Statistics
| Branch: | Tag: | Revision:

root / NEWS @ b3e1f9dd

History | View | Annotate | Download (99.1 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_disk_templates' is introduced. It will
15
  be used to manage the disk templates to be used by instances in the cluster.
16
  Initially, it will be set to a list that includes plain, drbd, if they were
17
  enabled by specifying a volume group name, and file and sharedfile, if those
18
  were enabled at configure time. Additionally, it will include all disk
19
  templates that are currently used by instances. The order of disk templates
20
  will be based on Ganeti's history of supporting them. In the future, the
21
  first entry of the list will be used as a default disk template on instance
22
  creation.
23
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
24
  configuration back to the previous stable version.
25
- The cluster option '--no-lvm-storage' was removed in favor of the new option
26
  '--enabled-disk-templates'.
27

    
28

    
29
Version 2.7.0 rc1
30
-----------------
31

    
32
*(unreleased)*
33

    
34
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
35

    
36

    
37
Version 2.7.0 beta2
38
-------------------
39

    
40
*(Released Tue, 2 Apr 2013)*
41

    
42
- Networks no longer have a "type" slot, since this information was
43
  unused in Ganeti: instead of it tags should be used.
44
- Diskless instances are now externally mirrored (Issue 237). This for
45
  now has only been tested in conjunction with explicit target nodes for
46
  migration/failover.
47
- The rapi client now has a ``target_node`` option to MigrateInstance.
48
- Fix early exit return code for hbal (Issue 386).
49
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
50
- Fix ``rbd showmapped`` output parsing (Issue 312).
51
- Networks are now referenced indexed by UUID, rather than name. This
52
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
53
  use.
54
- The OS environment now includes network information.
55
- Deleting of a network is now disallowed if any instance nic is using
56
  it, to prevent dangling references.
57
- External storage is now documented in man pages.
58
- The exclusive_storage flag can now only be set at nodegroup level.
59
- Hbal can now submit an explicit priority with its jobs.
60
- Many network related locking fixes.
61
- Bump up the required pylint version to 0.25.1.
62
- Fix the ``no_remember`` option in RAPI client.
63
- Many ipolicy related tests, qa, and fixes.
64
- Many documentation improvements and fixes.
65
- Fix building with ``--disable-file-storage``.
66
- Fix ``-q`` option in htools, which was broken if passed more than
67
  once.
68
- Some haskell/python interaction improvements and fixes.
69
- Fix iallocator in case of missing LVM storage.
70
- Fix confd config load in case of ``--no-lvm-storage``.
71
- The confd/query functionality is now mentioned in the security
72
  documentation.
73

    
74

    
75
Version 2.7.0 beta1
76
-------------------
77

    
78
*(Released Wed, 6 Feb 2013)*
79

    
80
- ``gnt-instance batch-create`` has been changed to use the bulk create
81
  opcode from Ganeti. This lead to incompatible changes in the format of
82
  the JSON file. It's now not a custom dict anymore but a dict
83
  compatible with the ``OpInstanceCreate`` opcode.
84
- Parent directories for file storage need to be listed in
85
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
86
  write the file automatically based on old configuration values, but it
87
  can not distribute it across all nodes and the file contents should be
88
  verified. Use ``gnt-cluster copyfile
89
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
90
  upgraded. The reason for requiring this list of paths now is that
91
  before it would have been possible to inject new paths via RPC,
92
  allowing files to be created in arbitrary locations. The RPC protocol
93
  is protected using SSL/X.509 certificates, but as a design principle
94
  Ganeti does not permit arbitrary paths to be passed.
95
- The parsing of the variants file for OSes (see
96
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
97
  lines and comment lines (starting with ``#``) are ignored for better
98
  readability.
99
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
100
  longer available. ``gnt-node add`` now invokes a new tool on the
101
  destination node, named ``prepare-node-join``, to configure the SSH
102
  daemon. Paramiko is no longer necessary to configure nodes' SSH
103
  daemons via ``gnt-node add``.
104
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
105
  for RAPI users. It allows granting permissions to query for
106
  information to a specific user without giving
107
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
108
- A new tool named ``node-cleanup`` has been added. It cleans remains of
109
  a cluster from a machine by stopping all daemons, removing
110
  certificates and ssconf files. Unless the ``--no-backup`` option is
111
  given, copies of the certificates are made.
112
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
113
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
114
  the flag is not reset after a master failover.
115
- Instance creations now support the use of opportunistic locking,
116
  potentially speeding up the (parallel) creation of multiple instances.
117
  This feature is currently only available via the :doc:`RAPI
118
  <rapi>` interface and when an instance allocator is used. If the
119
  ``opportunistic_locking`` parameter is set the opcode will try to
120
  acquire as many locks as possible, but will not wait for any locks
121
  held by other opcodes. If not enough resources can be found to
122
  allocate the instance, the temporary error code
123
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
124
  retried thereafter, with or without opportunistic locking.
125
- The functionality for allocating multiple instances at once has been
126
  overhauled and is now also available through :doc:`RAPI <rapi>`.
127
- Man pages can now be included when the documentation is built, in
128
  which case the output is in ``doc/man-html``. The configure-time
129
  option is ``--enable-manpages-in-doc``. Sphinx 1.0 or higher is
130
  required.
131
- A new htool called ``hroller`` helps scheduling parallel shutdown of
132
  nodes on clusters where all instances are shut down, but without ever
133
  rebooting both primary and secondary node for an instance at the same
134
  time.
135
- Queries not needing locks or RPC access to the node can now be
136
  performed by the confd daemon, making them independent from jobs, and
137
  thus faster to execute. This is selectable at configure time.
138
- The KVM hypervisor has been updated with many new hypervisor
139
  parameters, including a generic one for passing arbitrary command line
140
  values. See a complete list in :manpage:`gnt-instance(8)`.
141
- A new tool, called ``mon-collector``, is the stand-alone executor of
142
  the data collectors for a monitoring system. As of this version, it
143
  just includes the DRBD data collector, that can be executed by calling
144
  ``mon-collector`` using the ``drbd`` parameter. See
145
  :manpage:`mon-collector(7)`.
146
- New ``exclusive-storage`` node parameter added. When it's set to true,
147
  physical disks are assigned in an exclusive fashion to instances, as
148
  documented in :doc:`Partitioned Ganeti <design-partitioned>`.
149
  Currently, only instances using the ``plain`` disk template are
150
  supported.
151
- New network management functionality to support automatic allocation
152
  of IP addresses and managing of network parameters. See
153
  :manpage:`gnt-network(8)` for more details.
154
- New external storage backend, to allow managing arbitrary storage
155
  systems external to the cluster. See
156
  :manpage:`ganeti-extstorage-interface(7)`.
157
- Instance renames of LVM-based instances will now update the LV tags
158
  (which can be used to recover the instance-to-LV mapping in case of
159
  emergencies)
160
- ``hbal`` will now exit with status 0 if, during job execution over
161
  LUXI, early exit has been requested and all jobs are successful;
162
  before, exit status 1 was used, which cannot be differentiated from
163
  "job error" case
164

    
165

    
166
Version 2.6.2
167
-------------
168

    
169
*(Released Fri, 21 Dec 2012)*
170

    
171
Important behaviour change: hbal won't rebalance anymore instances which
172
have the ``auto_balance`` attribute set to false. This was the intention
173
all along, but until now it only skipped those from the N+1 memory
174
reservation (DRBD-specific).
175

    
176
A significant number of bug fixes in this release:
177

    
178
- Fixed disk adoption interaction with ipolicy checks.
179
- Fixed networking issues when instances are started, stopped or
180
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
181
- Fixed the warning in cluster verify for shared storage instances not
182
  being redundant.
183
- Fixed removal of storage directory on shared file storage (issue 262).
184
- Fixed validation of LVM volume group name in OpClusterSetParams
185
  (``gnt-cluster modify``) (issue 285).
186
- Fixed runtime memory increases (``gnt-instance modify -m``).
187
- Fixed live migration under Xen's ``xl`` mode.
188
- Fixed ``gnt-instance console`` with ``xl``.
189
- Fixed building with newer Haskell compiler/libraries.
190
- Fixed PID file writing in Haskell daemons (confd); this prevents
191
  restart issues if confd was launched manually (outside of
192
  ``daemon-util``) while another copy of it was running
193
- Fixed a type error when doing live migrations with KVM (issue 297) and
194
  the error messages for failing migrations have been improved.
195
- Fixed opcode validation for the out-of-band commands (``gnt-node
196
  power``).
197
- Fixed a type error when unsetting OS hypervisor parameters (issue
198
  311); now it's possible to unset all OS-specific hypervisor
199
  parameters.
200
- Fixed the ``dry-run`` mode for many operations: verification of
201
  results was over-zealous but didn't take into account the ``dry-run``
202
  operation, resulting in "wrong" failures.
203
- Fixed bash completion in ``gnt-job list`` when the job queue has
204
  hundreds of entries; especially with older ``bash`` versions, this
205
  results in significant CPU usage.
206

    
207
And lastly, a few other improvements have been made:
208

    
209
- Added option to force master-failover without voting (issue 282).
210
- Clarified error message on lock conflict (issue 287).
211
- Logging of newly submitted jobs has been improved (issue 290).
212
- Hostname checks have been made uniform between instance rename and
213
  create (issue 291).
214
- The ``--submit`` option is now supported by ``gnt-debug delay``.
215
- Shutting down the master daemon by sending SIGTERM now stops it from
216
  processing jobs waiting for locks; instead, those jobs will be started
217
  once again after the master daemon is started the next time (issue
218
  296).
219
- Support for Xen's ``xl`` program has been improved (besides the fixes
220
  above).
221
- Reduced logging noise in the Haskell confd daemon (only show one log
222
  entry for each config reload, instead of two).
223
- Several man page updates and typo fixes.
224

    
225

    
226
Version 2.6.1
227
-------------
228

    
229
*(Released Fri, 12 Oct 2012)*
230

    
231
A small bugfix release. Among the bugs fixed:
232

    
233
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
234
  made the command unusable.
235
- Commands that issue many jobs don't fail anymore just because some jobs
236
  take so long that other jobs are archived.
237
- Failures during ``gnt-instance reinstall`` are reflected by the exit
238
  status.
239
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
240
  DRBD is enabled.
241
- When ``always_failover`` is set, ``--allow-failover`` is not required
242
  in migrate commands anymore.
243
- ``bash_completion`` works even if extglob is disabled.
244
- Fixed bug with locks that made failover for RDB-based instances fail.
245
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
246
  a random node instead of one based on the allocator metric.
247
- Support for newer versions of pylint and pep8.
248
- Hail doesn't fail anymore when trying to add an instance of type
249
  ``file``, ``sharedfile`` or ``rbd``.
250
- Added new Makefile target to rebuild the whole distribution, so that
251
  all files are included.
252

    
253

    
254
Version 2.6.0
255
-------------
256

    
257
*(Released Fri, 27 Jul 2012)*
258

    
259

    
260
.. attention:: The ``LUXI`` protocol has been made more consistent
261
   regarding its handling of command arguments. This, however, leads to
262
   incompatibility issues with previous versions. Please ensure that you
263
   restart Ganeti daemons soon after the upgrade, otherwise most
264
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
265
   pausing/resuming the watcher, cancelling and archiving jobs, querying
266
   the cluster configuration) will fail.
267

    
268

    
269
New features
270
~~~~~~~~~~~~
271

    
272
Instance run status
273
+++++++++++++++++++
274

    
275
The current ``admin_up`` field, which used to denote whether an instance
276
should be running or not, has been removed. Instead, ``admin_state`` is
277
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
278

    
279
The rational behind this is that an instance being “down” can have
280
different meanings:
281

    
282
- it could be down during a reboot
283
- it could be temporarily be down for a reinstall
284
- or it could be down because it is deprecated and kept just for its
285
  disk
286

    
287
The previous Boolean state was making it difficult to do capacity
288
calculations: should Ganeti reserve memory for a down instance? Now, the
289
tri-state field makes it clear:
290

    
291
- in ``up`` and ``down`` state, all resources are reserved for the
292
  instance, and it can be at any time brought up if it is down
293
- in ``offline`` state, only disk space is reserved for it, but not
294
  memory or CPUs
295

    
296
The field can have an extra use: since the transition between ``up`` and
297
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
298
transition between ``offline`` and ``down`` is done via ``gnt-instance
299
modify``, it is possible to given different rights to users. For
300
example, owners of an instance could be allowed to start/stop it, but
301
not transition it out of the offline state.
302

    
303
Instance policies and specs
304
+++++++++++++++++++++++++++
305

    
306
In previous Ganeti versions, an instance creation request was not
307
limited on the minimum size and on the maximum size just by the cluster
308
resources. As such, any policy could be implemented only in third-party
309
clients (RAPI clients, or shell wrappers over ``gnt-*``
310
tools). Furthermore, calculating cluster capacity via ``hspace`` again
311
required external input with regards to instance sizes.
312

    
313
In order to improve these workflows and to allow for example better
314
per-node group differentiation, we introduced instance specs, which
315
allow declaring:
316

    
317
- minimum instance disk size, disk count, memory size, cpu count
318
- maximum values for the above metrics
319
- and “standard” values (used in ``hspace`` to calculate the standard
320
  sized instances)
321

    
322
The minimum/maximum values can be also customised at node-group level,
323
for example allowing more powerful hardware to support bigger instance
324
memory sizes.
325

    
326
Beside the instance specs, there are a few other settings belonging to
327
the instance policy framework. It is possible now to customise, per
328
cluster and node-group:
329

    
330
- the list of allowed disk templates
331
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
332
- the maximum ratio of instance to spindles (see below for more
333
  information) for local storage
334

    
335
All these together should allow all tools that talk to Ganeti to know
336
what are the ranges of allowed values for instances and the
337
over-subscription that is allowed.
338

    
339
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
340
instance configuration, and the physical CPU configuration from the
341
node. For the spindle ratios however, we didn't track before these
342
values, so new parameters have been added:
343

    
344
- a new node parameter ``spindle_count``, defaults to 1, customisable at
345
  node group or node level
346
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
347

    
348
Note that spindles in this context doesn't need to mean actual
349
mechanical hard-drives; it's just a relative number for both the node
350
I/O capacity and instance I/O consumption.
351

    
352
Instance migration behaviour
353
++++++++++++++++++++++++++++
354

    
355
While live-migration is in general desirable over failover, it is
356
possible that for some workloads it is actually worse, due to the
357
variable time of the “suspend” phase during live migration.
358

    
359
To allow the tools to work consistently over such instances (without
360
having to hard-code instance names), a new backend parameter
361
``always_failover`` has been added to control the migration/failover
362
behaviour. When set to True, all migration requests for an instance will
363
instead fall-back to failover.
364

    
365
Instance memory ballooning
366
++++++++++++++++++++++++++
367

    
368
Initial support for memory ballooning has been added. The memory for an
369
instance is no longer fixed (backend parameter ``memory``), but instead
370
can vary between minimum and maximum values (backend parameters
371
``minmem`` and ``maxmem``). Currently we only change an instance's
372
memory when:
373

    
374
- live migrating or failing over and instance and the target node
375
  doesn't have enough memory
376
- user requests changing the memory via ``gnt-instance modify
377
  --runtime-memory``
378

    
379
Instance CPU pinning
380
++++++++++++++++++++
381

    
382
In order to control the use of specific CPUs by instance, support for
383
controlling CPU pinning has been added for the Xen, HVM and LXC
384
hypervisors. This is controlled by a new hypervisor parameter
385
``cpu_mask``; details about possible values for this are in the
386
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
387
VCPU-to-CPU mapping) form will work well only when all nodes in your
388
cluster have the same amount of CPUs.
389

    
390
Disk parameters
391
+++++++++++++++
392

    
393
Another area in which Ganeti was not customisable were the parameters
394
used for storage configuration, e.g. how many stripes to use for LVM,
395
DRBD resync configuration, etc.
396

    
397
To improve this area, we've added disks parameters, which are
398
customisable at cluster and node group level, and which allow to
399
specify various parameters for disks (DRBD has the most parameters
400
currently), for example:
401

    
402
- DRBD resync algorithm and parameters (e.g. speed)
403
- the default VG for meta-data volumes for DRBD
404
- number of stripes for LVM (plain disk template)
405
- the RBD pool
406

    
407
These parameters can be modified via ``gnt-cluster modify -D …`` and
408
``gnt-group modify -D …``, and are used at either instance creation (in
409
case of LVM stripes, for example) or at disk “activation” time
410
(e.g. resync speed).
411

    
412
Rados block device support
413
++++++++++++++++++++++++++
414

    
415
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
416
denoted by the ``rbd`` disk template type. This is considered
417
experimental, feedback is welcome. For details on configuring it, see
418
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
419

    
420
Master IP setup
421
+++++++++++++++
422

    
423
The existing master IP functionality works well only in simple setups (a
424
single network shared by all nodes); however, if nodes belong to
425
different networks, then the ``/32`` setup and lack of routing
426
information is not enough.
427

    
428
To allow the master IP to function well in more complex cases, the
429
system was reworked as follows:
430

    
431
- a master IP netmask setting has been added
432
- the master IP activation/turn-down code was moved from the node daemon
433
  to a separate script
434
- whether to run the Ganeti-supplied master IP script or a user-supplied
435
  on is a ``gnt-cluster init`` setting
436

    
437
Details about the location of the standard and custom setup scripts are
438
in the man page :manpage:`gnt-cluster(8)`; for information about the
439
setup script protocol, look at the Ganeti-supplied script.
440

    
441
SPICE support
442
+++++++++++++
443

    
444
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
445
improved.
446

    
447
It is now possible to use TLS-protected connections, and when renewing
448
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
449
it is now possible to specify spice or spice CA certificates. Also, it
450
is possible to configure a password for SPICE sessions via the
451
hypervisor parameter ``spice_password_file``.
452

    
453
There are also new parameters to control the compression and streaming
454
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
455
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
456
for the spice parameters.
457

    
458
Lastly, it is now possible to see the SPICE connection information via
459
``gnt-instance console``.
460

    
461
OVF converter
462
+++++++++++++
463

    
464
A new tool (``tools/ovfconverter``) has been added that supports
465
conversion between Ganeti and the `Open Virtualization Format
466
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
467
from).
468

    
469
This relies on the ``qemu-img`` tool to convert the disk formats, so the
470
actual compatibility with other virtualization solutions depends on it.
471

    
472
Confd daemon changes
473
++++++++++++++++++++
474

    
475
The configuration query daemon (``ganeti-confd``) is now optional, and
476
has been rewritten in Haskell; whether to use the daemon at all, use the
477
Python (default) or the Haskell version is selectable at configure time
478
via the ``--enable-confd`` parameter, which can take one of the
479
``haskell``, ``python`` or ``no`` values. If not used, disabling the
480
daemon will result in a smaller footprint; for larger systems, we
481
welcome feedback on the Haskell version which might become the default
482
in future versions.
483

    
484
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
485
daemon running. The Python version doesn't implement the new call.
486

    
487

    
488
User interface changes
489
~~~~~~~~~~~~~~~~~~~~~~
490

    
491
We have replaced the ``--disks`` option of ``gnt-instance
492
replace-disks`` with a more flexible ``--disk`` option, which allows
493
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
494
disk size and mode can be changed upon recreation (via ``gnt-instance
495
recreate-disks``, which accepts the same ``--disk`` option).
496

    
497
As many people are used to a ``show`` command, we have added that as an
498
alias to ``info`` on all ``gnt-*`` commands.
499

    
500
The ``gnt-instance grow-disk`` command has a new mode in which it can
501
accept the target size of the disk, instead of the delta; this can be
502
more safe since two runs in absolute mode will be idempotent, and
503
sometimes it's also easier to specify the desired size directly.
504

    
505
Also the handling of instances with regard to offline secondaries has
506
been improved. Instance operations should not fail because one of it's
507
secondary nodes is offline, even though it's safe to proceed.
508

    
509
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
510
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
511
minors to instance name.
512

    
513
API changes
514
~~~~~~~~~~~
515

    
516
RAPI coverage has improved, with (for example) new resources for
517
recreate-disks, node power-cycle, etc.
518

    
519
Compatibility
520
~~~~~~~~~~~~~
521

    
522
There is partial support for ``xl`` in the Xen hypervisor; feedback is
523
welcome.
524

    
525
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
526
whether to still support Python 2.4 or move to Python 2.6 as minimum
527
required version.
528

    
529
Support for Fedora has been slightly improved; the provided example
530
init.d script should work better on it and the INSTALL file should
531
document the needed dependencies.
532

    
533
Internal changes
534
~~~~~~~~~~~~~~~~
535

    
536
The deprecated ``QueryLocks`` LUXI request has been removed. Use
537
``Query(what=QR_LOCK, ...)`` instead.
538

    
539
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
540
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
541
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
542
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
543
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
544

    
545
RAPI client: ``CertificateError`` now derives from
546
``GanetiApiError``. This should make it more easy to handle Ganeti
547
errors.
548

    
549
Deprecation warnings due to PyCrypto/paramiko import in
550
``tools/setup-ssh`` have been silenced, as usually they are safe; please
551
make sure to run an up-to-date paramiko version, if you use this tool.
552

    
553
The QA scripts now depend on Python 2.5 or above (the main code base
554
still works with Python 2.4).
555

    
556
The configuration file (``config.data``) is now written without
557
indentation for performance reasons; if you want to edit it, it can be
558
re-formatted via ``tools/fmtjson``.
559

    
560
A number of bugs has been fixed in the cluster merge tool.
561

    
562
``x509`` certification verification (used in import-export) has been
563
changed to allow the same clock skew as permitted by the cluster
564
verification. This will remove some rare but hard to diagnose errors in
565
import-export.
566

    
567

    
568
Version 2.6.0 rc4
569
-----------------
570

    
571
*(Released Thu, 19 Jul 2012)*
572

    
573
Very few changes from rc4 to the final release, only bugfixes:
574

    
575
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
576
  instance, fix CDROM booting for KVM instances)
577
- fixed node group modification of node parameters
578
- fixed issue in LUClusterVerifyGroup with multi-group clusters
579
- fixed generation of bash completion to ensure a stable ordering
580
- fixed a few typos
581

    
582

    
583
Version 2.6.0 rc3
584
-----------------
585

    
586
*(Released Fri, 13 Jul 2012)*
587

    
588
Third release candidate for 2.6. The following changes were done from
589
rc3 to rc4:
590

    
591
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
592
- Fixed an inconsistency in the LUXI protocol with the provided
593
  arguments (NOT backwards compatible)
594
- Fixed a bug with node groups ipolicy where ``min`` was greater than
595
  the cluster ``std`` value
596
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
597
  easier instance debugging on nodes (requires ``hconfd`` to work)
598

    
599

    
600
Version 2.6.0 rc2
601
-----------------
602

    
603
*(Released Tue, 03 Jul 2012)*
604

    
605
Second release candidate for 2.6. The following changes were done from
606
rc2 to rc3:
607

    
608
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
609
  master candidates
610
- Fixed a RAPI regression on missing beparams/memory
611
- Fixed redistribution of files on offline nodes
612
- Added possibility to run activate-disks even though secondaries are
613
  offline. With this change it relaxes also the strictness on some other
614
  commands which use activate disks internally:
615
  * ``gnt-instance start|reboot|rename|backup|export``
616
- Made it possible to remove safely an instance if its secondaries are
617
  offline
618
- Made it possible to reinstall even though secondaries are offline
619

    
620

    
621
Version 2.6.0 rc1
622
-----------------
623

    
624
*(Released Mon, 25 Jun 2012)*
625

    
626
First release candidate for 2.6. The following changes were done from
627
rc1 to rc2:
628

    
629
- Fixed bugs with disk parameters and ``rbd`` templates as well as
630
  ``instance_os_add``
631
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
632
  behaviour. It supports now the modify operation
633
- ``hcheck`` implemented to analyze cluster health and possibility of
634
  improving health by rebalance
635
- ``hbal`` has been improved in dealing with split instances
636

    
637

    
638
Version 2.6.0 beta2
639
-------------------
640

    
641
*(Released Mon, 11 Jun 2012)*
642

    
643
Second beta release of 2.6. The following changes were done from beta2
644
to rc1:
645

    
646
- Fixed ``daemon-util`` with non-root user models
647
- Fixed creation of plain instances with ``--no-wait-for-sync``
648
- Fix wrong iv_names when running ``cfgupgrade``
649
- Export more information in RAPI group queries
650
- Fixed bug when changing instance network interfaces
651
- Extended burnin to do NIC changes
652
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
653
- Changed default for DRBD barriers
654
- Fixed DRBD error reporting for syncer rate
655
- Verify the options on disk parameters
656

    
657
And of course various fixes to documentation and improved unittests and
658
QA.
659

    
660

    
661
Version 2.6.0 beta1
662
-------------------
663

    
664
*(Released Wed, 23 May 2012)*
665

    
666
First beta release of 2.6. The following changes were done from beta1 to
667
beta2:
668

    
669
- integrated patch for distributions without ``start-stop-daemon``
670
- adapted example init.d script to work on Fedora
671
- fixed log handling in Haskell daemons
672
- adapted checks in the watcher for pycurl linked against libnss
673
- add partial support for ``xl`` instead of ``xm`` for Xen
674
- fixed a type issue in cluster verification
675
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
676
  clusters)
677

    
678
Plus integrated fixes from the 2.5 branch:
679

    
680
- fixed ``kvm-ifup`` to use ``/bin/bash``
681
- fixed parallel build failures
682
- KVM live migration when using a custom keymap
683

    
684

    
685
Version 2.5.2
686
-------------
687

    
688
*(Released Tue, 24 Jul 2012)*
689

    
690
A small bugfix release, with no new features:
691

    
692
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
693
  different default shell (e.g. Debian, Ubuntu)
694
- fixed KVM startup and live migration with a custom keymap (fixes Issue
695
  243 and Debian bug #650664)
696
- fixed compatibility with KVM versions that don't support multiple boot
697
  devices (fixes Issue 230 and Debian bug #624256)
698

    
699
Additionally, a few fixes were done to the build system (fixed parallel
700
build failures) and to the unittests (fixed race condition in test for
701
FileID functions, and the default enable/disable mode for QA test is now
702
customisable).
703

    
704

    
705
Version 2.5.1
706
-------------
707

    
708
*(Released Fri, 11 May 2012)*
709

    
710
A small bugfix release.
711

    
712
The main issues solved are on the topic of compatibility with newer LVM
713
releases:
714

    
715
- fixed parsing of ``lv_attr`` field
716
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
717
  the ``--force`` flag is needed
718

    
719
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
720
to accept kernel 3.x too (was hardcoded to 2.6.*).
721

    
722
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
723
of hook results and that also made display of other errors suboptimal
724
was fixed; the code behaves now like 2.4 and earlier.
725

    
726
Another change in 2.5, the cleanup of the OS scripts environment, is too
727
aggressive: it removed even the ``PATH`` variable, which requires the OS
728
scripts to *always* need to export it. Since this is a bit too strict,
729
we now export a minimal PATH, the same that we export for hooks.
730

    
731
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
732
integrated into this release.
733

    
734
Finally, a few other miscellaneous changes were done (no new features,
735
just small improvements):
736

    
737
- Fix ``gnt-group --help`` display
738
- Fix hardcoded Xen kernel path
739
- Fix grow-disk handling of invalid units
740
- Update synopsis for ``gnt-cluster repair-disk-sizes``
741
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
742

    
743

    
744
Version 2.5.0
745
-------------
746

    
747
*(Released Thu, 12 Apr 2012)*
748

    
749
Incompatible/important changes and bugfixes
750
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
751

    
752
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
753
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
754
  to match the underlying LUXI interface.
755
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
756
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
757
  not maintain backwards-compatibility as the underlying operation
758
  changed in an incompatible way. The RAPI client can talk to old
759
  servers, but it needs to be told so as the return value changed.
760
- When creating file-based instances via RAPI, the ``file_driver``
761
  parameter no longer defaults to ``loop`` and must be specified.
762
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
763
  ``link`` instead.
764
- Support for the undocumented and deprecated RAPI instance creation
765
  request format version 0 has been dropped. Use version 1, supported
766
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
767
- Pyparsing 1.4.6 or above is required, see :doc:`installation
768
  documentation <install>`.
769
- The "cluster-verify" hooks are now executed per group by the
770
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
771
  if you just run ``gnt-cluster verify``, which generates one opcode per
772
  group.
773
- The environment as passed to the OS scripts is cleared, and thus no
774
  environment variables defined in the node daemon's environment will be
775
  inherited by the scripts.
776
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
777
  deprecated.
778
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
779
  support operations involving multiple node groups.
780
- Offline nodes are ignored when failing over an instance.
781
- Support for KVM version 1.0, which changed the version reporting format
782
  from 3 to 2 digits.
783
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
784
  removal.
785
- ``Makefile`` is now compatible with Automake 1.11.2
786
- Includes all bugfixes made in the 2.4 series
787

    
788
New features
789
~~~~~~~~~~~~
790

    
791
- The ganeti-htools project has been merged into the ganeti-core source
792
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
793
- Implemented support for :doc:`shared storage <design-shared-storage>`.
794
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
795
  GPT-style partition tables (requires `parted
796
  <http://www.gnu.org/s/parted/>`_).
797
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
798
- Allowed adding tags on instance creation.
799
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
800
- Allow instances to be started in a paused state, enabling the user to
801
  see the complete console output on boot using the console.
802
- Added new hypervisor flag to control default reboot behaviour
803
  (``reboot_behavior``).
804
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
805
- Improved out-of-band management support:
806

    
807
  - Added ``gnt-node health`` command reporting the health status of
808
    nodes.
809
  - Added ``gnt-node power`` command to manage power status of nodes.
810
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
811

    
812
- Instance migration can fall back to failover if instance is not
813
  running.
814
- Filters can be used when listing nodes, instances, groups and locks;
815
  see :manpage:`ganeti(7)` manpage.
816
- Added post-execution status as variables to :doc:`hooks <hooks>`
817
  environment.
818
- Instance tags are exported/imported together with the instance.
819
- When given an explicit job ID, ``gnt-job info`` will work for archived
820
  jobs.
821
- Jobs can define dependencies on other jobs (not yet supported via
822
  RAPI or command line, but used by internal commands and usable via
823
  LUXI).
824

    
825
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
826
    dependencies.
827

    
828
- Instance failover is now available as a RAPI resource
829
  (``/2/instances/[instance_name]/failover``).
830
- ``gnt-instance info`` defaults to static information if primary node
831
  is offline.
832
- Opcodes have a new ``comment`` attribute.
833
- Added basic SPICE support to KVM hypervisor.
834
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
835

    
836
Node group improvements
837
~~~~~~~~~~~~~~~~~~~~~~~
838

    
839
- ``gnt-cluster verify`` has been modified to check groups separately,
840
  thereby improving performance.
841
- Node group support has been added to ``gnt-cluster verify-disks``,
842
  which now operates per node group.
843
- Watcher has been changed to work better with node groups.
844

    
845
  - One process and state file per node group.
846
  - Slow watcher in one group doesn't block other group's watcher.
847

    
848
- Added new command, ``gnt-group evacuate``, to move all instances in a
849
  node group to other groups.
850
- Added ``gnt-instance change-group`` to move an instance to another
851
  node group.
852
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
853
  per-group operations.
854
- Node groups can be tagged.
855
- Some operations switch from an exclusive to a shared lock as soon as
856
  possible.
857
- Instance's primary and secondary nodes' groups are now available as
858
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
859
  and ``snodes.group.uuid``).
860

    
861
Misc
862
~~~~
863

    
864
- Numerous updates to documentation and manpages.
865

    
866
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
867
    descriptions.
868
  - Some opcode/job results are now also documented, see :doc:`RAPI
869
    <rapi>`.
870

    
871
- A lockset's internal lock is now also visible in lock monitor.
872
- Log messages from job queue workers now contain information about the
873
  opcode they're processing.
874
- ``gnt-instance console`` no longer requires the instance lock.
875
- A short delay when waiting for job changes reduces the number of LUXI
876
  requests significantly.
877
- DRBD metadata volumes are overwritten with zeros during disk creation.
878
- Out-of-band commands no longer acquire the cluster lock in exclusive
879
  mode.
880
- ``devel/upload`` now uses correct permissions for directories.
881

    
882

    
883
Version 2.5.0 rc6
884
-----------------
885

    
886
*(Released Fri, 23 Mar 2012)*
887

    
888
This was the sixth release candidate of the 2.5 series.
889

    
890

    
891
Version 2.5.0 rc5
892
-----------------
893

    
894
*(Released Mon, 9 Jan 2012)*
895

    
896
This was the fifth release candidate of the 2.5 series.
897

    
898

    
899
Version 2.5.0 rc4
900
-----------------
901

    
902
*(Released Thu, 27 Oct 2011)*
903

    
904
This was the fourth release candidate of the 2.5 series.
905

    
906

    
907
Version 2.5.0 rc3
908
-----------------
909

    
910
*(Released Wed, 26 Oct 2011)*
911

    
912
This was the third release candidate of the 2.5 series.
913

    
914

    
915
Version 2.5.0 rc2
916
-----------------
917

    
918
*(Released Tue, 18 Oct 2011)*
919

    
920
This was the second release candidate of the 2.5 series.
921

    
922

    
923
Version 2.5.0 rc1
924
-----------------
925

    
926
*(Released Tue, 4 Oct 2011)*
927

    
928
This was the first release candidate of the 2.5 series.
929

    
930

    
931
Version 2.5.0 beta3
932
-------------------
933

    
934
*(Released Wed, 31 Aug 2011)*
935

    
936
This was the third beta release of the 2.5 series.
937

    
938

    
939
Version 2.5.0 beta2
940
-------------------
941

    
942
*(Released Mon, 22 Aug 2011)*
943

    
944
This was the second beta release of the 2.5 series.
945

    
946

    
947
Version 2.5.0 beta1
948
-------------------
949

    
950
*(Released Fri, 12 Aug 2011)*
951

    
952
This was the first beta release of the 2.5 series.
953

    
954

    
955
Version 2.4.5
956
-------------
957

    
958
*(Released Thu, 27 Oct 2011)*
959

    
960
- Fixed bug when parsing command line parameter values ending in
961
  backslash
962
- Fixed assertion error after unclean master shutdown
963
- Disable HTTP client pool for RPC, significantly reducing memory usage
964
  of master daemon
965
- Fixed queue archive creation with wrong permissions
966

    
967

    
968
Version 2.4.4
969
-------------
970

    
971
*(Released Tue, 23 Aug 2011)*
972

    
973
Small bug-fixes:
974

    
975
- Fixed documentation for importing with ``--src-dir`` option
976
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
977
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
978

    
979

    
980
Version 2.4.3
981
-------------
982

    
983
*(Released Fri, 5 Aug 2011)*
984

    
985
Many bug-fixes and a few small features:
986

    
987
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
988
  issues when you tried to add an instance with two MAC addresses in one
989
  request
990
- KVM: fixed per-instance stored UID value
991
- KVM: configure bridged NICs at migration start
992
- KVM: Fix a bug where instance will not start with never KVM versions
993
  (>= 0.14)
994
- Added OS search path to ``gnt-cluster info``
995
- Fixed an issue with ``file_storage_dir`` where you were forced to
996
  provide an absolute path, but the documentation states it is a
997
  relative path, the documentation was right
998
- Added a new parameter to instance stop/start called ``--no-remember``
999
  that will make the state change to not be remembered
1000
- Implemented ``no_remember`` at RAPI level
1001
- Improved the documentation
1002
- Node evacuation: don't call IAllocator if node is already empty
1003
- Fixed bug in DRBD8 replace disks on current nodes
1004
- Fixed bug in recreate-disks for DRBD instances
1005
- Moved assertion checking locks in ``gnt-instance replace-disks``
1006
  causing it to abort with not owning the right locks for some situation
1007
- Job queue: Fixed potential race condition when cancelling queued jobs
1008
- Fixed off-by-one bug in job serial generation
1009
- ``gnt-node volumes``: Fix instance names
1010
- Fixed aliases in bash completion
1011
- Fixed a bug in reopening log files after being sent a SIGHUP
1012
- Added a flag to burnin to allow specifying VCPU count
1013
- Bugfixes to non-root Ganeti configuration
1014

    
1015

    
1016
Version 2.4.2
1017
-------------
1018

    
1019
*(Released Thu, 12 May 2011)*
1020

    
1021
Many bug-fixes and a few new small features:
1022

    
1023
- Fixed a bug related to log opening failures
1024
- Fixed a bug in instance listing with orphan instances
1025
- Fixed a bug which prevented resetting the cluster-level node parameter
1026
  ``oob_program`` to the default
1027
- Many fixes related to the ``cluster-merge`` tool
1028
- Fixed a race condition in the lock monitor, which caused failures
1029
  during (at least) creation of many instances in parallel
1030
- Improved output for gnt-job info
1031
- Removed the quiet flag on some ssh calls which prevented debugging
1032
  failures
1033
- Improved the N+1 failure messages in cluster verify by actually
1034
  showing the memory values (needed and available)
1035
- Increased lock attempt timeouts so that when executing long operations
1036
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1037
  too early and thus prevent the use of the 'fair' mechanism
1038
- Changed instance query data (``gnt-instance info``) to not acquire
1039
  locks unless needed, thus allowing its use on locked instance if only
1040
  static information is asked for
1041
- Improved behaviour with filesystems that do not support rename on an
1042
  opened file
1043
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1044
  kept locks on all nodes during the wipe, which is unneeded
1045
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1046
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1047
  10GiB) which caused the wipe to fail right at the end in some cases
1048
- Fixed master IP activation when doing master failover with no-voting
1049
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1050
  the master node itself
1051
- Fixed potential data-loss in under disk full conditions, where Ganeti
1052
  wouldn't check correctly the return code and would consider
1053
  partially-written files 'correct'
1054
- Fixed bug related to multiple VGs and DRBD disk replacing
1055
- Added new disk parameter ``metavg`` that allows placement of the meta
1056
  device for DRBD in a different volume group
1057
- Fixed error handling in the node daemon when the system libc doesn't
1058
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1059
- Fixed lock release during replace-disks, which kept cluster-wide locks
1060
  when doing disk replaces with an iallocator script
1061
- Added check for missing bridges in cluster verify
1062
- Handle EPIPE errors while writing to the terminal better, so that
1063
  piping the output to e.g. ``less`` doesn't cause a backtrace
1064
- Fixed rare case where a ^C during Luxi calls could have been
1065
  interpreted as server errors, instead of simply terminating
1066
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1067
  assign-nodes``)
1068
- Added a few more parameters to the KVM hypervisor, allowing a second
1069
  CDROM, custom disk type for CDROMs and a floppy image
1070
- Removed redundant message in instance rename when the name is given
1071
  already as a FQDN
1072
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1073
  disks on new nodes, allowing recreation when the original instance
1074
  nodes are completely gone
1075
- Added option when converting disk templates to DRBD to skip waiting
1076
  for the resync, in order to make the instance available sooner
1077
- Added two new variables to the OS scripts environment (containing the
1078
  instance's nodes)
1079
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1080
  to allow use of ``pvgrub`` as bootloader
1081
- Changed the instance memory modifications to only check out-of-memory
1082
  conditions on memory increases, and turned the secondary node warnings
1083
  into errors (they can still be overridden via ``--force``)
1084
- Fixed the handling of a corner case when the Python installation gets
1085
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1086
  to execute a command that doesn't exist
1087
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1088
  node of the instance returned failures during instance shutdown; this
1089
  adds the option ``--ignore-consistency`` to gnt-instance move
1090

    
1091
And as usual, various improvements to the error messages, documentation
1092
and man pages.
1093

    
1094

    
1095
Version 2.4.1
1096
-------------
1097

    
1098
*(Released Wed, 09 Mar 2011)*
1099

    
1100
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1101
the RAPI users file if run twice (even with ``--dry-run``).
1102

    
1103
The release fixes that bug (nothing else changed).
1104

    
1105

    
1106
Version 2.4.0
1107
-------------
1108

    
1109
*(Released Mon, 07 Mar 2011)*
1110

    
1111
Final 2.4.0 release. Just a few small fixes:
1112

    
1113
- Fixed RAPI node evacuate
1114
- Fixed the kvm-ifup script
1115
- Fixed internal error handling for special job cases
1116
- Updated man page to specify the escaping feature for options
1117

    
1118

    
1119
Version 2.4.0 rc3
1120
-----------------
1121

    
1122
*(Released Mon, 28 Feb 2011)*
1123

    
1124
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1125
that this feature wiped the disks of the wrong instance, leading to loss
1126
of data.
1127

    
1128
Other changes:
1129

    
1130
- Fixed title of query field containing instance name
1131
- Expanded the glossary in the documentation
1132
- Fixed one unittest (internal issue)
1133

    
1134

    
1135
Version 2.4.0 rc2
1136
-----------------
1137

    
1138
*(Released Mon, 21 Feb 2011)*
1139

    
1140
A number of bug fixes plus just a couple functionality changes.
1141

    
1142
On the user-visible side, the ``gnt-* list`` command output has changed
1143
with respect to "special" field states. The current rc1 style of display
1144
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1145
the default output mode special fields states are displayed as follows:
1146

    
1147
- Offline resource: ``*``
1148
- Unavailable/not applicable: ``-``
1149
- Data missing (RPC failure): ``?``
1150
- Unknown field: ``??``
1151

    
1152
Another user-visible change is the addition of ``--force-join`` to
1153
``gnt-node add``.
1154

    
1155
As for bug fixes:
1156

    
1157
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1158
- Fixed regression in RAPI/instance reinstall where all parameters were
1159
  required (instead of optional)
1160
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1161
- Fixed iallocator usage (offline nodes were not considered offline)
1162
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1163
- Fixed hypervisor and OS parameter validation with respect to
1164
  non-vm_capable nodes
1165
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1166
  cosmetic)
1167
- Fixed ``tools/listrunner`` with respect to agent-based usage
1168

    
1169

    
1170
Version 2.4.0 rc1
1171
-----------------
1172

    
1173
*(Released Fri,  4 Feb 2011)*
1174

    
1175
Many changes and fixes since the beta1 release. While there were some
1176
internal changes, the code has been mostly stabilised for the RC
1177
release.
1178

    
1179
Note: the dumb allocator was removed in this release, as it was not kept
1180
up-to-date with the IAllocator protocol changes. It is recommended to
1181
use the ``hail`` command from the ganeti-htools package.
1182

    
1183
Note: the 2.4 and up versions of Ganeti are not compatible with the
1184
0.2.x branch of ganeti-htools. You need to upgrade to
1185
ganeti-htools-0.3.0 (or later).
1186

    
1187
Regressions fixed from 2.3
1188
~~~~~~~~~~~~~~~~~~~~~~~~~~
1189

    
1190
- Fixed the ``gnt-cluster verify-disks`` command
1191
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1192
  serially on nodes)
1193
- Fixed disk adoption breakage
1194
- Fixed wrong headers in instance listing for field aliases
1195

    
1196
Other bugs fixed
1197
~~~~~~~~~~~~~~~~
1198

    
1199
- Fixed corner case in KVM handling of NICs
1200
- Fixed many cases of wrong handling of non-vm_capable nodes
1201
- Fixed a bug where a missing instance symlink was not possible to
1202
  recreate with any ``gnt-*`` command (now ``gnt-instance
1203
  activate-disks`` does it)
1204
- Fixed the volume group name as reported by ``gnt-cluster
1205
  verify-disks``
1206
- Increased timeouts for the import-export code, hopefully leading to
1207
  fewer aborts due network or instance timeouts
1208
- Fixed bug in ``gnt-node list-storage``
1209
- Fixed bug where not all daemons were started on cluster
1210
  initialisation, but only at the first watcher run
1211
- Fixed many bugs in the OOB implementation
1212
- Fixed watcher behaviour in presence of instances with offline
1213
  secondaries
1214
- Fixed instance list output for instances running on the wrong node
1215
- a few fixes to the cluster-merge tool, but it still cannot merge
1216
  multi-node groups (currently it is not recommended to use this tool)
1217

    
1218

    
1219
Improvements
1220
~~~~~~~~~~~~
1221

    
1222
- Improved network configuration for the KVM hypervisor
1223
- Added e1000 as a supported NIC for Xen-HVM
1224
- Improved the lvmstrap tool to also be able to use partitions, as
1225
  opposed to full disks
1226
- Improved speed of disk wiping (the cluster parameter
1227
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1228
  of instance creations
1229
- Added documentation for the OS parameters
1230
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1231
  hypervisor is not responding
1232
- Added display of blacklisted and hidden OS information in
1233
  ``gnt-cluster info``
1234
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1235
  NIC and node parameters, which might create problems with currently
1236
  invalid (but undetected) configuration files, but prevents validation
1237
  failures when unrelated parameters are modified
1238
- Changed cluster initialisation to wait for the master daemon to become
1239
  available
1240
- Expanded the RAPI interface:
1241

    
1242
  - Added config redistribution resource
1243
  - Added activation/deactivation of instance disks
1244
  - Added export of console information
1245

    
1246
- Implemented log file reopening on SIGHUP, which allows using
1247
  logrotate(8) for the Ganeti log files
1248
- Added a basic OOB helper script as an example
1249

    
1250

    
1251
Version 2.4.0 beta1
1252
-------------------
1253

    
1254
*(Released Fri, 14 Jan 2011)*
1255

    
1256
User-visible
1257
~~~~~~~~~~~~
1258

    
1259
- Fixed timezone issues when formatting timestamps
1260
- Added support for node groups, available via ``gnt-group`` and other
1261
  commands
1262
- Added out-of-band framework and management, see :doc:`design
1263
  document <design-oob>`
1264
- Removed support for roman numbers from ``gnt-node list`` and
1265
  ``gnt-instance list``.
1266
- Allowed modification of master network interface via ``gnt-cluster
1267
  modify --master-netdev``
1268
- Accept offline secondaries while shutting down instance disks
1269
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1270
- Added support for multiple LVM volume groups
1271
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1272
  requested
1273
- Added commands to list available fields:
1274

    
1275
  - ``gnt-node list-fields``
1276
  - ``gnt-group list-fields``
1277
  - ``gnt-instance list-fields``
1278

    
1279
- Updated documentation and man pages
1280

    
1281
Integration
1282
~~~~~~~~~~~
1283

    
1284
- Moved ``rapi_users`` file into separate directory, now named
1285
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1286
  symlink
1287
- Added new tool for running commands on many machines,
1288
  ``tools/ganeti-listrunner``
1289
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1290
  improving the ``gnt-instance console`` output
1291
- Allowed customisation of disk index separator at ``configure`` time
1292
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1293
- Added support for non-partitioned md disks in ``lvmstrap``
1294
- Added script to gracefully power off KVM instances
1295
- Split ``utils`` module into smaller parts
1296
- Changed query operations to return more detailed information, e.g.
1297
  whether an information is unavailable due to an offline node. To use
1298
  this new functionality, the LUXI call ``Query`` must be used. Field
1299
  information is now stored by the master daemon and can be retrieved
1300
  using ``QueryFields``. Instances, nodes and groups can also be queried
1301
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1302
  exposed via RAPI). The following commands make use of this
1303
  infrastructure change:
1304

    
1305
  - ``gnt-group list``
1306
  - ``gnt-group list-fields``
1307
  - ``gnt-node list``
1308
  - ``gnt-node list-fields``
1309
  - ``gnt-instance list``
1310
  - ``gnt-instance list-fields``
1311
  - ``gnt-debug locks``
1312

    
1313
Remote API
1314
~~~~~~~~~~
1315

    
1316
- New RAPI resources (see :doc:`rapi`):
1317

    
1318
  - ``/2/modify``
1319
  - ``/2/groups``
1320
  - ``/2/groups/[group_name]``
1321
  - ``/2/groups/[group_name]/assign-nodes``
1322
  - ``/2/groups/[group_name]/modify``
1323
  - ``/2/groups/[group_name]/rename``
1324
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1325

    
1326
- RAPI changes:
1327

    
1328
  - Implemented ``no_install`` for instance creation
1329
  - Implemented OS parameters for instance reinstallation, allowing
1330
    use of special settings on reinstallation (e.g. for preserving data)
1331

    
1332
Misc
1333
~~~~
1334

    
1335
- Added IPv6 support in import/export
1336
- Pause DRBD synchronization while wiping disks on instance creation
1337
- Updated unittests and QA scripts
1338
- Improved network parameters passed to KVM
1339
- Converted man pages from docbook to reStructuredText
1340

    
1341

    
1342
Version 2.3.1
1343
-------------
1344

    
1345
*(Released Mon, 20 Dec 2010)*
1346

    
1347
Released version 2.3.1~rc1 without any changes.
1348

    
1349

    
1350
Version 2.3.1 rc1
1351
-----------------
1352

    
1353
*(Released Wed, 1 Dec 2010)*
1354

    
1355
- impexpd: Disable OpenSSL compression in socat if possible (backport
1356
  from master, commit e90739d625b, see :doc:`installation guide
1357
  <install-quick>` for details)
1358
- Changed unittest coverage report to exclude test scripts
1359
- Added script to check version format
1360

    
1361

    
1362
Version 2.3.0
1363
-------------
1364

    
1365
*(Released Wed, 1 Dec 2010)*
1366

    
1367
Released version 2.3.0~rc1 without any changes.
1368

    
1369

    
1370
Version 2.3.0 rc1
1371
-----------------
1372

    
1373
*(Released Fri, 19 Nov 2010)*
1374

    
1375
A number of bugfixes and documentation updates:
1376

    
1377
- Update ganeti-os-interface documentation
1378
- Fixed a bug related to duplicate MACs or similar items which should be
1379
  unique
1380
- Fix breakage in OS state modify
1381
- Reinstall instance: disallow offline secondaries (fixes bug related to
1382
  OS changing but reinstall failing)
1383
- plus all the other fixes between 2.2.1 and 2.2.2
1384

    
1385

    
1386
Version 2.3.0 rc0
1387
-----------------
1388

    
1389
*(Released Tue, 2 Nov 2010)*
1390

    
1391
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1392
- Fixed master failover race with watcher
1393
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1394
  configuration
1395
- Accept previously stopped instance for export with instance removal
1396
- Simplify and extend the environment variables for instance OS scripts
1397
- Added new node flags, ``master_capable`` and ``vm_capable``
1398
- Added optional instance disk wiping prior during allocation. This is a
1399
  cluster-wide option and can be set/modified using
1400
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1401
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1402
  :doc:`install-quick`
1403
- Added a new watcher option (``--ignore-pause``)
1404
- Added option to ignore offline node on instance start/stop
1405
  (``--ignore-offline``)
1406
- Allow overriding OS parameters with ``gnt-instance reinstall``
1407
- Added ability to change node's secondary IP address using ``gnt-node
1408
  modify``
1409
- Implemented privilege separation for all daemons except
1410
  ``ganeti-noded``, see ``configure`` options
1411
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1412
  verify``
1413
- Implemented job priorities (see ``ganeti(7)`` manpage)
1414
- Ignore failures while shutting down instances during failover from
1415
  offline node
1416
- Exit daemon's bootstrap process only once daemon is ready
1417
- Export more information via ``LUInstanceQuery``/remote API
1418
- Improved documentation, QA and unittests
1419
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1420
  restart if the file was created or changed
1421
- Added LUXI protocol version sent with each request and response,
1422
  allowing detection of server/client mismatches
1423
- Moved the Python scripts among gnt-* and ganeti-* into modules
1424
- Moved all code related to setting up SSH to an external script,
1425
  ``setup-ssh``
1426
- Infrastructure changes for node group support in future versions
1427

    
1428

    
1429
Version 2.2.2
1430
-------------
1431

    
1432
*(Released Fri, 19 Nov 2010)*
1433

    
1434
A few small bugs fixed, and some improvements to the build system:
1435

    
1436
- Fix documentation regarding conversion to drbd
1437
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1438
  -B``)
1439
- Fix error handling in node modify with multiple changes
1440
- Allow remote imports without checked names
1441

    
1442

    
1443
Version 2.2.1
1444
-------------
1445

    
1446
*(Released Tue, 19 Oct 2010)*
1447

    
1448
- Disable SSL session ID cache in RPC client
1449

    
1450

    
1451
Version 2.2.1 rc1
1452
-----------------
1453

    
1454
*(Released Thu, 14 Oct 2010)*
1455

    
1456
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1457
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1458
  with GnuTLS
1459
- Fix problems with interaction between Curl and Python's HTTP server,
1460
  resulting in increased speed in many RPC calls
1461
- Improve our release script to prevent breakage with older aclocal and
1462
  Python 2.6
1463

    
1464

    
1465
Version 2.2.1 rc0
1466
-----------------
1467

    
1468
*(Released Thu, 7 Oct 2010)*
1469

    
1470
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1471
  value retrieved from master
1472
- Added support for blacklisted or hidden OS definitions
1473
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1474
- Added support for -mem-path in KVM hypervisor abstraction layer
1475
- Allow overriding instance parameters in tool for inter-cluster
1476
  instance moves (``tools/move-instance``)
1477
- Improved opcode summaries (e.g. in ``gnt-job list``)
1478
- Improve consistency of OS listing by sorting it
1479
- Documentation updates
1480

    
1481

    
1482
Version 2.2.0.1
1483
---------------
1484

    
1485
*(Released Fri, 8 Oct 2010)*
1486

    
1487
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1488

    
1489

    
1490
Version 2.2.0
1491
-------------
1492

    
1493
*(Released Mon, 4 Oct 2010)*
1494

    
1495
- Fixed regression in ``gnt-instance rename``
1496

    
1497

    
1498
Version 2.2.0 rc2
1499
-----------------
1500

    
1501
*(Released Wed, 22 Sep 2010)*
1502

    
1503
- Fixed OS_VARIANT variable for OS scripts
1504
- Fixed cluster tag operations via RAPI
1505
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1506
- Disabled RAPI CA checks in watcher
1507

    
1508

    
1509
Version 2.2.0 rc1
1510
-----------------
1511

    
1512
*(Released Mon, 23 Aug 2010)*
1513

    
1514
- Support DRBD versions of the format "a.b.c.d"
1515
- Updated manpages
1516
- Re-introduce support for usage from multiple threads in RAPI client
1517
- Instance renames and modify via RAPI
1518
- Work around race condition between processing and archival in job
1519
  queue
1520
- Mark opcodes following failed one as failed, too
1521
- Job field ``lock_status`` was removed due to difficulties making it
1522
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1523
  locks is expected for a later 2.2.x release
1524
- Fixed dry-run behaviour with many commands
1525
- Support ``ssh-agent`` again when adding nodes
1526
- Many additional bugfixes
1527

    
1528

    
1529
Version 2.2.0 rc0
1530
-----------------
1531

    
1532
*(Released Fri, 30 Jul 2010)*
1533

    
1534
Important change: the internal RPC mechanism between Ganeti nodes has
1535
changed from using a home-grown http library (based on the Python base
1536
libraries) to use the PycURL library. This requires that PycURL is
1537
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1538
against GnuTLS by default. cURL's support for GnuTLS had known issues
1539
before cURL 7.21.0 and we recommend using the latest cURL release or
1540
linking against OpenSSL. Most other distributions already link PycURL
1541
and cURL against OpenSSL. The command::
1542

    
1543
  python -c 'import pycurl; print pycurl.version'
1544

    
1545
can be used to determine the libraries PycURL and cURL are linked
1546
against.
1547

    
1548
Other significant changes:
1549

    
1550
- Rewrote much of the internals of the job queue, in order to achieve
1551
  better parallelism; this decouples job query operations from the job
1552
  processing, and it should allow much nicer behaviour of the master
1553
  daemon under load, and it also has uncovered some long-standing bugs
1554
  related to the job serialisation (now fixed)
1555
- Added a default iallocator setting to the cluster parameters,
1556
  eliminating the need to always pass nodes or an iallocator for
1557
  operations that require selection of new node(s)
1558
- Added experimental support for the LXC virtualization method
1559
- Added support for OS parameters, which allows the installation of
1560
  instances to pass parameter to OS scripts in order to customise the
1561
  instance
1562
- Added a hypervisor parameter controlling the migration type (live or
1563
  non-live), since hypervisors have various levels of reliability; this
1564
  has renamed the 'live' parameter to 'mode'
1565
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
1566
  logical volumes, meaning that cluster verify will ignore them and not
1567
  flag their presence as errors
1568
- The watcher will now reset the error count for failed instances after
1569
  8 hours, thus allowing self-healing if the problem that caused the
1570
  instances to be down/fail to start has cleared in the meantime
1571
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1572
  check for, and warn, if the drbd module parameter ``usermode_helper``
1573
  is not consistent with the cluster-wide setting; this is needed to
1574
  make diagnose easier of failed drbd creations
1575
- Started adding base IPv6 support, but this is not yet
1576
  enabled/available for use
1577
- Rename operations (cluster, instance) will now return the new name,
1578
  which is especially useful if a short name was passed in
1579
- Added support for instance migration in RAPI
1580
- Added a tool to pre-configure nodes for the SSH setup, before joining
1581
  them to the cluster; this will allow in the future a simplified model
1582
  for node joining (but not yet fully enabled in 2.2); this needs the
1583
  paramiko python library
1584
- Fixed handling of name-resolving errors
1585
- Fixed consistency of job results on the error path
1586
- Fixed master-failover race condition when executed multiple times in
1587
  sequence
1588
- Fixed many bugs related to the job queue (mostly introduced during the
1589
  2.2 development cycle, so not all are impacting 2.1)
1590
- Fixed instance migration with missing disk symlinks
1591
- Fixed handling of unknown jobs in ``gnt-job archive``
1592
- And many other small fixes/improvements
1593

    
1594
Internal changes:
1595

    
1596
- Enhanced both the unittest and the QA coverage
1597
- Switched the opcode validation to a generic model, and extended the
1598
  validation to all opcode parameters
1599
- Changed more parts of the code that write shell scripts to use the
1600
  same class for this
1601
- Switched the master daemon to use the asyncore library for the Luxi
1602
  server endpoint
1603

    
1604

    
1605
Version 2.2.0 beta0
1606
-------------------
1607

    
1608
*(Released Thu, 17 Jun 2010)*
1609

    
1610
- Added tool (``move-instance``) and infrastructure to move instances
1611
  between separate clusters (see :doc:`separate documentation
1612
  <move-instance>` and :doc:`design document <design-2.2>`)
1613
- Added per-request RPC timeout
1614
- RAPI now requires a Content-Type header for requests with a body (e.g.
1615
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1616
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
1617
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1618
  reachable
1619
- Implemented initial support for running Ganeti daemons as separate
1620
  users, see configure-time flags ``--with-user-prefix`` and
1621
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1622
  time)
1623
- Instances can be removed after export (``gnt-backup export
1624
  --remove-instance``)
1625
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1626
  key (instead of 1024 bit)
1627
- Added new cluster configuration file for cluster domain secret
1628
- Import/export now use SSL instead of SSH
1629
- Added support for showing estimated time when exporting an instance,
1630
  see the ``ganeti-os-interface(7)`` manpage and look for
1631
  ``EXP_SIZE_FD``
1632

    
1633

    
1634
Version 2.1.8
1635
-------------
1636

    
1637
*(Released Tue, 16 Nov 2010)*
1638

    
1639
Some more bugfixes. Unless critical bugs occur, this will be the last
1640
2.1 release:
1641

    
1642
- Fix case of MAC special-values
1643
- Fix mac checker regex
1644
- backend: Fix typo causing "out of range" error
1645
- Add missing --units in gnt-instance list man page
1646

    
1647

    
1648
Version 2.1.7
1649
-------------
1650

    
1651
*(Released Tue, 24 Aug 2010)*
1652

    
1653
Bugfixes only:
1654
  - Don't ignore secondary node silently on non-mirrored disk templates
1655
    (issue 113)
1656
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1657
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1658
  - Properly document the usb_mouse parameter
1659
  - Fix path in ganeti-rapi(8) (issue 116)
1660
  - Adjust error message when the ganeti user's .ssh directory is
1661
    missing
1662
  - Add same-node-check when changing the disk template to drbd
1663

    
1664

    
1665
Version 2.1.6
1666
-------------
1667

    
1668
*(Released Fri, 16 Jul 2010)*
1669

    
1670
Bugfixes only:
1671
  - Add an option to only select some reboot types during qa/burnin.
1672
    (on some hypervisors consequent reboots are not supported)
1673
  - Fix infrequent race condition in master failover. Sometimes the old
1674
    master ip address would be still detected as up for a short time
1675
    after it was removed, causing failover to fail.
1676
  - Decrease mlockall warnings when the ctypes module is missing. On
1677
    Python 2.4 we support running even if no ctypes module is installed,
1678
    but we were too verbose about this issue.
1679
  - Fix building on old distributions, on which man doesn't have a
1680
    --warnings option.
1681
  - Fix RAPI not to ignore the MAC address on instance creation
1682
  - Implement the old instance creation format in the RAPI client.
1683

    
1684

    
1685
Version 2.1.5
1686
-------------
1687

    
1688
*(Released Thu, 01 Jul 2010)*
1689

    
1690
A small bugfix release:
1691
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1692
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1693
    a non-existing option
1694
  - Fix instance create: the --force-variant option was ignored
1695
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
1696
  - Fix modify node storage with non-FQDN arguments
1697
  - Fix RAPI client to authenticate under Python 2.6 when used
1698
    for more than 5 requests needing authentication
1699
  - Fix gnt-instance modify -t (storage) giving a wrong error message
1700
    when converting a non-shutdown drbd instance to plain
1701

    
1702

    
1703
Version 2.1.4
1704
-------------
1705

    
1706
*(Released Fri, 18 Jun 2010)*
1707

    
1708
A small bugfix release:
1709

    
1710
  - Fix live migration of KVM instances started with older Ganeti
1711
    versions which had fewer hypervisor parameters
1712
  - Fix gnt-instance grow-disk on down instances
1713
  - Fix an error-reporting bug during instance migration
1714
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
1715
    silently ignoring broken ones
1716
  - Fix an RPC error reporting bug affecting, for example, RAPI client
1717
    users
1718
  - Fix bug triggered by different API version os-es on different nodes
1719
  - Fix a bug in instance startup with custom hvparams: OS level
1720
    parameters would fail to be applied.
1721
  - Fix the RAPI client under Python 2.6 (but more work is needed to
1722
    make it work completely well with OpenSSL)
1723
  - Fix handling of errors when resolving names from DNS
1724

    
1725

    
1726
Version 2.1.3
1727
-------------
1728

    
1729
*(Released Thu, 3 Jun 2010)*
1730

    
1731
A medium sized development cycle. Some new features, and some
1732
fixes/small improvements/cleanups.
1733

    
1734
Significant features
1735
~~~~~~~~~~~~~~~~~~~~
1736

    
1737
The node deamon now tries to mlock itself into memory, unless the
1738
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1739
its logs, and falls back to console logging. This allows emergency
1740
features such as ``gnt-node powercycle`` to work even in the event of a
1741
broken node disk (tested offlining the disk hosting the node's
1742
filesystem and dropping its memory caches; don't try this at home)
1743

    
1744
KVM: add vhost-net acceleration support. It can be tested with a new
1745
enough version of the kernel and of qemu-kvm.
1746

    
1747
KVM: Add instance chrooting feature. If you use privilege dropping for
1748
your VMs you can also now force them to chroot to an empty directory,
1749
before starting the emulated guest.
1750

    
1751
KVM: Add maximum migration bandwith and maximum downtime tweaking
1752
support (requires a new-enough version of qemu-kvm).
1753

    
1754
Cluster verify will now warn if the master node doesn't have the master
1755
ip configured on it.
1756

    
1757
Add a new (incompatible) instance creation request format to RAPI which
1758
supports all parameters (previously only a subset was supported, and it
1759
wasn't possible to extend the old format to accomodate all the new
1760
features. The old format is still supported, and a client can check for
1761
this feature, before using it, by checking for its presence in the
1762
``features`` RAPI resource.
1763

    
1764
Now with ancient latin support. Try it passing the ``--roman`` option to
1765
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1766
(requires the python-roman module to be installed, in order to work).
1767

    
1768
Other changes
1769
~~~~~~~~~~~~~
1770

    
1771
As usual many internal code refactorings, documentation updates, and
1772
such. Among others:
1773

    
1774
  - Lots of improvements and cleanups to the experimental Remote API
1775
    (RAPI) client library.
1776
  - A new unit test suite for the core daemon libraries.
1777
  - A fix to creating missing directories makes sure the umask is not
1778
    applied anymore. This enforces the same directory permissions
1779
    everywhere.
1780
  - Better handling terminating daemons with ctrl+c (used when running
1781
    them in debugging mode).
1782
  - Fix a race condition in live migrating a KVM instance, when stat()
1783
    on the old proc status file returned EINVAL, which is an unexpected
1784
    value.
1785
  - Fixed manpage checking with newer man and utf-8 charachters. But now
1786
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1787

    
1788

    
1789
Version 2.1.2.1
1790
---------------
1791

    
1792
*(Released Fri, 7 May 2010)*
1793

    
1794
Fix a bug which prevented untagged KVM instances from starting.
1795

    
1796

    
1797
Version 2.1.2
1798
-------------
1799

    
1800
*(Released Fri, 7 May 2010)*
1801

    
1802
Another release with a long development cycle, during which many
1803
different features were added.
1804

    
1805
Significant features
1806
~~~~~~~~~~~~~~~~~~~~
1807

    
1808
The KVM hypervisor now can run the individual instances as non-root, to
1809
reduce the impact of a VM being hijacked due to bugs in the
1810
hypervisor. It is possible to run all instances as a single (non-root)
1811
user, to manually specify a user for each instance, or to dynamically
1812
allocate a user out of a cluster-wide pool to each instance, with the
1813
guarantee that no two instances will run under the same user ID on any
1814
given node.
1815

    
1816
An experimental RAPI client library, that can be used standalone
1817
(without the other Ganeti libraries), is provided in the source tree as
1818
``lib/rapi/client.py``. Note this client might change its interface in
1819
the future, as we iterate on its capabilities.
1820

    
1821
A new command, ``gnt-cluster renew-crypto`` has been added to easily
1822
replace the cluster's certificates and crypto keys. This might help in
1823
case they have been compromised, or have simply expired.
1824

    
1825
A new disk option for instance creation has been added that allows one
1826
to "adopt" currently existing logical volumes, with data
1827
preservation. This should allow easier migration to Ganeti from
1828
unmanaged (or managed via other software) instances.
1829

    
1830
Another disk improvement is the possibility to convert between redundant
1831
(DRBD) and plain (LVM) disk configuration for an instance. This should
1832
allow better scalability (starting with one node and growing the
1833
cluster, or shrinking a two-node cluster to one node).
1834

    
1835
A new feature that could help with automated node failovers has been
1836
implemented: if a node sees itself as offline (by querying the master
1837
candidates), it will try to shutdown (hard) all instances and any active
1838
DRBD devices. This reduces the risk of duplicate instances if an
1839
external script automatically failovers the instances on such nodes. To
1840
enable this, the cluster parameter ``maintain_node_health`` should be
1841
enabled; in the future this option (per the name) will enable other
1842
automatic maintenance features.
1843

    
1844
Instance export/import now will reuse the original instance
1845
specifications for all parameters; that means exporting an instance,
1846
deleting it and the importing it back should give an almost identical
1847
instance. Note that the default import behaviour has changed from
1848
before, where it created only one NIC; now it recreates the original
1849
number of NICs.
1850

    
1851
Cluster verify has added a few new checks: SSL certificates validity,
1852
/etc/hosts consistency across the cluster, etc.
1853

    
1854
Other changes
1855
~~~~~~~~~~~~~
1856

    
1857
As usual, many internal changes were done, documentation fixes,
1858
etc. Among others:
1859

    
1860
- Fixed cluster initialization with disabled cluster storage (regression
1861
  introduced in 2.1.1)
1862
- File-based storage supports growing the disks
1863
- Fixed behaviour of node role changes
1864
- Fixed cluster verify for some corner cases, plus a general rewrite of
1865
  cluster verify to allow future extension with more checks
1866
- Fixed log spamming by watcher and node daemon (regression introduced
1867
  in 2.1.1)
1868
- Fixed possible validation issues when changing the list of enabled
1869
  hypervisors
1870
- Fixed cleanup of /etc/hosts during node removal
1871
- Fixed RAPI response for invalid methods
1872
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1873
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
1874
  from ide disks, etc.)
1875
- Allow OS changes without re-installation (to record a changed OS
1876
  outside of Ganeti, or to allow OS renames)
1877
- Allow instance creation without OS installation (useful for example if
1878
  the OS will be installed manually, or restored from a backup not in
1879
  Ganeti format)
1880
- Implemented option to make cluster ``copyfile`` use the replication
1881
  network
1882
- Added list of enabled hypervisors to ssconf (possibly useful for
1883
  external scripts)
1884
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1885
  1.2 clusters
1886
- A partial form of node re-IP is possible via node readd, which now
1887
  allows changed node primary IP
1888
- Command line utilities now show an informational message if the job is
1889
  waiting for a lock
1890
- The logs of the master daemon now show the PID/UID/GID of the
1891
  connected client
1892

    
1893

    
1894
Version 2.1.1
1895
-------------
1896

    
1897
*(Released Fri, 12 Mar 2010)*
1898

    
1899
During the 2.1.0 long release candidate cycle, a lot of improvements and
1900
changes have accumulated with were released later as 2.1.1.
1901

    
1902
Major changes
1903
~~~~~~~~~~~~~
1904

    
1905
The node evacuate command (``gnt-node evacuate``) was significantly
1906
rewritten, and as such the IAllocator protocol was changed - a new
1907
request type has been added. This unfortunate change during a stable
1908
series is designed to improve performance of node evacuations; on
1909
clusters with more than about five nodes and which are well-balanced,
1910
evacuation should proceed in parallel for all instances of the node
1911
being evacuated. As such, any existing IAllocator scripts need to be
1912
updated, otherwise the above command will fail due to the unknown
1913
request. The provided "dumb" allocator has not been updated; but the
1914
ganeti-htools package supports the new protocol since version 0.2.4.
1915

    
1916
Another important change is increased validation of node and instance
1917
names. This might create problems in special cases, if invalid host
1918
names are being used.
1919

    
1920
Also, a new layer of hypervisor parameters has been added, that sits at
1921
OS level between the cluster defaults and the instance ones. This allows
1922
customisation of virtualization parameters depending on the installed
1923
OS. For example instances with OS 'X' may have a different KVM kernel
1924
(or any other parameter) than the cluster defaults. This is intended to
1925
help managing a multiple OSes on the same cluster, without manual
1926
modification of each instance's parameters.
1927

    
1928
A tool for merging clusters, ``cluster-merge``, has been added in the
1929
tools sub-directory.
1930

    
1931
Bug fixes
1932
~~~~~~~~~
1933

    
1934
- Improved the int/float conversions that should make the code more
1935
  robust in face of errors from the node daemons
1936
- Fixed the remove node code in case of internal configuration errors
1937
- Fixed the node daemon behaviour in face of inconsistent queue
1938
  directory (e.g. read-only file-system where we can't open the files
1939
  read-write, etc.)
1940
- Fixed the behaviour of gnt-node modify for master candidate demotion;
1941
  now it either aborts cleanly or, if given the new "auto_promote"
1942
  parameter, will automatically promote other nodes as needed
1943
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1944
  completely prevent Ganeti from working
1945
- Fixed bug for instance export when not all disks were successfully
1946
  exported
1947
- Fixed behaviour of node add when the new node is slow in starting up
1948
  the node daemon
1949
- Fixed handling of signals in the LUXI client, which should improve
1950
  behaviour of command-line scripts
1951
- Added checks for invalid node/instance names in the configuration (now
1952
  flagged during cluster verify)
1953
- Fixed watcher behaviour for disk activation errors
1954
- Fixed two potentially endless loops in http library, which led to the
1955
  RAPI daemon hanging and consuming 100% CPU in some cases
1956
- Fixed bug in RAPI daemon related to hashed passwords
1957
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1958
  instances
1959
- Enhanced compatibility with non-Debian OSes, but not using absolute
1960
  path in some commands and allowing customisation of the ssh
1961
  configuration directory
1962
- Fixed possible future issue with new Python versions by abiding to the
1963
  proper use of ``__slots__`` attribute on classes
1964
- Added checks that should prevent directory traversal attacks
1965
- Many documentation fixes based on feedback from users
1966

    
1967
New features
1968
~~~~~~~~~~~~
1969

    
1970
- Added an "early_release" more for instance replace disks and node
1971
  evacuate, where we release locks earlier and thus allow higher
1972
  parallelism within the cluster
1973
- Added watcher hooks, intended to allow the watcher to restart other
1974
  daemons (e.g. from the ganeti-nbma project), but they can be used of
1975
  course for any other purpose
1976
- Added a compile-time disable for DRBD barriers, to increase
1977
  performance if the administrator trusts the power supply or the
1978
  storage system to not lose writes
1979
- Added the option of using syslog for logging instead of, or in
1980
  addition to, Ganeti's own log files
1981
- Removed boot restriction for paravirtual NICs for KVM, recent versions
1982
  can indeed boot from a paravirtual NIC
1983
- Added a generic debug level for many operations; while this is not
1984
  used widely yet, it allows one to pass the debug value all the way to
1985
  the OS scripts
1986
- Enhanced the hooks environment for instance moves (failovers,
1987
  migrations) where the primary/secondary nodes changed during the
1988
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1989
- Enhanced data validations for many user-supplied values; one important
1990
  item is the restrictions imposed on instance and node names, which
1991
  might reject some (invalid) host names
1992
- Add a configure-time option to disable file-based storage, if it's not
1993
  needed; this allows greater security separation between the master
1994
  node and the other nodes from the point of view of the inter-node RPC
1995
  protocol
1996
- Added user notification in interactive tools if job is waiting in the
1997
  job queue or trying to acquire locks
1998
- Added log messages when a job is waiting for locks
1999
- Added filtering by node tags in instance operations which admit
2000
  multiple instances (start, stop, reboot, reinstall)
2001
- Added a new tool for cluster mergers, ``cluster-merge``
2002
- Parameters from command line which are of the form ``a=b,c=d`` can now
2003
  use backslash escapes to pass in values which contain commas,
2004
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2005
  ``b,c``
2006
- For KVM, the instance name is the first parameter passed to KVM, so
2007
  that it's more visible in the process list
2008

    
2009

    
2010
Version 2.1.0
2011
-------------
2012

    
2013
*(Released Tue, 2 Mar 2010)*
2014

    
2015
Ganeti 2.1 brings many improvements with it. Major changes:
2016

    
2017
- Added infrastructure to ease automated disk repairs
2018
- Added new daemon to export configuration data in a cheaper way than
2019
  using the remote API
2020
- Instance NICs can now be routed instead of being associated with a
2021
  networking bridge
2022
- Improved job locking logic to reduce impact of jobs acquiring multiple
2023
  locks waiting for other long-running jobs
2024

    
2025
In-depth implementation details can be found in the Ganeti 2.1 design
2026
document.
2027

    
2028
Details
2029
~~~~~~~
2030

    
2031
- Added chroot hypervisor
2032
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2033
  ``device_model``)
2034
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2035
  ``bootloader_path`` and ``bootloader_args``)
2036
- Added the ``use_localtime`` option for the xen-hvm and kvm
2037
  hypervisors, and the default value for this has changed to false (in
2038
  2.0 xen-hvm always enabled it)
2039
- Added luxi call to submit multiple jobs in one go
2040
- Added cluster initialization option to not modify ``/etc/hosts``
2041
  file on nodes
2042
- Added network interface parameters
2043
- Added dry run mode to some LUs
2044
- Added RAPI resources:
2045

    
2046
  - ``/2/instances/[instance_name]/info``
2047
  - ``/2/instances/[instance_name]/replace-disks``
2048
  - ``/2/nodes/[node_name]/evacuate``
2049
  - ``/2/nodes/[node_name]/migrate``
2050
  - ``/2/nodes/[node_name]/role``
2051
  - ``/2/nodes/[node_name]/storage``
2052
  - ``/2/nodes/[node_name]/storage/modify``
2053
  - ``/2/nodes/[node_name]/storage/repair``
2054

    
2055
- Added OpCodes to evacuate or migrate all instances on a node
2056
- Added new command to list storage elements on nodes (``gnt-node
2057
  list-storage``) and modify them (``gnt-node modify-storage``)
2058
- Added new ssconf files with master candidate IP address
2059
  (``ssconf_master_candidates_ips``), node primary IP address
2060
  (``ssconf_node_primary_ips``) and node secondary IP address
2061
  (``ssconf_node_secondary_ips``)
2062
- Added ``ganeti-confd`` and a client library to query the Ganeti
2063
  configuration via UDP
2064
- Added ability to run hooks after cluster initialization and before
2065
  cluster destruction
2066
- Added automatic mode for disk replace (``gnt-instance replace-disks
2067
  --auto``)
2068
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2069
  after catastrophic data-loss
2070
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2071
  groups
2072
- Added ``gnt-instance move`` command to move instances
2073
- Added ``gnt-cluster watcher`` command to control watcher
2074
- Added ``gnt-node powercycle`` command to powercycle nodes
2075
- Added new job status field ``lock_status``
2076
- Added parseable error codes to cluster verification (``gnt-cluster
2077
  verify --error-codes``) and made output less verbose (use
2078
  ``--verbose`` to restore previous behaviour)
2079
- Added UUIDs to the main config entities (cluster, nodes, instances)
2080
- Added support for OS variants
2081
- Added support for hashed passwords in the Ganeti remote API users file
2082
  (``rapi_users``)
2083
- Added option to specify maximum timeout on instance shutdown
2084
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2085
- Added new helper script to start and stop Ganeti daemons
2086
  (``daemon-util``), with the intent to reduce the work necessary to
2087
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2088
  from one place
2089
- Added more unittests
2090
- Fixed critical bug in ganeti-masterd startup
2091
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2092
  now customisable at the cluster level for both the KVM and Xen
2093
  hypervisors using the new ``migration_port`` parameter
2094
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2095
  reinstalling an instance
2096
- Allowed ``@`` in tag names
2097
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2098
- Many documentation updates
2099
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2100
- ``gnt-instance reinstall`` can now reinstall multiple instances
2101
- Updated many command line parameters
2102
- Introduced new OS API version 15
2103
- No longer support a default hypervisor
2104
- Treat virtual LVs as inexistent
2105
- Improved job locking logic to reduce lock contention
2106
- Match instance and node names case insensitively
2107
- Reimplemented bash completion script to be more complete
2108
- Improved burnin
2109

    
2110

    
2111
Version 2.0.6
2112
-------------
2113

    
2114
*(Released Thu, 4 Feb 2010)*
2115

    
2116
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2117
- Fix a string formatting bug
2118
- Improve safety of the code in some error paths
2119
- Improve data validation in the master of values returned from nodes
2120

    
2121

    
2122
Version 2.0.5
2123
-------------
2124

    
2125
*(Released Thu, 17 Dec 2009)*
2126

    
2127
- Fix security issue due to missing validation of iallocator names; this
2128
  allows local and remote execution of arbitrary executables
2129
- Fix failure of gnt-node list during instance removal
2130
- Ship the RAPI documentation in the archive
2131

    
2132

    
2133
Version 2.0.4
2134
-------------
2135

    
2136
*(Released Wed, 30 Sep 2009)*
2137

    
2138
- Fixed many wrong messages
2139
- Fixed a few bugs related to the locking library
2140
- Fixed MAC checking at instance creation time
2141
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2142
- Fixed a few issues related to signal handling in both daemons and
2143
  scripts
2144
- Fixed the example startup script provided
2145
- Fixed insserv dependencies in the example startup script (patch from
2146
  Debian)
2147
- Fixed handling of drained nodes in the iallocator framework
2148
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2149
  #528618)
2150
- Fixed error related to invalid job IDs in job polling
2151
- Fixed job/opcode persistence on unclean master shutdown
2152
- Fixed handling of partial job processing after unclean master
2153
  shutdown
2154
- Fixed error reporting from LUs, previously all errors were converted
2155
  into execution errors
2156
- Fixed error reporting from burnin
2157
- Decreased significantly the memory usage of the job queue
2158
- Optimised slightly multi-job submission
2159
- Optimised slightly opcode loading
2160
- Backported the multi-job submit framework from the development
2161
  branch; multi-instance start and stop should be faster
2162
- Added script to clean archived jobs after 21 days; this will reduce
2163
  the size of the queue directory
2164
- Added some extra checks in disk size tracking
2165
- Added an example ethers hook script
2166
- Added a cluster parameter that prevents Ganeti from modifying of
2167
  /etc/hosts
2168
- Added more node information to RAPI responses
2169
- Added a ``gnt-job watch`` command that allows following the ouput of a
2170
  job
2171
- Added a bind-address option to ganeti-rapi
2172
- Added more checks to the configuration verify
2173
- Enhanced the burnin script such that some operations can be retried
2174
  automatically
2175
- Converted instance reinstall to multi-instance model
2176

    
2177

    
2178
Version 2.0.3
2179
-------------
2180

    
2181
*(Released Fri, 7 Aug 2009)*
2182

    
2183
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2184
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2185
  instances have mismatched disk sizes in the configuration
2186
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2187
  any configuration mismatches for disk sizes
2188
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2189
  failover to work on two-node clusters
2190
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2191
  unusable
2192
- Fixed detection of OS script errors in ``gnt-backup export``
2193
- Fixed exit code of ``gnt-backup export``
2194

    
2195

    
2196
Version 2.0.2
2197
-------------
2198

    
2199
*(Released Fri, 17 Jul 2009)*
2200

    
2201
- Added experimental support for stripped logical volumes; this should
2202
  enhance performance but comes with a higher complexity in the block
2203
  device handling; stripping is only enabled when passing
2204
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2205
  affected even in the non-stripped mode
2206
- Improved resiliency against transient failures at the end of DRBD
2207
  resyncs, and in general of DRBD resync checks
2208
- Fixed a couple of issues with exports and snapshot errors
2209
- Fixed a couple of issues in instance listing
2210
- Added display of the disk size in ``gnt-instance info``
2211
- Fixed checking for valid OSes in instance creation
2212
- Fixed handling of the "vcpus" parameter in instance listing and in
2213
  general of invalid parameters
2214
- Fixed http server library, and thus RAPI, to handle invalid
2215
  username/password combinations correctly; this means that now they
2216
  report unauthorized for queries too, not only for modifications,
2217
  allowing earlier detect of configuration problems
2218
- Added a new "role" node list field, equivalent to the master/master
2219
  candidate/drained/offline flags combinations
2220
- Fixed cluster modify and changes of candidate pool size
2221
- Fixed cluster verify error messages for wrong files on regular nodes
2222
- Fixed a couple of issues with node demotion from master candidate role
2223
- Fixed node readd issues
2224
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2225
- Added a new ``--no-voting`` option for masterfailover to fix failover
2226
  on two-nodes clusters when the former master node is unreachable
2227
- Added instance reinstall over RAPI
2228

    
2229

    
2230
Version 2.0.1
2231
-------------
2232

    
2233
*(Released Tue, 16 Jun 2009)*
2234

    
2235
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2236
  allow re-adding the start in single-user option (regression from 1.2)
2237
- the watcher writes the instance status to a file, to allow monitoring
2238
  to report the instance status (from the master) based on cached
2239
  results of the watcher's queries; while this can get stale if the
2240
  watcher is being locked due to other work on the cluster, this is
2241
  still an improvement
2242
- the watcher now also restarts the node daemon and the rapi daemon if
2243
  they died
2244
- fixed the watcher to handle full and drained queue cases
2245
- hooks export more instance data in the environment, which helps if
2246
  hook scripts need to take action based on the instance's properties
2247
  (no longer need to query back into ganeti)
2248
- instance failovers when the instance is stopped do not check for free
2249
  RAM, so that failing over a stopped instance is possible in low memory
2250
  situations
2251
- rapi uses queries for tags instead of jobs (for less job traffic), and
2252
  for cluster tags it won't talk to masterd at all but read them from
2253
  ssconf
2254
- a couple of error handling fixes in RAPI
2255
- drbd handling: improved the error handling of inconsistent disks after
2256
  resync to reduce the frequency of "there are some degraded disks for
2257
  this instance" messages
2258
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2259
  error handling path called a wrong function name)
2260

    
2261

    
2262
Version 2.0.0
2263
-------------
2264

    
2265
*(Released Wed, 27 May 2009)*
2266

    
2267
- no changes from rc5
2268

    
2269

    
2270
Version 2.0 rc5
2271
---------------
2272

    
2273
*(Released Wed, 20 May 2009)*
2274

    
2275
- fix a couple of bugs (validation, argument checks)
2276
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2277
- some small improvements to RAPI and IAllocator
2278
- make watcher automatically start the master daemon if down
2279

    
2280

    
2281
Version 2.0 rc4
2282
---------------
2283

    
2284
*(Released Mon, 27 Apr 2009)*
2285

    
2286
- change the OS list to not require locks; this helps with big clusters
2287
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2288
  volume group is broken
2289
- ``gnt-instance info``, without any arguments, doesn't run for all
2290
  instances anymore; either pass ``--all`` or pass the desired
2291
  instances; this helps against mistakes on big clusters where listing
2292
  the information for all instances takes a long time
2293
- miscellaneous doc and man pages fixes
2294

    
2295

    
2296
Version 2.0 rc3
2297
---------------
2298

    
2299
*(Released Wed, 8 Apr 2009)*
2300

    
2301
- Change the internal locking model of some ``gnt-node`` commands, in
2302
  order to reduce contention (and blocking of master daemon) when
2303
  batching many creation/reinstall jobs
2304
- Fixes to Xen soft reboot
2305
- No longer build documentation at build time, instead distribute it in
2306
  the archive, in order to reduce the need for the whole docbook/rst
2307
  toolchains
2308

    
2309

    
2310
Version 2.0 rc2
2311
---------------
2312

    
2313
*(Released Fri, 27 Mar 2009)*
2314

    
2315
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2316
- Fix watcher startup sequence, improves the behaviour of busy clusters
2317
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2318
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2319
  bind address and other places
2320
- Some documentation fixes and updates
2321

    
2322

    
2323
Version 2.0 rc1
2324
---------------
2325

    
2326
*(Released Mon, 2 Mar 2009)*
2327

    
2328
- More documentation updates, now all docs should be more-or-less
2329
  up-to-date
2330
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2331
  etc.)
2332
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2333
  and KVM)
2334
- Fix an issue related to $libdir/run/ganeti and cluster creation
2335

    
2336

    
2337
Version 2.0 beta2
2338
-----------------
2339

    
2340
*(Released Thu, 19 Feb 2009)*
2341

    
2342
- Xen PVM and KVM have switched the default value for the instance root
2343
  disk to the first partition on the first drive, instead of the whole
2344
  drive; this means that the OS installation scripts must be changed
2345
  accordingly
2346
- Man pages have been updated
2347
- RAPI has been switched by default to HTTPS, and the exported functions
2348
  should all work correctly
2349
- RAPI v1 has been removed
2350
- Many improvements to the KVM hypervisor
2351
- Block device errors are now better reported
2352
- Many other bugfixes and small improvements
2353

    
2354

    
2355
Version 2.0 beta1
2356
-----------------
2357

    
2358
*(Released Mon, 26 Jan 2009)*
2359

    
2360
- Version 2 is a general rewrite of the code and therefore the
2361
  differences are too many to list, see the design document for 2.0 in
2362
  the ``doc/`` subdirectory for more details
2363
- In this beta version there is not yet a migration path from 1.2 (there
2364
  will be one in the final 2.0 release)
2365
- A few significant changes are:
2366

    
2367
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2368
    various ``gnt-*`` commands are just front-ends to it
2369
  - all the commands are entered into, and executed from a job queue,
2370
    see the ``gnt-job(8)`` manpage
2371
  - the RAPI daemon supports read-write operations, secured by basic
2372
    HTTP authentication on top of HTTPS
2373
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2374
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2375
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2376
  - DRBD devices are using statically allocated minor numbers, which
2377
    will be assigned to existing instances during the migration process
2378
  - there is support for both Xen PVM and Xen HVM instances running on
2379
    the same cluster
2380
  - KVM virtualization is supported too
2381
  - file-based storage has been implemented, which means that it is
2382
    possible to run the cluster without LVM and DRBD storage, for
2383
    example using a shared filesystem exported from shared storage (and
2384
    still have live migration)
2385

    
2386

    
2387
Version 1.2.7
2388
-------------
2389

    
2390
*(Released Tue, 13 Jan 2009)*
2391

    
2392
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2393
- Reuse the old instance mac address by default on instance import, if
2394
  the instance name is the same.
2395
- Handle situations in which the node info rpc returns incomplete
2396
  results (issue 46)
2397
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2398
- Improved version of batcher:
2399

    
2400
  - state file support
2401
  - instance mac address support
2402
  - support for HVM clusters/instances
2403

    
2404
- Add an option to show the number of cpu sockets and nodes in
2405
  ``gnt-node list``
2406
- Support OSes that handle more than one version of the OS api (but do
2407
  not change the current API in any other way)
2408
- Fix ``gnt-node migrate``
2409
- ``gnt-debug`` man page
2410
- Fixes various more typos and small issues
2411
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2412

    
2413

    
2414
Version 1.2.6
2415
-------------
2416

    
2417
*(Released Wed, 24 Sep 2008)*
2418

    
2419
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2420
  type of disk exported to fully virtualized instances.
2421
- provide access to the serial console of HVM instances
2422
- instance auto_balance flag, set by default. If turned off it will
2423
  avoid warnings on cluster verify if there is not enough memory to fail
2424
  over an instance. in the future it will prevent automatically failing
2425
  it over when we will support that.
2426
- batcher tool for instance creation, see ``tools/README.batcher``
2427
- ``gnt-instance reinstall --select-os`` to interactively select a new
2428
  operating system when reinstalling an instance.
2429
- when changing the memory amount on instance modify a check has been
2430
  added that the instance will be able to start. also warnings are
2431
  emitted if the instance will not be able to fail over, if auto_balance
2432
  is true.
2433
- documentation fixes
2434
- sync fields between ``gnt-instance list/modify/add/import``
2435
- fix a race condition in drbd when the sync speed was set after giving
2436
  the device a remote peer.
2437

    
2438

    
2439
Version 1.2.5
2440
-------------
2441

    
2442
*(Released Tue, 22 Jul 2008)*
2443

    
2444
- note: the allowed size and number of tags per object were reduced
2445
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2446
- fixed twisted 8.x compatibility
2447
- fixed ``gnt-instance replace-disks`` with iallocator
2448
- add TCP keepalives on twisted connections to detect restarted nodes
2449
- disk increase support, see ``gnt-instance grow-disk``
2450
- implement bulk node/instance query for RAPI
2451
- add tags in node/instance listing (optional)
2452
- experimental migration (and live migration) support, read the man page
2453
  for ``gnt-instance migrate``
2454
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2455
  has some small improvements in handling its state file
2456

    
2457

    
2458
Version 1.2.4
2459
-------------
2460

    
2461
*(Released Fri, 13 Jun 2008)*
2462

    
2463
- Experimental readonly, REST-based remote API implementation;
2464
  automatically started on master node, TCP port 5080, if enabled by
2465
  ``--enable-rapi`` parameter to configure script.
2466
- Instance allocator support. Add and import instance accept a
2467
  ``--iallocator`` parameter, and call that instance allocator to decide
2468
  which node to use for the instance. The iallocator document describes
2469
  what's expected from an allocator script.
2470
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2471
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2472
  node is lost there is still enough memory to fail over the instances
2473
  that reside on it.
2474
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2475
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2476
  hooks will run, and their output, if any, will be displayed. Any
2477
  failing hook will make the verification return an error value.
2478
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2479
  primary and secondary interfaces
2480
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2481
  that is still declared as part of the cluster and has failed.
2482
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2483
  specifying output fields, that just adds the chosen fields to the
2484
  default ones.
2485
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2486
  export from the filesystem.
2487
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2488
  have been added. Using them you can enable/disable acpi and pae
2489
  support, and specify a path for a cd image to be exported to the
2490
  instance. These parameters as the name suggest only work on HVM
2491
  clusters.
2492
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2493
  PAE support will be set to the previously hardcoded values, but the
2494
  (previously hardcoded) path to the CDROM ISO image will be unset and
2495
  if required, needs to be set manually with ``gnt-instance modify``
2496
  after the upgrade.
2497
- The address to which an instance's VNC console is bound is now
2498
  selectable per-instance, rather than being cluster wide. Of course
2499
  this only applies to instances controlled via VNC, so currently just
2500
  applies to HVM clusters.
2501

    
2502

    
2503
Version 1.2.3
2504
-------------
2505

    
2506
*(Released Mon, 18 Feb 2008)*
2507

    
2508
- more tweaks to the disk activation code (especially helpful for DRBD)
2509
- change the default ``gnt-instance list`` output format, now there is
2510
  one combined status field (see the manpage for the exact values this
2511
  field will have)
2512
- some more fixes for the mac export to hooks change
2513
- make Ganeti not break with DRBD 8.2.x (which changed the version
2514
  format in ``/proc/drbd``) (issue 24)
2515
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2516
  template, allowing migration from DRBD0.7+MD to DRBD8
2517

    
2518

    
2519
Version 1.2.2
2520
-------------
2521

    
2522
*(Released Wed, 30 Jan 2008)*
2523

    
2524
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2525
  support (issue 23)
2526
- add command aliases infrastructure and a few aliases
2527
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2528
  man pages and the ``--help`` option of ``gnt-node
2529
  list``/``gnt-instance list``
2530
- fix ``gnt-backup list`` with down nodes (issue 21)
2531
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2532
- fix the dist archive and add a check for including svn/git files in
2533
  the future
2534
- some developer-related changes: improve the burnin and the QA suite,
2535
  add an upload script for testing during development
2536

    
2537

    
2538
Version 1.2.1
2539
-------------
2540

    
2541
*(Released Wed, 16 Jan 2008)*
2542

    
2543
- experimental HVM support, read the install document, section
2544
  "Initializing the cluster"
2545
- allow for the PVM hypervisor per-instance kernel and initrd paths
2546
- add a new command ``gnt-cluster verify-disks`` which uses a new
2547
  algorithm to improve the reconnection of the DRBD pairs if the device
2548
  on the secondary node has gone away
2549
- make logical volume code auto-activate LVs at disk activation time
2550
- slightly improve the speed of activating disks
2551
- allow specification of the MAC address at instance creation time, and
2552
  changing it later via ``gnt-instance modify``
2553
- fix handling of external commands that generate lots of output on
2554
  stderr
2555
- update documentation with regard to minimum version of DRBD8 supported
2556

    
2557

    
2558
Version 1.2.0
2559
-------------
2560

    
2561
*(Released Tue, 4 Dec 2007)*
2562

    
2563
- Log the ``xm create`` output to the node daemon log on failure (to
2564
  help diagnosing the error)
2565
- In debug mode, log all external commands output if failed to the logs
2566
- Change parsing of lvm commands to ignore stderr
2567

    
2568

    
2569
Version 1.2 beta3
2570
-----------------
2571

    
2572
*(Released Wed, 28 Nov 2007)*
2573

    
2574
- Another round of updates to the DRBD 8 code to deal with more failures
2575
  in the replace secondary node operation
2576
- Some more logging of failures in disk operations (lvm, drbd)
2577
- A few documentation updates
2578
- QA updates
2579

    
2580

    
2581
Version 1.2 beta2
2582
-----------------
2583

    
2584
*(Released Tue, 13 Nov 2007)*
2585

    
2586
- Change configuration file format from Python's Pickle to JSON.
2587
  Upgrading is possible using the cfgupgrade utility.
2588
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2589
  faster replace disks and is more stable (DRBD 8 has many improvements
2590
  compared to DRBD 0.7)
2591
- Added command line tags support (see man pages for ``gnt-instance``,
2592
  ``gnt-node``, ``gnt-cluster``)
2593
- Added instance rename support
2594
- Added multi-instance startup/shutdown
2595
- Added cluster rename support
2596
- Added ``gnt-node evacuate`` to simplify some node operations
2597
- Added instance reboot operation that can speedup reboot as compared to
2598
  stop and start
2599
- Soften the requirement that hostnames are in FQDN format
2600
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
2601
  reboots
2602
- Removed dependency on debian's patched fping that uses the
2603
  non-standard ``-S`` option
2604
- Now the OS definitions are searched for in multiple, configurable
2605
  paths (easier for distros to package)
2606
- Some changes to the hooks infrastructure (especially the new
2607
  post-configuration update hook)
2608
- Other small bugfixes
2609

    
2610
.. vim: set textwidth=72 syntax=rst :
2611
.. Local Variables:
2612
.. mode: rst
2613
.. fill-column: 72
2614
.. End: