Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 91fb0d18

History | View | Annotate | Download (97.2 kB)

1
News
2
====
3

    
4

    
5
Version 2.8.0 beta1
6
-------------------
7

    
8
*(unreleased)*
9

    
10
- The minimum Python version needed to run Ganeti is now 2.6.
11
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
12
  to always require authentication, ``--require-authentication``. It can
13
  be specified in ``$sysconfdir/default/ganeti``.
14
- A new cluster attribute 'enabled_storage_types' is introduced. It will
15
  be used to manage the storage types to be used by instances in the cluster.
16
  Initially, it will be set to a list that includes lvm, file and sharedfile
17
  if those are enabled. Additionally, it will include all storage types that
18
  are currently used by instances. The order of storage types will be based
19
  on Ganeti's history of supporting them. In the future, the first entry of
20
  the list will be used as a default storage type on instance creation.
21
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
22
  configuration back to the previous stable version.
23

    
24

    
25
Version 2.7.0 beta1
26
-------------------
27

    
28
*(Released Wed, 6 Feb 2013)*
29

    
30
- ``gnt-instance batch-create`` has been changed to use the bulk create
31
  opcode from Ganeti. This lead to incompatible changes in the format of
32
  the JSON file. It's now not a custom dict anymore but a dict
33
  compatible with the ``OpInstanceCreate`` opcode.
34
- Parent directories for file storage need to be listed in
35
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
36
  write the file automatically based on old configuration values, but it
37
  can not distribute it across all nodes and the file contents should be
38
  verified. Use ``gnt-cluster copyfile
39
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
40
  upgraded. The reason for requiring this list of paths now is that
41
  before it would have been possible to inject new paths via RPC,
42
  allowing files to be created in arbitrary locations. The RPC protocol
43
  is protected using SSL/X.509 certificates, but as a design principle
44
  Ganeti does not permit arbitrary paths to be passed.
45
- The parsing of the variants file for OSes (see
46
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
47
  lines and comment lines (starting with ``#``) are ignored for better
48
  readability.
49
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
50
  longer available. ``gnt-node add`` now invokes a new tool on the
51
  destination node, named ``prepare-node-join``, to configure the SSH
52
  daemon. Paramiko is no longer necessary to configure nodes' SSH
53
  daemons via ``gnt-node add``.
54
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
55
  for RAPI users. It allows granting permissions to query for
56
  information to a specific user without giving
57
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
58
- A new tool named ``node-cleanup`` has been added. It cleans remains of
59
  a cluster from a machine by stopping all daemons, removing
60
  certificates and ssconf files. Unless the ``--no-backup`` option is
61
  given, copies of the certificates are made.
62
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
63
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
64
  the flag is not reset after a master failover.
65
- Instance creations now support the use of opportunistic locking,
66
  potentially speeding up the (parallel) creation of multiple instances.
67
  This feature is currently only available via the :doc:`RAPI
68
  <rapi>` interface and when an instance allocator is used. If the
69
  ``opportunistic_locking`` parameter is set the opcode will try to
70
  acquire as many locks as possible, but will not wait for any locks
71
  held by other opcodes. If not enough resources can be found to
72
  allocate the instance, the temporary error code
73
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
74
  retried thereafter, with or without opportunistic locking.
75
- The functionality for allocating multiple instances at once has been
76
  overhauled and is now also available through :doc:`RAPI <rapi>`.
77
- Man pages can now be included when the documentation is built, in
78
  which case the output is in ``doc/man-html``. The configure-time
79
  option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
80
  required.
81
- A new htool called ``hroller`` helps scheduling parallel shutdown of
82
  nodes on clusters where all instances are shut down, but without ever
83
  rebooting both primary and secondary node for an instance at the same
84
  time.
85
- Queries not needing locks or RPC access to the node can now be
86
  performed by the confd daemon, making them independent from jobs, and
87
  thus faster to execute. This is selectable at configure time.
88
- The KVM hypervisor has been updated with many new hypervisor
89
  parameters, including a generic one for passing arbitrary command line
90
  values. See a complete list in :manpage:`gnt-instance(8)`.
91
- A new tool, called ``mon-collector``, is the stand-alone executor of
92
  the data collectors for a monitoring system. As of this version, it
93
  just includes the DRBD data collector, that can be executed by calling
94
  ``mon-collector`` using the ``drbd`` parameter. See
95
  :manpage:`mon-collector(7)`.
96
- New ``exclusive-storage`` node parameter added. When it's set to true,
97
  physical disks are assigned in an exclusive fashion to instances, as
98
  documented in :doc:`Partitioned Ganeti <design-partitioned>`.
99
  Currently, only instances using the ``plain`` disk template are
100
  supported.
101
- New network management functionality to support automatic allocation
102
  of IP addresses and managing of network parameters. See
103
  :manpage:`gnt-network(8)` for more details.
104
- New external storage backend, to allow managing arbitrary storage
105
  systems external to the cluster. See
106
  :manpage:`ganeti-extstorage-interface(7)`.
107
- Instance renames of LVM-based instances will now update the LV tags
108
  (which can be used to recover the instance-to-LV mapping in case of
109
  emergencies)
110
- ``hbal`` will now exit with status 0 if, during job execution over
111
  LUXI, early exit has been requested and all jobs are successful;
112
  before, exit status 1 was used, which cannot be differentiated from
113
  "job error" case
114

    
115

    
116
Version 2.6.2
117
-------------
118

    
119
*(Released Fri, 21 Dec 2012)*
120

    
121
Important behaviour change: hbal won't rebalance anymore instances which
122
have the ``auto_balance`` attribute set to false. This was the intention
123
all along, but until now it only skipped those from the N+1 memory
124
reservation (DRBD-specific).
125

    
126
A significant number of bug fixes in this release:
127

    
128
- Fixed disk adoption interaction with ipolicy checks.
129
- Fixed networking issues when instances are started, stopped or
130
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
131
- Fixed the warning in cluster verify for shared storage instances not
132
  being redundant.
133
- Fixed removal of storage directory on shared file storage (issue 262).
134
- Fixed validation of LVM volume group name in OpClusterSetParams
135
  (``gnt-cluster modify``) (issue 285).
136
- Fixed runtime memory increases (``gnt-instance modify -m``).
137
- Fixed live migration under Xen's ``xl`` mode.
138
- Fixed ``gnt-instance console`` with ``xl``.
139
- Fixed building with newer Haskell compiler/libraries.
140
- Fixed PID file writing in Haskell daemons (confd); this prevents
141
  restart issues if confd was launched manually (outside of
142
  ``daemon-util``) while another copy of it was running
143
- Fixed a type error when doing live migrations with KVM (issue 297) and
144
  the error messages for failing migrations have been improved.
145
- Fixed opcode validation for the out-of-band commands (``gnt-node
146
  power``).
147
- Fixed a type error when unsetting OS hypervisor parameters (issue
148
  311); now it's possible to unset all OS-specific hypervisor
149
  parameters.
150
- Fixed the ``dry-run`` mode for many operations: verification of
151
  results was over-zealous but didn't take into account the ``dry-run``
152
  operation, resulting in "wrong" failures.
153
- Fixed bash completion in ``gnt-job list`` when the job queue has
154
  hundreds of entries; especially with older ``bash`` versions, this
155
  results in significant CPU usage.
156

    
157
And lastly, a few other improvements have been made:
158

    
159
- Added option to force master-failover without voting (issue 282).
160
- Clarified error message on lock conflict (issue 287).
161
- Logging of newly submitted jobs has been improved (issue 290).
162
- Hostname checks have been made uniform between instance rename and
163
  create (issue 291).
164
- The ``--submit`` option is now supported by ``gnt-debug delay``.
165
- Shutting down the master daemon by sending SIGTERM now stops it from
166
  processing jobs waiting for locks; instead, those jobs will be started
167
  once again after the master daemon is started the next time (issue
168
  296).
169
- Support for Xen's ``xl`` program has been improved (besides the fixes
170
  above).
171
- Reduced logging noise in the Haskell confd daemon (only show one log
172
  entry for each config reload, instead of two).
173
- Several man page updates and typo fixes.
174

    
175

    
176
Version 2.6.1
177
-------------
178

    
179
*(Released Fri, 12 Oct 2012)*
180

    
181
A small bugfix release. Among the bugs fixed:
182

    
183
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
184
  made the command unusable.
185
- Commands that issue many jobs don't fail anymore just because some jobs
186
  take so long that other jobs are archived.
187
- Failures during ``gnt-instance reinstall`` are reflected by the exit
188
  status.
189
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
190
  DRBD is enabled.
191
- When ``always_failover`` is set, ``--allow-failover`` is not required
192
  in migrate commands anymore.
193
- ``bash_completion`` works even if extglob is disabled.
194
- Fixed bug with locks that made failover for RDB-based instances fail.
195
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
196
  a random node instead of one based on the allocator metric.
197
- Support for newer versions of pylint and pep8.
198
- Hail doesn't fail anymore when trying to add an instance of type
199
  ``file``, ``sharedfile`` or ``rbd``.
200
- Added new Makefile target to rebuild the whole distribution, so that
201
  all files are included.
202

    
203

    
204
Version 2.6.0
205
-------------
206

    
207
*(Released Fri, 27 Jul 2012)*
208

    
209

    
210
.. attention:: The ``LUXI`` protocol has been made more consistent
211
   regarding its handling of command arguments. This, however, leads to
212
   incompatibility issues with previous versions. Please ensure that you
213
   restart Ganeti daemons soon after the upgrade, otherwise most
214
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
215
   pausing/resuming the watcher, cancelling and archiving jobs, querying
216
   the cluster configuration) will fail.
217

    
218

    
219
New features
220
~~~~~~~~~~~~
221

    
222
Instance run status
223
+++++++++++++++++++
224

    
225
The current ``admin_up`` field, which used to denote whether an instance
226
should be running or not, has been removed. Instead, ``admin_state`` is
227
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
228

    
229
The rational behind this is that an instance being “down” can have
230
different meanings:
231

    
232
- it could be down during a reboot
233
- it could be temporarily be down for a reinstall
234
- or it could be down because it is deprecated and kept just for its
235
  disk
236

    
237
The previous Boolean state was making it difficult to do capacity
238
calculations: should Ganeti reserve memory for a down instance? Now, the
239
tri-state field makes it clear:
240

    
241
- in ``up`` and ``down`` state, all resources are reserved for the
242
  instance, and it can be at any time brought up if it is down
243
- in ``offline`` state, only disk space is reserved for it, but not
244
  memory or CPUs
245

    
246
The field can have an extra use: since the transition between ``up`` and
247
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
248
transition between ``offline`` and ``down`` is done via ``gnt-instance
249
modify``, it is possible to given different rights to users. For
250
example, owners of an instance could be allowed to start/stop it, but
251
not transition it out of the offline state.
252

    
253
Instance policies and specs
254
+++++++++++++++++++++++++++
255

    
256
In previous Ganeti versions, an instance creation request was not
257
limited on the minimum size and on the maximum size just by the cluster
258
resources. As such, any policy could be implemented only in third-party
259
clients (RAPI clients, or shell wrappers over ``gnt-*``
260
tools). Furthermore, calculating cluster capacity via ``hspace`` again
261
required external input with regards to instance sizes.
262

    
263
In order to improve these workflows and to allow for example better
264
per-node group differentiation, we introduced instance specs, which
265
allow declaring:
266

    
267
- minimum instance disk size, disk count, memory size, cpu count
268
- maximum values for the above metrics
269
- and “standard” values (used in ``hspace`` to calculate the standard
270
  sized instances)
271

    
272
The minimum/maximum values can be also customised at node-group level,
273
for example allowing more powerful hardware to support bigger instance
274
memory sizes.
275

    
276
Beside the instance specs, there are a few other settings belonging to
277
the instance policy framework. It is possible now to customise, per
278
cluster and node-group:
279

    
280
- the list of allowed disk templates
281
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
282
- the maximum ratio of instance to spindles (see below for more
283
  information) for local storage
284

    
285
All these together should allow all tools that talk to Ganeti to know
286
what are the ranges of allowed values for instances and the
287
over-subscription that is allowed.
288

    
289
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
290
instance configuration, and the physical CPU configuration from the
291
node. For the spindle ratios however, we didn't track before these
292
values, so new parameters have been added:
293

    
294
- a new node parameter ``spindle_count``, defaults to 1, customisable at
295
  node group or node level
296
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
297

    
298
Note that spindles in this context doesn't need to mean actual
299
mechanical hard-drives; it's just a relative number for both the node
300
I/O capacity and instance I/O consumption.
301

    
302
Instance migration behaviour
303
++++++++++++++++++++++++++++
304

    
305
While live-migration is in general desirable over failover, it is
306
possible that for some workloads it is actually worse, due to the
307
variable time of the “suspend” phase during live migration.
308

    
309
To allow the tools to work consistently over such instances (without
310
having to hard-code instance names), a new backend parameter
311
``always_failover`` has been added to control the migration/failover
312
behaviour. When set to True, all migration requests for an instance will
313
instead fall-back to failover.
314

    
315
Instance memory ballooning
316
++++++++++++++++++++++++++
317

    
318
Initial support for memory ballooning has been added. The memory for an
319
instance is no longer fixed (backend parameter ``memory``), but instead
320
can vary between minimum and maximum values (backend parameters
321
``minmem`` and ``maxmem``). Currently we only change an instance's
322
memory when:
323

    
324
- live migrating or failing over and instance and the target node
325
  doesn't have enough memory
326
- user requests changing the memory via ``gnt-instance modify
327
  --runtime-memory``
328

    
329
Instance CPU pinning
330
++++++++++++++++++++
331

    
332
In order to control the use of specific CPUs by instance, support for
333
controlling CPU pinning has been added for the Xen, HVM and LXC
334
hypervisors. This is controlled by a new hypervisor parameter
335
``cpu_mask``; details about possible values for this are in the
336
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
337
VCPU-to-CPU mapping) form will work well only when all nodes in your
338
cluster have the same amount of CPUs.
339

    
340
Disk parameters
341
+++++++++++++++
342

    
343
Another area in which Ganeti was not customisable were the parameters
344
used for storage configuration, e.g. how many stripes to use for LVM,
345
DRBD resync configuration, etc.
346

    
347
To improve this area, we've added disks parameters, which are
348
customisable at cluster and node group level, and which allow to
349
specify various parameters for disks (DRBD has the most parameters
350
currently), for example:
351

    
352
- DRBD resync algorithm and parameters (e.g. speed)
353
- the default VG for meta-data volumes for DRBD
354
- number of stripes for LVM (plain disk template)
355
- the RBD pool
356

    
357
These parameters can be modified via ``gnt-cluster modify -D …`` and
358
``gnt-group modify -D …``, and are used at either instance creation (in
359
case of LVM stripes, for example) or at disk “activation” time
360
(e.g. resync speed).
361

    
362
Rados block device support
363
++++++++++++++++++++++++++
364

    
365
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
366
denoted by the ``rbd`` disk template type. This is considered
367
experimental, feedback is welcome. For details on configuring it, see
368
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
369

    
370
Master IP setup
371
+++++++++++++++
372

    
373
The existing master IP functionality works well only in simple setups (a
374
single network shared by all nodes); however, if nodes belong to
375
different networks, then the ``/32`` setup and lack of routing
376
information is not enough.
377

    
378
To allow the master IP to function well in more complex cases, the
379
system was reworked as follows:
380

    
381
- a master IP netmask setting has been added
382
- the master IP activation/turn-down code was moved from the node daemon
383
  to a separate script
384
- whether to run the Ganeti-supplied master IP script or a user-supplied
385
  on is a ``gnt-cluster init`` setting
386

    
387
Details about the location of the standard and custom setup scripts are
388
in the man page :manpage:`gnt-cluster(8)`; for information about the
389
setup script protocol, look at the Ganeti-supplied script.
390

    
391
SPICE support
392
+++++++++++++
393

    
394
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
395
improved.
396

    
397
It is now possible to use TLS-protected connections, and when renewing
398
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
399
it is now possible to specify spice or spice CA certificates. Also, it
400
is possible to configure a password for SPICE sessions via the
401
hypervisor parameter ``spice_password_file``.
402

    
403
There are also new parameters to control the compression and streaming
404
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
405
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
406
for the spice parameters.
407

    
408
Lastly, it is now possible to see the SPICE connection information via
409
``gnt-instance console``.
410

    
411
OVF converter
412
+++++++++++++
413

    
414
A new tool (``tools/ovfconverter``) has been added that supports
415
conversion between Ganeti and the `Open Virtualization Format
416
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
417
from).
418

    
419
This relies on the ``qemu-img`` tool to convert the disk formats, so the
420
actual compatibility with other virtualization solutions depends on it.
421

    
422
Confd daemon changes
423
++++++++++++++++++++
424

    
425
The configuration query daemon (``ganeti-confd``) is now optional, and
426
has been rewritten in Haskell; whether to use the daemon at all, use the
427
Python (default) or the Haskell version is selectable at configure time
428
via the ``--enable-confd`` parameter, which can take one of the
429
``haskell``, ``python`` or ``no`` values. If not used, disabling the
430
daemon will result in a smaller footprint; for larger systems, we
431
welcome feedback on the Haskell version which might become the default
432
in future versions.
433

    
434
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
435
daemon running. The Python version doesn't implement the new call.
436

    
437

    
438
User interface changes
439
~~~~~~~~~~~~~~~~~~~~~~
440

    
441
We have replaced the ``--disks`` option of ``gnt-instance
442
replace-disks`` with a more flexible ``--disk`` option, which allows
443
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
444
disk size and mode can be changed upon recreation (via ``gnt-instance
445
recreate-disks``, which accepts the same ``--disk`` option).
446

    
447
As many people are used to a ``show`` command, we have added that as an
448
alias to ``info`` on all ``gnt-*`` commands.
449

    
450
The ``gnt-instance grow-disk`` command has a new mode in which it can
451
accept the target size of the disk, instead of the delta; this can be
452
more safe since two runs in absolute mode will be idempotent, and
453
sometimes it's also easier to specify the desired size directly.
454

    
455
Also the handling of instances with regard to offline secondaries has
456
been improved. Instance operations should not fail because one of it's
457
secondary nodes is offline, even though it's safe to proceed.
458

    
459
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
460
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
461
minors to instance name.
462

    
463
API changes
464
~~~~~~~~~~~
465

    
466
RAPI coverage has improved, with (for example) new resources for
467
recreate-disks, node power-cycle, etc.
468

    
469
Compatibility
470
~~~~~~~~~~~~~
471

    
472
There is partial support for ``xl`` in the Xen hypervisor; feedback is
473
welcome.
474

    
475
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
476
whether to still support Python 2.4 or move to Python 2.6 as minimum
477
required version.
478

    
479
Support for Fedora has been slightly improved; the provided example
480
init.d script should work better on it and the INSTALL file should
481
document the needed dependencies.
482

    
483
Internal changes
484
~~~~~~~~~~~~~~~~
485

    
486
The deprecated ``QueryLocks`` LUXI request has been removed. Use
487
``Query(what=QR_LOCK, ...)`` instead.
488

    
489
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
490
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
491
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
492
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
493
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
494

    
495
RAPI client: ``CertificateError`` now derives from
496
``GanetiApiError``. This should make it more easy to handle Ganeti
497
errors.
498

    
499
Deprecation warnings due to PyCrypto/paramiko import in
500
``tools/setup-ssh`` have been silenced, as usually they are safe; please
501
make sure to run an up-to-date paramiko version, if you use this tool.
502

    
503
The QA scripts now depend on Python 2.5 or above (the main code base
504
still works with Python 2.4).
505

    
506
The configuration file (``config.data``) is now written without
507
indentation for performance reasons; if you want to edit it, it can be
508
re-formatted via ``tools/fmtjson``.
509

    
510
A number of bugs has been fixed in the cluster merge tool.
511

    
512
``x509`` certification verification (used in import-export) has been
513
changed to allow the same clock skew as permitted by the cluster
514
verification. This will remove some rare but hard to diagnose errors in
515
import-export.
516

    
517

    
518
Version 2.6.0 rc4
519
-----------------
520

    
521
*(Released Thu, 19 Jul 2012)*
522

    
523
Very few changes from rc4 to the final release, only bugfixes:
524

    
525
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
526
  instance, fix CDROM booting for KVM instances)
527
- fixed node group modification of node parameters
528
- fixed issue in LUClusterVerifyGroup with multi-group clusters
529
- fixed generation of bash completion to ensure a stable ordering
530
- fixed a few typos
531

    
532

    
533
Version 2.6.0 rc3
534
-----------------
535

    
536
*(Released Fri, 13 Jul 2012)*
537

    
538
Third release candidate for 2.6. The following changes were done from
539
rc3 to rc4:
540

    
541
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
542
- Fixed an inconsistency in the LUXI protocol with the provided
543
  arguments (NOT backwards compatible)
544
- Fixed a bug with node groups ipolicy where ``min`` was greater than
545
  the cluster ``std`` value
546
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
547
  easier instance debugging on nodes (requires ``hconfd`` to work)
548

    
549

    
550
Version 2.6.0 rc2
551
-----------------
552

    
553
*(Released Tue, 03 Jul 2012)*
554

    
555
Second release candidate for 2.6. The following changes were done from
556
rc2 to rc3:
557

    
558
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
559
  master candidates
560
- Fixed a RAPI regression on missing beparams/memory
561
- Fixed redistribution of files on offline nodes
562
- Added possibility to run activate-disks even though secondaries are
563
  offline. With this change it relaxes also the strictness on some other
564
  commands which use activate disks internally:
565
  * ``gnt-instance start|reboot|rename|backup|export``
566
- Made it possible to remove safely an instance if its secondaries are
567
  offline
568
- Made it possible to reinstall even though secondaries are offline
569

    
570

    
571
Version 2.6.0 rc1
572
-----------------
573

    
574
*(Released Mon, 25 Jun 2012)*
575

    
576
First release candidate for 2.6. The following changes were done from
577
rc1 to rc2:
578

    
579
- Fixed bugs with disk parameters and ``rbd`` templates as well as
580
  ``instance_os_add``
581
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
582
  behaviour. It supports now the modify operation
583
- ``hcheck`` implemented to analyze cluster health and possibility of
584
  improving health by rebalance
585
- ``hbal`` has been improved in dealing with split instances
586

    
587

    
588
Version 2.6.0 beta2
589
-------------------
590

    
591
*(Released Mon, 11 Jun 2012)*
592

    
593
Second beta release of 2.6. The following changes were done from beta2
594
to rc1:
595

    
596
- Fixed ``daemon-util`` with non-root user models
597
- Fixed creation of plain instances with ``--no-wait-for-sync``
598
- Fix wrong iv_names when running ``cfgupgrade``
599
- Export more information in RAPI group queries
600
- Fixed bug when changing instance network interfaces
601
- Extended burnin to do NIC changes
602
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
603
- Changed default for DRBD barriers
604
- Fixed DRBD error reporting for syncer rate
605
- Verify the options on disk parameters
606

    
607
And of course various fixes to documentation and improved unittests and
608
QA.
609

    
610

    
611
Version 2.6.0 beta1
612
-------------------
613

    
614
*(Released Wed, 23 May 2012)*
615

    
616
First beta release of 2.6. The following changes were done from beta1 to
617
beta2:
618

    
619
- integrated patch for distributions without ``start-stop-daemon``
620
- adapted example init.d script to work on Fedora
621
- fixed log handling in Haskell daemons
622
- adapted checks in the watcher for pycurl linked against libnss
623
- add partial support for ``xl`` instead of ``xm`` for Xen
624
- fixed a type issue in cluster verification
625
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
626
  clusters)
627

    
628
Plus integrated fixes from the 2.5 branch:
629

    
630
- fixed ``kvm-ifup`` to use ``/bin/bash``
631
- fixed parallel build failures
632
- KVM live migration when using a custom keymap
633

    
634

    
635
Version 2.5.2
636
-------------
637

    
638
*(Released Tue, 24 Jul 2012)*
639

    
640
A small bugfix release, with no new features:
641

    
642
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
643
  different default shell (e.g. Debian, Ubuntu)
644
- fixed KVM startup and live migration with a custom keymap (fixes Issue
645
  243 and Debian bug #650664)
646
- fixed compatibility with KVM versions that don't support multiple boot
647
  devices (fixes Issue 230 and Debian bug #624256)
648

    
649
Additionally, a few fixes were done to the build system (fixed parallel
650
build failures) and to the unittests (fixed race condition in test for
651
FileID functions, and the default enable/disable mode for QA test is now
652
customisable).
653

    
654

    
655
Version 2.5.1
656
-------------
657

    
658
*(Released Fri, 11 May 2012)*
659

    
660
A small bugfix release.
661

    
662
The main issues solved are on the topic of compatibility with newer LVM
663
releases:
664

    
665
- fixed parsing of ``lv_attr`` field
666
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
667
  the ``--force`` flag is needed
668

    
669
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
670
to accept kernel 3.x too (was hardcoded to 2.6.*).
671

    
672
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
673
of hook results and that also made display of other errors suboptimal
674
was fixed; the code behaves now like 2.4 and earlier.
675

    
676
Another change in 2.5, the cleanup of the OS scripts environment, is too
677
aggressive: it removed even the ``PATH`` variable, which requires the OS
678
scripts to *always* need to export it. Since this is a bit too strict,
679
we now export a minimal PATH, the same that we export for hooks.
680

    
681
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
682
integrated into this release.
683

    
684
Finally, a few other miscellaneous changes were done (no new features,
685
just small improvements):
686

    
687
- Fix ``gnt-group --help`` display
688
- Fix hardcoded Xen kernel path
689
- Fix grow-disk handling of invalid units
690
- Update synopsis for ``gnt-cluster repair-disk-sizes``
691
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
692

    
693

    
694
Version 2.5.0
695
-------------
696

    
697
*(Released Thu, 12 Apr 2012)*
698

    
699
Incompatible/important changes and bugfixes
700
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
701

    
702
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
703
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
704
  to match the underlying LUXI interface.
705
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
706
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
707
  not maintain backwards-compatibility as the underlying operation
708
  changed in an incompatible way. The RAPI client can talk to old
709
  servers, but it needs to be told so as the return value changed.
710
- When creating file-based instances via RAPI, the ``file_driver``
711
  parameter no longer defaults to ``loop`` and must be specified.
712
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
713
  ``link`` instead.
714
- Support for the undocumented and deprecated RAPI instance creation
715
  request format version 0 has been dropped. Use version 1, supported
716
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
717
- Pyparsing 1.4.6 or above is required, see :doc:`installation
718
  documentation <install>`.
719
- The "cluster-verify" hooks are now executed per group by the
720
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
721
  if you just run ``gnt-cluster verify``, which generates one opcode per
722
  group.
723
- The environment as passed to the OS scripts is cleared, and thus no
724
  environment variables defined in the node daemon's environment will be
725
  inherited by the scripts.
726
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
727
  deprecated.
728
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
729
  support operations involving multiple node groups.
730
- Offline nodes are ignored when failing over an instance.
731
- Support for KVM version 1.0, which changed the version reporting format
732
  from 3 to 2 digits.
733
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
734
  removal.
735
- ``Makefile`` is now compatible with Automake 1.11.2
736
- Includes all bugfixes made in the 2.4 series
737

    
738
New features
739
~~~~~~~~~~~~
740

    
741
- The ganeti-htools project has been merged into the ganeti-core source
742
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
743
- Implemented support for :doc:`shared storage <design-shared-storage>`.
744
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
745
  GPT-style partition tables (requires `parted
746
  <http://www.gnu.org/s/parted/>`_).
747
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
748
- Allowed adding tags on instance creation.
749
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
750
- Allow instances to be started in a paused state, enabling the user to
751
  see the complete console output on boot using the console.
752
- Added new hypervisor flag to control default reboot behaviour
753
  (``reboot_behavior``).
754
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
755
- Improved out-of-band management support:
756

    
757
  - Added ``gnt-node health`` command reporting the health status of
758
    nodes.
759
  - Added ``gnt-node power`` command to manage power status of nodes.
760
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
761

    
762
- Instance migration can fall back to failover if instance is not
763
  running.
764
- Filters can be used when listing nodes, instances, groups and locks;
765
  see :manpage:`ganeti(7)` manpage.
766
- Added post-execution status as variables to :doc:`hooks <hooks>`
767
  environment.
768
- Instance tags are exported/imported together with the instance.
769
- When given an explicit job ID, ``gnt-job info`` will work for archived
770
  jobs.
771
- Jobs can define dependencies on other jobs (not yet supported via
772
  RAPI or command line, but used by internal commands and usable via
773
  LUXI).
774

    
775
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
776
    dependencies.
777

    
778
- Instance failover is now available as a RAPI resource
779
  (``/2/instances/[instance_name]/failover``).
780
- ``gnt-instance info`` defaults to static information if primary node
781
  is offline.
782
- Opcodes have a new ``comment`` attribute.
783
- Added basic SPICE support to KVM hypervisor.
784
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
785

    
786
Node group improvements
787
~~~~~~~~~~~~~~~~~~~~~~~
788

    
789
- ``gnt-cluster verify`` has been modified to check groups separately,
790
  thereby improving performance.
791
- Node group support has been added to ``gnt-cluster verify-disks``,
792
  which now operates per node group.
793
- Watcher has been changed to work better with node groups.
794

    
795
  - One process and state file per node group.
796
  - Slow watcher in one group doesn't block other group's watcher.
797

    
798
- Added new command, ``gnt-group evacuate``, to move all instances in a
799
  node group to other groups.
800
- Added ``gnt-instance change-group`` to move an instance to another
801
  node group.
802
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
803
  per-group operations.
804
- Node groups can be tagged.
805
- Some operations switch from an exclusive to a shared lock as soon as
806
  possible.
807
- Instance's primary and secondary nodes' groups are now available as
808
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
809
  and ``snodes.group.uuid``).
810

    
811
Misc
812
~~~~
813

    
814
- Numerous updates to documentation and manpages.
815

    
816
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
817
    descriptions.
818
  - Some opcode/job results are now also documented, see :doc:`RAPI
819
    <rapi>`.
820

    
821
- A lockset's internal lock is now also visible in lock monitor.
822
- Log messages from job queue workers now contain information about the
823
  opcode they're processing.
824
- ``gnt-instance console`` no longer requires the instance lock.
825
- A short delay when waiting for job changes reduces the number of LUXI
826
  requests significantly.
827
- DRBD metadata volumes are overwritten with zeros during disk creation.
828
- Out-of-band commands no longer acquire the cluster lock in exclusive
829
  mode.
830
- ``devel/upload`` now uses correct permissions for directories.
831

    
832

    
833
Version 2.5.0 rc6
834
-----------------
835

    
836
*(Released Fri, 23 Mar 2012)*
837

    
838
This was the sixth release candidate of the 2.5 series.
839

    
840

    
841
Version 2.5.0 rc5
842
-----------------
843

    
844
*(Released Mon, 9 Jan 2012)*
845

    
846
This was the fifth release candidate of the 2.5 series.
847

    
848

    
849
Version 2.5.0 rc4
850
-----------------
851

    
852
*(Released Thu, 27 Oct 2011)*
853

    
854
This was the fourth release candidate of the 2.5 series.
855

    
856

    
857
Version 2.5.0 rc3
858
-----------------
859

    
860
*(Released Wed, 26 Oct 2011)*
861

    
862
This was the third release candidate of the 2.5 series.
863

    
864

    
865
Version 2.5.0 rc2
866
-----------------
867

    
868
*(Released Tue, 18 Oct 2011)*
869

    
870
This was the second release candidate of the 2.5 series.
871

    
872

    
873
Version 2.5.0 rc1
874
-----------------
875

    
876
*(Released Tue, 4 Oct 2011)*
877

    
878
This was the first release candidate of the 2.5 series.
879

    
880

    
881
Version 2.5.0 beta3
882
-------------------
883

    
884
*(Released Wed, 31 Aug 2011)*
885

    
886
This was the third beta release of the 2.5 series.
887

    
888

    
889
Version 2.5.0 beta2
890
-------------------
891

    
892
*(Released Mon, 22 Aug 2011)*
893

    
894
This was the second beta release of the 2.5 series.
895

    
896

    
897
Version 2.5.0 beta1
898
-------------------
899

    
900
*(Released Fri, 12 Aug 2011)*
901

    
902
This was the first beta release of the 2.5 series.
903

    
904

    
905
Version 2.4.5
906
-------------
907

    
908
*(Released Thu, 27 Oct 2011)*
909

    
910
- Fixed bug when parsing command line parameter values ending in
911
  backslash
912
- Fixed assertion error after unclean master shutdown
913
- Disable HTTP client pool for RPC, significantly reducing memory usage
914
  of master daemon
915
- Fixed queue archive creation with wrong permissions
916

    
917

    
918
Version 2.4.4
919
-------------
920

    
921
*(Released Tue, 23 Aug 2011)*
922

    
923
Small bug-fixes:
924

    
925
- Fixed documentation for importing with ``--src-dir`` option
926
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
927
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
928

    
929

    
930
Version 2.4.3
931
-------------
932

    
933
*(Released Fri, 5 Aug 2011)*
934

    
935
Many bug-fixes and a few small features:
936

    
937
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
938
  issues when you tried to add an instance with two MAC addresses in one
939
  request
940
- KVM: fixed per-instance stored UID value
941
- KVM: configure bridged NICs at migration start
942
- KVM: Fix a bug where instance will not start with never KVM versions
943
  (>= 0.14)
944
- Added OS search path to ``gnt-cluster info``
945
- Fixed an issue with ``file_storage_dir`` where you were forced to
946
  provide an absolute path, but the documentation states it is a
947
  relative path, the documentation was right
948
- Added a new parameter to instance stop/start called ``--no-remember``
949
  that will make the state change to not be remembered
950
- Implemented ``no_remember`` at RAPI level
951
- Improved the documentation
952
- Node evacuation: don't call IAllocator if node is already empty
953
- Fixed bug in DRBD8 replace disks on current nodes
954
- Fixed bug in recreate-disks for DRBD instances
955
- Moved assertion checking locks in ``gnt-instance replace-disks``
956
  causing it to abort with not owning the right locks for some situation
957
- Job queue: Fixed potential race condition when cancelling queued jobs
958
- Fixed off-by-one bug in job serial generation
959
- ``gnt-node volumes``: Fix instance names
960
- Fixed aliases in bash completion
961
- Fixed a bug in reopening log files after being sent a SIGHUP
962
- Added a flag to burnin to allow specifying VCPU count
963
- Bugfixes to non-root Ganeti configuration
964

    
965

    
966
Version 2.4.2
967
-------------
968

    
969
*(Released Thu, 12 May 2011)*
970

    
971
Many bug-fixes and a few new small features:
972

    
973
- Fixed a bug related to log opening failures
974
- Fixed a bug in instance listing with orphan instances
975
- Fixed a bug which prevented resetting the cluster-level node parameter
976
  ``oob_program`` to the default
977
- Many fixes related to the ``cluster-merge`` tool
978
- Fixed a race condition in the lock monitor, which caused failures
979
  during (at least) creation of many instances in parallel
980
- Improved output for gnt-job info
981
- Removed the quiet flag on some ssh calls which prevented debugging
982
  failures
983
- Improved the N+1 failure messages in cluster verify by actually
984
  showing the memory values (needed and available)
985
- Increased lock attempt timeouts so that when executing long operations
986
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
987
  too early and thus prevent the use of the 'fair' mechanism
988
- Changed instance query data (``gnt-instance info``) to not acquire
989
  locks unless needed, thus allowing its use on locked instance if only
990
  static information is asked for
991
- Improved behaviour with filesystems that do not support rename on an
992
  opened file
993
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
994
  kept locks on all nodes during the wipe, which is unneeded
995
- Fixed ``gnt-watcher`` handling of errors during hooks execution
996
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
997
  10GiB) which caused the wipe to fail right at the end in some cases
998
- Fixed master IP activation when doing master failover with no-voting
999
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1000
  the master node itself
1001
- Fixed potential data-loss in under disk full conditions, where Ganeti
1002
  wouldn't check correctly the return code and would consider
1003
  partially-written files 'correct'
1004
- Fixed bug related to multiple VGs and DRBD disk replacing
1005
- Added new disk parameter ``metavg`` that allows placement of the meta
1006
  device for DRBD in a different volume group
1007
- Fixed error handling in the node daemon when the system libc doesn't
1008
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1009
- Fixed lock release during replace-disks, which kept cluster-wide locks
1010
  when doing disk replaces with an iallocator script
1011
- Added check for missing bridges in cluster verify
1012
- Handle EPIPE errors while writing to the terminal better, so that
1013
  piping the output to e.g. ``less`` doesn't cause a backtrace
1014
- Fixed rare case where a ^C during Luxi calls could have been
1015
  interpreted as server errors, instead of simply terminating
1016
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1017
  assign-nodes``)
1018
- Added a few more parameters to the KVM hypervisor, allowing a second
1019
  CDROM, custom disk type for CDROMs and a floppy image
1020
- Removed redundant message in instance rename when the name is given
1021
  already as a FQDN
1022
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1023
  disks on new nodes, allowing recreation when the original instance
1024
  nodes are completely gone
1025
- Added option when converting disk templates to DRBD to skip waiting
1026
  for the resync, in order to make the instance available sooner
1027
- Added two new variables to the OS scripts environment (containing the
1028
  instance's nodes)
1029
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1030
  to allow use of ``pvgrub`` as bootloader
1031
- Changed the instance memory modifications to only check out-of-memory
1032
  conditions on memory increases, and turned the secondary node warnings
1033
  into errors (they can still be overridden via ``--force``)
1034
- Fixed the handling of a corner case when the Python installation gets
1035
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1036
  to execute a command that doesn't exist
1037
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1038
  node of the instance returned failures during instance shutdown; this
1039
  adds the option ``--ignore-consistency`` to gnt-instance move
1040

    
1041
And as usual, various improvements to the error messages, documentation
1042
and man pages.
1043

    
1044

    
1045
Version 2.4.1
1046
-------------
1047

    
1048
*(Released Wed, 09 Mar 2011)*
1049

    
1050
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1051
the RAPI users file if run twice (even with ``--dry-run``).
1052

    
1053
The release fixes that bug (nothing else changed).
1054

    
1055

    
1056
Version 2.4.0
1057
-------------
1058

    
1059
*(Released Mon, 07 Mar 2011)*
1060

    
1061
Final 2.4.0 release. Just a few small fixes:
1062

    
1063
- Fixed RAPI node evacuate
1064
- Fixed the kvm-ifup script
1065
- Fixed internal error handling for special job cases
1066
- Updated man page to specify the escaping feature for options
1067

    
1068

    
1069
Version 2.4.0 rc3
1070
-----------------
1071

    
1072
*(Released Mon, 28 Feb 2011)*
1073

    
1074
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1075
that this feature wiped the disks of the wrong instance, leading to loss
1076
of data.
1077

    
1078
Other changes:
1079

    
1080
- Fixed title of query field containing instance name
1081
- Expanded the glossary in the documentation
1082
- Fixed one unittest (internal issue)
1083

    
1084

    
1085
Version 2.4.0 rc2
1086
-----------------
1087

    
1088
*(Released Mon, 21 Feb 2011)*
1089

    
1090
A number of bug fixes plus just a couple functionality changes.
1091

    
1092
On the user-visible side, the ``gnt-* list`` command output has changed
1093
with respect to "special" field states. The current rc1 style of display
1094
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1095
the default output mode special fields states are displayed as follows:
1096

    
1097
- Offline resource: ``*``
1098
- Unavailable/not applicable: ``-``
1099
- Data missing (RPC failure): ``?``
1100
- Unknown field: ``??``
1101

    
1102
Another user-visible change is the addition of ``--force-join`` to
1103
``gnt-node add``.
1104

    
1105
As for bug fixes:
1106

    
1107
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1108
- Fixed regression in RAPI/instance reinstall where all parameters were
1109
  required (instead of optional)
1110
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1111
- Fixed iallocator usage (offline nodes were not considered offline)
1112
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1113
- Fixed hypervisor and OS parameter validation with respect to
1114
  non-vm_capable nodes
1115
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1116
  cosmetic)
1117
- Fixed ``tools/listrunner`` with respect to agent-based usage
1118

    
1119

    
1120
Version 2.4.0 rc1
1121
-----------------
1122

    
1123
*(Released Fri,  4 Feb 2011)*
1124

    
1125
Many changes and fixes since the beta1 release. While there were some
1126
internal changes, the code has been mostly stabilised for the RC
1127
release.
1128

    
1129
Note: the dumb allocator was removed in this release, as it was not kept
1130
up-to-date with the IAllocator protocol changes. It is recommended to
1131
use the ``hail`` command from the ganeti-htools package.
1132

    
1133
Note: the 2.4 and up versions of Ganeti are not compatible with the
1134
0.2.x branch of ganeti-htools. You need to upgrade to
1135
ganeti-htools-0.3.0 (or later).
1136

    
1137
Regressions fixed from 2.3
1138
~~~~~~~~~~~~~~~~~~~~~~~~~~
1139

    
1140
- Fixed the ``gnt-cluster verify-disks`` command
1141
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1142
  serially on nodes)
1143
- Fixed disk adoption breakage
1144
- Fixed wrong headers in instance listing for field aliases
1145

    
1146
Other bugs fixed
1147
~~~~~~~~~~~~~~~~
1148

    
1149
- Fixed corner case in KVM handling of NICs
1150
- Fixed many cases of wrong handling of non-vm_capable nodes
1151
- Fixed a bug where a missing instance symlink was not possible to
1152
  recreate with any ``gnt-*`` command (now ``gnt-instance
1153
  activate-disks`` does it)
1154
- Fixed the volume group name as reported by ``gnt-cluster
1155
  verify-disks``
1156
- Increased timeouts for the import-export code, hopefully leading to
1157
  fewer aborts due network or instance timeouts
1158
- Fixed bug in ``gnt-node list-storage``
1159
- Fixed bug where not all daemons were started on cluster
1160
  initialisation, but only at the first watcher run
1161
- Fixed many bugs in the OOB implementation
1162
- Fixed watcher behaviour in presence of instances with offline
1163
  secondaries
1164
- Fixed instance list output for instances running on the wrong node
1165
- a few fixes to the cluster-merge tool, but it still cannot merge
1166
  multi-node groups (currently it is not recommended to use this tool)
1167

    
1168

    
1169
Improvements
1170
~~~~~~~~~~~~
1171

    
1172
- Improved network configuration for the KVM hypervisor
1173
- Added e1000 as a supported NIC for Xen-HVM
1174
- Improved the lvmstrap tool to also be able to use partitions, as
1175
  opposed to full disks
1176
- Improved speed of disk wiping (the cluster parameter
1177
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1178
  of instance creations
1179
- Added documentation for the OS parameters
1180
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1181
  hypervisor is not responding
1182
- Added display of blacklisted and hidden OS information in
1183
  ``gnt-cluster info``
1184
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1185
  NIC and node parameters, which might create problems with currently
1186
  invalid (but undetected) configuration files, but prevents validation
1187
  failures when unrelated parameters are modified
1188
- Changed cluster initialisation to wait for the master daemon to become
1189
  available
1190
- Expanded the RAPI interface:
1191

    
1192
  - Added config redistribution resource
1193
  - Added activation/deactivation of instance disks
1194
  - Added export of console information
1195

    
1196
- Implemented log file reopening on SIGHUP, which allows using
1197
  logrotate(8) for the Ganeti log files
1198
- Added a basic OOB helper script as an example
1199

    
1200

    
1201
Version 2.4.0 beta1
1202
-------------------
1203

    
1204
*(Released Fri, 14 Jan 2011)*
1205

    
1206
User-visible
1207
~~~~~~~~~~~~
1208

    
1209
- Fixed timezone issues when formatting timestamps
1210
- Added support for node groups, available via ``gnt-group`` and other
1211
  commands
1212
- Added out-of-band framework and management, see :doc:`design
1213
  document <design-oob>`
1214
- Removed support for roman numbers from ``gnt-node list`` and
1215
  ``gnt-instance list``.
1216
- Allowed modification of master network interface via ``gnt-cluster
1217
  modify --master-netdev``
1218
- Accept offline secondaries while shutting down instance disks
1219
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1220
- Added support for multiple LVM volume groups
1221
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1222
  requested
1223
- Added commands to list available fields:
1224

    
1225
  - ``gnt-node list-fields``
1226
  - ``gnt-group list-fields``
1227
  - ``gnt-instance list-fields``
1228

    
1229
- Updated documentation and man pages
1230

    
1231
Integration
1232
~~~~~~~~~~~
1233

    
1234
- Moved ``rapi_users`` file into separate directory, now named
1235
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1236
  symlink
1237
- Added new tool for running commands on many machines,
1238
  ``tools/ganeti-listrunner``
1239
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1240
  improving the ``gnt-instance console`` output
1241
- Allowed customisation of disk index separator at ``configure`` time
1242
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1243
- Added support for non-partitioned md disks in ``lvmstrap``
1244
- Added script to gracefully power off KVM instances
1245
- Split ``utils`` module into smaller parts
1246
- Changed query operations to return more detailed information, e.g.
1247
  whether an information is unavailable due to an offline node. To use
1248
  this new functionality, the LUXI call ``Query`` must be used. Field
1249
  information is now stored by the master daemon and can be retrieved
1250
  using ``QueryFields``. Instances, nodes and groups can also be queried
1251
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1252
  exposed via RAPI). The following commands make use of this
1253
  infrastructure change:
1254

    
1255
  - ``gnt-group list``
1256
  - ``gnt-group list-fields``
1257
  - ``gnt-node list``
1258
  - ``gnt-node list-fields``
1259
  - ``gnt-instance list``
1260
  - ``gnt-instance list-fields``
1261
  - ``gnt-debug locks``
1262

    
1263
Remote API
1264
~~~~~~~~~~
1265

    
1266
- New RAPI resources (see :doc:`rapi`):
1267

    
1268
  - ``/2/modify``
1269
  - ``/2/groups``
1270
  - ``/2/groups/[group_name]``
1271
  - ``/2/groups/[group_name]/assign-nodes``
1272
  - ``/2/groups/[group_name]/modify``
1273
  - ``/2/groups/[group_name]/rename``
1274
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1275

    
1276
- RAPI changes:
1277

    
1278
  - Implemented ``no_install`` for instance creation
1279
  - Implemented OS parameters for instance reinstallation, allowing
1280
    use of special settings on reinstallation (e.g. for preserving data)
1281

    
1282
Misc
1283
~~~~
1284

    
1285
- Added IPv6 support in import/export
1286
- Pause DRBD synchronization while wiping disks on instance creation
1287
- Updated unittests and QA scripts
1288
- Improved network parameters passed to KVM
1289
- Converted man pages from docbook to reStructuredText
1290

    
1291

    
1292
Version 2.3.1
1293
-------------
1294

    
1295
*(Released Mon, 20 Dec 2010)*
1296

    
1297
Released version 2.3.1~rc1 without any changes.
1298

    
1299

    
1300
Version 2.3.1 rc1
1301
-----------------
1302

    
1303
*(Released Wed, 1 Dec 2010)*
1304

    
1305
- impexpd: Disable OpenSSL compression in socat if possible (backport
1306
  from master, commit e90739d625b, see :doc:`installation guide
1307
  <install-quick>` for details)
1308
- Changed unittest coverage report to exclude test scripts
1309
- Added script to check version format
1310

    
1311

    
1312
Version 2.3.0
1313
-------------
1314

    
1315
*(Released Wed, 1 Dec 2010)*
1316

    
1317
Released version 2.3.0~rc1 without any changes.
1318

    
1319

    
1320
Version 2.3.0 rc1
1321
-----------------
1322

    
1323
*(Released Fri, 19 Nov 2010)*
1324

    
1325
A number of bugfixes and documentation updates:
1326

    
1327
- Update ganeti-os-interface documentation
1328
- Fixed a bug related to duplicate MACs or similar items which should be
1329
  unique
1330
- Fix breakage in OS state modify
1331
- Reinstall instance: disallow offline secondaries (fixes bug related to
1332
  OS changing but reinstall failing)
1333
- plus all the other fixes between 2.2.1 and 2.2.2
1334

    
1335

    
1336
Version 2.3.0 rc0
1337
-----------------
1338

    
1339
*(Released Tue, 2 Nov 2010)*
1340

    
1341
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1342
- Fixed master failover race with watcher
1343
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1344
  configuration
1345
- Accept previously stopped instance for export with instance removal
1346
- Simplify and extend the environment variables for instance OS scripts
1347
- Added new node flags, ``master_capable`` and ``vm_capable``
1348
- Added optional instance disk wiping prior during allocation. This is a
1349
  cluster-wide option and can be set/modified using
1350
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1351
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1352
  :doc:`install-quick`
1353
- Added a new watcher option (``--ignore-pause``)
1354
- Added option to ignore offline node on instance start/stop
1355
  (``--ignore-offline``)
1356
- Allow overriding OS parameters with ``gnt-instance reinstall``
1357
- Added ability to change node's secondary IP address using ``gnt-node
1358
  modify``
1359
- Implemented privilege separation for all daemons except
1360
  ``ganeti-noded``, see ``configure`` options
1361
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1362
  verify``
1363
- Implemented job priorities (see ``ganeti(7)`` manpage)
1364
- Ignore failures while shutting down instances during failover from
1365
  offline node
1366
- Exit daemon's bootstrap process only once daemon is ready
1367
- Export more information via ``LUInstanceQuery``/remote API
1368
- Improved documentation, QA and unittests
1369
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1370
  restart if the file was created or changed
1371
- Added LUXI protocol version sent with each request and response,
1372
  allowing detection of server/client mismatches
1373
- Moved the Python scripts among gnt-* and ganeti-* into modules
1374
- Moved all code related to setting up SSH to an external script,
1375
  ``setup-ssh``
1376
- Infrastructure changes for node group support in future versions
1377

    
1378

    
1379
Version 2.2.2
1380
-------------
1381

    
1382
*(Released Fri, 19 Nov 2010)*
1383

    
1384
A few small bugs fixed, and some improvements to the build system:
1385

    
1386
- Fix documentation regarding conversion to drbd
1387
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1388
  -B``)
1389
- Fix error handling in node modify with multiple changes
1390
- Allow remote imports without checked names
1391

    
1392

    
1393
Version 2.2.1
1394
-------------
1395

    
1396
*(Released Tue, 19 Oct 2010)*
1397

    
1398
- Disable SSL session ID cache in RPC client
1399

    
1400

    
1401
Version 2.2.1 rc1
1402
-----------------
1403

    
1404
*(Released Thu, 14 Oct 2010)*
1405

    
1406
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1407
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1408
  with GnuTLS
1409
- Fix problems with interaction between Curl and Python's HTTP server,
1410
  resulting in increased speed in many RPC calls
1411
- Improve our release script to prevent breakage with older aclocal and
1412
  Python 2.6
1413

    
1414

    
1415
Version 2.2.1 rc0
1416
-----------------
1417

    
1418
*(Released Thu, 7 Oct 2010)*
1419

    
1420
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1421
  value retrieved from master
1422
- Added support for blacklisted or hidden OS definitions
1423
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1424
- Added support for -mem-path in KVM hypervisor abstraction layer
1425
- Allow overriding instance parameters in tool for inter-cluster
1426
  instance moves (``tools/move-instance``)
1427
- Improved opcode summaries (e.g. in ``gnt-job list``)
1428
- Improve consistency of OS listing by sorting it
1429
- Documentation updates
1430

    
1431

    
1432
Version 2.2.0.1
1433
---------------
1434

    
1435
*(Released Fri, 8 Oct 2010)*
1436

    
1437
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1438

    
1439

    
1440
Version 2.2.0
1441
-------------
1442

    
1443
*(Released Mon, 4 Oct 2010)*
1444

    
1445
- Fixed regression in ``gnt-instance rename``
1446

    
1447

    
1448
Version 2.2.0 rc2
1449
-----------------
1450

    
1451
*(Released Wed, 22 Sep 2010)*
1452

    
1453
- Fixed OS_VARIANT variable for OS scripts
1454
- Fixed cluster tag operations via RAPI
1455
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1456
- Disabled RAPI CA checks in watcher
1457

    
1458

    
1459
Version 2.2.0 rc1
1460
-----------------
1461

    
1462
*(Released Mon, 23 Aug 2010)*
1463

    
1464
- Support DRBD versions of the format "a.b.c.d"
1465
- Updated manpages
1466
- Re-introduce support for usage from multiple threads in RAPI client
1467
- Instance renames and modify via RAPI
1468
- Work around race condition between processing and archival in job
1469
  queue
1470
- Mark opcodes following failed one as failed, too
1471
- Job field ``lock_status`` was removed due to difficulties making it
1472
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1473
  locks is expected for a later 2.2.x release
1474
- Fixed dry-run behaviour with many commands
1475
- Support ``ssh-agent`` again when adding nodes
1476
- Many additional bugfixes
1477

    
1478

    
1479
Version 2.2.0 rc0
1480
-----------------
1481

    
1482
*(Released Fri, 30 Jul 2010)*
1483

    
1484
Important change: the internal RPC mechanism between Ganeti nodes has
1485
changed from using a home-grown http library (based on the Python base
1486
libraries) to use the PycURL library. This requires that PycURL is
1487
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1488
against GnuTLS by default. cURL's support for GnuTLS had known issues
1489
before cURL 7.21.0 and we recommend using the latest cURL release or
1490
linking against OpenSSL. Most other distributions already link PycURL
1491
and cURL against OpenSSL. The command::
1492

    
1493
  python -c 'import pycurl; print pycurl.version'
1494

    
1495
can be used to determine the libraries PycURL and cURL are linked
1496
against.
1497

    
1498
Other significant changes:
1499

    
1500
- Rewrote much of the internals of the job queue, in order to achieve
1501
  better parallelism; this decouples job query operations from the job
1502
  processing, and it should allow much nicer behaviour of the master
1503
  daemon under load, and it also has uncovered some long-standing bugs
1504
  related to the job serialisation (now fixed)
1505
- Added a default iallocator setting to the cluster parameters,
1506
  eliminating the need to always pass nodes or an iallocator for
1507
  operations that require selection of new node(s)
1508
- Added experimental support for the LXC virtualization method
1509
- Added support for OS parameters, which allows the installation of
1510
  instances to pass parameter to OS scripts in order to customise the
1511
  instance
1512
- Added a hypervisor parameter controlling the migration type (live or
1513
  non-live), since hypervisors have various levels of reliability; this
1514
  has renamed the 'live' parameter to 'mode'
1515
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
1516
  logical volumes, meaning that cluster verify will ignore them and not
1517
  flag their presence as errors
1518
- The watcher will now reset the error count for failed instances after
1519
  8 hours, thus allowing self-healing if the problem that caused the
1520
  instances to be down/fail to start has cleared in the meantime
1521
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1522
  check for, and warn, if the drbd module parameter ``usermode_helper``
1523
  is not consistent with the cluster-wide setting; this is needed to
1524
  make diagnose easier of failed drbd creations
1525
- Started adding base IPv6 support, but this is not yet
1526
  enabled/available for use
1527
- Rename operations (cluster, instance) will now return the new name,
1528
  which is especially useful if a short name was passed in
1529
- Added support for instance migration in RAPI
1530
- Added a tool to pre-configure nodes for the SSH setup, before joining
1531
  them to the cluster; this will allow in the future a simplified model
1532
  for node joining (but not yet fully enabled in 2.2); this needs the
1533
  paramiko python library
1534
- Fixed handling of name-resolving errors
1535
- Fixed consistency of job results on the error path
1536
- Fixed master-failover race condition when executed multiple times in
1537
  sequence
1538
- Fixed many bugs related to the job queue (mostly introduced during the
1539
  2.2 development cycle, so not all are impacting 2.1)
1540
- Fixed instance migration with missing disk symlinks
1541
- Fixed handling of unknown jobs in ``gnt-job archive``
1542
- And many other small fixes/improvements
1543

    
1544
Internal changes:
1545

    
1546
- Enhanced both the unittest and the QA coverage
1547
- Switched the opcode validation to a generic model, and extended the
1548
  validation to all opcode parameters
1549
- Changed more parts of the code that write shell scripts to use the
1550
  same class for this
1551
- Switched the master daemon to use the asyncore library for the Luxi
1552
  server endpoint
1553

    
1554

    
1555
Version 2.2.0 beta0
1556
-------------------
1557

    
1558
*(Released Thu, 17 Jun 2010)*
1559

    
1560
- Added tool (``move-instance``) and infrastructure to move instances
1561
  between separate clusters (see :doc:`separate documentation
1562
  <move-instance>` and :doc:`design document <design-2.2>`)
1563
- Added per-request RPC timeout
1564
- RAPI now requires a Content-Type header for requests with a body (e.g.
1565
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1566
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
1567
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1568
  reachable
1569
- Implemented initial support for running Ganeti daemons as separate
1570
  users, see configure-time flags ``--with-user-prefix`` and
1571
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1572
  time)
1573
- Instances can be removed after export (``gnt-backup export
1574
  --remove-instance``)
1575
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1576
  key (instead of 1024 bit)
1577
- Added new cluster configuration file for cluster domain secret
1578
- Import/export now use SSL instead of SSH
1579
- Added support for showing estimated time when exporting an instance,
1580
  see the ``ganeti-os-interface(7)`` manpage and look for
1581
  ``EXP_SIZE_FD``
1582

    
1583

    
1584
Version 2.1.8
1585
-------------
1586

    
1587
*(Released Tue, 16 Nov 2010)*
1588

    
1589
Some more bugfixes. Unless critical bugs occur, this will be the last
1590
2.1 release:
1591

    
1592
- Fix case of MAC special-values
1593
- Fix mac checker regex
1594
- backend: Fix typo causing "out of range" error
1595
- Add missing --units in gnt-instance list man page
1596

    
1597

    
1598
Version 2.1.7
1599
-------------
1600

    
1601
*(Released Tue, 24 Aug 2010)*
1602

    
1603
Bugfixes only:
1604
  - Don't ignore secondary node silently on non-mirrored disk templates
1605
    (issue 113)
1606
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1607
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1608
  - Properly document the usb_mouse parameter
1609
  - Fix path in ganeti-rapi(8) (issue 116)
1610
  - Adjust error message when the ganeti user's .ssh directory is
1611
    missing
1612
  - Add same-node-check when changing the disk template to drbd
1613

    
1614

    
1615
Version 2.1.6
1616
-------------
1617

    
1618
*(Released Fri, 16 Jul 2010)*
1619

    
1620
Bugfixes only:
1621
  - Add an option to only select some reboot types during qa/burnin.
1622
    (on some hypervisors consequent reboots are not supported)
1623
  - Fix infrequent race condition in master failover. Sometimes the old
1624
    master ip address would be still detected as up for a short time
1625
    after it was removed, causing failover to fail.
1626
  - Decrease mlockall warnings when the ctypes module is missing. On
1627
    Python 2.4 we support running even if no ctypes module is installed,
1628
    but we were too verbose about this issue.
1629
  - Fix building on old distributions, on which man doesn't have a
1630
    --warnings option.
1631
  - Fix RAPI not to ignore the MAC address on instance creation
1632
  - Implement the old instance creation format in the RAPI client.
1633

    
1634

    
1635
Version 2.1.5
1636
-------------
1637

    
1638
*(Released Thu, 01 Jul 2010)*
1639

    
1640
A small bugfix release:
1641
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1642
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1643
    a non-existing option
1644
  - Fix instance create: the --force-variant option was ignored
1645
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
1646
  - Fix modify node storage with non-FQDN arguments
1647
  - Fix RAPI client to authenticate under Python 2.6 when used
1648
    for more than 5 requests needing authentication
1649
  - Fix gnt-instance modify -t (storage) giving a wrong error message
1650
    when converting a non-shutdown drbd instance to plain
1651

    
1652

    
1653
Version 2.1.4
1654
-------------
1655

    
1656
*(Released Fri, 18 Jun 2010)*
1657

    
1658
A small bugfix release:
1659

    
1660
  - Fix live migration of KVM instances started with older Ganeti
1661
    versions which had fewer hypervisor parameters
1662
  - Fix gnt-instance grow-disk on down instances
1663
  - Fix an error-reporting bug during instance migration
1664
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
1665
    silently ignoring broken ones
1666
  - Fix an RPC error reporting bug affecting, for example, RAPI client
1667
    users
1668
  - Fix bug triggered by different API version os-es on different nodes
1669
  - Fix a bug in instance startup with custom hvparams: OS level
1670
    parameters would fail to be applied.
1671
  - Fix the RAPI client under Python 2.6 (but more work is needed to
1672
    make it work completely well with OpenSSL)
1673
  - Fix handling of errors when resolving names from DNS
1674

    
1675

    
1676
Version 2.1.3
1677
-------------
1678

    
1679
*(Released Thu, 3 Jun 2010)*
1680

    
1681
A medium sized development cycle. Some new features, and some
1682
fixes/small improvements/cleanups.
1683

    
1684
Significant features
1685
~~~~~~~~~~~~~~~~~~~~
1686

    
1687
The node deamon now tries to mlock itself into memory, unless the
1688
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1689
its logs, and falls back to console logging. This allows emergency
1690
features such as ``gnt-node powercycle`` to work even in the event of a
1691
broken node disk (tested offlining the disk hosting the node's
1692
filesystem and dropping its memory caches; don't try this at home)
1693

    
1694
KVM: add vhost-net acceleration support. It can be tested with a new
1695
enough version of the kernel and of qemu-kvm.
1696

    
1697
KVM: Add instance chrooting feature. If you use privilege dropping for
1698
your VMs you can also now force them to chroot to an empty directory,
1699
before starting the emulated guest.
1700

    
1701
KVM: Add maximum migration bandwith and maximum downtime tweaking
1702
support (requires a new-enough version of qemu-kvm).
1703

    
1704
Cluster verify will now warn if the master node doesn't have the master
1705
ip configured on it.
1706

    
1707
Add a new (incompatible) instance creation request format to RAPI which
1708
supports all parameters (previously only a subset was supported, and it
1709
wasn't possible to extend the old format to accomodate all the new
1710
features. The old format is still supported, and a client can check for
1711
this feature, before using it, by checking for its presence in the
1712
``features`` RAPI resource.
1713

    
1714
Now with ancient latin support. Try it passing the ``--roman`` option to
1715
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1716
(requires the python-roman module to be installed, in order to work).
1717

    
1718
Other changes
1719
~~~~~~~~~~~~~
1720

    
1721
As usual many internal code refactorings, documentation updates, and
1722
such. Among others:
1723

    
1724
  - Lots of improvements and cleanups to the experimental Remote API
1725
    (RAPI) client library.
1726
  - A new unit test suite for the core daemon libraries.
1727
  - A fix to creating missing directories makes sure the umask is not
1728
    applied anymore. This enforces the same directory permissions
1729
    everywhere.
1730
  - Better handling terminating daemons with ctrl+c (used when running
1731
    them in debugging mode).
1732
  - Fix a race condition in live migrating a KVM instance, when stat()
1733
    on the old proc status file returned EINVAL, which is an unexpected
1734
    value.
1735
  - Fixed manpage checking with newer man and utf-8 charachters. But now
1736
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1737

    
1738

    
1739
Version 2.1.2.1
1740
---------------
1741

    
1742
*(Released Fri, 7 May 2010)*
1743

    
1744
Fix a bug which prevented untagged KVM instances from starting.
1745

    
1746

    
1747
Version 2.1.2
1748
-------------
1749

    
1750
*(Released Fri, 7 May 2010)*
1751

    
1752
Another release with a long development cycle, during which many
1753
different features were added.
1754

    
1755
Significant features
1756
~~~~~~~~~~~~~~~~~~~~
1757

    
1758
The KVM hypervisor now can run the individual instances as non-root, to
1759
reduce the impact of a VM being hijacked due to bugs in the
1760
hypervisor. It is possible to run all instances as a single (non-root)
1761
user, to manually specify a user for each instance, or to dynamically
1762
allocate a user out of a cluster-wide pool to each instance, with the
1763
guarantee that no two instances will run under the same user ID on any
1764
given node.
1765

    
1766
An experimental RAPI client library, that can be used standalone
1767
(without the other Ganeti libraries), is provided in the source tree as
1768
``lib/rapi/client.py``. Note this client might change its interface in
1769
the future, as we iterate on its capabilities.
1770

    
1771
A new command, ``gnt-cluster renew-crypto`` has been added to easily
1772
replace the cluster's certificates and crypto keys. This might help in
1773
case they have been compromised, or have simply expired.
1774

    
1775
A new disk option for instance creation has been added that allows one
1776
to "adopt" currently existing logical volumes, with data
1777
preservation. This should allow easier migration to Ganeti from
1778
unmanaged (or managed via other software) instances.
1779

    
1780
Another disk improvement is the possibility to convert between redundant
1781
(DRBD) and plain (LVM) disk configuration for an instance. This should
1782
allow better scalability (starting with one node and growing the
1783
cluster, or shrinking a two-node cluster to one node).
1784

    
1785
A new feature that could help with automated node failovers has been
1786
implemented: if a node sees itself as offline (by querying the master
1787
candidates), it will try to shutdown (hard) all instances and any active
1788
DRBD devices. This reduces the risk of duplicate instances if an
1789
external script automatically failovers the instances on such nodes. To
1790
enable this, the cluster parameter ``maintain_node_health`` should be
1791
enabled; in the future this option (per the name) will enable other
1792
automatic maintenance features.
1793

    
1794
Instance export/import now will reuse the original instance
1795
specifications for all parameters; that means exporting an instance,
1796
deleting it and the importing it back should give an almost identical
1797
instance. Note that the default import behaviour has changed from
1798
before, where it created only one NIC; now it recreates the original
1799
number of NICs.
1800

    
1801
Cluster verify has added a few new checks: SSL certificates validity,
1802
/etc/hosts consistency across the cluster, etc.
1803

    
1804
Other changes
1805
~~~~~~~~~~~~~
1806

    
1807
As usual, many internal changes were done, documentation fixes,
1808
etc. Among others:
1809

    
1810
- Fixed cluster initialization with disabled cluster storage (regression
1811
  introduced in 2.1.1)
1812
- File-based storage supports growing the disks
1813
- Fixed behaviour of node role changes
1814
- Fixed cluster verify for some corner cases, plus a general rewrite of
1815
  cluster verify to allow future extension with more checks
1816
- Fixed log spamming by watcher and node daemon (regression introduced
1817
  in 2.1.1)
1818
- Fixed possible validation issues when changing the list of enabled
1819
  hypervisors
1820
- Fixed cleanup of /etc/hosts during node removal
1821
- Fixed RAPI response for invalid methods
1822
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1823
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
1824
  from ide disks, etc.)
1825
- Allow OS changes without re-installation (to record a changed OS
1826
  outside of Ganeti, or to allow OS renames)
1827
- Allow instance creation without OS installation (useful for example if
1828
  the OS will be installed manually, or restored from a backup not in
1829
  Ganeti format)
1830
- Implemented option to make cluster ``copyfile`` use the replication
1831
  network
1832
- Added list of enabled hypervisors to ssconf (possibly useful for
1833
  external scripts)
1834
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1835
  1.2 clusters
1836
- A partial form of node re-IP is possible via node readd, which now
1837
  allows changed node primary IP
1838
- Command line utilities now show an informational message if the job is
1839
  waiting for a lock
1840
- The logs of the master daemon now show the PID/UID/GID of the
1841
  connected client
1842

    
1843

    
1844
Version 2.1.1
1845
-------------
1846

    
1847
*(Released Fri, 12 Mar 2010)*
1848

    
1849
During the 2.1.0 long release candidate cycle, a lot of improvements and
1850
changes have accumulated with were released later as 2.1.1.
1851

    
1852
Major changes
1853
~~~~~~~~~~~~~
1854

    
1855
The node evacuate command (``gnt-node evacuate``) was significantly
1856
rewritten, and as such the IAllocator protocol was changed - a new
1857
request type has been added. This unfortunate change during a stable
1858
series is designed to improve performance of node evacuations; on
1859
clusters with more than about five nodes and which are well-balanced,
1860
evacuation should proceed in parallel for all instances of the node
1861
being evacuated. As such, any existing IAllocator scripts need to be
1862
updated, otherwise the above command will fail due to the unknown
1863
request. The provided "dumb" allocator has not been updated; but the
1864
ganeti-htools package supports the new protocol since version 0.2.4.
1865

    
1866
Another important change is increased validation of node and instance
1867
names. This might create problems in special cases, if invalid host
1868
names are being used.
1869

    
1870
Also, a new layer of hypervisor parameters has been added, that sits at
1871
OS level between the cluster defaults and the instance ones. This allows
1872
customisation of virtualization parameters depending on the installed
1873
OS. For example instances with OS 'X' may have a different KVM kernel
1874
(or any other parameter) than the cluster defaults. This is intended to
1875
help managing a multiple OSes on the same cluster, without manual
1876
modification of each instance's parameters.
1877

    
1878
A tool for merging clusters, ``cluster-merge``, has been added in the
1879
tools sub-directory.
1880

    
1881
Bug fixes
1882
~~~~~~~~~
1883

    
1884
- Improved the int/float conversions that should make the code more
1885
  robust in face of errors from the node daemons
1886
- Fixed the remove node code in case of internal configuration errors
1887
- Fixed the node daemon behaviour in face of inconsistent queue
1888
  directory (e.g. read-only file-system where we can't open the files
1889
  read-write, etc.)
1890
- Fixed the behaviour of gnt-node modify for master candidate demotion;
1891
  now it either aborts cleanly or, if given the new "auto_promote"
1892
  parameter, will automatically promote other nodes as needed
1893
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1894
  completely prevent Ganeti from working
1895
- Fixed bug for instance export when not all disks were successfully
1896
  exported
1897
- Fixed behaviour of node add when the new node is slow in starting up
1898
  the node daemon
1899
- Fixed handling of signals in the LUXI client, which should improve
1900
  behaviour of command-line scripts
1901
- Added checks for invalid node/instance names in the configuration (now
1902
  flagged during cluster verify)
1903
- Fixed watcher behaviour for disk activation errors
1904
- Fixed two potentially endless loops in http library, which led to the
1905
  RAPI daemon hanging and consuming 100% CPU in some cases
1906
- Fixed bug in RAPI daemon related to hashed passwords
1907
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1908
  instances
1909
- Enhanced compatibility with non-Debian OSes, but not using absolute
1910
  path in some commands and allowing customisation of the ssh
1911
  configuration directory
1912
- Fixed possible future issue with new Python versions by abiding to the
1913
  proper use of ``__slots__`` attribute on classes
1914
- Added checks that should prevent directory traversal attacks
1915
- Many documentation fixes based on feedback from users
1916

    
1917
New features
1918
~~~~~~~~~~~~
1919

    
1920
- Added an "early_release" more for instance replace disks and node
1921
  evacuate, where we release locks earlier and thus allow higher
1922
  parallelism within the cluster
1923
- Added watcher hooks, intended to allow the watcher to restart other
1924
  daemons (e.g. from the ganeti-nbma project), but they can be used of
1925
  course for any other purpose
1926
- Added a compile-time disable for DRBD barriers, to increase
1927
  performance if the administrator trusts the power supply or the
1928
  storage system to not lose writes
1929
- Added the option of using syslog for logging instead of, or in
1930
  addition to, Ganeti's own log files
1931
- Removed boot restriction for paravirtual NICs for KVM, recent versions
1932
  can indeed boot from a paravirtual NIC
1933
- Added a generic debug level for many operations; while this is not
1934
  used widely yet, it allows one to pass the debug value all the way to
1935
  the OS scripts
1936
- Enhanced the hooks environment for instance moves (failovers,
1937
  migrations) where the primary/secondary nodes changed during the
1938
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1939
- Enhanced data validations for many user-supplied values; one important
1940
  item is the restrictions imposed on instance and node names, which
1941
  might reject some (invalid) host names
1942
- Add a configure-time option to disable file-based storage, if it's not
1943
  needed; this allows greater security separation between the master
1944
  node and the other nodes from the point of view of the inter-node RPC
1945
  protocol
1946
- Added user notification in interactive tools if job is waiting in the
1947
  job queue or trying to acquire locks
1948
- Added log messages when a job is waiting for locks
1949
- Added filtering by node tags in instance operations which admit
1950
  multiple instances (start, stop, reboot, reinstall)
1951
- Added a new tool for cluster mergers, ``cluster-merge``
1952
- Parameters from command line which are of the form ``a=b,c=d`` can now
1953
  use backslash escapes to pass in values which contain commas,
1954
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1955
  ``b,c``
1956
- For KVM, the instance name is the first parameter passed to KVM, so
1957
  that it's more visible in the process list
1958

    
1959

    
1960
Version 2.1.0
1961
-------------
1962

    
1963
*(Released Tue, 2 Mar 2010)*
1964

    
1965
Ganeti 2.1 brings many improvements with it. Major changes:
1966

    
1967
- Added infrastructure to ease automated disk repairs
1968
- Added new daemon to export configuration data in a cheaper way than
1969
  using the remote API
1970
- Instance NICs can now be routed instead of being associated with a
1971
  networking bridge
1972
- Improved job locking logic to reduce impact of jobs acquiring multiple
1973
  locks waiting for other long-running jobs
1974

    
1975
In-depth implementation details can be found in the Ganeti 2.1 design
1976
document.
1977

    
1978
Details
1979
~~~~~~~
1980

    
1981
- Added chroot hypervisor
1982
- Added more options to xen-hvm hypervisor (``kernel_path`` and
1983
  ``device_model``)
1984
- Added more options to xen-pvm hypervisor (``use_bootloader``,
1985
  ``bootloader_path`` and ``bootloader_args``)
1986
- Added the ``use_localtime`` option for the xen-hvm and kvm
1987
  hypervisors, and the default value for this has changed to false (in
1988
  2.0 xen-hvm always enabled it)
1989
- Added luxi call to submit multiple jobs in one go
1990
- Added cluster initialization option to not modify ``/etc/hosts``
1991
  file on nodes
1992
- Added network interface parameters
1993
- Added dry run mode to some LUs
1994
- Added RAPI resources:
1995

    
1996
  - ``/2/instances/[instance_name]/info``
1997
  - ``/2/instances/[instance_name]/replace-disks``
1998
  - ``/2/nodes/[node_name]/evacuate``
1999
  - ``/2/nodes/[node_name]/migrate``
2000
  - ``/2/nodes/[node_name]/role``
2001
  - ``/2/nodes/[node_name]/storage``
2002
  - ``/2/nodes/[node_name]/storage/modify``
2003
  - ``/2/nodes/[node_name]/storage/repair``
2004

    
2005
- Added OpCodes to evacuate or migrate all instances on a node
2006
- Added new command to list storage elements on nodes (``gnt-node
2007
  list-storage``) and modify them (``gnt-node modify-storage``)
2008
- Added new ssconf files with master candidate IP address
2009
  (``ssconf_master_candidates_ips``), node primary IP address
2010
  (``ssconf_node_primary_ips``) and node secondary IP address
2011
  (``ssconf_node_secondary_ips``)
2012
- Added ``ganeti-confd`` and a client library to query the Ganeti
2013
  configuration via UDP
2014
- Added ability to run hooks after cluster initialization and before
2015
  cluster destruction
2016
- Added automatic mode for disk replace (``gnt-instance replace-disks
2017
  --auto``)
2018
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2019
  after catastrophic data-loss
2020
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2021
  groups
2022
- Added ``gnt-instance move`` command to move instances
2023
- Added ``gnt-cluster watcher`` command to control watcher
2024
- Added ``gnt-node powercycle`` command to powercycle nodes
2025
- Added new job status field ``lock_status``
2026
- Added parseable error codes to cluster verification (``gnt-cluster
2027
  verify --error-codes``) and made output less verbose (use
2028
  ``--verbose`` to restore previous behaviour)
2029
- Added UUIDs to the main config entities (cluster, nodes, instances)
2030
- Added support for OS variants
2031
- Added support for hashed passwords in the Ganeti remote API users file
2032
  (``rapi_users``)
2033
- Added option to specify maximum timeout on instance shutdown
2034
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2035
- Added new helper script to start and stop Ganeti daemons
2036
  (``daemon-util``), with the intent to reduce the work necessary to
2037
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2038
  from one place
2039
- Added more unittests
2040
- Fixed critical bug in ganeti-masterd startup
2041
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2042
  now customisable at the cluster level for both the KVM and Xen
2043
  hypervisors using the new ``migration_port`` parameter
2044
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2045
  reinstalling an instance
2046
- Allowed ``@`` in tag names
2047
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2048
- Many documentation updates
2049
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2050
- ``gnt-instance reinstall`` can now reinstall multiple instances
2051
- Updated many command line parameters
2052
- Introduced new OS API version 15
2053
- No longer support a default hypervisor
2054
- Treat virtual LVs as inexistent
2055
- Improved job locking logic to reduce lock contention
2056
- Match instance and node names case insensitively
2057
- Reimplemented bash completion script to be more complete
2058
- Improved burnin
2059

    
2060

    
2061
Version 2.0.6
2062
-------------
2063

    
2064
*(Released Thu, 4 Feb 2010)*
2065

    
2066
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2067
- Fix a string formatting bug
2068
- Improve safety of the code in some error paths
2069
- Improve data validation in the master of values returned from nodes
2070

    
2071

    
2072
Version 2.0.5
2073
-------------
2074

    
2075
*(Released Thu, 17 Dec 2009)*
2076

    
2077
- Fix security issue due to missing validation of iallocator names; this
2078
  allows local and remote execution of arbitrary executables
2079
- Fix failure of gnt-node list during instance removal
2080
- Ship the RAPI documentation in the archive
2081

    
2082

    
2083
Version 2.0.4
2084
-------------
2085

    
2086
*(Released Wed, 30 Sep 2009)*
2087

    
2088
- Fixed many wrong messages
2089
- Fixed a few bugs related to the locking library
2090
- Fixed MAC checking at instance creation time
2091
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2092
- Fixed a few issues related to signal handling in both daemons and
2093
  scripts
2094
- Fixed the example startup script provided
2095
- Fixed insserv dependencies in the example startup script (patch from
2096
  Debian)
2097
- Fixed handling of drained nodes in the iallocator framework
2098
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2099
  #528618)
2100
- Fixed error related to invalid job IDs in job polling
2101
- Fixed job/opcode persistence on unclean master shutdown
2102
- Fixed handling of partial job processing after unclean master
2103
  shutdown
2104
- Fixed error reporting from LUs, previously all errors were converted
2105
  into execution errors
2106
- Fixed error reporting from burnin
2107
- Decreased significantly the memory usage of the job queue
2108
- Optimised slightly multi-job submission
2109
- Optimised slightly opcode loading
2110
- Backported the multi-job submit framework from the development
2111
  branch; multi-instance start and stop should be faster
2112
- Added script to clean archived jobs after 21 days; this will reduce
2113
  the size of the queue directory
2114
- Added some extra checks in disk size tracking
2115
- Added an example ethers hook script
2116
- Added a cluster parameter that prevents Ganeti from modifying of
2117
  /etc/hosts
2118
- Added more node information to RAPI responses
2119
- Added a ``gnt-job watch`` command that allows following the ouput of a
2120
  job
2121
- Added a bind-address option to ganeti-rapi
2122
- Added more checks to the configuration verify
2123
- Enhanced the burnin script such that some operations can be retried
2124
  automatically
2125
- Converted instance reinstall to multi-instance model
2126

    
2127

    
2128
Version 2.0.3
2129
-------------
2130

    
2131
*(Released Fri, 7 Aug 2009)*
2132

    
2133
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2134
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2135
  instances have mismatched disk sizes in the configuration
2136
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2137
  any configuration mismatches for disk sizes
2138
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2139
  failover to work on two-node clusters
2140
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2141
  unusable
2142
- Fixed detection of OS script errors in ``gnt-backup export``
2143
- Fixed exit code of ``gnt-backup export``
2144

    
2145

    
2146
Version 2.0.2
2147
-------------
2148

    
2149
*(Released Fri, 17 Jul 2009)*
2150

    
2151
- Added experimental support for stripped logical volumes; this should
2152
  enhance performance but comes with a higher complexity in the block
2153
  device handling; stripping is only enabled when passing
2154
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2155
  affected even in the non-stripped mode
2156
- Improved resiliency against transient failures at the end of DRBD
2157
  resyncs, and in general of DRBD resync checks
2158
- Fixed a couple of issues with exports and snapshot errors
2159
- Fixed a couple of issues in instance listing
2160
- Added display of the disk size in ``gnt-instance info``
2161
- Fixed checking for valid OSes in instance creation
2162
- Fixed handling of the "vcpus" parameter in instance listing and in
2163
  general of invalid parameters
2164
- Fixed http server library, and thus RAPI, to handle invalid
2165
  username/password combinations correctly; this means that now they
2166
  report unauthorized for queries too, not only for modifications,
2167
  allowing earlier detect of configuration problems
2168
- Added a new "role" node list field, equivalent to the master/master
2169
  candidate/drained/offline flags combinations
2170
- Fixed cluster modify and changes of candidate pool size
2171
- Fixed cluster verify error messages for wrong files on regular nodes
2172
- Fixed a couple of issues with node demotion from master candidate role
2173
- Fixed node readd issues
2174
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2175
- Added a new ``--no-voting`` option for masterfailover to fix failover
2176
  on two-nodes clusters when the former master node is unreachable
2177
- Added instance reinstall over RAPI
2178

    
2179

    
2180
Version 2.0.1
2181
-------------
2182

    
2183
*(Released Tue, 16 Jun 2009)*
2184

    
2185
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2186
  allow re-adding the start in single-user option (regression from 1.2)
2187
- the watcher writes the instance status to a file, to allow monitoring
2188
  to report the instance status (from the master) based on cached
2189
  results of the watcher's queries; while this can get stale if the
2190
  watcher is being locked due to other work on the cluster, this is
2191
  still an improvement
2192
- the watcher now also restarts the node daemon and the rapi daemon if
2193
  they died
2194
- fixed the watcher to handle full and drained queue cases
2195
- hooks export more instance data in the environment, which helps if
2196
  hook scripts need to take action based on the instance's properties
2197
  (no longer need to query back into ganeti)
2198
- instance failovers when the instance is stopped do not check for free
2199
  RAM, so that failing over a stopped instance is possible in low memory
2200
  situations
2201
- rapi uses queries for tags instead of jobs (for less job traffic), and
2202
  for cluster tags it won't talk to masterd at all but read them from
2203
  ssconf
2204
- a couple of error handling fixes in RAPI
2205
- drbd handling: improved the error handling of inconsistent disks after
2206
  resync to reduce the frequency of "there are some degraded disks for
2207
  this instance" messages
2208
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2209
  error handling path called a wrong function name)
2210

    
2211

    
2212
Version 2.0.0
2213
-------------
2214

    
2215
*(Released Wed, 27 May 2009)*
2216

    
2217
- no changes from rc5
2218

    
2219

    
2220
Version 2.0 rc5
2221
---------------
2222

    
2223
*(Released Wed, 20 May 2009)*
2224

    
2225
- fix a couple of bugs (validation, argument checks)
2226
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2227
- some small improvements to RAPI and IAllocator
2228
- make watcher automatically start the master daemon if down
2229

    
2230

    
2231
Version 2.0 rc4
2232
---------------
2233

    
2234
*(Released Mon, 27 Apr 2009)*
2235

    
2236
- change the OS list to not require locks; this helps with big clusters
2237
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2238
  volume group is broken
2239
- ``gnt-instance info``, without any arguments, doesn't run for all
2240
  instances anymore; either pass ``--all`` or pass the desired
2241
  instances; this helps against mistakes on big clusters where listing
2242
  the information for all instances takes a long time
2243
- miscellaneous doc and man pages fixes
2244

    
2245

    
2246
Version 2.0 rc3
2247
---------------
2248

    
2249
*(Released Wed, 8 Apr 2009)*
2250

    
2251
- Change the internal locking model of some ``gnt-node`` commands, in
2252
  order to reduce contention (and blocking of master daemon) when
2253
  batching many creation/reinstall jobs
2254
- Fixes to Xen soft reboot
2255
- No longer build documentation at build time, instead distribute it in
2256
  the archive, in order to reduce the need for the whole docbook/rst
2257
  toolchains
2258

    
2259

    
2260
Version 2.0 rc2
2261
---------------
2262

    
2263
*(Released Fri, 27 Mar 2009)*
2264

    
2265
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2266
- Fix watcher startup sequence, improves the behaviour of busy clusters
2267
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2268
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2269
  bind address and other places
2270
- Some documentation fixes and updates
2271

    
2272

    
2273
Version 2.0 rc1
2274
---------------
2275

    
2276
*(Released Mon, 2 Mar 2009)*
2277

    
2278
- More documentation updates, now all docs should be more-or-less
2279
  up-to-date
2280
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2281
  etc.)
2282
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2283
  and KVM)
2284
- Fix an issue related to $libdir/run/ganeti and cluster creation
2285

    
2286

    
2287
Version 2.0 beta2
2288
-----------------
2289

    
2290
*(Released Thu, 19 Feb 2009)*
2291

    
2292
- Xen PVM and KVM have switched the default value for the instance root
2293
  disk to the first partition on the first drive, instead of the whole
2294
  drive; this means that the OS installation scripts must be changed
2295
  accordingly
2296
- Man pages have been updated
2297
- RAPI has been switched by default to HTTPS, and the exported functions
2298
  should all work correctly
2299
- RAPI v1 has been removed
2300
- Many improvements to the KVM hypervisor
2301
- Block device errors are now better reported
2302
- Many other bugfixes and small improvements
2303

    
2304

    
2305
Version 2.0 beta1
2306
-----------------
2307

    
2308
*(Released Mon, 26 Jan 2009)*
2309

    
2310
- Version 2 is a general rewrite of the code and therefore the
2311
  differences are too many to list, see the design document for 2.0 in
2312
  the ``doc/`` subdirectory for more details
2313
- In this beta version there is not yet a migration path from 1.2 (there
2314
  will be one in the final 2.0 release)
2315
- A few significant changes are:
2316

    
2317
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2318
    various ``gnt-*`` commands are just front-ends to it
2319
  - all the commands are entered into, and executed from a job queue,
2320
    see the ``gnt-job(8)`` manpage
2321
  - the RAPI daemon supports read-write operations, secured by basic
2322
    HTTP authentication on top of HTTPS
2323
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2324
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2325
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2326
  - DRBD devices are using statically allocated minor numbers, which
2327
    will be assigned to existing instances during the migration process
2328
  - there is support for both Xen PVM and Xen HVM instances running on
2329
    the same cluster
2330
  - KVM virtualization is supported too
2331
  - file-based storage has been implemented, which means that it is
2332
    possible to run the cluster without LVM and DRBD storage, for
2333
    example using a shared filesystem exported from shared storage (and
2334
    still have live migration)
2335

    
2336

    
2337
Version 1.2.7
2338
-------------
2339

    
2340
*(Released Tue, 13 Jan 2009)*
2341

    
2342
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2343
- Reuse the old instance mac address by default on instance import, if
2344
  the instance name is the same.
2345
- Handle situations in which the node info rpc returns incomplete
2346
  results (issue 46)
2347
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2348
- Improved version of batcher:
2349

    
2350
  - state file support
2351
  - instance mac address support
2352
  - support for HVM clusters/instances
2353

    
2354
- Add an option to show the number of cpu sockets and nodes in
2355
  ``gnt-node list``
2356
- Support OSes that handle more than one version of the OS api (but do
2357
  not change the current API in any other way)
2358
- Fix ``gnt-node migrate``
2359
- ``gnt-debug`` man page
2360
- Fixes various more typos and small issues
2361
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2362

    
2363

    
2364
Version 1.2.6
2365
-------------
2366

    
2367
*(Released Wed, 24 Sep 2008)*
2368

    
2369
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2370
  type of disk exported to fully virtualized instances.
2371
- provide access to the serial console of HVM instances
2372
- instance auto_balance flag, set by default. If turned off it will
2373
  avoid warnings on cluster verify if there is not enough memory to fail
2374
  over an instance. in the future it will prevent automatically failing
2375
  it over when we will support that.
2376
- batcher tool for instance creation, see ``tools/README.batcher``
2377
- ``gnt-instance reinstall --select-os`` to interactively select a new
2378
  operating system when reinstalling an instance.
2379
- when changing the memory amount on instance modify a check has been
2380
  added that the instance will be able to start. also warnings are
2381
  emitted if the instance will not be able to fail over, if auto_balance
2382
  is true.
2383
- documentation fixes
2384
- sync fields between ``gnt-instance list/modify/add/import``
2385
- fix a race condition in drbd when the sync speed was set after giving
2386
  the device a remote peer.
2387

    
2388

    
2389
Version 1.2.5
2390
-------------
2391

    
2392
*(Released Tue, 22 Jul 2008)*
2393

    
2394
- note: the allowed size and number of tags per object were reduced
2395
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2396
- fixed twisted 8.x compatibility
2397
- fixed ``gnt-instance replace-disks`` with iallocator
2398
- add TCP keepalives on twisted connections to detect restarted nodes
2399
- disk increase support, see ``gnt-instance grow-disk``
2400
- implement bulk node/instance query for RAPI
2401
- add tags in node/instance listing (optional)
2402
- experimental migration (and live migration) support, read the man page
2403
  for ``gnt-instance migrate``
2404
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2405
  has some small improvements in handling its state file
2406

    
2407

    
2408
Version 1.2.4
2409
-------------
2410

    
2411
*(Released Fri, 13 Jun 2008)*
2412

    
2413
- Experimental readonly, REST-based remote API implementation;
2414
  automatically started on master node, TCP port 5080, if enabled by
2415
  ``--enable-rapi`` parameter to configure script.
2416
- Instance allocator support. Add and import instance accept a
2417
  ``--iallocator`` parameter, and call that instance allocator to decide
2418
  which node to use for the instance. The iallocator document describes
2419
  what's expected from an allocator script.
2420
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2421
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2422
  node is lost there is still enough memory to fail over the instances
2423
  that reside on it.
2424
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2425
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2426
  hooks will run, and their output, if any, will be displayed. Any
2427
  failing hook will make the verification return an error value.
2428
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2429
  primary and secondary interfaces
2430
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2431
  that is still declared as part of the cluster and has failed.
2432
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2433
  specifying output fields, that just adds the chosen fields to the
2434
  default ones.
2435
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2436
  export from the filesystem.
2437
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2438
  have been added. Using them you can enable/disable acpi and pae
2439
  support, and specify a path for a cd image to be exported to the
2440
  instance. These parameters as the name suggest only work on HVM
2441
  clusters.
2442
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2443
  PAE support will be set to the previously hardcoded values, but the
2444
  (previously hardcoded) path to the CDROM ISO image will be unset and
2445
  if required, needs to be set manually with ``gnt-instance modify``
2446
  after the upgrade.
2447
- The address to which an instance's VNC console is bound is now
2448
  selectable per-instance, rather than being cluster wide. Of course
2449
  this only applies to instances controlled via VNC, so currently just
2450
  applies to HVM clusters.
2451

    
2452

    
2453
Version 1.2.3
2454
-------------
2455

    
2456
*(Released Mon, 18 Feb 2008)*
2457

    
2458
- more tweaks to the disk activation code (especially helpful for DRBD)
2459
- change the default ``gnt-instance list`` output format, now there is
2460
  one combined status field (see the manpage for the exact values this
2461
  field will have)
2462
- some more fixes for the mac export to hooks change
2463
- make Ganeti not break with DRBD 8.2.x (which changed the version
2464
  format in ``/proc/drbd``) (issue 24)
2465
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2466
  template, allowing migration from DRBD0.7+MD to DRBD8
2467

    
2468

    
2469
Version 1.2.2
2470
-------------
2471

    
2472
*(Released Wed, 30 Jan 2008)*
2473

    
2474
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2475
  support (issue 23)
2476
- add command aliases infrastructure and a few aliases
2477
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2478
  man pages and the ``--help`` option of ``gnt-node
2479
  list``/``gnt-instance list``
2480
- fix ``gnt-backup list`` with down nodes (issue 21)
2481
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2482
- fix the dist archive and add a check for including svn/git files in
2483
  the future
2484
- some developer-related changes: improve the burnin and the QA suite,
2485
  add an upload script for testing during development
2486

    
2487

    
2488
Version 1.2.1
2489
-------------
2490

    
2491
*(Released Wed, 16 Jan 2008)*
2492

    
2493
- experimental HVM support, read the install document, section
2494
  "Initializing the cluster"
2495
- allow for the PVM hypervisor per-instance kernel and initrd paths
2496
- add a new command ``gnt-cluster verify-disks`` which uses a new
2497
  algorithm to improve the reconnection of the DRBD pairs if the device
2498
  on the secondary node has gone away
2499
- make logical volume code auto-activate LVs at disk activation time
2500
- slightly improve the speed of activating disks
2501
- allow specification of the MAC address at instance creation time, and
2502
  changing it later via ``gnt-instance modify``
2503
- fix handling of external commands that generate lots of output on
2504
  stderr
2505
- update documentation with regard to minimum version of DRBD8 supported
2506

    
2507

    
2508
Version 1.2.0
2509
-------------
2510

    
2511
*(Released Tue, 4 Dec 2007)*
2512

    
2513
- Log the ``xm create`` output to the node daemon log on failure (to
2514
  help diagnosing the error)
2515
- In debug mode, log all external commands output if failed to the logs
2516
- Change parsing of lvm commands to ignore stderr
2517

    
2518

    
2519
Version 1.2 beta3
2520
-----------------
2521

    
2522
*(Released Wed, 28 Nov 2007)*
2523

    
2524
- Another round of updates to the DRBD 8 code to deal with more failures
2525
  in the replace secondary node operation
2526
- Some more logging of failures in disk operations (lvm, drbd)
2527
- A few documentation updates
2528
- QA updates
2529

    
2530

    
2531
Version 1.2 beta2
2532
-----------------
2533

    
2534
*(Released Tue, 13 Nov 2007)*
2535

    
2536
- Change configuration file format from Python's Pickle to JSON.
2537
  Upgrading is possible using the cfgupgrade utility.
2538
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2539
  faster replace disks and is more stable (DRBD 8 has many improvements
2540
  compared to DRBD 0.7)
2541
- Added command line tags support (see man pages for ``gnt-instance``,
2542
  ``gnt-node``, ``gnt-cluster``)
2543
- Added instance rename support
2544
- Added multi-instance startup/shutdown
2545
- Added cluster rename support
2546
- Added ``gnt-node evacuate`` to simplify some node operations
2547
- Added instance reboot operation that can speedup reboot as compared to
2548
  stop and start
2549
- Soften the requirement that hostnames are in FQDN format
2550
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
2551
  reboots
2552
- Removed dependency on debian's patched fping that uses the
2553
  non-standard ``-S`` option
2554
- Now the OS definitions are searched for in multiple, configurable
2555
  paths (easier for distros to package)
2556
- Some changes to the hooks infrastructure (especially the new
2557
  post-configuration update hook)
2558
- Other small bugfixes
2559

    
2560
.. vim: set textwidth=72 syntax=rst :
2561
.. Local Variables:
2562
.. mode: rst
2563
.. fill-column: 72
2564
.. End: