Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 56372573

History | View | Annotate | Download (61.8 kB)

1
News
2
====
3

    
4
Version 2.5.0 beta1
5
-------------------
6

    
7
*(unreleased)*
8

    
9
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
10
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
11
  to match the underlying LUXI interface
12
- When creating file-based instances via RAPI, the ``file_driver``
13
  parameter no longer defaults to ``loop`` and must be specified
14
- The deprecated "bridge" nic parameter is no longer supported. Use
15
  "link" instead.
16
- Support for the undocumented and deprecated RAPI instance creation
17
  request format version 0 has been dropped. Use version 1, supported
18
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
19
- Pyparsing 1.4.6 or above is required, see :doc:`installation
20
  documentation <install>`
21
- The "cluster-verify" hooks are now executed per group by the
22
  OP_CLUSTER_VERIFY_GROUP opcode. This maintains the same behavior if
23
  you just run "gnt-cluster verify", which generates one op per group.
24

    
25

    
26
Version 2.4.3
27
-------------
28

    
29
*(unreleased)*
30

    
31
- Added a new parameter to instance stop/start called ``--no-remember``
32
  that will make the state change to not be remembered
33

    
34

    
35
Version 2.4.2
36
-------------
37

    
38
*(Released Thu, 12 May 2011)*
39

    
40
Many bug-fixes and a few new small features:
41

    
42
- Fixed a bug related to log opening failures
43
- Fixed a bug in instance listing with orphan instances
44
- Fixed a bug which prevented resetting the cluster-level node parameter
45
  ``oob_program`` to the default
46
- Many fixes related to the ``cluster-merge`` tool
47
- Fixed a race condition in the lock monitor, which caused failures
48
  during (at least) creation of many instances in parallel
49
- Improved output for gnt-job info
50
- Removed the quiet flag on some ssh calls which prevented debugging
51
  failures
52
- Improved the N+1 failure messages in cluster verify by actually
53
  showing the memory values (needed and available)
54
- Increased lock attempt timeouts so that when executing long operations
55
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
56
  too early and thus prevent the use of the 'fair' mechanism
57
- Changed instance query data (``gnt-instance info``) to not acquire
58
  locks unless needed, thus allowing its use on locked instance if only
59
  static information is asked for
60
- Improved behaviour with filesystems that do not support rename on an
61
  opened file
62
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
63
  kept locks on all nodes during the wipe, which is unneeded
64
- Fixed ``gnt-watcher`` handling of errors during hooks execution
65
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
66
  10GiB) which caused the wipe to fail right at the end in some cases
67
- Fixed master IP activation when doing master failover with no-voting
68
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
69
  the master node itself
70
- Fixed potential data-loss in under disk full conditions, where Ganeti
71
  wouldn't check correctly the return code and would consider
72
  partially-written files 'correct'
73
- Fixed bug related to multiple VGs and DRBD disk replacing
74
- Added new disk parameter ``metavg`` that allows placement of the meta
75
  device for DRBD in a different volume group
76
- Fixed error handling in the node daemon when the system libc doesn't
77
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
78
- Fixed lock release during replace-disks, which kept cluster-wide locks
79
  when doing disk replaces with an iallocator script
80
- Added check for missing bridges in cluster verify
81
- Handle EPIPE errors while writing to the terminal better, so that
82
  piping the output to e.g. ``less`` doesn't cause a backtrace
83
- Fixed rare case where a ^C during Luxi calls could have been
84
  interpreted as server errors, instead of simply terminating
85
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
86
  assign-nodes``)
87
- Added a few more parameters to the KVM hypervisor, allowing a second
88
  CDROM, custom disk type for CDROMs and a floppy image
89
- Removed redundant message in instance rename when the name is given
90
  already as a FQDN
91
- Added option to ``gnt-instance recreate-disks`` to allow creating the
92
  disks on new nodes, allowing recreation when the original instance
93
  nodes are completely gone
94
- Added option when converting disk templates to DRBD to skip waiting
95
  for the resync, in order to make the instance available sooner
96
- Added two new variables to the OS scripts environment (containing the
97
  instance's nodes)
98
- Made the root_path and optional parameter for the xen-pvm hypervisor,
99
  to allow use of ``pvgrub`` as bootloader
100
- Changed the instance memory modifications to only check out-of-memory
101
  conditions on memory increases, and turned the secondary node warnings
102
  into errors (they can still be overridden via ``--force``)
103
- Fixed the handling of a corner case when the Python installation gets
104
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
105
  to execute a command that doesn't exist
106
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
107
  node of the instance returned failures during instance shutdown; this
108
  adds the option ``--ignore-consistency`` to gnt-instance move
109

    
110
And as usual, various improvements to the error messages, documentation
111
and man pages.
112

    
113
Version 2.4.1
114
-------------
115

    
116
*(Released Wed, 09 Mar 2011)*
117

    
118
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
119
the RAPI users file if run twice (even with ``-dry-run``).
120

    
121
The release fixes that bug (nothing else changed).
122

    
123

    
124
Version 2.4.0
125
-------------
126

    
127
*(Released Mon, 07 Mar 2011)*
128

    
129
Final 2.4.0 release. Just a few small fixes:
130

    
131
- Fixed RAPI node evacuate
132
- Fixed the kvm-ifup script
133
- Fixed internal error handling for special job cases
134
- Updated man page to specify the escaping feature for options
135

    
136

    
137
Version 2.4.0 rc3
138
-----------------
139

    
140
*(Released Mon, 28 Feb 2011)*
141

    
142
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
143
that this feature wiped the disks of the wrong instance, leading to loss
144
of data.
145

    
146
Other changes:
147

    
148
- Fixed title of query field containing instance name
149
- Expanded the glossary in the documentation
150
- Fixed one unittest (internal issue)
151

    
152

    
153
Version 2.4.0 rc2
154
-----------------
155

    
156
*(Released Mon, 21 Feb 2011)*
157

    
158
A number of bug fixes plus just a couple functionality changes.
159

    
160
On the user-visible side, the ``gnt-* list`` command output has changed
161
with respect to "special" field states. The current rc1 style of display
162
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
163
the default output mode special fields states are displayed as follows:
164

    
165
- Offline resource: ``*``
166
- Unavailable/not applicable: ``-``
167
- Data missing (RPC failure): ``?``
168
- Unknown field: ``??``
169

    
170
Another user-visible change is the addition of ``--force-join`` to
171
``gnt-node add``.
172

    
173
As for bug fixes:
174

    
175
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
176
- Fixed regression in RAPI/instance reinstall where all parameters were
177
  required (instead of optional)
178
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
179
- Fixed iallocator usage (offline nodes were not considered offline)
180
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
181
- Fixed hypervisor and OS parameter validation with respect to
182
  non-vm_capable nodes
183
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
184
  cosmetic)
185
- Fixed ``tools/listrunner`` with respect to agent-based usage
186

    
187

    
188
Version 2.4.0 rc1
189
-----------------
190

    
191
*(Released Fri,  4 Feb 2011)*
192

    
193
Many changes and fixes since the beta1 release. While there were some
194
internal changes, the code has been mostly stabilised for the RC
195
release.
196

    
197
Note: the dumb allocator was removed in this release, as it was not kept
198
up-to-date with the IAllocator protocol changes. It is recommended to
199
use the ``hail`` command from the ganeti-htools package.
200

    
201
Note: the 2.4 and up versions of Ganeti are not compatible with the
202
0.2.x branch of ganeti-htools. You need to upgrade to
203
ganeti-htools-0.3.0 (or later).
204

    
205
Regressions fixed from 2.3
206
~~~~~~~~~~~~~~~~~~~~~~~~~~
207

    
208
- Fixed the ``gnt-cluster verify-disks`` command
209
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
210
  serially on nodes)
211
- Fixed disk adoption breakage
212
- Fixed wrong headers in instance listing for field aliases
213

    
214
Other bugs fixed
215
~~~~~~~~~~~~~~~~
216

    
217
- Fixed corner case in KVM handling of NICs
218
- Fixed many cases of wrong handling of non-vm_capable nodes
219
- Fixed a bug where a missing instance symlink was not possible to
220
  recreate with any ``gnt-*`` command (now ``gnt-instance
221
  activate-disks`` does it)
222
- Fixed the volume group name as reported by ``gnt-cluster
223
  verify-disks``
224
- Increased timeouts for the import-export code, hopefully leading to
225
  fewer aborts due network or instance timeouts
226
- Fixed bug in ``gnt-node list-storage``
227
- Fixed bug where not all daemons were started on cluster
228
  initialisation, but only at the first watcher run
229
- Fixed many bugs in the OOB implementation
230
- Fixed watcher behaviour in presence of instances with offline
231
  secondaries
232
- Fixed instance list output for instances running on the wrong node
233
- a few fixes to the cluster-merge tool, but it still cannot merge
234
  multi-node groups (currently it is not recommended to use this tool)
235

    
236

    
237
Improvements
238
~~~~~~~~~~~~
239

    
240
- Improved network configuration for the KVM hypervisor
241
- Added e1000 as a supported NIC for Xen-HVM
242
- Improved the lvmstrap tool to also be able to use partitions, as
243
  opposed to full disks
244
- Improved speed of disk wiping (the cluster parameter
245
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
246
  of instance creations
247
- Added documentation for the OS parameters
248
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
249
  hypervisor is not responding
250
- Added display of blacklisted and hidden OS information in
251
  ``gnt-cluster info``
252
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
253
  NIC and node parameters, which might create problems with currently
254
  invalid (but undetected) configuration files, but prevents validation
255
  failures when unrelated parameters are modified
256
- Changed cluster initialisation to wait for the master daemon to become
257
  available
258
- Expanded the RAPI interface:
259

    
260
  - Added config redistribution resource
261
  - Added activation/deactivation of instance disks
262
  - Added export of console information
263

    
264
- Implemented log file reopening on SIGHUP, which allows using
265
  logrotate(8) for the Ganeti log files
266
- Added a basic OOB helper script as an example
267

    
268

    
269
Version 2.4.0 beta1
270
-------------------
271

    
272
*(Released Fri, 14 Jan 2011)*
273

    
274
User-visible
275
~~~~~~~~~~~~
276

    
277
- Fixed timezone issues when formatting timestamps
278
- Added support for node groups, available via ``gnt-group`` and other
279
  commands
280
- Added out-of-band framework and management, see :doc:`design
281
  document <design-oob>`
282
- Removed support for roman numbers from ``gnt-node list`` and
283
  ``gnt-instance list``.
284
- Allowed modification of master network interface via ``gnt-cluster
285
  modify --master-netdev``
286
- Accept offline secondaries while shutting down instance disks
287
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
288
- Added support for multiple LVM volume groups
289
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
290
  requested
291
- Added commands to list available fields:
292

    
293
  - ``gnt-node list-fields``
294
  - ``gnt-group list-fields``
295
  - ``gnt-instance list-fields``
296

    
297
- Updated documentation and man pages
298

    
299
Integration
300
~~~~~~~~~~~
301

    
302
- Moved ``rapi_users`` file into separate directory, now named
303
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
304
  symlink
305
- Added new tool for running commands on many machines,
306
  ``tools/ganeti-listrunner``
307
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
308
  improving the ``gnt-instance console`` output
309
- Allowed customisation of disk index separator at ``configure`` time
310
- Export node group allocation policy to :doc:`iallocator <iallocator>`
311
- Added support for non-partitioned md disks in ``lvmstrap``
312
- Added script to gracefully power off KVM instances
313
- Split ``utils`` module into smaller parts
314
- Changed query operations to return more detailed information, e.g.
315
  whether an information is unavailable due to an offline node. To use
316
  this new functionality, the LUXI call ``Query`` must be used. Field
317
  information is now stored by the master daemon and can be retrieved
318
  using ``QueryFields``. Instances, nodes and groups can also be queried
319
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
320
  exposed via RAPI). The following commands make use of this
321
  infrastructure change:
322

    
323
  - ``gnt-group list``
324
  - ``gnt-group list-fields``
325
  - ``gnt-node list``
326
  - ``gnt-node list-fields``
327
  - ``gnt-instance list``
328
  - ``gnt-instance list-fields``
329
  - ``gnt-debug locks``
330

    
331
Remote API
332
~~~~~~~~~~
333

    
334
- New RAPI resources (see :doc:`rapi`):
335

    
336
  - ``/2/modify``
337
  - ``/2/groups``
338
  - ``/2/groups/[group_name]``
339
  - ``/2/groups/[group_name]/assign-nodes``
340
  - ``/2/groups/[group_name]/modify``
341
  - ``/2/groups/[group_name]/rename``
342
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
343

    
344
- RAPI changes:
345

    
346
  - Implemented ``no_install`` for instance creation
347
  - Implemented OS parameters for instance reinstallation, allowing
348
    use of special settings on reinstallation (e.g. for preserving data)
349

    
350
Misc
351
~~~~
352

    
353
- Added IPv6 support in import/export
354
- Pause DRBD synchronization while wiping disks on instance creation
355
- Updated unittests and QA scripts
356
- Improved network parameters passed to KVM
357
- Converted man pages from docbook to reStructuredText
358

    
359

    
360
Version 2.3.1
361
-------------
362

    
363
*(Released Mon, 20 Dec 2010)*
364

    
365
Released version 2.3.1~rc1 without any changes.
366

    
367

    
368
Version 2.3.1 rc1
369
-----------------
370

    
371
*(Released Wed, 1 Dec 2010)*
372

    
373
- impexpd: Disable OpenSSL compression in socat if possible (backport
374
  from master, commit e90739d625b, see :doc:`installation guide
375
  <install-quick>` for details)
376
- Changed unittest coverage report to exclude test scripts
377
- Added script to check version format
378

    
379

    
380
Version 2.3.0
381
-------------
382

    
383
*(Released Wed, 1 Dec 2010)*
384

    
385
Released version 2.3.0~rc1 without any changes.
386

    
387

    
388
Version 2.3.0 rc1
389
-----------------
390

    
391
*(Released Fri, 19 Nov 2010)*
392

    
393
A number of bugfixes and documentation updates:
394

    
395
- Update ganeti-os-interface documentation
396
- Fixed a bug related to duplicate MACs or similar items which should be
397
  unique
398
- Fix breakage in OS state modify
399
- Reinstall instance: disallow offline secondaries (fixes bug related to
400
  OS changing but reinstall failing)
401
- plus all the other fixes between 2.2.1 and 2.2.2
402

    
403

    
404
Version 2.3.0 rc0
405
-----------------
406

    
407
*(Released Tue, 2 Nov 2010)*
408

    
409
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
410
- Fixed master failover race with watcher
411
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
412
  configuration
413
- Accept previously stopped instance for export with instance removal
414
- Simplify and extend the environment variables for instance OS scripts
415
- Added new node flags, ``master_capable`` and ``vm_capable``
416
- Added optional instance disk wiping prior during allocation. This is a
417
  cluster-wide option and can be set/modified using
418
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
419
- Added IPv6 support, see :doc:`design document <design-2.3>` and
420
  :doc:`install-quick`
421
- Added a new watcher option (``--ignore-pause``)
422
- Added option to ignore offline node on instance start/stop
423
  (``--ignore-offline``)
424
- Allow overriding OS parameters with ``gnt-instance reinstall``
425
- Added ability to change node's secondary IP address using ``gnt-node
426
  modify``
427
- Implemented privilege separation for all daemons except
428
  ``ganeti-noded``, see ``configure`` options
429
- Complain if an instance's disk is marked faulty in ``gnt-cluster
430
  verify``
431
- Implemented job priorities (see ``ganeti(7)`` manpage)
432
- Ignore failures while shutting down instances during failover from
433
  offline node
434
- Exit daemon's bootstrap process only once daemon is ready
435
- Export more information via ``LUInstanceQuery``/remote API
436
- Improved documentation, QA and unittests
437
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
438
  restart if the file was created or changed
439
- Added LUXI protocol version sent with each request and response,
440
  allowing detection of server/client mismatches
441
- Moved the Python scripts among gnt-* and ganeti-* into modules
442
- Moved all code related to setting up SSH to an external script,
443
  ``setup-ssh``
444
- Infrastructure changes for node group support in future versions
445

    
446

    
447
Version 2.2.2
448
-------------
449

    
450
*(Released Fri, 19 Nov 2010)*
451

    
452
A few small bugs fixed, and some improvements to the build system:
453

    
454
- Fix documentation regarding conversion to drbd
455
- Fix validation of parameters in cluster modify (``gnt-cluster modify
456
  -B``)
457
- Fix error handling in node modify with multiple changes
458
- Allow remote imports without checked names
459

    
460

    
461
Version 2.2.1
462
-------------
463

    
464
*(Released Tue, 19 Oct 2010)*
465

    
466
- Disable SSL session ID cache in RPC client
467

    
468

    
469
Version 2.2.1 rc1
470
-----------------
471

    
472
*(Released Thu, 14 Oct 2010)*
473

    
474
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
475
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
476
  with GnuTLS
477
- Fix problems with interaction between Curl and Python's HTTP server,
478
  resulting in increased speed in many RPC calls
479
- Improve our release script to prevent breakage with older aclocal and
480
  Python 2.6
481

    
482

    
483
Version 2.2.1 rc0
484
-----------------
485

    
486
*(Released Thu, 7 Oct 2010)*
487

    
488
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
489
  value retrieved from master
490
- Added support for blacklisted or hidden OS definitions
491
- Added simple lock monitor (accessible via (``gnt-debug locks``)
492
- Added support for -mem-path in KVM hypervisor abstraction layer
493
- Allow overriding instance parameters in tool for inter-cluster
494
  instance moves (``tools/move-instance``)
495
- Improved opcode summaries (e.g. in ``gnt-job list``)
496
- Improve consistency of OS listing by sorting it
497
- Documentation updates
498

    
499

    
500
Version 2.2.0.1
501
---------------
502

    
503
*(Released Fri, 8 Oct 2010)*
504

    
505
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
506

    
507

    
508
Version 2.2.0
509
-------------
510

    
511
*(Released Mon, 4 Oct 2010)*
512

    
513
- Fixed regression in ``gnt-instance rename``
514

    
515

    
516
Version 2.2.0 rc2
517
-----------------
518

    
519
*(Released Wed, 22 Sep 2010)*
520

    
521
- Fixed OS_VARIANT variable for OS scripts
522
- Fixed cluster tag operations via RAPI
523
- Made ``setup-ssh`` exit with non-zero code if an error occurred
524
- Disabled RAPI CA checks in watcher
525

    
526

    
527
Version 2.2.0 rc1
528
-----------------
529

    
530
*(Released Mon, 23 Aug 2010)*
531

    
532
- Support DRBD versions of the format "a.b.c.d"
533
- Updated manpages
534
- Re-introduce support for usage from multiple threads in RAPI client
535
- Instance renames and modify via RAPI
536
- Work around race condition between processing and archival in job
537
  queue
538
- Mark opcodes following failed one as failed, too
539
- Job field ``lock_status`` was removed due to difficulties making it
540
  work with the changed job queue in Ganeti 2.2; a better way to monitor
541
  locks is expected for a later 2.2.x release
542
- Fixed dry-run behaviour with many commands
543
- Support ``ssh-agent`` again when adding nodes
544
- Many additional bugfixes
545

    
546

    
547
Version 2.2.0 rc0
548
-----------------
549

    
550
*(Released Fri, 30 Jul 2010)*
551

    
552
Important change: the internal RPC mechanism between Ganeti nodes has
553
changed from using a home-grown http library (based on the Python base
554
libraries) to use the PycURL library. This requires that PycURL is
555
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
556
against GnuTLS by default. cURL's support for GnuTLS had known issues
557
before cURL 7.21.0 and we recommend using the latest cURL release or
558
linking against OpenSSL. Most other distributions already link PycURL
559
and cURL against OpenSSL. The command::
560

    
561
  python -c 'import pycurl; print pycurl.version'
562

    
563
can be used to determine the libraries PycURL and cURL are linked
564
against.
565

    
566
Other significant changes:
567

    
568
- Rewrote much of the internals of the job queue, in order to achieve
569
  better parallelism; this decouples job query operations from the job
570
  processing, and it should allow much nicer behaviour of the master
571
  daemon under load, and it also has uncovered some long-standing bugs
572
  related to the job serialisation (now fixed)
573
- Added a default iallocator setting to the cluster parameters,
574
  eliminating the need to always pass nodes or an iallocator for
575
  operations that require selection of new node(s)
576
- Added experimental support for the LXC virtualization method
577
- Added support for OS parameters, which allows the installation of
578
  instances to pass parameter to OS scripts in order to customise the
579
  instance
580
- Added a hypervisor parameter controlling the migration type (live or
581
  non-live), since hypervisors have various levels of reliability; this
582
  has renamed the 'live' parameter to 'mode'
583
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
584
  logical volumes, meaning that cluster verify will ignore them and not
585
  flag their presence as errors
586
- The watcher will now reset the error count for failed instances after
587
  8 hours, thus allowing self-healing if the problem that caused the
588
  instances to be down/fail to start has cleared in the meantime
589
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
590
  check for, and warn, if the drbd module parameter ``usermode_helper``
591
  is not consistent with the cluster-wide setting; this is needed to
592
  make diagnose easier of failed drbd creations
593
- Started adding base IPv6 support, but this is not yet
594
  enabled/available for use
595
- Rename operations (cluster, instance) will now return the new name,
596
  which is especially useful if a short name was passed in
597
- Added support for instance migration in RAPI
598
- Added a tool to pre-configure nodes for the SSH setup, before joining
599
  them to the cluster; this will allow in the future a simplified model
600
  for node joining (but not yet fully enabled in 2.2); this needs the
601
  paramiko python library
602
- Fixed handling of name-resolving errors
603
- Fixed consistency of job results on the error path
604
- Fixed master-failover race condition when executed multiple times in
605
  sequence
606
- Fixed many bugs related to the job queue (mostly introduced during the
607
  2.2 development cycle, so not all are impacting 2.1)
608
- Fixed instance migration with missing disk symlinks
609
- Fixed handling of unknown jobs in ``gnt-job archive``
610
- And many other small fixes/improvements
611

    
612
Internal changes:
613

    
614
- Enhanced both the unittest and the QA coverage
615
- Switched the opcode validation to a generic model, and extended the
616
  validation to all opcode parameters
617
- Changed more parts of the code that write shell scripts to use the
618
  same class for this
619
- Switched the master daemon to use the asyncore library for the Luxi
620
  server endpoint
621

    
622

    
623
Version 2.2.0 beta 0
624
--------------------
625

    
626
*(Released Thu, 17 Jun 2010)*
627

    
628
- Added tool (``move-instance``) and infrastructure to move instances
629
  between separate clusters (see :doc:`separate documentation
630
  <move-instance>` and :doc:`design document <design-2.2>`)
631
- Added per-request RPC timeout
632
- RAPI now requires a Content-Type header for requests with a body (e.g.
633
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
634
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
635
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
636
  reachable
637
- Implemented initial support for running Ganeti daemons as separate
638
  users, see configure-time flags ``--with-user-prefix`` and
639
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
640
  time)
641
- Instances can be removed after export (``gnt-backup export
642
  --remove-instance``)
643
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
644
  key (instead of 1024 bit)
645
- Added new cluster configuration file for cluster domain secret
646
- Import/export now use SSL instead of SSH
647
- Added support for showing estimated time when exporting an instance,
648
  see the ``ganeti-os-interface(7)`` manpage and look for
649
  ``EXP_SIZE_FD``
650

    
651

    
652
Version 2.1.8
653
-------------
654

    
655
*(Released Tue, 16 Nov 2010)*
656

    
657
Some more bugfixes. Unless critical bugs occur, this will be the last
658
2.1 release:
659

    
660
- Fix case of MAC special-values
661
- Fix mac checker regex
662
- backend: Fix typo causing "out of range" error
663
- Add missing --units in gnt-instance list man page
664

    
665

    
666
Version 2.1.7
667
-------------
668

    
669
*(Released Tue, 24 Aug 2010)*
670

    
671
Bugfixes only:
672
  - Don't ignore secondary node silently on non-mirrored disk templates
673
    (issue 113)
674
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
675
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
676
  - Properly document the usb_mouse parameter
677
  - Fix path in ganeti-rapi(8) (issue 116)
678
  - Adjust error message when the ganeti user's .ssh directory is
679
    missing
680
  - Add same-node-check when changing the disk template to drbd
681

    
682

    
683
Version 2.1.6
684
-------------
685

    
686
*(Released Fri, 16 Jul 2010)*
687

    
688
Bugfixes only:
689
  - Add an option to only select some reboot types during qa/burnin.
690
    (on some hypervisors consequent reboots are not supported)
691
  - Fix infrequent race condition in master failover. Sometimes the old
692
    master ip address would be still detected as up for a short time
693
    after it was removed, causing failover to fail.
694
  - Decrease mlockall warnings when the ctypes module is missing. On
695
    Python 2.4 we support running even if no ctypes module is installed,
696
    but we were too verbose about this issue.
697
  - Fix building on old distributions, on which man doesn't have a
698
    --warnings option.
699
  - Fix RAPI not to ignore the MAC address on instance creation
700
  - Implement the old instance creation format in the RAPI client.
701

    
702

    
703
Version 2.1.5
704
-------------
705

    
706
*(Released Thu, 01 Jul 2010)*
707

    
708
A small bugfix release:
709
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
710
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
711
    a non-existing option
712
  - Fix instance create: the --force-variant option was ignored
713
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
714
  - Fix modify node storage with non-FQDN arguments
715
  - Fix RAPI client to authenticate under Python 2.6 when used
716
    for more than 5 requests needing authentication
717
  - Fix gnt-instance modify -t (storage) giving a wrong error message
718
    when converting a non-shutdown drbd instance to plain
719

    
720

    
721
Version 2.1.4
722
-------------
723

    
724
*(Released Fri, 18 Jun 2010)*
725

    
726
A small bugfix release:
727

    
728
  - Fix live migration of KVM instances started with older Ganeti
729
    versions which had fewer hypervisor parameters
730
  - Fix gnt-instance grow-disk on down instances
731
  - Fix an error-reporting bug during instance migration
732
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
733
    silently ignoring broken ones
734
  - Fix an RPC error reporting bug affecting, for example, RAPI client
735
    users
736
  - Fix bug triggered by different API version os-es on different nodes
737
  - Fix a bug in instance startup with custom hvparams: OS level
738
    parameters would fail to be applied.
739
  - Fix the RAPI client under Python 2.6 (but more work is needed to
740
    make it work completely well with OpenSSL)
741
  - Fix handling of errors when resolving names from DNS
742

    
743

    
744
Version 2.1.3
745
-------------
746

    
747
*(Released Thu, 3 Jun 2010)*
748

    
749
A medium sized development cycle. Some new features, and some
750
fixes/small improvements/cleanups.
751

    
752
Significant features
753
~~~~~~~~~~~~~~~~~~~~
754

    
755
The node deamon now tries to mlock itself into memory, unless the
756
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
757
its logs, and falls back to console logging. This allows emergency
758
features such as ``gnt-node powercycle`` to work even in the event of a
759
broken node disk (tested offlining the disk hosting the node's
760
filesystem and dropping its memory caches; don't try this at home)
761

    
762
KVM: add vhost-net acceleration support. It can be tested with a new
763
enough version of the kernel and of qemu-kvm.
764

    
765
KVM: Add instance chrooting feature. If you use privilege dropping for
766
your VMs you can also now force them to chroot to an empty directory,
767
before starting the emulated guest.
768

    
769
KVM: Add maximum migration bandwith and maximum downtime tweaking
770
support (requires a new-enough version of qemu-kvm).
771

    
772
Cluster verify will now warn if the master node doesn't have the master
773
ip configured on it.
774

    
775
Add a new (incompatible) instance creation request format to RAPI which
776
supports all parameters (previously only a subset was supported, and it
777
wasn't possible to extend the old format to accomodate all the new
778
features. The old format is still supported, and a client can check for
779
this feature, before using it, by checking for its presence in the
780
``features`` RAPI resource.
781

    
782
Now with ancient latin support. Try it passing the ``--roman`` option to
783
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
784
(requires the python-roman module to be installed, in order to work).
785

    
786
Other changes
787
~~~~~~~~~~~~~
788

    
789
As usual many internal code refactorings, documentation updates, and
790
such. Among others:
791

    
792
  - Lots of improvements and cleanups to the experimental Remote API
793
    (RAPI) client library.
794
  - A new unit test suite for the core daemon libraries.
795
  - A fix to creating missing directories makes sure the umask is not
796
    applied anymore. This enforces the same directory permissions
797
    everywhere.
798
  - Better handling terminating daemons with ctrl+c (used when running
799
    them in debugging mode).
800
  - Fix a race condition in live migrating a KVM instance, when stat()
801
    on the old proc status file returned EINVAL, which is an unexpected
802
    value.
803
  - Fixed manpage checking with newer man and utf-8 charachters. But now
804
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
805

    
806

    
807
Version 2.1.2.1
808
---------------
809

    
810
*(Released Fri, 7 May 2010)*
811

    
812
Fix a bug which prevented untagged KVM instances from starting.
813

    
814

    
815
Version 2.1.2
816
-------------
817

    
818
*(Released Fri, 7 May 2010)*
819

    
820
Another release with a long development cycle, during which many
821
different features were added.
822

    
823
Significant features
824
~~~~~~~~~~~~~~~~~~~~
825

    
826
The KVM hypervisor now can run the individual instances as non-root, to
827
reduce the impact of a VM being hijacked due to bugs in the
828
hypervisor. It is possible to run all instances as a single (non-root)
829
user, to manually specify a user for each instance, or to dynamically
830
allocate a user out of a cluster-wide pool to each instance, with the
831
guarantee that no two instances will run under the same user ID on any
832
given node.
833

    
834
An experimental RAPI client library, that can be used standalone
835
(without the other Ganeti libraries), is provided in the source tree as
836
``lib/rapi/client.py``. Note this client might change its interface in
837
the future, as we iterate on its capabilities.
838

    
839
A new command, ``gnt-cluster renew-crypto`` has been added to easily
840
replace the cluster's certificates and crypto keys. This might help in
841
case they have been compromised, or have simply expired.
842

    
843
A new disk option for instance creation has been added that allows one
844
to "adopt" currently existing logical volumes, with data
845
preservation. This should allow easier migration to Ganeti from
846
unmanaged (or managed via other software) instances.
847

    
848
Another disk improvement is the possibility to convert between redundant
849
(DRBD) and plain (LVM) disk configuration for an instance. This should
850
allow better scalability (starting with one node and growing the
851
cluster, or shrinking a two-node cluster to one node).
852

    
853
A new feature that could help with automated node failovers has been
854
implemented: if a node sees itself as offline (by querying the master
855
candidates), it will try to shutdown (hard) all instances and any active
856
DRBD devices. This reduces the risk of duplicate instances if an
857
external script automatically failovers the instances on such nodes. To
858
enable this, the cluster parameter ``maintain_node_health`` should be
859
enabled; in the future this option (per the name) will enable other
860
automatic maintenance features.
861

    
862
Instance export/import now will reuse the original instance
863
specifications for all parameters; that means exporting an instance,
864
deleting it and the importing it back should give an almost identical
865
instance. Note that the default import behaviour has changed from
866
before, where it created only one NIC; now it recreates the original
867
number of NICs.
868

    
869
Cluster verify has added a few new checks: SSL certificates validity,
870
/etc/hosts consistency across the cluster, etc.
871

    
872
Other changes
873
~~~~~~~~~~~~~
874

    
875
As usual, many internal changes were done, documentation fixes,
876
etc. Among others:
877

    
878
- Fixed cluster initialization with disabled cluster storage (regression
879
  introduced in 2.1.1)
880
- File-based storage supports growing the disks
881
- Fixed behaviour of node role changes
882
- Fixed cluster verify for some corner cases, plus a general rewrite of
883
  cluster verify to allow future extension with more checks
884
- Fixed log spamming by watcher and node daemon (regression introduced
885
  in 2.1.1)
886
- Fixed possible validation issues when changing the list of enabled
887
  hypervisors
888
- Fixed cleanup of /etc/hosts during node removal
889
- Fixed RAPI response for invalid methods
890
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
891
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
892
  from ide disks, etc.)
893
- Allow OS changes without re-installation (to record a changed OS
894
  outside of Ganeti, or to allow OS renames)
895
- Allow instance creation without OS installation (useful for example if
896
  the OS will be installed manually, or restored from a backup not in
897
  Ganeti format)
898
- Implemented option to make cluster ``copyfile`` use the replication
899
  network
900
- Added list of enabled hypervisors to ssconf (possibly useful for
901
  external scripts)
902
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
903
  1.2 clusters
904
- A partial form of node re-IP is possible via node readd, which now
905
  allows changed node primary IP
906
- Command line utilities now show an informational message if the job is
907
  waiting for a lock
908
- The logs of the master daemon now show the PID/UID/GID of the
909
  connected client
910

    
911

    
912
Version 2.1.1
913
-------------
914

    
915
*(Released Fri, 12 Mar 2010)*
916

    
917
During the 2.1.0 long release candidate cycle, a lot of improvements and
918
changes have accumulated with were released later as 2.1.1.
919

    
920
Major changes
921
~~~~~~~~~~~~~
922

    
923
The node evacuate command (``gnt-node evacuate``) was significantly
924
rewritten, and as such the IAllocator protocol was changed - a new
925
request type has been added. This unfortunate change during a stable
926
series is designed to improve performance of node evacuations; on
927
clusters with more than about five nodes and which are well-balanced,
928
evacuation should proceed in parallel for all instances of the node
929
being evacuated. As such, any existing IAllocator scripts need to be
930
updated, otherwise the above command will fail due to the unknown
931
request. The provided "dumb" allocator has not been updated; but the
932
ganeti-htools package supports the new protocol since version 0.2.4.
933

    
934
Another important change is increased validation of node and instance
935
names. This might create problems in special cases, if invalid host
936
names are being used.
937

    
938
Also, a new layer of hypervisor parameters has been added, that sits at
939
OS level between the cluster defaults and the instance ones. This allows
940
customisation of virtualization parameters depending on the installed
941
OS. For example instances with OS 'X' may have a different KVM kernel
942
(or any other parameter) than the cluster defaults. This is intended to
943
help managing a multiple OSes on the same cluster, without manual
944
modification of each instance's parameters.
945

    
946
A tool for merging clusters, ``cluster-merge``, has been added in the
947
tools sub-directory.
948

    
949
Bug fixes
950
~~~~~~~~~
951

    
952
- Improved the int/float conversions that should make the code more
953
  robust in face of errors from the node daemons
954
- Fixed the remove node code in case of internal configuration errors
955
- Fixed the node daemon behaviour in face of inconsistent queue
956
  directory (e.g. read-only file-system where we can't open the files
957
  read-write, etc.)
958
- Fixed the behaviour of gnt-node modify for master candidate demotion;
959
  now it either aborts cleanly or, if given the new "auto_promote"
960
  parameter, will automatically promote other nodes as needed
961
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
962
  completely prevent Ganeti from working
963
- Fixed bug for instance export when not all disks were successfully
964
  exported
965
- Fixed behaviour of node add when the new node is slow in starting up
966
  the node daemon
967
- Fixed handling of signals in the LUXI client, which should improve
968
  behaviour of command-line scripts
969
- Added checks for invalid node/instance names in the configuration (now
970
  flagged during cluster verify)
971
- Fixed watcher behaviour for disk activation errors
972
- Fixed two potentially endless loops in http library, which led to the
973
  RAPI daemon hanging and consuming 100% CPU in some cases
974
- Fixed bug in RAPI daemon related to hashed passwords
975
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
976
  instances
977
- Enhanced compatibility with non-Debian OSes, but not using absolute
978
  path in some commands and allowing customisation of the ssh
979
  configuration directory
980
- Fixed possible future issue with new Python versions by abiding to the
981
  proper use of ``__slots__`` attribute on classes
982
- Added checks that should prevent directory traversal attacks
983
- Many documentation fixes based on feedback from users
984

    
985
New features
986
~~~~~~~~~~~~
987

    
988
- Added an "early_release" more for instance replace disks and node
989
  evacuate, where we release locks earlier and thus allow higher
990
  parallelism within the cluster
991
- Added watcher hooks, intended to allow the watcher to restart other
992
  daemons (e.g. from the ganeti-nbma project), but they can be used of
993
  course for any other purpose
994
- Added a compile-time disable for DRBD barriers, to increase
995
  performance if the administrator trusts the power supply or the
996
  storage system to not lose writes
997
- Added the option of using syslog for logging instead of, or in
998
  addition to, Ganeti's own log files
999
- Removed boot restriction for paravirtual NICs for KVM, recent versions
1000
  can indeed boot from a paravirtual NIC
1001
- Added a generic debug level for many operations; while this is not
1002
  used widely yet, it allows one to pass the debug value all the way to
1003
  the OS scripts
1004
- Enhanced the hooks environment for instance moves (failovers,
1005
  migrations) where the primary/secondary nodes changed during the
1006
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1007
- Enhanced data validations for many user-supplied values; one important
1008
  item is the restrictions imposed on instance and node names, which
1009
  might reject some (invalid) host names
1010
- Add a configure-time option to disable file-based storage, if it's not
1011
  needed; this allows greater security separation between the master
1012
  node and the other nodes from the point of view of the inter-node RPC
1013
  protocol
1014
- Added user notification in interactive tools if job is waiting in the
1015
  job queue or trying to acquire locks
1016
- Added log messages when a job is waiting for locks
1017
- Added filtering by node tags in instance operations which admit
1018
  multiple instances (start, stop, reboot, reinstall)
1019
- Added a new tool for cluster mergers, ``cluster-merge``
1020
- Parameters from command line which are of the form ``a=b,c=d`` can now
1021
  use backslash escapes to pass in values which contain commas,
1022
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1023
  ``b,c``
1024
- For KVM, the instance name is the first parameter passed to KVM, so
1025
  that it's more visible in the process list
1026

    
1027

    
1028
Version 2.1.0
1029
-------------
1030

    
1031
*(Released Tue, 2 Mar 2010)*
1032

    
1033
Ganeti 2.1 brings many improvements with it. Major changes:
1034

    
1035
- Added infrastructure to ease automated disk repairs
1036
- Added new daemon to export configuration data in a cheaper way than
1037
  using the remote API
1038
- Instance NICs can now be routed instead of being associated with a
1039
  networking bridge
1040
- Improved job locking logic to reduce impact of jobs acquiring multiple
1041
  locks waiting for other long-running jobs
1042

    
1043
In-depth implementation details can be found in the Ganeti 2.1 design
1044
document.
1045

    
1046
Details
1047
~~~~~~~
1048

    
1049
- Added chroot hypervisor
1050
- Added more options to xen-hvm hypervisor (``kernel_path`` and
1051
  ``device_model``)
1052
- Added more options to xen-pvm hypervisor (``use_bootloader``,
1053
  ``bootloader_path`` and ``bootloader_args``)
1054
- Added the ``use_localtime`` option for the xen-hvm and kvm
1055
  hypervisors, and the default value for this has changed to false (in
1056
  2.0 xen-hvm always enabled it)
1057
- Added luxi call to submit multiple jobs in one go
1058
- Added cluster initialization option to not modify ``/etc/hosts``
1059
  file on nodes
1060
- Added network interface parameters
1061
- Added dry run mode to some LUs
1062
- Added RAPI resources:
1063

    
1064
  - ``/2/instances/[instance_name]/info``
1065
  - ``/2/instances/[instance_name]/replace-disks``
1066
  - ``/2/nodes/[node_name]/evacuate``
1067
  - ``/2/nodes/[node_name]/migrate``
1068
  - ``/2/nodes/[node_name]/role``
1069
  - ``/2/nodes/[node_name]/storage``
1070
  - ``/2/nodes/[node_name]/storage/modify``
1071
  - ``/2/nodes/[node_name]/storage/repair``
1072

    
1073
- Added OpCodes to evacuate or migrate all instances on a node
1074
- Added new command to list storage elements on nodes (``gnt-node
1075
  list-storage``) and modify them (``gnt-node modify-storage``)
1076
- Added new ssconf files with master candidate IP address
1077
  (``ssconf_master_candidates_ips``), node primary IP address
1078
  (``ssconf_node_primary_ips``) and node secondary IP address
1079
  (``ssconf_node_secondary_ips``)
1080
- Added ``ganeti-confd`` and a client library to query the Ganeti
1081
  configuration via UDP
1082
- Added ability to run hooks after cluster initialization and before
1083
  cluster destruction
1084
- Added automatic mode for disk replace (``gnt-instance replace-disks
1085
  --auto``)
1086
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1087
  after catastrophic data-loss
1088
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1089
  groups
1090
- Added ``gnt-instance move`` command to move instances
1091
- Added ``gnt-cluster watcher`` command to control watcher
1092
- Added ``gnt-node powercycle`` command to powercycle nodes
1093
- Added new job status field ``lock_status``
1094
- Added parseable error codes to cluster verification (``gnt-cluster
1095
  verify --error-codes``) and made output less verbose (use
1096
  ``--verbose`` to restore previous behaviour)
1097
- Added UUIDs to the main config entities (cluster, nodes, instances)
1098
- Added support for OS variants
1099
- Added support for hashed passwords in the Ganeti remote API users file
1100
  (``rapi_users``)
1101
- Added option to specify maximum timeout on instance shutdown
1102
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
1103
- Added new helper script to start and stop Ganeti daemons
1104
  (``daemon-util``), with the intent to reduce the work necessary to
1105
  adjust Ganeti for non-Debian distributions and to start/stop daemons
1106
  from one place
1107
- Added more unittests
1108
- Fixed critical bug in ganeti-masterd startup
1109
- Removed the configure-time ``kvm-migration-port`` parameter, this is
1110
  now customisable at the cluster level for both the KVM and Xen
1111
  hypervisors using the new ``migration_port`` parameter
1112
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1113
  reinstalling an instance
1114
- Allowed ``@`` in tag names
1115
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1116
- Many documentation updates
1117
- Distribute hypervisor files on ``gnt-cluster redist-conf``
1118
- ``gnt-instance reinstall`` can now reinstall multiple instances
1119
- Updated many command line parameters
1120
- Introduced new OS API version 15
1121
- No longer support a default hypervisor
1122
- Treat virtual LVs as inexistent
1123
- Improved job locking logic to reduce lock contention
1124
- Match instance and node names case insensitively
1125
- Reimplemented bash completion script to be more complete
1126
- Improved burnin
1127

    
1128

    
1129
Version 2.0.6
1130
-------------
1131

    
1132
*(Released Thu, 4 Feb 2010)*
1133

    
1134
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1135
- Fix a string formatting bug
1136
- Improve safety of the code in some error paths
1137
- Improve data validation in the master of values returned from nodes
1138

    
1139

    
1140
Version 2.0.5
1141
-------------
1142

    
1143
*(Released Thu, 17 Dec 2009)*
1144

    
1145
- Fix security issue due to missing validation of iallocator names; this
1146
  allows local and remote execution of arbitrary executables
1147
- Fix failure of gnt-node list during instance removal
1148
- Ship the RAPI documentation in the archive
1149

    
1150

    
1151
Version 2.0.4
1152
-------------
1153

    
1154
*(Released Wed, 30 Sep 2009)*
1155

    
1156
- Fixed many wrong messages
1157
- Fixed a few bugs related to the locking library
1158
- Fixed MAC checking at instance creation time
1159
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
1160
- Fixed a few issues related to signal handling in both daemons and
1161
  scripts
1162
- Fixed the example startup script provided
1163
- Fixed insserv dependencies in the example startup script (patch from
1164
  Debian)
1165
- Fixed handling of drained nodes in the iallocator framework
1166
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1167
  #528618)
1168
- Fixed error related to invalid job IDs in job polling
1169
- Fixed job/opcode persistence on unclean master shutdown
1170
- Fixed handling of partial job processing after unclean master
1171
  shutdown
1172
- Fixed error reporting from LUs, previously all errors were converted
1173
  into execution errors
1174
- Fixed error reporting from burnin
1175
- Decreased significantly the memory usage of the job queue
1176
- Optimised slightly multi-job submission
1177
- Optimised slightly opcode loading
1178
- Backported the multi-job submit framework from the development
1179
  branch; multi-instance start and stop should be faster
1180
- Added script to clean archived jobs after 21 days; this will reduce
1181
  the size of the queue directory
1182
- Added some extra checks in disk size tracking
1183
- Added an example ethers hook script
1184
- Added a cluster parameter that prevents Ganeti from modifying of
1185
  /etc/hosts
1186
- Added more node information to RAPI responses
1187
- Added a ``gnt-job watch`` command that allows following the ouput of a
1188
  job
1189
- Added a bind-address option to ganeti-rapi
1190
- Added more checks to the configuration verify
1191
- Enhanced the burnin script such that some operations can be retried
1192
  automatically
1193
- Converted instance reinstall to multi-instance model
1194

    
1195

    
1196
Version 2.0.3
1197
-------------
1198

    
1199
*(Released Fri, 7 Aug 2009)*
1200

    
1201
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1202
  to allow using the pre-2.0.2 behaviour in activation, if any existing
1203
  instances have mismatched disk sizes in the configuration
1204
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
1205
  any configuration mismatches for disk sizes
1206
- Added ``gnt-master cluste-failover --no-voting`` to allow master
1207
  failover to work on two-node clusters
1208
- Fixed the ``--net`` option of ``gnt-backup import``, which was
1209
  unusable
1210
- Fixed detection of OS script errors in ``gnt-backup export``
1211
- Fixed exit code of ``gnt-backup export``
1212

    
1213

    
1214
Version 2.0.2
1215
-------------
1216

    
1217
*(Released Fri, 17 Jul 2009)*
1218

    
1219
- Added experimental support for stripped logical volumes; this should
1220
  enhance performance but comes with a higher complexity in the block
1221
  device handling; stripping is only enabled when passing
1222
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1223
  affected even in the non-stripped mode
1224
- Improved resiliency against transient failures at the end of DRBD
1225
  resyncs, and in general of DRBD resync checks
1226
- Fixed a couple of issues with exports and snapshot errors
1227
- Fixed a couple of issues in instance listing
1228
- Added display of the disk size in ``gnt-instance info``
1229
- Fixed checking for valid OSes in instance creation
1230
- Fixed handling of the "vcpus" parameter in instance listing and in
1231
  general of invalid parameters
1232
- Fixed http server library, and thus RAPI, to handle invalid
1233
  username/password combinations correctly; this means that now they
1234
  report unauthorized for queries too, not only for modifications,
1235
  allowing earlier detect of configuration problems
1236
- Added a new "role" node list field, equivalent to the master/master
1237
  candidate/drained/offline flags combinations
1238
- Fixed cluster modify and changes of candidate pool size
1239
- Fixed cluster verify error messages for wrong files on regular nodes
1240
- Fixed a couple of issues with node demotion from master candidate role
1241
- Fixed node readd issues
1242
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1243
- Added a new ``--no-voting`` option for masterfailover to fix failover
1244
  on two-nodes clusters when the former master node is unreachable
1245
- Added instance reinstall over RAPI
1246

    
1247

    
1248
Version 2.0.1
1249
-------------
1250

    
1251
*(Released Tue, 16 Jun 2009)*
1252

    
1253
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1254
  allow re-adding the start in single-user option (regression from 1.2)
1255
- the watcher writes the instance status to a file, to allow monitoring
1256
  to report the instance status (from the master) based on cached
1257
  results of the watcher's queries; while this can get stale if the
1258
  watcher is being locked due to other work on the cluster, this is
1259
  still an improvement
1260
- the watcher now also restarts the node daemon and the rapi daemon if
1261
  they died
1262
- fixed the watcher to handle full and drained queue cases
1263
- hooks export more instance data in the environment, which helps if
1264
  hook scripts need to take action based on the instance's properties
1265
  (no longer need to query back into ganeti)
1266
- instance failovers when the instance is stopped do not check for free
1267
  RAM, so that failing over a stopped instance is possible in low memory
1268
  situations
1269
- rapi uses queries for tags instead of jobs (for less job traffic), and
1270
  for cluster tags it won't talk to masterd at all but read them from
1271
  ssconf
1272
- a couple of error handling fixes in RAPI
1273
- drbd handling: improved the error handling of inconsistent disks after
1274
  resync to reduce the frequency of "there are some degraded disks for
1275
  this instance" messages
1276
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
1277
  error handling path called a wrong function name)
1278

    
1279

    
1280
Version 2.0.0 final
1281
-------------------
1282

    
1283
*(Released Wed, 27 May 2009)*
1284

    
1285
- no changes from rc5
1286

    
1287

    
1288
Version 2.0 release candidate 5
1289
-------------------------------
1290

    
1291
*(Released Wed, 20 May 2009)*
1292

    
1293
- fix a couple of bugs (validation, argument checks)
1294
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1295
- some small improvements to RAPI and IAllocator
1296
- make watcher automatically start the master daemon if down
1297

    
1298

    
1299
Version 2.0 release candidate 4
1300
-------------------------------
1301

    
1302
*(Released Mon, 27 Apr 2009)*
1303

    
1304
- change the OS list to not require locks; this helps with big clusters
1305
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1306
  volume group is broken
1307
- ``gnt-instance info``, without any arguments, doesn't run for all
1308
  instances anymore; either pass ``--all`` or pass the desired
1309
  instances; this helps against mistakes on big clusters where listing
1310
  the information for all instances takes a long time
1311
- miscellaneous doc and man pages fixes
1312

    
1313

    
1314
Version 2.0 release candidate 3
1315
-------------------------------
1316

    
1317
*(Released Wed, 8 Apr 2009)*
1318

    
1319
- Change the internal locking model of some ``gnt-node`` commands, in
1320
  order to reduce contention (and blocking of master daemon) when
1321
  batching many creation/reinstall jobs
1322
- Fixes to Xen soft reboot
1323
- No longer build documentation at build time, instead distribute it in
1324
  the archive, in order to reduce the need for the whole docbook/rst
1325
  toolchains
1326

    
1327

    
1328
Version 2.0 release candidate 2
1329
-------------------------------
1330

    
1331
*(Released Fri, 27 Mar 2009)*
1332

    
1333
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1334
- Fix watcher startup sequence, improves the behaviour of busy clusters
1335
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1336
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1337
  bind address and other places
1338
- Some documentation fixes and updates
1339

    
1340

    
1341
Version 2.0 release candidate 1
1342
-------------------------------
1343

    
1344
*(Released Mon, 2 Mar 2009)*
1345

    
1346
- More documentation updates, now all docs should be more-or-less
1347
  up-to-date
1348
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
1349
  etc.)
1350
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1351
  and KVM)
1352
- Fix an issue related to $libdir/run/ganeti and cluster creation
1353

    
1354

    
1355
Version 2.0 beta 2
1356
------------------
1357

    
1358
*(Released Thu, 19 Feb 2009)*
1359

    
1360
- Xen PVM and KVM have switched the default value for the instance root
1361
  disk to the first partition on the first drive, instead of the whole
1362
  drive; this means that the OS installation scripts must be changed
1363
  accordingly
1364
- Man pages have been updated
1365
- RAPI has been switched by default to HTTPS, and the exported functions
1366
  should all work correctly
1367
- RAPI v1 has been removed
1368
- Many improvements to the KVM hypervisor
1369
- Block device errors are now better reported
1370
- Many other bugfixes and small improvements
1371

    
1372

    
1373
Version 2.0 beta 1
1374
------------------
1375

    
1376
*(Released Mon, 26 Jan 2009)*
1377

    
1378
- Version 2 is a general rewrite of the code and therefore the
1379
  differences are too many to list, see the design document for 2.0 in
1380
  the ``doc/`` subdirectory for more details
1381
- In this beta version there is not yet a migration path from 1.2 (there
1382
  will be one in the final 2.0 release)
1383
- A few significant changes are:
1384

    
1385
  - all commands are executed by a daemon (``ganeti-masterd``) and the
1386
    various ``gnt-*`` commands are just front-ends to it
1387
  - all the commands are entered into, and executed from a job queue,
1388
    see the ``gnt-job(8)`` manpage
1389
  - the RAPI daemon supports read-write operations, secured by basic
1390
    HTTP authentication on top of HTTPS
1391
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
1392
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
1393
    to migrate to DRBD 8 first while still running Ganeti 1.2)
1394
  - DRBD devices are using statically allocated minor numbers, which
1395
    will be assigned to existing instances during the migration process
1396
  - there is support for both Xen PVM and Xen HVM instances running on
1397
    the same cluster
1398
  - KVM virtualization is supported too
1399
  - file-based storage has been implemented, which means that it is
1400
    possible to run the cluster without LVM and DRBD storage, for
1401
    example using a shared filesystem exported from shared storage (and
1402
    still have live migration)
1403

    
1404

    
1405
Version 1.2.7
1406
-------------
1407

    
1408
*(Released Tue, 13 Jan 2009)*
1409

    
1410
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
1411
- Reuse the old instance mac address by default on instance import, if
1412
  the instance name is the same.
1413
- Handle situations in which the node info rpc returns incomplete
1414
  results (issue 46)
1415
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1416
- Improved version of batcher:
1417

    
1418
  - state file support
1419
  - instance mac address support
1420
  - support for HVM clusters/instances
1421

    
1422
- Add an option to show the number of cpu sockets and nodes in
1423
  ``gnt-node list``
1424
- Support OSes that handle more than one version of the OS api (but do
1425
  not change the current API in any other way)
1426
- Fix ``gnt-node migrate``
1427
- ``gnt-debug`` man page
1428
- Fixes various more typos and small issues
1429
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1430

    
1431

    
1432
Version 1.2.6
1433
-------------
1434

    
1435
*(Released Wed, 24 Sep 2008)*
1436

    
1437
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1438
  type of disk exported to fully virtualized instances.
1439
- provide access to the serial console of HVM instances
1440
- instance auto_balance flag, set by default. If turned off it will
1441
  avoid warnings on cluster verify if there is not enough memory to fail
1442
  over an instance. in the future it will prevent automatically failing
1443
  it over when we will support that.
1444
- batcher tool for instance creation, see ``tools/README.batcher``
1445
- ``gnt-instance reinstall --select-os`` to interactively select a new
1446
  operating system when reinstalling an instance.
1447
- when changing the memory amount on instance modify a check has been
1448
  added that the instance will be able to start. also warnings are
1449
  emitted if the instance will not be able to fail over, if auto_balance
1450
  is true.
1451
- documentation fixes
1452
- sync fields between ``gnt-instance list/modify/add/import``
1453
- fix a race condition in drbd when the sync speed was set after giving
1454
  the device a remote peer.
1455

    
1456

    
1457
Version 1.2.5
1458
-------------
1459

    
1460
*(Released Tue, 22 Jul 2008)*
1461

    
1462
- note: the allowed size and number of tags per object were reduced
1463
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1464
- fixed twisted 8.x compatibility
1465
- fixed ``gnt-instance replace-disks`` with iallocator
1466
- add TCP keepalives on twisted connections to detect restarted nodes
1467
- disk increase support, see ``gnt-instance grow-disk``
1468
- implement bulk node/instance query for RAPI
1469
- add tags in node/instance listing (optional)
1470
- experimental migration (and live migration) support, read the man page
1471
  for ``gnt-instance migrate``
1472
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1473
  has some small improvements in handling its state file
1474

    
1475

    
1476
Version 1.2.4
1477
-------------
1478

    
1479
*(Released Fri, 13 Jun 2008)*
1480

    
1481
- Experimental readonly, REST-based remote API implementation;
1482
  automatically started on master node, TCP port 5080, if enabled by
1483
  ``--enable-rapi`` parameter to configure script.
1484
- Instance allocator support. Add and import instance accept a
1485
  ``--iallocator`` parameter, and call that instance allocator to decide
1486
  which node to use for the instance. The iallocator document describes
1487
  what's expected from an allocator script.
1488
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1489
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1490
  node is lost there is still enough memory to fail over the instances
1491
  that reside on it.
1492
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1493
  ``gnt-cluster verify``, to check for site-specific compliance. All the
1494
  hooks will run, and their output, if any, will be displayed. Any
1495
  failing hook will make the verification return an error value.
1496
- ``gnt-cluster verify`` now checks that its peers are reachable on the
1497
  primary and secondary interfaces
1498
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1499
  that is still declared as part of the cluster and has failed.
1500
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
1501
  specifying output fields, that just adds the chosen fields to the
1502
  default ones.
1503
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
1504
  export from the filesystem.
1505
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1506
  have been added. Using them you can enable/disable acpi and pae
1507
  support, and specify a path for a cd image to be exported to the
1508
  instance. These parameters as the name suggest only work on HVM
1509
  clusters.
1510
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1511
  PAE support will be set to the previously hardcoded values, but the
1512
  (previously hardcoded) path to the CDROM ISO image will be unset and
1513
  if required, needs to be set manually with ``gnt-instance modify``
1514
  after the upgrade.
1515
- The address to which an instance's VNC console is bound is now
1516
  selectable per-instance, rather than being cluster wide. Of course
1517
  this only applies to instances controlled via VNC, so currently just
1518
  applies to HVM clusters.
1519

    
1520

    
1521
Version 1.2.3
1522
-------------
1523

    
1524
*(Released Mon, 18 Feb 2008)*
1525

    
1526
- more tweaks to the disk activation code (especially helpful for DRBD)
1527
- change the default ``gnt-instance list`` output format, now there is
1528
  one combined status field (see the manpage for the exact values this
1529
  field will have)
1530
- some more fixes for the mac export to hooks change
1531
- make Ganeti not break with DRBD 8.2.x (which changed the version
1532
  format in ``/proc/drbd``) (issue 24)
1533
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1534
  template, allowing migration from DRBD0.7+MD to DRBD8
1535

    
1536

    
1537
Version 1.2.2
1538
-------------
1539

    
1540
*(Released Wed, 30 Jan 2008)*
1541

    
1542
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1543
  support (issue 23)
1544
- add command aliases infrastructure and a few aliases
1545
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
1546
  man pages and the ``--help`` option of ``gnt-node
1547
  list``/``gnt-instance list``
1548
- fix ``gnt-backup list`` with down nodes (issue 21)
1549
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1550
- fix the dist archive and add a check for including svn/git files in
1551
  the future
1552
- some developer-related changes: improve the burnin and the QA suite,
1553
  add an upload script for testing during development
1554

    
1555

    
1556
Version 1.2.1
1557
-------------
1558

    
1559
*(Released Wed, 16 Jan 2008)*
1560

    
1561
- experimental HVM support, read the install document, section
1562
  "Initializing the cluster"
1563
- allow for the PVM hypervisor per-instance kernel and initrd paths
1564
- add a new command ``gnt-cluster verify-disks`` which uses a new
1565
  algorithm to improve the reconnection of the DRBD pairs if the device
1566
  on the secondary node has gone away
1567
- make logical volume code auto-activate LVs at disk activation time
1568
- slightly improve the speed of activating disks
1569
- allow specification of the MAC address at instance creation time, and
1570
  changing it later via ``gnt-instance modify``
1571
- fix handling of external commands that generate lots of output on
1572
  stderr
1573
- update documentation with regard to minimum version of DRBD8 supported
1574

    
1575

    
1576
Version 1.2.0
1577
-------------
1578

    
1579
*(Released Tue, 4 Dec 2007)*
1580

    
1581
- Log the ``xm create`` output to the node daemon log on failure (to
1582
  help diagnosing the error)
1583
- In debug mode, log all external commands output if failed to the logs
1584
- Change parsing of lvm commands to ignore stderr
1585

    
1586

    
1587
Version 1.2b3
1588
-------------
1589

    
1590
*(Released Wed, 28 Nov 2007)*
1591

    
1592
- Another round of updates to the DRBD 8 code to deal with more failures
1593
  in the replace secondary node operation
1594
- Some more logging of failures in disk operations (lvm, drbd)
1595
- A few documentation updates
1596
- QA updates
1597

    
1598

    
1599
Version 1.2b2
1600
-------------
1601

    
1602
*(Released Tue, 13 Nov 2007)*
1603

    
1604
- Change configuration file format from Python's Pickle to JSON.
1605
  Upgrading is possible using the cfgupgrade utility.
1606
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1607
  faster replace disks and is more stable (DRBD 8 has many improvements
1608
  compared to DRBD 0.7)
1609
- Added command line tags support (see man pages for ``gnt-instance``,
1610
  ``gnt-node``, ``gnt-cluster``)
1611
- Added instance rename support
1612
- Added multi-instance startup/shutdown
1613
- Added cluster rename support
1614
- Added ``gnt-node evacuate`` to simplify some node operations
1615
- Added instance reboot operation that can speedup reboot as compared to
1616
  stop and start
1617
- Soften the requirement that hostnames are in FQDN format
1618
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
1619
  reboots
1620
- Removed dependency on debian's patched fping that uses the
1621
  non-standard ``-S`` option
1622
- Now the OS definitions are searched for in multiple, configurable
1623
  paths (easier for distros to package)
1624
- Some changes to the hooks infrastructure (especially the new
1625
  post-configuration update hook)
1626
- Other small bugfixes
1627

    
1628
.. vim: set textwidth=72 syntax=rst :
1629
.. Local Variables:
1630
.. mode: rst
1631
.. fill-column: 72
1632
.. End: