when allocating nodes. This led to an
231
  update of the IAllocator protocol. See the IAllocator documentation for
232
  details.
233
- confd now only answers static configuration request over the network. luxid
234
  was extracted, listens on the local LUXI socket and responds to live queries.
235
  This allows finer grained permissions if using separate users.
236

    
237
New features
238
~~~~~~~~~~~~
239

    
240
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
241
  to always require authentication, ``--require-authentication``. It can
242
  be specified in ``$sysconfdir/default/ganeti``.
243
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
244
  be used to manage the disk templates to be used by instances in the cluster.
245
  Initially, it will be set to a list that includes plain, drbd, if they were
246
  enabled by specifying a volume group name, and file and sharedfile, if those
247
  were enabled at configure time. Additionally, it will include all disk
248
  templates that are currently used by instances. The order of disk templates
249
  will be based on Ganeti's history of supporting them. In the future, the
250
  first entry of the list will be used as a default disk template on instance
251
  creation.
252
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
253
  configuration back to the previous stable version.
254
- Disk templates in group ipolicy can be restored to the default value.
255
- Initial support for diskless instances and virtual clusters in QA.
256
- More QA and unit tests for instance policies.
257
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
258
  describing why the opcode itself was executed.
259
- The monitoring daemon is now available. It allows users to query the cluster
260
  for obtaining information about the status of the system. The daemon is only
261
  responsible for providing the information over the network: the actual data
262
  gathering is performed by data collectors (currently, only the DRBD status
263
  collector is available).
264
- In order to help developers work on Ganeti, a new script
265
  (``devel/build_chroot``) is provided, for building a chroot that contains all
266
  the required development libraries and tools for compiling Ganeti on a Debian
267
  Squeeze system.
268
- A new tool, ``harep``, for performing self-repair and recreation of instances
269
  in Ganeti has been added.
270
- Split queries are enabled for tags, network, exports, cluster info, groups,
271
  jobs, nodes.
272
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
273
  It prints the command line to set the current policies, to ease
274
  changing them.
275
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
276
  devices for KVM virtio-net interfaces will get created with VNET_HDR
277
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
278
  virtio-net interfaces, which prevents host kernel tainting and log
279
  flooding, when dealing with broken or malicious virtio-net drivers.
280
  It's set to true by default.
281
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
282
  failures.
283
- Support 'viridian' parameter in Xen HVM
284
- Support DSA SSH keys in bootstrap
285
- To simplify the work of packaging frameworks that want to add the needed users
286
  and groups in a split-user setup themselves, at build time three files in
287
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
288
  the groups to be generated, the ``users`` file contains, one per line, the
289
  users to be generated, optionally followed by their primary group, where
290
  important. The ``groupmemberships`` file contains, one per line, additional
291
  user-group membership relations that need to be established. The syntax of
292
  these files will remain stable in all future versions.
293

    
294

    
295
New dependencies
296
~~~~~~~~~~~~~~~~
297
The following new dependencies have been added:
298

    
299
For Haskell:
300
- The ``curl`` library is not optional anymore for compiling the Haskell code.
301
- ``snap-server`` library (if monitoring is enabled).
302

    
303
For Python:
304
- The minimum Python version needed to run Ganeti is now 2.6.
305
- ``yaml`` library (only for running the QA).
306

    
307
Since 2.8.0 rc3
308
~~~~~~~~~~~~~~~
309
- Perform proper cleanup on termination of Haskell daemons
310
- Fix corner-case in handling of remaining retry time
311

    
312

    
313
Version 2.8.0 rc3
314
-----------------
315

    
316
*(Released Tue, 17 Sep 2013)*
317

    
318
- To simplify the work of packaging frameworks that want to add the needed users
319
  and groups in a split-user setup themselves, at build time three files in
320
  ``doc/users`` will be generated. The ``groups`` files contains, one per line,
321
  the groups to be generated, the ``users`` file contains, one per line, the
322
  users to be generated, optionally followed by their primary group, where
323
  important. The ``groupmemberships`` file contains, one per line, additional
324
  user-group membership relations that need to be established. The syntax of
325
  these files will remain stable in all future versions.
326
- Add a default to file-driver when unspecified over RAPI (Issue 571)
327
- Mark the DSA host pubkey as optional, and remove it during config downgrade
328
  (Issue 560)
329
- Some documentation fixes
330

    
331

    
332
Version 2.8.0 rc2
333
-----------------
334

    
335
*(Released Tue, 27 Aug 2013)*
336

    
337
The second release candidate of the 2.8 series. Since 2.8.0. rc1:
338

    
339
- Support 'viridian' parameter in Xen HVM (Issue 233)
340
- Include VCS version in ``gnt-cluster version``
341
- Support DSA SSH keys in bootstrap (Issue 338)
342
- Fix batch creation of instances
343
- Use FQDN to check master node status (Issue 551)
344
- Make the DRBD collector more failure-resilient
345

    
346

    
347
Version 2.8.0 rc1
348
-----------------
349

    
350
*(Released Fri, 2 Aug 2013)*
351

    
352
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
353

    
354
- Fix upgrading/downgrading from 2.7
355
- Increase maximum RAPI message size
356
- Documentation updates
357
- Split ``confd`` between ``luxid`` and ``confd``
358
- Merge 2.7 series up to the 2.7.1 release
359
- Allow the ``modify_etc_hosts`` option to be changed
360
- Add better debugging for ``luxid`` queries
361
- Expose bulk parameter for GetJobs in RAPI client
362
- Expose missing ``network`` fields in RAPI
363
- Add some ``cluster verify`` tests
364
- Some unittest fixes
365
- Fix a malfunction in ``hspace``'s tiered allocation
366
- Fix query compatibility between haskell and python implementations
367
- Add the ``vnet_hdr`` HV parameter for KVM
368
- Add ``--cleanup`` to instance failover
369
- Change the connected groups format in ``gnt-network info`` output; it
370
  was previously displayed as a raw list by mistake. (Merged from 2.7)
371

    
372

    
373
Version 2.8.0 beta1
374
-------------------
375

    
376
*(Released Mon, 24 Jun 2013)*
377

    
378
This was the first beta release of the 2.8 series. All important changes
379
are listed in the latest 2.8 entry.
380

    
381

    
382
Version 2.7.2
383
-------------
384

    
385
*(Released Thu, 26 Sep 2013)*
386

    
387
- Change the connected groups format in ``gnt-network info`` output; it
388
  was previously displayed as a raw list by mistake
389
- Check disk template in right dict when copying
390
- Support multi-instance allocs without iallocator
391
- Fix some errors in the documentation
392
- Fix formatting of tuple in an error message
393

    
394

    
395
Version 2.7.1
396
-------------
397

    
398
*(Released Thu, 25 Jul 2013)*
399

    
400
- Add logrotate functionality in daemon-util
401
- Add logrotate example file
402
- Add missing fields to network queries over rapi
403
- Fix network object timestamps
404
- Add support for querying network timestamps
405
- Fix a typo in the example crontab
406
- Fix a documentation typo
407

    
408

    
409
Version 2.7.0
410
-------------
411

    
412
*(Released Thu, 04 Jul 2013)*
413

    
414
Incompatible/important changes
415
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
416

    
417
- Instance policies for disk size were documented to be on a per-disk
418
  basis, but hail applied them to the sum of all disks. This has been
419
  fixed.
420
- ``hbal`` will now exit with status 0 if, during job execution over
421
  LUXI, early exit has been requested and all jobs are successful;
422
  before, exit status 1 was used, which cannot be differentiated from
423
  "job error" case
424
- Compatibility with newer versions of rbd has been fixed
425
- ``gnt-instance batch-create`` has been changed to use the bulk create
426
  opcode from Ganeti. This lead to incompatible changes in the format of
427
  the JSON file. It's now not a custom dict anymore but a dict
428
  compatible with the ``OpInstanceCreate`` opcode.
429
- Parent directories for file storage need to be listed in
430
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
431
  write the file automatically based on old configuration values, but it
432
  can not distribute it across all nodes and the file contents should be
433
  verified. Use ``gnt-cluster copyfile
434
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
435
  upgraded. The reason for requiring this list of paths now is that
436
  before it would have been possible to inject new paths via RPC,
437
  allowing files to be created in arbitrary locations. The RPC protocol
438
  is protected using SSL/X.509 certificates, but as a design principle
439
  Ganeti does not permit arbitrary paths to be passed.
440
- The parsing of the variants file for OSes (see
441
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
442
  lines and comment lines (starting with ``#``) are ignored for better
443
  readability.
444
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
445
  longer available. ``gnt-node add`` now invokes a new tool on the
446
  destination node, named ``prepare-node-join``, to configure the SSH
447
  daemon. Paramiko is no longer necessary to configure nodes' SSH
448
  daemons via ``gnt-node add``.
449
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
450
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
451
  the flag is not reset after a master failover.
452
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
453
  is recommended. 2.6 will be mandatory from the 2.8 series.
454

    
455

    
456
New features
457
~~~~~~~~~~~~
458

    
459
- New network management functionality to support automatic allocation
460
  of IP addresses and managing of network parameters. See
461
  :manpage:`gnt-network(8)` for more details.
462
- New external storage backend, to allow managing arbitrary storage
463
  systems external to the cluster. See
464
  :manpage:`ganeti-extstorage-interface(7)`.
465
- New ``exclusive-storage`` node parameter added, restricted to
466
  nodegroup level. When it's set to true, physical disks are assigned in
467
  an exclusive fashion to instances, as documented in :doc:`Partitioned
468
  Ganeti <design-partitioned>`.  Currently, only instances using the
469
  ``plain`` disk template are supported.
470
- The KVM hypervisor has been updated with many new hypervisor
471
  parameters, including a generic one for passing arbitrary command line
472
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
473
  compatible up to qemu 1.4.
474
- A new tool, called ``mon-collector``, is the stand-alone executor of
475
  the data collectors for a monitoring system. As of this version, it
476
  just includes the DRBD data collector, that can be executed by calling
477
  ``mon-collector`` using the ``drbd`` parameter. See
478
  :manpage:`mon-collector(7)`.
479
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
480
  for RAPI users. It allows granting permissions to query for
481
  information to a specific user without giving
482
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
483
- A new tool named ``node-cleanup`` has been added. It cleans remains of
484
  a cluster from a machine by stopping all daemons, removing
485
  certificates and ssconf files. Unless the ``--no-backup`` option is
486
  given, copies of the certificates are made.
487
- Instance creations now support the use of opportunistic locking,
488
  potentially speeding up the (parallel) creation of multiple instances.
489
  This feature is currently only available via the :doc:`RAPI
490
  <rapi>` interface and when an instance allocator is used. If the
491
  ``opportunistic_locking`` parameter is set the opcode will try to
492
  acquire as many locks as possible, but will not wait for any locks
493
  held by other opcodes. If not enough resources can be found to
494
  allocate the instance, the temporary error code
495
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
496
  retried thereafter, with or without opportunistic locking.
497
- New experimental linux-ha resource scripts.
498
- Restricted-commands support: ganeti can now be asked (via command line
499
  or rapi) to perform commands on a node. These are passed via ganeti
500
  RPC rather than ssh. This functionality is restricted to commands
501
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
502
  reasons. The file is not copied automatically.
503

    
504

    
505
Misc changes
506
~~~~~~~~~~~~
507

    
508
- Diskless instances are now externally mirrored (Issue 237). This for
509
  now has only been tested in conjunction with explicit target nodes for
510
  migration/failover.
511
- Queries not needing locks or RPC access to the node can now be
512
  performed by the confd daemon, making them independent from jobs, and
513
  thus faster to execute. This is selectable at configure time.
514
- The functionality for allocating multiple instances at once has been
515
  overhauled and is now also available through :doc:`RAPI <rapi>`.
516

    
517
There are no significant changes from version 2.7.0~rc3.
518

    
519

    
520
Version 2.7.0 rc3
521
-----------------
522

    
523
*(Released Tue, 25 Jun 2013)*
524

    
525
- Fix permissions on the confd query socket (Issue 477)
526
- Fix permissions on the job archive dir (Issue 498)
527
- Fix handling of an internal exception in replace-disks (Issue 472)
528
- Fix gnt-node info handling of shortened names (Issue 497)
529
- Fix gnt-instance grow-disk when wiping is enabled
530
- Documentation improvements, and support for newer pandoc
531
- Fix hspace honoring ipolicy for disks (Issue 484)
532
- Improve handling of the ``kvm_extra`` HV parameter
533

    
534

    
535
Version 2.7.0 rc2
536
-----------------
537

    
538
*(Released Fri, 24 May 2013)*
539

    
540
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
541
  symlink.
542
- Disks added through ``gnt-instance modify`` or created through
543
  ``gnt-instance recreate-disks`` are wiped, if the
544
  ``prealloc_wipe_disks`` flag is set.
545
- If wiping newly created disks fails, the disks are removed. Also,
546
  partial failures in creating disks through ``gnt-instance modify``
547
  triggers a cleanup of the partially-created disks.
548
- Removing the master IP address doesn't fail if the address has been
549
  already removed.
550
- Fix ownership of the OS log dir
551
- Workaround missing SO_PEERCRED constant (Issue 191)
552

    
553

    
554
Version 2.7.0 rc1
555
-----------------
556

    
557
*(Released Fri, 3 May 2013)*
558

    
559
This was the first release candidate of the 2.7 series. Since beta3:
560

    
561
- Fix kvm compatibility with qemu 1.4 (Issue 389)
562
- Documentation updates (admin guide, upgrade notes, install
563
  instructions) (Issue 372)
564
- Fix gnt-group list nodes and instances count (Issue 436)
565
- Fix compilation without non-mandatory libraries (Issue 441)
566
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
567
- Make confd logging more verbose at INFO level (Issue 435)
568
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
569
- Fix failure path for instance storage type conversion (Issue 229)
570
- Update htools text backend documentation
571
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
572
- Disable inter-cluster instance move for file-based instances, because
573
  it is dependant on instance export, which is not supported for
574
  file-based instances. (Issue 414)
575
- Fix gnt-job crashes on non-ascii characters (Issue 427)
576
- Fix volume group checks on non-vm-capable nodes (Issue 432)
577

    
578

    
579
Version 2.7.0 beta3
580
-------------------
581

    
582
*(Released Mon, 22 Apr 2013)*
583

    
584
This was the third beta release of the 2.7 series. Since beta2:
585

    
586
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
587
- Fix data loss on wrong usage of ``gnt-instance move``
588
- Properly export errors in confd-based job queries
589
- Add ``users-setup`` tool
590
- Fix iallocator protocol to report 0 as a disk size for diskless
591
  instances. This avoids hail breaking when a diskless instance is
592
  present.
593
- Fix job queue directory permission problem that made confd job queries
594
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
595
  for access to archived jobs (Issue 406).
596
- Limit the sizes of networks supported by ``gnt-network`` to something
597
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
598
- Fix bugs in instance disk template conversion
599
- Fix GHC 7 compatibility
600
- Fix ``burnin`` install path (Issue 426).
601
- Allow very small disk grows (Issue 347).
602
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
603
  that noded doesn't import masterd code (Issue 419).
604
- Make sure the default metavg at cluster init is the same as the vg, if
605
  unspecified (Issue 358).
606
- Fix cleanup of partially created disks (part of Issue 416)
607

    
608

    
609
Version 2.7.0 beta2
610
-------------------
611

    
612
*(Released Tue, 2 Apr 2013)*
613

    
614
This was the second beta release of the 2.7 series. Since beta1:
615

    
616
- Networks no longer have a "type" slot, since this information was
617
  unused in Ganeti: instead of it tags should be used.
618
- The rapi client now has a ``target_node`` option to MigrateInstance.
619
- Fix early exit return code for hbal (Issue 386).
620
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
621
- Fix ``rbd showmapped`` output parsing (Issue 312).
622
- Networks are now referenced indexed by UUID, rather than name. This
623
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
624
  use.
625
- The OS environment now includes network information.
626
- Deleting of a network is now disallowed if any instance nic is using
627
  it, to prevent dangling references.
628
- External storage is now documented in man pages.
629
- The exclusive_storage flag can now only be set at nodegroup level.
630
- Hbal can now submit an explicit priority with its jobs.
631
- Many network related locking fixes.
632
- Bump up the required pylint version to 0.25.1.
633
- Fix the ``no_remember`` option in RAPI client.
634
- Many ipolicy related tests, qa, and fixes.
635
- Many documentation improvements and fixes.
636
- Fix building with ``--disable-file-storage``.
637
- Fix ``-q`` option in htools, which was broken if passed more than
638
  once.
639
- Some haskell/python interaction improvements and fixes.
640
- Fix iallocator in case of missing LVM storage.
641
- Fix confd config load in case of ``--no-lvm-storage``.
642
- The confd/query functionality is now mentioned in the security
643
  documentation.
644

    
645

    
646
Version 2.7.0 beta1
647
-------------------
648

    
649
*(Released Wed, 6 Feb 2013)*
650

    
651
This was the first beta release of the 2.7 series. All important changes
652
are listed in the latest 2.7 entry.
653

    
654

    
655
Version 2.6.2
656
-------------
657

    
658
*(Released Fri, 21 Dec 2012)*
659

    
660
Important behaviour change: hbal won't rebalance anymore instances which
661
have the ``auto_balance`` attribute set to false. This was the intention
662
all along, but until now it only skipped those from the N+1 memory
663
reservation (DRBD-specific).
664

    
665
A significant number of bug fixes in this release:
666

    
667
- Fixed disk adoption interaction with ipolicy checks.
668
- Fixed networking issues when instances are started, stopped or
669
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
670
- Fixed the warning in cluster verify for shared storage instances not
671
  being redundant.
672
- Fixed removal of storage directory on shared file storage (issue 262).
673
- Fixed validation of LVM volume group name in OpClusterSetParams
674
  (``gnt-cluster modify``) (issue 285).
675
- Fixed runtime memory increases (``gnt-instance modify -m``).
676
- Fixed live migration under Xen's ``xl`` mode.
677
- Fixed ``gnt-instance console`` with ``xl``.
678
- Fixed building with newer Haskell compiler/libraries.
679
- Fixed PID file writing in Haskell daemons (confd); this prevents
680
  restart issues if confd was launched manually (outside of
681
  ``daemon-util``) while another copy of it was running
682
- Fixed a type error when doing live migrations with KVM (issue 297) and
683
  the error messages for failing migrations have been improved.
684
- Fixed opcode validation for the out-of-band commands (``gnt-node
685
  power``).
686
- Fixed a type error when unsetting OS hypervisor parameters (issue
687
  311); now it's possible to unset all OS-specific hypervisor
688
  parameters.
689
- Fixed the ``dry-run`` mode for many operations: verification of
690
  results was over-zealous but didn't take into account the ``dry-run``
691
  operation, resulting in "wrong" failures.
692
- Fixed bash completion in ``gnt-job list`` when the job queue has
693
  hundreds of entries; especially with older ``bash`` versions, this
694
  results in significant CPU usage.
695

    
696
And lastly, a few other improvements have been made:
697

    
698
- Added option to force master-failover without voting (issue 282).
699
- Clarified error message on lock conflict (issue 287).
700
- Logging of newly submitted jobs has been improved (issue 290).
701
- Hostname checks have been made uniform between instance rename and
702
  create (issue 291).
703
- The ``--submit`` option is now supported by ``gnt-debug delay``.
704
- Shutting down the master daemon by sending SIGTERM now stops it from
705
  processing jobs waiting for locks; instead, those jobs will be started
706
  once again after the master daemon is started the next time (issue
707
  296).
708
- Support for Xen's ``xl`` program has been improved (besides the fixes
709
  above).
710
- Reduced logging noise in the Haskell confd daemon (only show one log
711
  entry for each config reload, instead of two).
712
- Several man page updates and typo fixes.
713

    
714

    
715
Version 2.6.1
716
-------------
717

    
718
*(Released Fri, 12 Oct 2012)*
719

    
720
A small bugfix release. Among the bugs fixed:
721

    
722
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
723
  made the command unusable.
724
- Commands that issue many jobs don't fail anymore just because some jobs
725
  take so long that other jobs are archived.
726
- Failures during ``gnt-instance reinstall`` are reflected by the exit
727
  status.
728
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
729
  DRBD is enabled.
730
- When ``always_failover`` is set, ``--allow-failover`` is not required
731
  in migrate commands anymore.
732
- ``bash_completion`` works even if extglob is disabled.
733
- Fixed bug with locks that made failover for RDB-based instances fail.
734
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
735
  a random node instead of one based on the allocator metric.
736
- Support for newer versions of pylint and pep8.
737
- Hail doesn't fail anymore when trying to add an instance of type
738
  ``file``, ``sharedfile`` or ``rbd``.
739
- Added new Makefile target to rebuild the whole distribution, so that
740
  all files are included.
741

    
742

    
743
Version 2.6.0
744
-------------
745

    
746
*(Released Fri, 27 Jul 2012)*
747

    
748

    
749
.. attention:: The ``LUXI`` protocol has been made more consistent
750
   regarding its handling of command arguments. This, however, leads to
751
   incompatibility issues with previous versions. Please ensure that you
752
   restart Ganeti daemons soon after the upgrade, otherwise most
753
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
754
   pausing/resuming the watcher, cancelling and archiving jobs, querying
755
   the cluster configuration) will fail.
756

    
757

    
758
New features
759
~~~~~~~~~~~~
760

    
761
Instance run status
762
+++++++++++++++++++
763

    
764
The current ``admin_up`` field, which used to denote whether an instance
765
should be running or not, has been removed. Instead, ``admin_state`` is
766
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
767

    
768
The rational behind this is that an instance being “down” can have
769
different meanings:
770

    
771
- it could be down during a reboot
772
- it could be temporarily be down for a reinstall
773
- or it could be down because it is deprecated and kept just for its
774
  disk
775

    
776
The previous Boolean state was making it difficult to do capacity
777
calculations: should Ganeti reserve memory for a down instance? Now, the
778
tri-state field makes it clear:
779

    
780
- in ``up`` and ``down`` state, all resources are reserved for the
781
  instance, and it can be at any time brought up if it is down
782
- in ``offline`` state, only disk space is reserved for it, but not
783
  memory or CPUs
784

    
785
The field can have an extra use: since the transition between ``up`` and
786
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
787
transition between ``offline`` and ``down`` is done via ``gnt-instance
788
modify``, it is possible to given different rights to users. For
789
example, owners of an instance could be allowed to start/stop it, but
790
not transition it out of the offline state.
791

    
792
Instance policies and specs
793
+++++++++++++++++++++++++++
794

    
795
In previous Ganeti versions, an instance creation request was not
796
limited on the minimum size and on the maximum size just by the cluster
797
resources. As such, any policy could be implemented only in third-party
798
clients (RAPI clients, or shell wrappers over ``gnt-*``
799
tools). Furthermore, calculating cluster capacity via ``hspace`` again
800
required external input with regards to instance sizes.
801

    
802
In order to improve these workflows and to allow for example better
803
per-node group differentiation, we introduced instance specs, which
804
allow declaring:
805

    
806
- minimum instance disk size, disk count, memory size, cpu count
807
- maximum values for the above metrics
808
- and “standard” values (used in ``hspace`` to calculate the standard
809
  sized instances)
810

    
811
The minimum/maximum values can be also customised at node-group level,
812
for example allowing more powerful hardware to support bigger instance
813
memory sizes.
814

    
815
Beside the instance specs, there are a few other settings belonging to
816
the instance policy framework. It is possible now to customise, per
817
cluster and node-group:
818

    
819
- the list of allowed disk templates
820
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
821
- the maximum ratio of instance to spindles (see below for more
822
  information) for local storage
823

    
824
All these together should allow all tools that talk to Ganeti to know
825
what are the ranges of allowed values for instances and the
826
over-subscription that is allowed.
827

    
828
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
829
instance configuration, and the physical CPU configuration from the
830
node. For the spindle ratios however, we didn't track before these
831
values, so new parameters have been added:
832

    
833
- a new node parameter ``spindle_count``, defaults to 1, customisable at
834
  node group or node level
835
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
836

    
837
Note that spindles in this context doesn't need to mean actual
838
mechanical hard-drives; it's just a relative number for both the node
839
I/O capacity and instance I/O consumption.
840

    
841
Instance migration behaviour
842
++++++++++++++++++++++++++++
843

    
844
While live-migration is in general desirable over failover, it is
845
possible that for some workloads it is actually worse, due to the
846
variable time of the “suspend” phase during live migration.
847

    
848
To allow the tools to work consistently over such instances (without
849
having to hard-code instance names), a new backend parameter
850
``always_failover`` has been added to control the migration/failover
851
behaviour. When set to True, all migration requests for an instance will
852
instead fall-back to failover.
853

    
854
Instance memory ballooning
855
++++++++++++++++++++++++++
856

    
857
Initial support for memory ballooning has been added. The memory for an
858
instance is no longer fixed (backend parameter ``memory``), but instead
859
can vary between minimum and maximum values (backend parameters
860
``minmem`` and ``maxmem``). Currently we only change an instance's
861
memory when:
862

    
863
- live migrating or failing over and instance and the target node
864
  doesn't have enough memory
865
- user requests changing the memory via ``gnt-instance modify
866
  --runtime-memory``
867

    
868
Instance CPU pinning
869
++++++++++++++++++++
870

    
871
In order to control the use of specific CPUs by instance, support for
872
controlling CPU pinning has been added for the Xen, HVM and LXC
873
hypervisors. This is controlled by a new hypervisor parameter
874
``cpu_mask``; details about possible values for this are in the
875
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
876
VCPU-to-CPU mapping) form will work well only when all nodes in your
877
cluster have the same amount of CPUs.
878

    
879
Disk parameters
880
+++++++++++++++
881

    
882
Another area in which Ganeti was not customisable were the parameters
883
used for storage configuration, e.g. how many stripes to use for LVM,
884
DRBD resync configuration, etc.
885

    
886
To improve this area, we've added disks parameters, which are
887
customisable at cluster and node group level, and which allow to
888
specify various parameters for disks (DRBD has the most parameters
889
currently), for example:
890

    
891
- DRBD resync algorithm and parameters (e.g. speed)
892
- the default VG for meta-data volumes for DRBD
893
- number of stripes for LVM (plain disk template)
894
- the RBD pool
895

    
896
These parameters can be modified via ``gnt-cluster modify -D …`` and
897
``gnt-group modify -D …``, and are used at either instance creation (in
898
case of LVM stripes, for example) or at disk “activation” time
899
(e.g. resync speed).
900

    
901
Rados block device support
902
++++++++++++++++++++++++++
903

    
904
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
905
denoted by the ``rbd`` disk template type. This is considered
906
experimental, feedback is welcome. For details on configuring it, see
907
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
908

    
909
Master IP setup
910
+++++++++++++++
911

    
912
The existing master IP functionality works well only in simple setups (a
913
single network shared by all nodes); however, if nodes belong to
914
different networks, then the ``/32`` setup and lack of routing
915
information is not enough.
916

    
917
To allow the master IP to function well in more complex cases, the
918
system was reworked as follows:
919

    
920
- a master IP netmask setting has been added
921
- the master IP activation/turn-down code was moved from the node daemon
922
  to a separate script
923
- whether to run the Ganeti-supplied master IP script or a user-supplied
924
  on is a ``gnt-cluster init`` setting
925

    
926
Details about the location of the standard and custom setup scripts are
927
in the man page :manpage:`gnt-cluster(8)`; for information about the
928
setup script protocol, look at the Ganeti-supplied script.
929

    
930
SPICE support
931
+++++++++++++
932

    
933
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
934
improved.
935

    
936
It is now possible to use TLS-protected connections, and when renewing
937
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
938
it is now possible to specify spice or spice CA certificates. Also, it
939
is possible to configure a password for SPICE sessions via the
940
hypervisor parameter ``spice_password_file``.
941

    
942
There are also new parameters to control the compression and streaming
943
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
944
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
945
for the spice parameters.
946

    
947
Lastly, it is now possible to see the SPICE connection information via
948
``gnt-instance console``.
949

    
950
OVF converter
951
+++++++++++++
952

    
953
A new tool (``tools/ovfconverter``) has been added that supports
954
conversion between Ganeti and the `Open Virtualization Format
955
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
956
from).
957

    
958
This relies on the ``qemu-img`` tool to convert the disk formats, so the
959
actual compatibility with other virtualization solutions depends on it.
960

    
961
Confd daemon changes
962
++++++++++++++++++++
963

    
964
The configuration query daemon (``ganeti-confd``) is now optional, and
965
has been rewritten in Haskell; whether to use the daemon at all, use the
966
Python (default) or the Haskell version is selectable at configure time
967
via the ``--enable-confd`` parameter, which can take one of the
968
``haskell``, ``python`` or ``no`` values. If not used, disabling the
969
daemon will result in a smaller footprint; for larger systems, we
970
welcome feedback on the Haskell version which might become the default
971
in future versions.
972

    
973
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
974
daemon running. The Python version doesn't implement the new call.
975

    
976

    
977
User interface changes
978
~~~~~~~~~~~~~~~~~~~~~~
979

    
980
We have replaced the ``--disks`` option of ``gnt-instance
981
replace-disks`` with a more flexible ``--disk`` option, which allows
982
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
983
disk size and mode can be changed upon recreation (via ``gnt-instance
984
recreate-disks``, which accepts the same ``--disk`` option).
985

    
986
As many people are used to a ``show`` command, we have added that as an
987
alias to ``info`` on all ``gnt-*`` commands.
988

    
989
The ``gnt-instance grow-disk`` command has a new mode in which it can
990
accept the target size of the disk, instead of the delta; this can be
991
more safe since two runs in absolute mode will be idempotent, and
992
sometimes it's also easier to specify the desired size directly.
993

    
994
Also the handling of instances with regard to offline secondaries has
995
been improved. Instance operations should not fail because one of it's
996
secondary nodes is offline, even though it's safe to proceed.
997

    
998
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
999
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
1000
minors to instance name.
1001

    
1002
API changes
1003
~~~~~~~~~~~
1004

    
1005
RAPI coverage has improved, with (for example) new resources for
1006
recreate-disks, node power-cycle, etc.
1007

    
1008
Compatibility
1009
~~~~~~~~~~~~~
1010

    
1011
There is partial support for ``xl`` in the Xen hypervisor; feedback is
1012
welcome.
1013

    
1014
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
1015
whether to still support Python 2.4 or move to Python 2.6 as minimum
1016
required version.
1017

    
1018
Support for Fedora has been slightly improved; the provided example
1019
init.d script should work better on it and the INSTALL file should
1020
document the needed dependencies.
1021

    
1022
Internal changes
1023
~~~~~~~~~~~~~~~~
1024

    
1025
The deprecated ``QueryLocks`` LUXI request has been removed. Use
1026
``Query(what=QR_LOCK, ...)`` instead.
1027

    
1028
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
1029
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
1030
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
1031
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
1032
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
1033

    
1034
RAPI client: ``CertificateError`` now derives from
1035
``GanetiApiError``. This should make it more easy to handle Ganeti
1036
errors.
1037

    
1038
Deprecation warnings due to PyCrypto/paramiko import in
1039
``tools/setup-ssh`` have been silenced, as usually they are safe; please
1040
make sure to run an up-to-date paramiko version, if you use this tool.
1041

    
1042
The QA scripts now depend on Python 2.5 or above (the main code base
1043
still works with Python 2.4).
1044

    
1045
The configuration file (``config.data``) is now written without
1046
indentation for performance reasons; if you want to edit it, it can be
1047
re-formatted via ``tools/fmtjson``.
1048

    
1049
A number of bugs has been fixed in the cluster merge tool.
1050

    
1051
``x509`` certification verification (used in import-export) has been
1052
changed to allow the same clock skew as permitted by the cluster
1053
verification. This will remove some rare but hard to diagnose errors in
1054
import-export.
1055

    
1056

    
1057
Version 2.6.0 rc4
1058
-----------------
1059

    
1060
*(Released Thu, 19 Jul 2012)*
1061

    
1062
Very few changes from rc4 to the final release, only bugfixes:
1063

    
1064
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
1065
  instance, fix CDROM booting for KVM instances)
1066
- fixed node group modification of node parameters
1067
- fixed issue in LUClusterVerifyGroup with multi-group clusters
1068
- fixed generation of bash completion to ensure a stable ordering
1069
- fixed a few typos
1070

    
1071

    
1072
Version 2.6.0 rc3
1073
-----------------
1074

    
1075
*(Released Fri, 13 Jul 2012)*
1076

    
1077
Third release candidate for 2.6. The following changes were done from
1078
rc3 to rc4:
1079

    
1080
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
1081
- Fixed an inconsistency in the LUXI protocol with the provided
1082
  arguments (NOT backwards compatible)
1083
- Fixed a bug with node groups ipolicy where ``min`` was greater than
1084
  the cluster ``std`` value
1085
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1086
  easier instance debugging on nodes (requires ``hconfd`` to work)
1087

    
1088

    
1089
Version 2.6.0 rc2
1090
-----------------
1091

    
1092
*(Released Tue, 03 Jul 2012)*
1093

    
1094
Second release candidate for 2.6. The following changes were done from
1095
rc2 to rc3:
1096

    
1097
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1098
  master candidates
1099
- Fixed a RAPI regression on missing beparams/memory
1100
- Fixed redistribution of files on offline nodes
1101
- Added possibility to run activate-disks even though secondaries are
1102
  offline. With this change it relaxes also the strictness on some other
1103
  commands which use activate disks internally:
1104
  * ``gnt-instance start|reboot|rename|backup|export``
1105
- Made it possible to remove safely an instance if its secondaries are
1106
  offline
1107
- Made it possible to reinstall even though secondaries are offline
1108

    
1109

    
1110
Version 2.6.0 rc1
1111
-----------------
1112

    
1113
*(Released Mon, 25 Jun 2012)*
1114

    
1115
First release candidate for 2.6. The following changes were done from
1116
rc1 to rc2:
1117

    
1118
- Fixed bugs with disk parameters and ``rbd`` templates as well as
1119
  ``instance_os_add``
1120
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1121
  behaviour. It supports now the modify operation
1122
- ``hcheck`` implemented to analyze cluster health and possibility of
1123
  improving health by rebalance
1124
- ``hbal`` has been improved in dealing with split instances
1125

    
1126

    
1127
Version 2.6.0 beta2
1128
-------------------
1129

    
1130
*(Released Mon, 11 Jun 2012)*
1131

    
1132
Second beta release of 2.6. The following changes were done from beta2
1133
to rc1:
1134

    
1135
- Fixed ``daemon-util`` with non-root user models
1136
- Fixed creation of plain instances with ``--no-wait-for-sync``
1137
- Fix wrong iv_names when running ``cfgupgrade``
1138
- Export more information in RAPI group queries
1139
- Fixed bug when changing instance network interfaces
1140
- Extended burnin to do NIC changes
1141
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1142
- Changed default for DRBD barriers
1143
- Fixed DRBD error reporting for syncer rate
1144
- Verify the options on disk parameters
1145

    
1146
And of course various fixes to documentation and improved unittests and
1147
QA.
1148

    
1149

    
1150
Version 2.6.0 beta1
1151
-------------------
1152

    
1153
*(Released Wed, 23 May 2012)*
1154

    
1155
First beta release of 2.6. The following changes were done from beta1 to
1156
beta2:
1157

    
1158
- integrated patch for distributions without ``start-stop-daemon``
1159
- adapted example init.d script to work on Fedora
1160
- fixed log handling in Haskell daemons
1161
- adapted checks in the watcher for pycurl linked against libnss
1162
- add partial support for ``xl`` instead of ``xm`` for Xen
1163
- fixed a type issue in cluster verification
1164
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1165
  clusters)
1166

    
1167
Plus integrated fixes from the 2.5 branch:
1168

    
1169
- fixed ``kvm-ifup`` to use ``/bin/bash``
1170
- fixed parallel build failures
1171
- KVM live migration when using a custom keymap
1172

    
1173

    
1174
Version 2.5.2
1175
-------------
1176

    
1177
*(Released Tue, 24 Jul 2012)*
1178

    
1179
A small bugfix release, with no new features:
1180

    
1181
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1182
  different default shell (e.g. Debian, Ubuntu)
1183
- fixed KVM startup and live migration with a custom keymap (fixes Issue
1184
  243 and Debian bug #650664)
1185
- fixed compatibility with KVM versions that don't support multiple boot
1186
  devices (fixes Issue 230 and Debian bug #624256)
1187

    
1188
Additionally, a few fixes were done to the build system (fixed parallel
1189
build failures) and to the unittests (fixed race condition in test for
1190
FileID functions, and the default enable/disable mode for QA test is now
1191
customisable).
1192

    
1193

    
1194
Version 2.5.1
1195
-------------
1196

    
1197
*(Released Fri, 11 May 2012)*
1198

    
1199
A small bugfix release.
1200

    
1201
The main issues solved are on the topic of compatibility with newer LVM
1202
releases:
1203

    
1204
- fixed parsing of ``lv_attr`` field
1205
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1206
  the ``--force`` flag is needed
1207

    
1208
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1209
to accept kernel 3.x too (was hardcoded to 2.6.*).
1210

    
1211
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1212
of hook results and that also made display of other errors suboptimal
1213
was fixed; the code behaves now like 2.4 and earlier.
1214

    
1215
Another change in 2.5, the cleanup of the OS scripts environment, is too
1216
aggressive: it removed even the ``PATH`` variable, which requires the OS
1217
scripts to *always* need to export it. Since this is a bit too strict,
1218
we now export a minimal PATH, the same that we export for hooks.
1219

    
1220
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1221
integrated into this release.
1222

    
1223
Finally, a few other miscellaneous changes were done (no new features,
1224
just small improvements):
1225

    
1226
- Fix ``gnt-group --help`` display
1227
- Fix hardcoded Xen kernel path
1228
- Fix grow-disk handling of invalid units
1229
- Update synopsis for ``gnt-cluster repair-disk-sizes``
1230
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1231

    
1232

    
1233
Version 2.5.0
1234
-------------
1235

    
1236
*(Released Thu, 12 Apr 2012)*
1237

    
1238
Incompatible/important changes and bugfixes
1239
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1240

    
1241
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
1242
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
1243
  to match the underlying LUXI interface.
1244
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1245
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1246
  not maintain backwards-compatibility as the underlying operation
1247
  changed in an incompatible way. The RAPI client can talk to old
1248
  servers, but it needs to be told so as the return value changed.
1249
- When creating file-based instances via RAPI, the ``file_driver``
1250
  parameter no longer defaults to ``loop`` and must be specified.
1251
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
1252
  ``link`` instead.
1253
- Support for the undocumented and deprecated RAPI instance creation
1254
  request format version 0 has been dropped. Use version 1, supported
1255
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1256
- Pyparsing 1.4.6 or above is required, see :doc:`installation
1257
  documentation <install>`.
1258
- The "cluster-verify" hooks are now executed per group by the
1259
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1260
  if you just run ``gnt-cluster verify``, which generates one opcode per
1261
  group.
1262
- The environment as passed to the OS scripts is cleared, and thus no
1263
  environment variables defined in the node daemon's environment will be
1264
  inherited by the scripts.
1265
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1266
  deprecated.
1267
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1268
  support operations involving multiple node groups.
1269
- Offline nodes are ignored when failing over an instance.
1270
- Support for KVM version 1.0, which changed the version reporting format
1271
  from 3 to 2 digits.
1272
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1273
  removal.
1274
- ``Makefile`` is now compatible with Automake 1.11.2
1275
- Includes all bugfixes made in the 2.4 series
1276

    
1277
New features
1278
~~~~~~~~~~~~
1279

    
1280
- The ganeti-htools project has been merged into the ganeti-core source
1281
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1282
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1283
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1284
  GPT-style partition tables (requires `parted
1285
  <http://www.gnu.org/s/parted/>`_).
1286
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1287
- Allowed adding tags on instance creation.
1288
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1289
- Allow instances to be started in a paused state, enabling the user to
1290
  see the complete console output on boot using the console.
1291
- Added new hypervisor flag to control default reboot behaviour
1292
  (``reboot_behavior``).
1293
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1294
- Improved out-of-band management support:
1295

    
1296
  - Added ``gnt-node health`` command reporting the health status of
1297
    nodes.
1298
  - Added ``gnt-node power`` command to manage power status of nodes.
1299
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1300

    
1301
- Instance migration can fall back to failover if instance is not
1302
  running.
1303
- Filters can be used when listing nodes, instances, groups and locks;
1304
  see :manpage:`ganeti(7)` manpage.
1305
- Added post-execution status as variables to :doc:`hooks <hooks>`
1306
  environment.
1307
- Instance tags are exported/imported together with the instance.
1308
- When given an explicit job ID, ``gnt-job info`` will work for archived
1309
  jobs.
1310
- Jobs can define dependencies on other jobs (not yet supported via
1311
  RAPI or command line, but used by internal commands and usable via
1312
  LUXI).
1313

    
1314
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1315
    dependencies.
1316

    
1317
- Instance failover is now available as a RAPI resource
1318
  (``/2/instances/[instance_name]/failover``).
1319
- ``gnt-instance info`` defaults to static information if primary node
1320
  is offline.
1321
- Opcodes have a new ``comment`` attribute.
1322
- Added basic SPICE support to KVM hypervisor.
1323
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1324

    
1325
Node group improvements
1326
~~~~~~~~~~~~~~~~~~~~~~~
1327

    
1328
- ``gnt-cluster verify`` has been modified to check groups separately,
1329
  thereby improving performance.
1330
- Node group support has been added to ``gnt-cluster verify-disks``,
1331
  which now operates per node group.
1332
- Watcher has been changed to work better with node groups.
1333

    
1334
  - One process and state file per node group.
1335
  - Slow watcher in one group doesn't block other group's watcher.
1336

    
1337
- Added new command, ``gnt-group evacuate``, to move all instances in a
1338
  node group to other groups.
1339
- Added ``gnt-instance change-group`` to move an instance to another
1340
  node group.
1341
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1342
  per-group operations.
1343
- Node groups can be tagged.
1344
- Some operations switch from an exclusive to a shared lock as soon as
1345
  possible.
1346
- Instance's primary and secondary nodes' groups are now available as
1347
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1348
  and ``snodes.group.uuid``).
1349

    
1350
Misc
1351
~~~~
1352

    
1353
- Numerous updates to documentation and manpages.
1354

    
1355
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1356
    descriptions.
1357
  - Some opcode/job results are now also documented, see :doc:`RAPI
1358
    <rapi>`.
1359

    
1360
- A lockset's internal lock is now also visible in lock monitor.
1361
- Log messages from job queue workers now contain information about the
1362
  opcode they're processing.
1363
- ``gnt-instance console`` no longer requires the instance lock.
1364
- A short delay when waiting for job changes reduces the number of LUXI
1365
  requests significantly.
1366
- DRBD metadata volumes are overwritten with zeros during disk creation.
1367
- Out-of-band commands no longer acquire the cluster lock in exclusive
1368
  mode.
1369
- ``devel/upload`` now uses correct permissions for directories.
1370

    
1371

    
1372
Version 2.5.0 rc6
1373
-----------------
1374

    
1375
*(Released Fri, 23 Mar 2012)*
1376

    
1377
This was the sixth release candidate of the 2.5 series.
1378

    
1379

    
1380
Version 2.5.0 rc5
1381
-----------------
1382

    
1383
*(Released Mon, 9 Jan 2012)*
1384

    
1385
This was the fifth release candidate of the 2.5 series.
1386

    
1387

    
1388
Version 2.5.0 rc4
1389
-----------------
1390

    
1391
*(Released Thu, 27 Oct 2011)*
1392

    
1393
This was the fourth release candidate of the 2.5 series.
1394

    
1395

    
1396
Version 2.5.0 rc3
1397
-----------------
1398

    
1399
*(Released Wed, 26 Oct 2011)*
1400

    
1401
This was the third release candidate of the 2.5 series.
1402

    
1403

    
1404
Version 2.5.0 rc2
1405
-----------------
1406

    
1407
*(Released Tue, 18 Oct 2011)*
1408

    
1409
This was the second release candidate of the 2.5 series.
1410

    
1411

    
1412
Version 2.5.0 rc1
1413
-----------------
1414

    
1415
*(Released Tue, 4 Oct 2011)*
1416

    
1417
This was the first release candidate of the 2.5 series.
1418

    
1419

    
1420
Version 2.5.0 beta3
1421
-------------------
1422

    
1423
*(Released Wed, 31 Aug 2011)*
1424

    
1425
This was the third beta release of the 2.5 series.
1426

    
1427

    
1428
Version 2.5.0 beta2
1429
-------------------
1430

    
1431
*(Released Mon, 22 Aug 2011)*
1432

    
1433
This was the second beta release of the 2.5 series.
1434

    
1435

    
1436
Version 2.5.0 beta1
1437
-------------------
1438

    
1439
*(Released Fri, 12 Aug 2011)*
1440

    
1441
This was the first beta release of the 2.5 series.
1442

    
1443

    
1444
Version 2.4.5
1445
-------------
1446

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

    
1449
- Fixed bug when parsing command line parameter values ending in
1450
  backslash
1451
- Fixed assertion error after unclean master shutdown
1452
- Disable HTTP client pool for RPC, significantly reducing memory usage
1453
  of master daemon
1454
- Fixed queue archive creation with wrong permissions
1455

    
1456

    
1457
Version 2.4.4
1458
-------------
1459

    
1460
*(Released Tue, 23 Aug 2011)*
1461

    
1462
Small bug-fixes:
1463

    
1464
- Fixed documentation for importing with ``--src-dir`` option
1465
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1466
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1467

    
1468

    
1469
Version 2.4.3
1470
-------------
1471

    
1472
*(Released Fri, 5 Aug 2011)*
1473

    
1474
Many bug-fixes and a few small features:
1475

    
1476
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1477
  issues when you tried to add an instance with two MAC addresses in one
1478
  request
1479
- KVM: fixed per-instance stored UID value
1480
- KVM: configure bridged NICs at migration start
1481
- KVM: Fix a bug where instance will not start with never KVM versions
1482
  (>= 0.14)
1483
- Added OS search path to ``gnt-cluster info``
1484
- Fixed an issue with ``file_storage_dir`` where you were forced to
1485
  provide an absolute path, but the documentation states it is a
1486
  relative path, the documentation was right
1487
- Added a new parameter to instance stop/start called ``--no-remember``
1488
  that will make the state change to not be remembered
1489
- Implemented ``no_remember`` at RAPI level
1490
- Improved the documentation
1491
- Node evacuation: don't call IAllocator if node is already empty
1492
- Fixed bug in DRBD8 replace disks on current nodes
1493
- Fixed bug in recreate-disks for DRBD instances
1494
- Moved assertion checking locks in ``gnt-instance replace-disks``
1495
  causing it to abort with not owning the right locks for some situation
1496
- Job queue: Fixed potential race condition when cancelling queued jobs
1497
- Fixed off-by-one bug in job serial generation
1498
- ``gnt-node volumes``: Fix instance names
1499
- Fixed aliases in bash completion
1500
- Fixed a bug in reopening log files after being sent a SIGHUP
1501
- Added a flag to burnin to allow specifying VCPU count
1502
- Bugfixes to non-root Ganeti configuration
1503

    
1504

    
1505
Version 2.4.2
1506
-------------
1507

    
1508
*(Released Thu, 12 May 2011)*
1509

    
1510
Many bug-fixes and a few new small features:
1511

    
1512
- Fixed a bug related to log opening failures
1513
- Fixed a bug in instance listing with orphan instances
1514
- Fixed a bug which prevented resetting the cluster-level node parameter
1515
  ``oob_program`` to the default
1516
- Many fixes related to the ``cluster-merge`` tool
1517
- Fixed a race condition in the lock monitor, which caused failures
1518
  during (at least) creation of many instances in parallel
1519
- Improved output for gnt-job info
1520
- Removed the quiet flag on some ssh calls which prevented debugging
1521
  failures
1522
- Improved the N+1 failure messages in cluster verify by actually
1523
  showing the memory values (needed and available)
1524
- Increased lock attempt timeouts so that when executing long operations
1525
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1526
  too early and thus prevent the use of the 'fair' mechanism
1527
- Changed instance query data (``gnt-instance info``) to not acquire
1528
  locks unless needed, thus allowing its use on locked instance if only
1529
  static information is asked for
1530
- Improved behaviour with filesystems that do not support rename on an
1531
  opened file
1532
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1533
  kept locks on all nodes during the wipe, which is unneeded
1534
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1535
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1536
  10GiB) which caused the wipe to fail right at the end in some cases
1537
- Fixed master IP activation when doing master failover with no-voting
1538
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1539
  the master node itself
1540
- Fixed potential data-loss in under disk full conditions, where Ganeti
1541
  wouldn't check correctly the return code and would consider
1542
  partially-written files 'correct'
1543
- Fixed bug related to multiple VGs and DRBD disk replacing
1544
- Added new disk parameter ``metavg`` that allows placement of the meta
1545
  device for DRBD in a different volume group
1546
- Fixed error handling in the node daemon when the system libc doesn't
1547
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1548
- Fixed lock release during replace-disks, which kept cluster-wide locks
1549
  when doing disk replaces with an iallocator script
1550
- Added check for missing bridges in cluster verify
1551
- Handle EPIPE errors while writing to the terminal better, so that
1552
  piping the output to e.g. ``less`` doesn't cause a backtrace
1553
- Fixed rare case where a ^C during Luxi calls could have been
1554
  interpreted as server errors, instead of simply terminating
1555
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1556
  assign-nodes``)
1557
- Added a few more parameters to the KVM hypervisor, allowing a second
1558
  CDROM, custom disk type for CDROMs and a floppy image
1559
- Removed redundant message in instance rename when the name is given
1560
  already as a FQDN
1561
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1562
  disks on new nodes, allowing recreation when the original instance
1563
  nodes are completely gone
1564
- Added option when converting disk templates to DRBD to skip waiting
1565
  for the resync, in order to make the instance available sooner
1566
- Added two new variables to the OS scripts environment (containing the
1567
  instance's nodes)
1568
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1569
  to allow use of ``pvgrub`` as bootloader
1570
- Changed the instance memory modifications to only check out-of-memory
1571
  conditions on memory increases, and turned the secondary node warnings
1572
  into errors (they can still be overridden via ``--force``)
1573
- Fixed the handling of a corner case when the Python installation gets
1574
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1575
  to execute a command that doesn't exist
1576
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1577
  node of the instance returned failures during instance shutdown; this
1578
  adds the option ``--ignore-consistency`` to gnt-instance move
1579

    
1580
And as usual, various improvements to the error messages, documentation
1581
and man pages.
1582

    
1583

    
1584
Version 2.4.1
1585
-------------
1586

    
1587
*(Released Wed, 09 Mar 2011)*
1588

    
1589
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1590
the RAPI users file if run twice (even with ``--dry-run``).
1591

    
1592
The release fixes that bug (nothing else changed).
1593

    
1594

    
1595
Version 2.4.0
1596
-------------
1597

    
1598
*(Released Mon, 07 Mar 2011)*
1599

    
1600
Final 2.4.0 release. Just a few small fixes:
1601

    
1602
- Fixed RAPI node evacuate
1603
- Fixed the kvm-ifup script
1604
- Fixed internal error handling for special job cases
1605
- Updated man page to specify the escaping feature for options
1606

    
1607

    
1608
Version 2.4.0 rc3
1609
-----------------
1610

    
1611
*(Released Mon, 28 Feb 2011)*
1612

    
1613
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1614
that this feature wiped the disks of the wrong instance, leading to loss
1615
of data.
1616

    
1617
Other changes:
1618

    
1619
- Fixed title of query field containing instance name
1620
- Expanded the glossary in the documentation
1621
- Fixed one unittest (internal issue)
1622

    
1623

    
1624
Version 2.4.0 rc2
1625
-----------------
1626

    
1627
*(Released Mon, 21 Feb 2011)*
1628

    
1629
A number of bug fixes plus just a couple functionality changes.
1630

    
1631
On the user-visible side, the ``gnt-* list`` command output has changed
1632
with respect to "special" field states. The current rc1 style of display
1633
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1634
the default output mode special fields states are displayed as follows:
1635

    
1636
- Offline resource: ``*``
1637
- Unavailable/not applicable: ``-``
1638
- Data missing (RPC failure): ``?``
1639
- Unknown field: ``??``
1640

    
1641
Another user-visible change is the addition of ``--force-join`` to
1642
``gnt-node add``.
1643

    
1644
As for bug fixes:
1645

    
1646
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1647
- Fixed regression in RAPI/instance reinstall where all parameters were
1648
  required (instead of optional)
1649
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1650
- Fixed iallocator usage (offline nodes were not considered offline)
1651
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1652
- Fixed hypervisor and OS parameter validation with respect to
1653
  non-vm_capable nodes
1654
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1655
  cosmetic)
1656
- Fixed ``tools/listrunner`` with respect to agent-based usage
1657

    
1658

    
1659
Version 2.4.0 rc1
1660
-----------------
1661

    
1662
*(Released Fri,  4 Feb 2011)*
1663

    
1664
Many changes and fixes since the beta1 release. While there were some
1665
internal changes, the code has been mostly stabilised for the RC
1666
release.
1667

    
1668
Note: the dumb allocator was removed in this release, as it was not kept
1669
up-to-date with the IAllocator protocol changes. It is recommended to
1670
use the ``hail`` command from the ganeti-htools package.
1671

    
1672
Note: the 2.4 and up versions of Ganeti are not compatible with the
1673
0.2.x branch of ganeti-htools. You need to upgrade to
1674
ganeti-htools-0.3.0 (or later).
1675

    
1676
Regressions fixed from 2.3
1677
~~~~~~~~~~~~~~~~~~~~~~~~~~
1678

    
1679
- Fixed the ``gnt-cluster verify-disks`` command
1680
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1681
  serially on nodes)
1682
- Fixed disk adoption breakage
1683
- Fixed wrong headers in instance listing for field aliases
1684

    
1685
Other bugs fixed
1686
~~~~~~~~~~~~~~~~
1687

    
1688
- Fixed corner case in KVM handling of NICs
1689
- Fixed many cases of wrong handling of non-vm_capable nodes
1690
- Fixed a bug where a missing instance symlink was not possible to
1691
  recreate with any ``gnt-*`` command (now ``gnt-instance
1692
  activate-disks`` does it)
1693
- Fixed the volume group name as reported by ``gnt-cluster
1694
  verify-disks``
1695
- Increased timeouts for the import-export code, hopefully leading to
1696
  fewer aborts due network or instance timeouts
1697
- Fixed bug in ``gnt-node list-storage``
1698
- Fixed bug where not all daemons were started on cluster
1699
  initialisation, but only at the first watcher run
1700
- Fixed many bugs in the OOB implementation
1701
- Fixed watcher behaviour in presence of instances with offline
1702
  secondaries
1703
- Fixed instance list output for instances running on the wrong node
1704
- a few fixes to the cluster-merge tool, but it still cannot merge
1705
  multi-node groups (currently it is not recommended to use this tool)
1706

    
1707

    
1708
Improvements
1709
~~~~~~~~~~~~
1710

    
1711
- Improved network configuration for the KVM hypervisor
1712
- Added e1000 as a supported NIC for Xen-HVM
1713
- Improved the lvmstrap tool to also be able to use partitions, as
1714
  opposed to full disks
1715
- Improved speed of disk wiping (the cluster parameter
1716
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1717
  of instance creations
1718
- Added documentation for the OS parameters
1719
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1720
  hypervisor is not responding
1721
- Added display of blacklisted and hidden OS information in
1722
  ``gnt-cluster info``
1723
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1724
  NIC and node parameters, which might create problems with currently
1725
  invalid (but undetected) configuration files, but prevents validation
1726
  failures when unrelated parameters are modified
1727
- Changed cluster initialisation to wait for the master daemon to become
1728
  available
1729
- Expanded the RAPI interface:
1730

    
1731
  - Added config redistribution resource
1732
  - Added activation/deactivation of instance disks
1733
  - Added export of console information
1734

    
1735
- Implemented log file reopening on SIGHUP, which allows using
1736
  logrotate(8) for the Ganeti log files
1737
- Added a basic OOB helper script as an example
1738

    
1739

    
1740
Version 2.4.0 beta1
1741
-------------------
1742

    
1743
*(Released Fri, 14 Jan 2011)*
1744

    
1745
User-visible
1746
~~~~~~~~~~~~
1747

    
1748
- Fixed timezone issues when formatting timestamps
1749
- Added support for node groups, available via ``gnt-group`` and other
1750
  commands
1751
- Added out-of-band framework and management, see :doc:`design
1752
  document <design-oob>`
1753
- Removed support for roman numbers from ``gnt-node list`` and
1754
  ``gnt-instance list``.
1755
- Allowed modification of master network interface via ``gnt-cluster
1756
  modify --master-netdev``
1757
- Accept offline secondaries while shutting down instance disks
1758
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1759
- Added support for multiple LVM volume groups
1760
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1761
  requested
1762
- Added commands to list available fields:
1763

    
1764
  - ``gnt-node list-fields``
1765
  - ``gnt-group list-fields``
1766
  - ``gnt-instance list-fields``
1767

    
1768
- Updated documentation and man pages
1769

    
1770
Integration
1771
~~~~~~~~~~~
1772

    
1773
- Moved ``rapi_users`` file into separate directory, now named
1774
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1775
  symlink
1776
- Added new tool for running commands on many machines,
1777
  ``tools/ganeti-listrunner``
1778
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1779
  improving the ``gnt-instance console`` output
1780
- Allowed customisation of disk index separator at ``configure`` time
1781
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1782
- Added support for non-partitioned md disks in ``lvmstrap``
1783
- Added script to gracefully power off KVM instances
1784
- Split ``utils`` module into smaller parts
1785
- Changed query operations to return more detailed information, e.g.
1786
  whether an information is unavailable due to an offline node. To use
1787
  this new functionality, the LUXI call ``Query`` must be used. Field
1788
  information is now stored by the master daemon and can be retrieved
1789
  using ``QueryFields``. Instances, nodes and groups can also be queried
1790
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1791
  exposed via RAPI). The following commands make use of this
1792
  infrastructure change:
1793

    
1794
  - ``gnt-group list``
1795
  - ``gnt-group list-fields``
1796
  - ``gnt-node list``
1797
  - ``gnt-node list-fields``
1798
  - ``gnt-instance list``
1799
  - ``gnt-instance list-fields``
1800
  - ``gnt-debug locks``
1801

    
1802
Remote API
1803
~~~~~~~~~~
1804

    
1805
- New RAPI resources (see :doc:`rapi`):
1806

    
1807
  - ``/2/modify``
1808
  - ``/2/groups``
1809
  - ``/2/groups/[group_name]``
1810
  - ``/2/groups/[group_name]/assign-nodes``
1811
  - ``/2/groups/[group_name]/modify``
1812
  - ``/2/groups/[group_name]/rename``
1813
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1814

    
1815
- RAPI changes:
1816

    
1817
  - Implemented ``no_install`` for instance creation
1818
  - Implemented OS parameters for instance reinstallation, allowing
1819
    use of special settings on reinstallation (e.g. for preserving data)
1820

    
1821
Misc
1822
~~~~
1823

    
1824
- Added IPv6 support in import/export
1825
- Pause DRBD synchronization while wiping disks on instance creation
1826
- Updated unittests and QA scripts
1827
- Improved network parameters passed to KVM
1828
- Converted man pages from docbook to reStructuredText
1829

    
1830

    
1831
Version 2.3.1
1832
-------------
1833

    
1834
*(Released Mon, 20 Dec 2010)*
1835

    
1836
Released version 2.3.1~rc1 without any changes.
1837

    
1838

    
1839
Version 2.3.1 rc1
1840
-----------------
1841

    
1842
*(Released Wed, 1 Dec 2010)*
1843

    
1844
- impexpd: Disable OpenSSL compression in socat if possible (backport
1845
  from master, commit e90739d625b, see :doc:`installation guide
1846
  <install-quick>` for details)
1847
- Changed unittest coverage report to exclude test scripts
1848
- Added script to check version format
1849

    
1850

    
1851
Version 2.3.0
1852
-------------
1853

    
1854
*(Released Wed, 1 Dec 2010)*
1855

    
1856
Released version 2.3.0~rc1 without any changes.
1857

    
1858

    
1859
Version 2.3.0 rc1
1860
-----------------
1861

    
1862
*(Released Fri, 19 Nov 2010)*
1863

    
1864
A number of bugfixes and documentation updates:
1865

    
1866
- Update ganeti-os-interface documentation
1867
- Fixed a bug related to duplicate MACs or similar items which should be
1868
  unique
1869
- Fix breakage in OS state modify
1870
- Reinstall instance: disallow offline secondaries (fixes bug related to
1871
  OS changing but reinstall failing)
1872
- plus all the other fixes between 2.2.1 and 2.2.2
1873

    
1874

    
1875
Version 2.3.0 rc0
1876
-----------------
1877

    
1878
*(Released Tue, 2 Nov 2010)*
1879

    
1880
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1881
- Fixed master failover race with watcher
1882
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1883
  configuration
1884
- Accept previously stopped instance for export with instance removal
1885
- Simplify and extend the environment variables for instance OS scripts
1886
- Added new node flags, ``master_capable`` and ``vm_capable``
1887
- Added optional instance disk wiping prior during allocation. This is a
1888
  cluster-wide option and can be set/modified using
1889
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1890
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1891
  :doc:`install-quick`
1892
- Added a new watcher option (``--ignore-pause``)
1893
- Added option to ignore offline node on instance start/stop
1894
  (``--ignore-offline``)
1895
- Allow overriding OS parameters with ``gnt-instance reinstall``
1896
- Added ability to change node's secondary IP address using ``gnt-node
1897
  modify``
1898
- Implemented privilege separation for all daemons except
1899
  ``ganeti-noded``, see ``configure`` options
1900
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1901
  verify``
1902
- Implemented job priorities (see ``ganeti(7)`` manpage)
1903
- Ignore failures while shutting down instances during failover from
1904
  offline node
1905
- Exit daemon's bootstrap process only once daemon is ready
1906
- Export more information via ``LUInstanceQuery``/remote API
1907
- Improved documentation, QA and unittests
1908
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1909
  restart if the file was created or changed
1910
- Added LUXI protocol version sent with each request and response,
1911
  allowing detection of server/client mismatches
1912
- Moved the Python scripts among gnt-* and ganeti-* into modules
1913
- Moved all code related to setting up SSH to an external script,
1914
  ``setup-ssh``
1915
- Infrastructure changes for node group support in future versions
1916

    
1917

    
1918
Version 2.2.2
1919
-------------
1920

    
1921
*(Released Fri, 19 Nov 2010)*
1922

    
1923
A few small bugs fixed, and some improvements to the build system:
1924

    
1925
- Fix documentation regarding conversion to drbd
1926
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1927
  -B``)
1928
- Fix error handling in node modify with multiple changes
1929
- Allow remote imports without checked names
1930

    
1931

    
1932
Version 2.2.1
1933
-------------
1934

    
1935
*(Released Tue, 19 Oct 2010)*
1936

    
1937
- Disable SSL session ID cache in RPC client
1938

    
1939

    
1940
Version 2.2.1 rc1
1941
-----------------
1942

    
1943
*(Released Thu, 14 Oct 2010)*
1944

    
1945
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1946
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1947
  with GnuTLS
1948
- Fix problems with interaction between Curl and Python's HTTP server,
1949
  resulting in increased speed in many RPC calls
1950
- Improve our release script to prevent breakage with older aclocal and
1951
  Python 2.6
1952

    
1953

    
1954
Version 2.2.1 rc0
1955
-----------------
1956

    
1957
*(Released Thu, 7 Oct 2010)*
1958

    
1959
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1960
  value retrieved from master
1961
- Added support for blacklisted or hidden OS definitions
1962
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1963
- Added support for -mem-path in KVM hypervisor abstraction layer
1964
- Allow overriding instance parameters in tool for inter-cluster
1965
  instance moves (``tools/move-instance``)
1966
- Improved opcode summaries (e.g. in ``gnt-job list``)
1967
- Improve consistency of OS listing by sorting it
1968
- Documentation updates
1969

    
1970

    
1971
Version 2.2.0.1
1972
---------------
1973

    
1974
*(Released Fri, 8 Oct 2010)*
1975

    
1976
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1977

    
1978

    
1979
Version 2.2.0
1980
-------------
1981

    
1982
*(Released Mon, 4 Oct 2010)*
1983

    
1984
- Fixed regression in ``gnt-instance rename``
1985

    
1986

    
1987
Version 2.2.0 rc2
1988
-----------------
1989

    
1990
*(Released Wed, 22 Sep 2010)*
1991

    
1992
- Fixed OS_VARIANT variable for OS scripts
1993
- Fixed cluster tag operations via RAPI
1994
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1995
- Disabled RAPI CA checks in watcher
1996

    
1997

    
1998
Version 2.2.0 rc1
1999
-----------------
2000

    
2001
*(Released Mon, 23 Aug 2010)*
2002

    
2003
- Support DRBD versions of the format "a.b.c.d"
2004
- Updated manpages
2005
- Re-introduce support for usage from multiple threads in RAPI client
2006
- Instance renames and modify via RAPI
2007
- Work around race condition between processing and archival in job
2008
  queue
2009
- Mark opcodes following failed one as failed, too
2010
- Job field ``lock_status`` was removed due to difficulties making it
2011
  work with the changed job queue in Ganeti 2.2; a better way to monitor
2012
  locks is expected for a later 2.2.x release
2013
- Fixed dry-run behaviour with many commands
2014
- Support ``ssh-agent`` again when adding nodes
2015
- Many additional bugfixes
2016

    
2017

    
2018
Version 2.2.0 rc0
2019
-----------------
2020

    
2021
*(Released Fri, 30 Jul 2010)*
2022

    
2023
Important change: the internal RPC mechanism between Ganeti nodes has
2024
changed from using a home-grown http library (based on the Python base
2025
libraries) to use the PycURL library. This requires that PycURL is
2026
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
2027
against GnuTLS by default. cURL's support for GnuTLS had known issues
2028
before cURL 7.21.0 and we recommend using the latest cURL release or
2029
linking against OpenSSL. Most other distributions already link PycURL
2030
and cURL against OpenSSL. The command::
2031

    
2032
  python -c 'import pycurl; print pycurl.version'
2033

    
2034
can be used to determine the libraries PycURL and cURL are linked
2035
against.
2036

    
2037
Other significant changes:
2038

    
2039
- Rewrote much of the internals of the job queue, in order to achieve
2040
  better parallelism; this decouples job query operations from the job
2041
  processing, and it should allow much nicer behaviour of the master
2042
  daemon under load, and it also has uncovered some long-standing bugs
2043
  related to the job serialisation (now fixed)
2044
- Added a default iallocator setting to the cluster parameters,
2045
  eliminating the need to always pass nodes or an iallocator for
2046
  operations that require selection of new node(s)
2047
- Added experimental support for the LXC virtualization method
2048
- Added support for OS parameters, which allows the installation of
2049
  instances to pass parameter to OS scripts in order to customise the
2050
  instance
2051
- Added a hypervisor parameter controlling the migration type (live or
2052
  non-live), since hypervisors have various levels of reliability; this
2053
  has renamed the 'live' parameter to 'mode'
2054
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
2055
  logical volumes, meaning that cluster verify will ignore them and not
2056
  flag their presence as errors
2057
- The watcher will now reset the error count for failed instances after
2058
  8 hours, thus allowing self-healing if the problem that caused the
2059
  instances to be down/fail to start has cleared in the meantime
2060
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
2061
  check for, and warn, if the drbd module parameter ``usermode_helper``
2062
  is not consistent with the cluster-wide setting; this is needed to
2063
  make diagnose easier of failed drbd creations
2064
- Started adding base IPv6 support, but this is not yet
2065
  enabled/available for use
2066
- Rename operations (cluster, instance) will now return the new name,
2067
  which is especially useful if a short name was passed in
2068
- Added support for instance migration in RAPI
2069
- Added a tool to pre-configure nodes for the SSH setup, before joining
2070
  them to the cluster; this will allow in the future a simplified model
2071
  for node joining (but not yet fully enabled in 2.2); this needs the
2072
  paramiko python library
2073
- Fixed handling of name-resolving errors
2074
- Fixed consistency of job results on the error path
2075
- Fixed master-failover race condition when executed multiple times in
2076
  sequence
2077
- Fixed many bugs related to the job queue (mostly introduced during the
2078
  2.2 development cycle, so not all are impacting 2.1)
2079
- Fixed instance migration with missing disk symlinks
2080
- Fixed handling of unknown jobs in ``gnt-job archive``
2081
- And many other small fixes/improvements
2082

    
2083
Internal changes:
2084

    
2085
- Enhanced both the unittest and the QA coverage
2086
- Switched the opcode validation to a generic model, and extended the
2087
  validation to all opcode parameters
2088
- Changed more parts of the code that write shell scripts to use the
2089
  same class for this
2090
- Switched the master daemon to use the asyncore library for the Luxi
2091
  server endpoint
2092

    
2093

    
2094
Version 2.2.0 beta0
2095
-------------------
2096

    
2097
*(Released Thu, 17 Jun 2010)*
2098

    
2099
- Added tool (``move-instance``) and infrastructure to move instances
2100
  between separate clusters (see :doc:`separate documentation
2101
  <move-instance>` and :doc:`design document <design-2.2>`)
2102
- Added per-request RPC timeout
2103
- RAPI now requires a Content-Type header for requests with a body (e.g.
2104
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2105
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
2106
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2107
  reachable
2108
- Implemented initial support for running Ganeti daemons as separate
2109
  users, see configure-time flags ``--with-user-prefix`` and
2110
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2111
  time)
2112
- Instances can be removed after export (``gnt-backup export
2113
  --remove-instance``)
2114
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2115
  key (instead of 1024 bit)
2116
- Added new cluster configuration file for cluster domain secret
2117
- Import/export now use SSL instead of SSH
2118
- Added support for showing estimated time when exporting an instance,
2119
  see the ``ganeti-os-interface(7)`` manpage and look for
2120
  ``EXP_SIZE_FD``
2121

    
2122

    
2123
Version 2.1.8
2124
-------------
2125

    
2126
*(Released Tue, 16 Nov 2010)*
2127

    
2128
Some more bugfixes. Unless critical bugs occur, this will be the last
2129
2.1 release:
2130

    
2131
- Fix case of MAC special-values
2132
- Fix mac checker regex
2133
- backend: Fix typo causing "out of range" error
2134
- Add missing --units in gnt-instance list man page
2135

    
2136

    
2137
Version 2.1.7
2138
-------------
2139

    
2140
*(Released Tue, 24 Aug 2010)*
2141

    
2142
Bugfixes only:
2143
  - Don't ignore secondary node silently on non-mirrored disk templates
2144
    (issue 113)
2145
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2146
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2147
  - Properly document the usb_mouse parameter
2148
  - Fix path in ganeti-rapi(8) (issue 116)
2149
  - Adjust error message when the ganeti user's .ssh directory is
2150
    missing
2151
  - Add same-node-check when changing the disk template to drbd
2152

    
2153

    
2154
Version 2.1.6
2155
-------------
2156

    
2157
*(Released Fri, 16 Jul 2010)*
2158

    
2159
Bugfixes only:
2160
  - Add an option to only select some reboot types during qa/burnin.
2161
    (on some hypervisors consequent reboots are not supported)
2162
  - Fix infrequent race condition in master failover. Sometimes the old
2163
    master ip address would be still detected as up for a short time
2164
    after it was removed, causing failover to fail.
2165
  - Decrease mlockall warnings when the ctypes module is missing. On
2166
    Python 2.4 we support running even if no ctypes module is installed,
2167
    but we were too verbose about this issue.
2168
  - Fix building on old distributions, on which man doesn't have a
2169
    --warnings option.
2170
  - Fix RAPI not to ignore the MAC address on instance creation
2171
  - Implement the old instance creation format in the RAPI client.
2172

    
2173

    
2174
Version 2.1.5
2175
-------------
2176

    
2177
*(Released Thu, 01 Jul 2010)*
2178

    
2179
A small bugfix release:
2180
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2181
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2182
    a non-existing option
2183
  - Fix instance create: the --force-variant option was ignored
2184
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
2185
  - Fix modify node storage with non-FQDN arguments
2186
  - Fix RAPI client to authenticate under Python 2.6 when used
2187
    for more than 5 requests needing authentication
2188
  - Fix gnt-instance modify -t (storage) giving a wrong error message
2189
    when converting a non-shutdown drbd instance to plain
2190

    
2191

    
2192
Version 2.1.4
2193
-------------
2194

    
2195
*(Released Fri, 18 Jun 2010)*
2196

    
2197
A small bugfix release:
2198

    
2199
  - Fix live migration of KVM instances started with older Ganeti
2200
    versions which had fewer hypervisor parameters
2201
  - Fix gnt-instance grow-disk on down instances
2202
  - Fix an error-reporting bug during instance migration
2203
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
2204
    silently ignoring broken ones
2205
  - Fix an RPC error reporting bug affecting, for example, RAPI client
2206
    users
2207
  - Fix bug triggered by different API version os-es on different nodes
2208
  - Fix a bug in instance startup with custom hvparams: OS level
2209
    parameters would fail to be applied.
2210
  - Fix the RAPI client under Python 2.6 (but more work is needed to
2211
    make it work completely well with OpenSSL)
2212
  - Fix handling of errors when resolving names from DNS
2213

    
2214

    
2215
Version 2.1.3
2216
-------------
2217

    
2218
*(Released Thu, 3 Jun 2010)*
2219

    
2220
A medium sized development cycle. Some new features, and some
2221
fixes/small improvements/cleanups.
2222

    
2223
Significant features
2224
~~~~~~~~~~~~~~~~~~~~
2225

    
2226
The node deamon now tries to mlock itself into memory, unless the
2227
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2228
its logs, and falls back to console logging. This allows emergency
2229
features such as ``gnt-node powercycle`` to work even in the event of a
2230
broken node disk (tested offlining the disk hosting the node's
2231
filesystem and dropping its memory caches; don't try this at home)
2232

    
2233
KVM: add vhost-net acceleration support. It can be tested with a new
2234
enough version of the kernel and of qemu-kvm.
2235

    
2236
KVM: Add instance chrooting feature. If you use privilege dropping for
2237
your VMs you can also now force them to chroot to an empty directory,
2238
before starting the emulated guest.
2239

    
2240
KVM: Add maximum migration bandwith and maximum downtime tweaking
2241
support (requires a new-enough version of qemu-kvm).
2242

    
2243
Cluster verify will now warn if the master node doesn't have the master
2244
ip configured on it.
2245

    
2246
Add a new (incompatible) instance creation request format to RAPI which
2247
supports all parameters (previously only a subset was supported, and it
2248
wasn't possible to extend the old format to accomodate all the new
2249
features. The old format is still supported, and a client can check for
2250
this feature, before using it, by checking for its presence in the
2251
``features`` RAPI resource.
2252

    
2253
Now with ancient latin support. Try it passing the ``--roman`` option to
2254
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2255
(requires the python-roman module to be installed, in order to work).
2256

    
2257
Other changes
2258
~~~~~~~~~~~~~
2259

    
2260
As usual many internal code refactorings, documentation updates, and
2261
such. Among others:
2262

    
2263
  - Lots of improvements and cleanups to the experimental Remote API
2264
    (RAPI) client library.
2265
  - A new unit test suite for the core daemon libraries.
2266
  - A fix to creating missing directories makes sure the umask is not
2267
    applied anymore. This enforces the same directory permissions
2268
    everywhere.
2269
  - Better handling terminating daemons with ctrl+c (used when running
2270
    them in debugging mode).
2271
  - Fix a race condition in live migrating a KVM instance, when stat()
2272
    on the old proc status file returned EINVAL, which is an unexpected
2273
    value.
2274
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2275
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2276

    
2277

    
2278
Version 2.1.2.1
2279
---------------
2280

    
2281
*(Released Fri, 7 May 2010)*
2282

    
2283
Fix a bug which prevented untagged KVM instances from starting.
2284

    
2285

    
2286
Version 2.1.2
2287
-------------
2288

    
2289
*(Released Fri, 7 May 2010)*
2290

    
2291
Another release with a long development cycle, during which many
2292
different features were added.
2293

    
2294
Significant features
2295
~~~~~~~~~~~~~~~~~~~~
2296

    
2297
The KVM hypervisor now can run the individual instances as non-root, to
2298
reduce the impact of a VM being hijacked due to bugs in the
2299
hypervisor. It is possible to run all instances as a single (non-root)
2300
user, to manually specify a user for each instance, or to dynamically
2301
allocate a user out of a cluster-wide pool to each instance, with the
2302
guarantee that no two instances will run under the same user ID on any
2303
given node.
2304

    
2305
An experimental RAPI client library, that can be used standalone
2306
(without the other Ganeti libraries), is provided in the source tree as
2307
``lib/rapi/client.py``. Note this client might change its interface in
2308
the future, as we iterate on its capabilities.
2309

    
2310
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2311
replace the cluster's certificates and crypto keys. This might help in
2312
case they have been compromised, or have simply expired.
2313

    
2314
A new disk option for instance creation has been added that allows one
2315
to "adopt" currently existing logical volumes, with data
2316
preservation. This should allow easier migration to Ganeti from
2317
unmanaged (or managed via other software) instances.
2318

    
2319
Another disk improvement is the possibility to convert between redundant
2320
(DRBD) and plain (LVM) disk configuration for an instance. This should
2321
allow better scalability (starting with one node and growing the
2322
cluster, or shrinking a two-node cluster to one node).
2323

    
2324
A new feature that could help with automated node failovers has been
2325
implemented: if a node sees itself as offline (by querying the master
2326
candidates), it will try to shutdown (hard) all instances and any active
2327
DRBD devices. This reduces the risk of duplicate instances if an
2328
external script automatically failovers the instances on such nodes. To
2329
enable this, the cluster parameter ``maintain_node_health`` should be
2330
enabled; in the future this option (per the name) will enable other
2331
automatic maintenance features.
2332

    
2333
Instance export/import now will reuse the original instance
2334
specifications for all parameters; that means exporting an instance,
2335
deleting it and the importing it back should give an almost identical
2336
instance. Note that the default import behaviour has changed from
2337
before, where it created only one NIC; now it recreates the original
2338
number of NICs.
2339

    
2340
Cluster verify has added a few new checks: SSL certificates validity,
2341
/etc/hosts consistency across the cluster, etc.
2342

    
2343
Other changes
2344
~~~~~~~~~~~~~
2345

    
2346
As usual, many internal changes were done, documentation fixes,
2347
etc. Among others:
2348

    
2349
- Fixed cluster initialization with disabled cluster storage (regression
2350
  introduced in 2.1.1)
2351
- File-based storage supports growing the disks
2352
- Fixed behaviour of node role changes
2353
- Fixed cluster verify for some corner cases, plus a general rewrite of
2354
  cluster verify to allow future extension with more checks
2355
- Fixed log spamming by watcher and node daemon (regression introduced
2356
  in 2.1.1)
2357
- Fixed possible validation issues when changing the list of enabled
2358
  hypervisors
2359
- Fixed cleanup of /etc/hosts during node removal
2360
- Fixed RAPI response for invalid methods
2361
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2362
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2363
  from ide disks, etc.)
2364
- Allow OS changes without re-installation (to record a changed OS
2365
  outside of Ganeti, or to allow OS renames)
2366
- Allow instance creation without OS installation (useful for example if
2367
  the OS will be installed manually, or restored from a backup not in
2368
  Ganeti format)
2369
- Implemented option to make cluster ``copyfile`` use the replication
2370
  network
2371
- Added list of enabled hypervisors to ssconf (possibly useful for
2372
  external scripts)
2373
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2374
  1.2 clusters
2375
- A partial form of node re-IP is possible via node readd, which now
2376
  allows changed node primary IP
2377
- Command line utilities now show an informational message if the job is
2378
  waiting for a lock
2379
- The logs of the master daemon now show the PID/UID/GID of the
2380
  connected client
2381

    
2382

    
2383
Version 2.1.1
2384
-------------
2385

    
2386
*(Released Fri, 12 Mar 2010)*
2387

    
2388
During the 2.1.0 long release candidate cycle, a lot of improvements and
2389
changes have accumulated with were released later as 2.1.1.
2390

    
2391
Major changes
2392
~~~~~~~~~~~~~
2393

    
2394
The node evacuate command (``gnt-node evacuate``) was significantly
2395
rewritten, and as such the IAllocator protocol was changed - a new
2396
request type has been added. This unfortunate change during a stable
2397
series is designed to improve performance of node evacuations; on
2398
clusters with more than about five nodes and which are well-balanced,
2399
evacuation should proceed in parallel for all instances of the node
2400
being evacuated. As such, any existing IAllocator scripts need to be
2401
updated, otherwise the above command will fail due to the unknown
2402
request. The provided "dumb" allocator has not been updated; but the
2403
ganeti-htools package supports the new protocol since version 0.2.4.
2404

    
2405
Another important change is increased validation of node and instance
2406
names. This might create problems in special cases, if invalid host
2407
names are being used.
2408

    
2409
Also, a new layer of hypervisor parameters has been added, that sits at
2410
OS level between the cluster defaults and the instance ones. This allows
2411
customisation of virtualization parameters depending on the installed
2412
OS. For example instances with OS 'X' may have a different KVM kernel
2413
(or any other parameter) than the cluster defaults. This is intended to
2414
help managing a multiple OSes on the same cluster, without manual
2415
modification of each instance's parameters.
2416

    
2417
A tool for merging clusters, ``cluster-merge``, has been added in the
2418
tools sub-directory.
2419

    
2420
Bug fixes
2421
~~~~~~~~~
2422

    
2423
- Improved the int/float conversions that should make the code more
2424
  robust in face of errors from the node daemons
2425
- Fixed the remove node code in case of internal configuration errors
2426
- Fixed the node daemon behaviour in face of inconsistent queue
2427
  directory (e.g. read-only file-system where we can't open the files
2428
  read-write, etc.)
2429
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2430
  now it either aborts cleanly or, if given the new "auto_promote"
2431
  parameter, will automatically promote other nodes as needed
2432
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2433
  completely prevent Ganeti from working
2434
- Fixed bug for instance export when not all disks were successfully
2435
  exported
2436
- Fixed behaviour of node add when the new node is slow in starting up
2437
  the node daemon
2438
- Fixed handling of signals in the LUXI client, which should improve
2439
  behaviour of command-line scripts
2440
- Added checks for invalid node/instance names in the configuration (now
2441
  flagged during cluster verify)
2442
- Fixed watcher behaviour for disk activation errors
2443
- Fixed two potentially endless loops in http library, which led to the
2444
  RAPI daemon hanging and consuming 100% CPU in some cases
2445
- Fixed bug in RAPI daemon related to hashed passwords
2446
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2447
  instances
2448
- Enhanced compatibility with non-Debian OSes, but not using absolute
2449
  path in some commands and allowing customisation of the ssh
2450
  configuration directory
2451
- Fixed possible future issue with new Python versions by abiding to the
2452
  proper use of ``__slots__`` attribute on classes
2453
- Added checks that should prevent directory traversal attacks
2454
- Many documentation fixes based on feedback from users
2455

    
2456
New features
2457
~~~~~~~~~~~~
2458

    
2459
- Added an "early_release" more for instance replace disks and node
2460
  evacuate, where we release locks earlier and thus allow higher
2461
  parallelism within the cluster
2462
- Added watcher hooks, intended to allow the watcher to restart other
2463
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2464
  course for any other purpose
2465
- Added a compile-time disable for DRBD barriers, to increase
2466
  performance if the administrator trusts the power supply or the
2467
  storage system to not lose writes
2468
- Added the option of using syslog for logging instead of, or in
2469
  addition to, Ganeti's own log files
2470
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2471
  can indeed boot from a paravirtual NIC
2472
- Added a generic debug level for many operations; while this is not
2473
  used widely yet, it allows one to pass the debug value all the way to
2474
  the OS scripts
2475
- Enhanced the hooks environment for instance moves (failovers,
2476
  migrations) where the primary/secondary nodes changed during the
2477
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2478
- Enhanced data validations for many user-supplied values; one important
2479
  item is the restrictions imposed on instance and node names, which
2480
  might reject some (invalid) host names
2481
- Add a configure-time option to disable file-based storage, if it's not
2482
  needed; this allows greater security separation between the master
2483
  node and the other nodes from the point of view of the inter-node RPC
2484
  protocol
2485
- Added user notification in interactive tools if job is waiting in the
2486
  job queue or trying to acquire locks
2487
- Added log messages when a job is waiting for locks
2488
- Added filtering by node tags in instance operations which admit
2489
  multiple instances (start, stop, reboot, reinstall)
2490
- Added a new tool for cluster mergers, ``cluster-merge``
2491
- Parameters from command line which are of the form ``a=b,c=d`` can now
2492
  use backslash escapes to pass in values which contain commas,
2493
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2494
  ``b,c``
2495
- For KVM, the instance name is the first parameter passed to KVM, so
2496
  that it's more visible in the process list
2497

    
2498

    
2499
Version 2.1.0
2500
-------------
2501

    
2502
*(Released Tue, 2 Mar 2010)*
2503

    
2504
Ganeti 2.1 brings many improvements with it. Major changes:
2505

    
2506
- Added infrastructure to ease automated disk repairs
2507
- Added new daemon to export configuration data in a cheaper way than
2508
  using the remote API
2509
- Instance NICs can now be routed instead of being associated with a
2510
  networking bridge
2511
- Improved job locking logic to reduce impact of jobs acquiring multiple
2512
  locks waiting for other long-running jobs
2513

    
2514
In-depth implementation details can be found in the Ganeti 2.1 design
2515
document.
2516

    
2517
Details
2518
~~~~~~~
2519

    
2520
- Added chroot hypervisor
2521
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2522
  ``device_model``)
2523
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2524
  ``bootloader_path`` and ``bootloader_args``)
2525
- Added the ``use_localtime`` option for the xen-hvm and kvm
2526
  hypervisors, and the default value for this has changed to false (in
2527
  2.0 xen-hvm always enabled it)
2528
- Added luxi call to submit multiple jobs in one go
2529
- Added cluster initialization option to not modify ``/etc/hosts``
2530
  file on nodes
2531
- Added network interface parameters
2532
- Added dry run mode to some LUs
2533
- Added RAPI resources:
2534

    
2535
  - ``/2/instances/[instance_name]/info``
2536
  - ``/2/instances/[instance_name]/replace-disks``
2537
  - ``/2/nodes/[node_name]/evacuate``
2538
  - ``/2/nodes/[node_name]/migrate``
2539
  - ``/2/nodes/[node_name]/role``
2540
  - ``/2/nodes/[node_name]/storage``
2541
  - ``/2/nodes/[node_name]/storage/modify``
2542
  - ``/2/nodes/[node_name]/storage/repair``
2543

    
2544
- Added OpCodes to evacuate or migrate all instances on a node
2545
- Added new command to list storage elements on nodes (``gnt-node
2546
  list-storage``) and modify them (``gnt-node modify-storage``)
2547
- Added new ssconf files with master candidate IP address
2548
  (``ssconf_master_candidates_ips``), node primary IP address
2549
  (``ssconf_node_primary_ips``) and node secondary IP address
2550
  (``ssconf_node_secondary_ips``)
2551
- Added ``ganeti-confd`` and a client library to query the Ganeti
2552
  configuration via UDP
2553
- Added ability to run hooks after cluster initialization and before
2554
  cluster destruction
2555
- Added automatic mode for disk replace (``gnt-instance replace-disks
2556
  --auto``)
2557
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2558
  after catastrophic data-loss
2559
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2560
  groups
2561
- Added ``gnt-instance move`` command to move instances
2562
- Added ``gnt-cluster watcher`` command to control watcher
2563
- Added ``gnt-node powercycle`` command to powercycle nodes
2564
- Added new job status field ``lock_status``
2565
- Added parseable error codes to cluster verification (``gnt-cluster
2566
  verify --error-codes``) and made output less verbose (use
2567
  ``--verbose`` to restore previous behaviour)
2568
- Added UUIDs to the main config entities (cluster, nodes, instances)
2569
- Added support for OS variants
2570
- Added support for hashed passwords in the Ganeti remote API users file
2571
  (``rapi_users``)
2572
- Added option to specify maximum timeout on instance shutdown
2573
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2574
- Added new helper script to start and stop Ganeti daemons
2575
  (``daemon-util``), with the intent to reduce the work necessary to
2576
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2577
  from one place
2578
- Added more unittests
2579
- Fixed critical bug in ganeti-masterd startup
2580
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2581
  now customisable at the cluster level for both the KVM and Xen
2582
  hypervisors using the new ``migration_port`` parameter
2583
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2584
  reinstalling an instance
2585
- Allowed ``@`` in tag names
2586
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2587
- Many documentation updates
2588
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2589
- ``gnt-instance reinstall`` can now reinstall multiple instances
2590
- Updated many command line parameters
2591
- Introduced new OS API version 15
2592
- No longer support a default hypervisor
2593
- Treat virtual LVs as inexistent
2594
- Improved job locking logic to reduce lock contention
2595
- Match instance and node names case insensitively
2596
- Reimplemented bash completion script to be more complete
2597
- Improved burnin
2598

    
2599

    
2600
Version 2.0.6
2601
-------------
2602

    
2603
*(Released Thu, 4 Feb 2010)*
2604

    
2605
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2606
- Fix a string formatting bug
2607
- Improve safety of the code in some error paths
2608
- Improve data validation in the master of values returned from nodes
2609

    
2610

    
2611
Version 2.0.5
2612
-------------
2613

    
2614
*(Released Thu, 17 Dec 2009)*
2615

    
2616
- Fix security issue due to missing validation of iallocator names; this
2617
  allows local and remote execution of arbitrary executables
2618
- Fix failure of gnt-node list during instance removal
2619
- Ship the RAPI documentation in the archive
2620

    
2621

    
2622
Version 2.0.4
2623
-------------
2624

    
2625
*(Released Wed, 30 Sep 2009)*
2626

    
2627
- Fixed many wrong messages
2628
- Fixed a few bugs related to the locking library
2629
- Fixed MAC checking at instance creation time
2630
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2631
- Fixed a few issues related to signal handling in both daemons and
2632
  scripts
2633
- Fixed the example startup script provided
2634
- Fixed insserv dependencies in the example startup script (patch from
2635
  Debian)
2636
- Fixed handling of drained nodes in the iallocator framework
2637
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2638
  #528618)
2639
- Fixed error related to invalid job IDs in job polling
2640
- Fixed job/opcode persistence on unclean master shutdown
2641
- Fixed handling of partial job processing after unclean master
2642
  shutdown
2643
- Fixed error reporting from LUs, previously all errors were converted
2644
  into execution errors
2645
- Fixed error reporting from burnin
2646
- Decreased significantly the memory usage of the job queue
2647
- Optimised slightly multi-job submission
2648
- Optimised slightly opcode loading
2649
- Backported the multi-job submit framework from the development
2650
  branch; multi-instance start and stop should be faster
2651
- Added script to clean archived jobs after 21 days; this will reduce
2652
  the size of the queue directory
2653
- Added some extra checks in disk size tracking
2654
- Added an example ethers hook script
2655
- Added a cluster parameter that prevents Ganeti from modifying of
2656
  /etc/hosts
2657
- Added more node information to RAPI responses
2658
- Added a ``gnt-job watch`` command that allows following the ouput of a
2659
  job
2660
- Added a bind-address option to ganeti-rapi
2661
- Added more checks to the configuration verify
2662
- Enhanced the burnin script such that some operations can be retried
2663
  automatically
2664
- Converted instance reinstall to multi-instance model
2665

    
2666

    
2667
Version 2.0.3
2668
-------------
2669

    
2670
*(Released Fri, 7 Aug 2009)*
2671

    
2672
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2673
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2674
  instances have mismatched disk sizes in the configuration
2675
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2676
  any configuration mismatches for disk sizes
2677
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2678
  failover to work on two-node clusters
2679
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2680
  unusable
2681
- Fixed detection of OS script errors in ``gnt-backup export``
2682
- Fixed exit code of ``gnt-backup export``
2683

    
2684

    
2685
Version 2.0.2
2686
-------------
2687

    
2688
*(Released Fri, 17 Jul 2009)*
2689

    
2690
- Added experimental support for stripped logical volumes; this should
2691
  enhance performance but comes with a higher complexity in the block
2692
  device handling; stripping is only enabled when passing
2693
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2694
  affected even in the non-stripped mode
2695
- Improved resiliency against transient failures at the end of DRBD
2696
  resyncs, and in general of DRBD resync checks
2697
- Fixed a couple of issues with exports and snapshot errors
2698
- Fixed a couple of issues in instance listing
2699
- Added display of the disk size in ``gnt-instance info``
2700
- Fixed checking for valid OSes in instance creation
2701
- Fixed handling of the "vcpus" parameter in instance listing and in
2702
  general of invalid parameters
2703
- Fixed http server library, and thus RAPI, to handle invalid
2704
  username/password combinations correctly; this means that now they
2705
  report unauthorized for queries too, not only for modifications,
2706
  allowing earlier detect of configuration problems
2707
- Added a new "role" node list field, equivalent to the master/master
2708
  candidate/drained/offline flags combinations
2709
- Fixed cluster modify and changes of candidate pool size
2710
- Fixed cluster verify error messages for wrong files on regular nodes
2711
- Fixed a couple of issues with node demotion from master candidate role
2712
- Fixed node readd issues
2713
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2714
- Added a new ``--no-voting`` option for masterfailover to fix failover
2715
  on two-nodes clusters when the former master node is unreachable
2716
- Added instance reinstall over RAPI
2717

    
2718

    
2719
Version 2.0.1
2720
-------------
2721

    
2722
*(Released Tue, 16 Jun 2009)*
2723

    
2724
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2725
  allow re-adding the start in single-user option (regression from 1.2)
2726
- the watcher writes the instance status to a file, to allow monitoring
2727
  to report the instance status (from the master) based on cached
2728
  results of the watcher's queries; while this can get stale if the
2729
  watcher is being locked due to other work on the cluster, this is
2730
  still an improvement
2731
- the watcher now also restarts the node daemon and the rapi daemon if
2732
  they died
2733
- fixed the watcher to handle full and drained queue cases
2734
- hooks export more instance data in the environment, which helps if
2735
  hook scripts need to take action based on the instance's properties
2736
  (no longer need to query back into ganeti)
2737
- instance failovers when the instance is stopped do not check for free
2738
  RAM, so that failing over a stopped instance is possible in low memory
2739
  situations
2740
- rapi uses queries for tags instead of jobs (for less job traffic), and
2741
  for cluster tags it won't talk to masterd at all but read them from
2742
  ssconf
2743
- a couple of error handling fixes in RAPI
2744
- drbd handling: improved the error handling of inconsistent disks after
2745
  resync to reduce the frequency of "there are some degraded disks for
2746
  this instance" messages
2747
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2748
  error handling path called a wrong function name)
2749

    
2750

    
2751
Version 2.0.0
2752
-------------
2753

    
2754
*(Released Wed, 27 May 2009)*
2755

    
2756
- no changes from rc5
2757

    
2758

    
2759
Version 2.0 rc5
2760
---------------
2761

    
2762
*(Released Wed, 20 May 2009)*
2763

    
2764
- fix a couple of bugs (validation, argument checks)
2765
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2766
- some small improvements to RAPI and IAllocator
2767
- make watcher automatically start the master daemon if down
2768

    
2769

    
2770
Version 2.0 rc4
2771
---------------
2772

    
2773
*(Released Mon, 27 Apr 2009)*
2774

    
2775
- change the OS list to not require locks; this helps with big clusters
2776
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2777
  volume group is broken
2778
- ``gnt-instance info``, without any arguments, doesn't run for all
2779
  instances anymore; either pass ``--all`` or pass the desired
2780
  instances; this helps against mistakes on big clusters where listing
2781
  the information for all instances takes a long time
2782
- miscellaneous doc and man pages fixes
2783

    
2784

    
2785
Version 2.0 rc3
2786
---------------
2787

    
2788
*(Released Wed, 8 Apr 2009)*
2789

    
2790
- Change the internal locking model of some ``gnt-node`` commands, in
2791
  order to reduce contention (and blocking of master daemon) when
2792
  batching many creation/reinstall jobs
2793
- Fixes to Xen soft reboot
2794
- No longer build documentation at build time, instead distribute it in
2795
  the archive, in order to reduce the need for the whole docbook/rst
2796
  toolchains
2797

    
2798

    
2799
Version 2.0 rc2
2800
---------------
2801

    
2802
*(Released Fri, 27 Mar 2009)*
2803

    
2804
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2805
- Fix watcher startup sequence, improves the behaviour of busy clusters
2806
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2807
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2808
  bind address and other places
2809
- Some documentation fixes and updates
2810

    
2811

    
2812
Version 2.0 rc1
2813
---------------
2814

    
2815
*(Released Mon, 2 Mar 2009)*
2816

    
2817
- More documentation updates, now all docs should be more-or-less
2818
  up-to-date
2819
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2820
  etc.)
2821
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2822
  and KVM)
2823
- Fix an issue related to $libdir/run/ganeti and cluster creation
2824

    
2825

    
2826
Version 2.0 beta2
2827
-----------------
2828

    
2829
*(Released Thu, 19 Feb 2009)*
2830

    
2831
- Xen PVM and KVM have switched the default value for the instance root
2832
  disk to the first partition on the first drive, instead of the whole
2833
  drive; this means that the OS installation scripts must be changed
2834
  accordingly
2835
- Man pages have been updated
2836
- RAPI has been switched by default to HTTPS, and the exported functions
2837
  should all work correctly
2838
- RAPI v1 has been removed
2839
- Many improvements to the KVM hypervisor
2840
- Block device errors are now better reported
2841
- Many other bugfixes and small improvements
2842

    
2843

    
2844
Version 2.0 beta1
2845
-----------------
2846

    
2847
*(Released Mon, 26 Jan 2009)*
2848

    
2849
- Version 2 is a general rewrite of the code and therefore the
2850
  differences are too many to list, see the design document for 2.0 in
2851
  the ``doc/`` subdirectory for more details
2852
- In this beta version there is not yet a migration path from 1.2 (there
2853
  will be one in the final 2.0 release)
2854
- A few significant changes are:
2855

    
2856
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2857
    various ``gnt-*`` commands are just front-ends to it
2858
  - all the commands are entered into, and executed from a job queue,
2859
    see the ``gnt-job(8)`` manpage
2860
  - the RAPI daemon supports read-write operations, secured by basic
2861
    HTTP authentication on top of HTTPS
2862
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2863
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2864
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2865
  - DRBD devices are using statically allocated minor numbers, which
2866
    will be assigned to existing instances during the migration process
2867
  - there is support for both Xen PVM and Xen HVM instances running on
2868
    the same cluster
2869
  - KVM virtualization is supported too
2870
  - file-based storage has been implemented, which means that it is
2871
    possible to run the cluster without LVM and DRBD storage, for
2872
    example using a shared filesystem exported from shared storage (and
2873
    still have live migration)
2874

    
2875

    
2876
Version 1.2.7
2877
-------------
2878

    
2879
*(Released Tue, 13 Jan 2009)*
2880

    
2881
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2882
- Reuse the old instance mac address by default on instance import, if
2883
  the instance name is the same.
2884
- Handle situations in which the node info rpc returns incomplete
2885
  results (issue 46)
2886
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2887
- Improved version of batcher:
2888

    
2889
  - state file support
2890
  - instance mac address support
2891
  - support for HVM clusters/instances
2892

    
2893
- Add an option to show the number of cpu sockets and nodes in
2894
  ``gnt-node list``
2895
- Support OSes that handle more than one version of the OS api (but do
2896
  not change the current API in any other way)
2897
- Fix ``gnt-node migrate``
2898
- ``gnt-debug`` man page
2899
- Fixes various more typos and small issues
2900
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2901

    
2902

    
2903
Version 1.2.6
2904
-------------
2905

    
2906
*(Released Wed, 24 Sep 2008)*
2907

    
2908
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2909
  type of disk exported to fully virtualized instances.
2910
- provide access to the serial console of HVM instances
2911
- instance auto_balance flag, set by default. If turned off it will
2912
  avoid warnings on cluster verify if there is not enough memory to fail
2913
  over an instance. in the future it will prevent automatically failing
2914
  it over when we will support that.
2915
- batcher tool for instance creation, see ``tools/README.batcher``
2916
- ``gnt-instance reinstall --select-os`` to interactively select a new
2917
  operating system when reinstalling an instance.
2918
- when changing the memory amount on instance modify a check has been
2919
  added that the instance will be able to start. also warnings are
2920
  emitted if the instance will not be able to fail over, if auto_balance
2921
  is true.
2922
- documentation fixes
2923
- sync fields between ``gnt-instance list/modify/add/import``
2924
- fix a race condition in drbd when the sync speed was set after giving
2925
  the device a remote peer.
2926

    
2927

    
2928
Version 1.2.5
2929
-------------
2930

    
2931
*(Released Tue, 22 Jul 2008)*
2932

    
2933
- note: the allowed size and number of tags per object were reduced
2934
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2935
- fixed twisted 8.x compatibility
2936
- fixed ``gnt-instance replace-disks`` with iallocator
2937
- add TCP keepalives on twisted connections to detect restarted nodes
2938
- disk increase support, see ``gnt-instance grow-disk``
2939
- implement bulk node/instance query for RAPI
2940
- add tags in node/instance listing (optional)
2941
- experimental migration (and live migration) support, read the man page
2942
  for ``gnt-instance migrate``
2943
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2944
  has some small improvements in handling its state file
2945

    
2946

    
2947
Version 1.2.4
2948
-------------
2949

    
2950
*(Released Fri, 13 Jun 2008)*
2951

    
2952
- Experimental readonly, REST-based remote API implementation;
2953
  automatically started on master node, TCP port 5080, if enabled by
2954
  ``--enable-rapi`` parameter to configure script.
2955
- Instance allocator support. Add and import instance accept a
2956
  ``--iallocator`` parameter, and call that instance allocator to decide
2957
  which node to use for the instance. The iallocator document describes
2958
  what's expected from an allocator script.
2959
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2960
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2961
  node is lost there is still enough memory to fail over the instances
2962
  that reside on it.
2963
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2964
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2965
  hooks will run, and their output, if any, will be displayed. Any
2966
  failing hook will make the verification return an error value.
2967
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2968
  primary and secondary interfaces
2969
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2970
  that is still declared as part of the cluster and has failed.
2971
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2972
  specifying output fields, that just adds the chosen fields to the
2973
  default ones.
2974
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2975
  export from the filesystem.
2976
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2977
  have been added. Using them you can enable/disable acpi and pae
2978
  support, and specify a path for a cd image to be exported to the
2979
  instance. These parameters as the name suggest only work on HVM
2980
  clusters.
2981
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2982
  PAE support will be set to the previously hardcoded values, but the
2983
  (previously hardcoded) path to the CDROM ISO image will be unset and
2984
  if required, needs to be set manually with ``gnt-instance modify``
2985
  after the upgrade.
2986
- The address to which an instance's VNC console is bound is now
2987
  selectable per-instance, rather than being cluster wide. Of course
2988
  this only applies to instances controlled via VNC, so currently just
2989
  applies to HVM clusters.
2990

    
2991

    
2992
Version 1.2.3
2993
-------------
2994

    
2995
*(Released Mon, 18 Feb 2008)*
2996

    
2997
- more tweaks to the disk activation code (especially helpful for DRBD)
2998
- change the default ``gnt-instance list`` output format, now there is
2999
  one combined status field (see the manpage for the exact values this
3000
  field will have)
3001
- some more fixes for the mac export to hooks change
3002
- make Ganeti not break with DRBD 8.2.x (which changed the version
3003
  format in ``/proc/drbd``) (issue 24)
3004
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
3005
  template, allowing migration from DRBD0.7+MD to DRBD8
3006

    
3007

    
3008
Version 1.2.2
3009
-------------
3010

    
3011
*(Released Wed, 30 Jan 2008)*
3012

    
3013
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
3014
  support (issue 23)
3015
- add command aliases infrastructure and a few aliases
3016
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
3017
  man pages and the ``--help`` option of ``gnt-node
3018
  list``/``gnt-instance list``
3019
- fix ``gnt-backup list`` with down nodes (issue 21)
3020
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
3021
- fix the dist archive and add a check for including svn/git files in
3022
  the future
3023
- some developer-related changes: improve the burnin and the QA suite,
3024
  add an upload script for testing during development
3025

    
3026

    
3027
Version 1.2.1
3028
-------------
3029

    
3030
*(Released Wed, 16 Jan 2008)*
3031

    
3032
- experimental HVM support, read the install document, section
3033
  "Initializing the cluster"
3034
- allow for the PVM hypervisor per-instance kernel and initrd paths
3035
- add a new command ``gnt-cluster verify-disks`` which uses a new
3036
  algorithm to improve the reconnection of the DRBD pairs if the device
3037
  on the secondary node has gone away
3038
- make logical volume code auto-activate LVs at disk activation time
3039
- slightly improve the speed of activating disks
3040
- allow specification of the MAC address at instance creation time, and
3041
  changing it later via ``gnt-instance modify``
3042
- fix handling of external commands that generate lots of output on
3043
  stderr
3044
- update documentation with regard to minimum version of DRBD8 supported
3045

    
3046

    
3047
Version 1.2.0
3048
-------------
3049

    
3050
*(Released Tue, 4 Dec 2007)*
3051

    
3052
- Log the ``xm create`` output to the node daemon log on failure (to
3053
  help diagnosing the error)
3054
- In debug mode, log all external commands output if failed to the logs
3055
- Change parsing of lvm commands to ignore stderr
3056

    
3057

    
3058
Version 1.2 beta3
3059
-----------------
3060

    
3061
*(Released Wed, 28 Nov 2007)*
3062

    
3063
- Another round of updates to the DRBD 8 code to deal with more failures
3064
  in the replace secondary node operation
3065
- Some more logging of failures in disk operations (lvm, drbd)
3066
- A few documentation updates
3067
- QA updates
3068

    
3069

    
3070
Version 1.2 beta2
3071
-----------------
3072

    
3073
*(Released Tue, 13 Nov 2007)*
3074

    
3075
- Change configuration file format from Python's Pickle to JSON.
3076
  Upgrading is possible using the cfgupgrade utility.
3077
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
3078
  faster replace disks and is more stable (DRBD 8 has many improvements
3079
  compared to DRBD 0.7)
3080
- Added command line tags support (see man pages for ``gnt-instance``,
3081
  ``gnt-node``, ``gnt-cluster``)
3082
- Added instance rename support
3083
- Added multi-instance startup/shutdown
3084
- Added cluster rename support
3085
- Added ``gnt-node evacuate`` to simplify some node operations
3086
- Added instance reboot operation that can speedup reboot as compared to
3087
  stop and start
3088
- Soften the requirement that hostnames are in FQDN format
3089
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
3090
  reboots
3091
- Removed dependency on debian's patched fping that uses the
3092
  non-standard ``-S`` option
3093
- Now the OS definitions are searched for in multiple, configurable
3094
  paths (easier for distros to package)
3095
- Some changes to the hooks infrastructure (especially the new
3096
  post-configuration update hook)
3097
- Other small bugfixes
3098

    
3099
.. vim: set textwidth=72 syntax=rst :
3100
.. Local Variables:
3101
.. mode: rst
3102
.. fill-column: 72
3103
.. End: