Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 0865101d

History | View | Annotate | Download (107.7 kB)

1
News
2
====
3

    
4

    
5
Version 2.8.0 rc2
6
-----------------
7

    
8
*(Released Tue, 27 Aug 2013)*
9

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

    
13
- Instance policy can contain multiple instance specs, as described in
14
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
15
  <design-partitioned>`. As a consequence, it's not possible to partially change
16
  or override instance specs. Bounding specs (min and max) can be specified as a
17
  whole using the new option ``--ipolicy-bounds-specs``, while standard
18
  specs use the new option ``--ipolicy-std-specs``.
19
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
20
  gnt-instance is a valid YAML object.
21
- hail now honors network restrictions when allocating nodes. This led to an
22
  update of the IAllocator protocol. See the IAllocator documentation for
23
  details.
24
- confd now only answers static configuration request over the network. luxid
25
  was extracted, listens on the local LUXI socket and responds to live queries.
26
  This allows finer grained permissions if using separate users.
27

    
28
New features
29
~~~~~~~~~~~~
30

    
31
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
32
  to always require authentication, ``--require-authentication``. It can
33
  be specified in ``$sysconfdir/default/ganeti``.
34
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
35
  be used to manage the disk templates to be used by instances in the cluster.
36
  Initially, it will be set to a list that includes plain, drbd, if they were
37
  enabled by specifying a volume group name, and file and sharedfile, if those
38
  were enabled at configure time. Additionally, it will include all disk
39
  templates that are currently used by instances. The order of disk templates
40
  will be based on Ganeti's history of supporting them. In the future, the
41
  first entry of the list will be used as a default disk template on instance
42
  creation.
43
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
44
  configuration back to the previous stable version.
45
- Disk templates in group ipolicy can be restored to the default value.
46
- Initial support for diskless instances and virtual clusters in QA.
47
- More QA and unit tests for instance policies.
48
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
49
  describing why the opcode itself was executed.
50
- The monitoring daemon is now available. It allows users to query the cluster
51
  for obtaining information about the status of the system. The daemon is only
52
  responsible for providing the information over the network: the actual data
53
  gathering is performed by data collectors (currently, only the DRBD status
54
  collector is available).
55
- In order to help developers work on Ganeti, a new script
56
  (``devel/build_chroot``) is provided, for building a chroot that contains all
57
  the required development libraries and tools for compiling Ganeti on a Debian
58
  Squeeze system.
59
- A new tool, ``harep``, for performing self-repair and recreation of instances
60
  in Ganeti has been added.
61
- Split queries are enabled for tags, network, exports, cluster info, groups,
62
  jobs, nodes.
63
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
64
  It prints the command line to set the current policies, to ease
65
  changing them.
66
- Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
67
  devices for KVM virtio-net interfaces will get created with VNET_HDR
68
  (IFF_VNET_HDR) support. If set to false, it disables offloading on the
69
  virtio-net interfaces, which prevents host kernel tainting and log
70
  flooding, when dealing with broken or malicious virtio-net drivers.
71
  It's set to true by default.
72
- Instance failover now supports a ``--cleanup`` parameter for fixing previous
73
  failures.
74
- Support 'viridian' parameter in Xen HVM
75
- Support DSA SSH keys in bootstrap
76

    
77
New dependencies
78
~~~~~~~~~~~~~~~~
79
The following new dependencies have been added:
80

    
81
For Haskell:
82
- The ``curl`` library is not optional anymore for compiling the Haskell code.
83
- ``snap-server`` library (if monitoring is enabled).
84

    
85
For Python:
86
- The minimum Python version needed to run Ganeti is now 2.6.
87
- ``yaml`` library (only for running the QA).
88

    
89
Since 2.8.0 rc1
90
~~~~~~~~~~~~~~~
91

    
92
- Support 'viridian' parameter in Xen HVM (Issue 233)
93
- Include VCS version in ``gnt-cluster version``
94
- Support DSA SSH keys in bootstrap (Issue 338)
95
- Fix batch creation of instances
96
- Use FQDN to check master node status (Issue 551)
97
- Make the DRBD collector more failure-resilient
98

    
99

    
100
Version 2.8.0 rc1
101
-----------------
102

    
103
*(Released Fri, 2 Aug 2013)*
104

    
105
The first release candidate of the 2.8 series. Since 2.8.0 beta1:
106

    
107
- Fix upgrading/downgrading from 2.7
108
- Increase maximum RAPI message size
109
- Documentation updates
110
- Split ``confd`` between ``luxid`` and ``confd``
111
- Merge 2.7 series up to the 2.7.1 release
112
- Allow the ``modify_etc_hosts`` option to be changed
113
- Add better debugging for ``luxid`` queries
114
- Expose bulk parameter for GetJobs in RAPI client
115
- Expose missing ``network`` fields in RAPI
116
- Add some ``cluster verify`` tests
117
- Some unittest fixes
118
- Fix a malfunction in ``hspace``'s tiered allocation
119
- Fix query compatibility between haskell and python implementations
120
- Add the ``vnet_hdr`` HV parameter for KVM
121
- Add ``--cleanup`` to instance failover
122
- Change the connected groups format in ``gnt-network info`` output; it
123
  was previously displayed as a raw list by mistake. (Merged from 2.7)
124

    
125

    
126
Version 2.8.0 beta1
127
-------------------
128

    
129
*(Released Mon, 24 Jun 2013)*
130

    
131
This was the first beta release of the 2.8 series. All important changes
132
are listed in the latest 2.8 entry.
133

    
134

    
135
Version 2.7.1
136
-------------
137

    
138
*(Released Thu, 25 Jul 2013)*
139

    
140
- Add logrotate functionality in daemon-util
141
- Add logrotate example file
142
- Add missing fields to network queries over rapi
143
- Fix network object timestamps
144
- Add support for querying network timestamps
145
- Fix a typo in the example crontab
146
- Fix a documentation typo
147

    
148

    
149
Version 2.7.0
150
-------------
151

    
152
*(Released Thu, 04 Jul 2013)*
153

    
154
Incompatible/important changes
155
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
156

    
157
- Instance policies for disk size were documented to be on a per-disk
158
  basis, but hail applied them to the sum of all disks. This has been
159
  fixed.
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
- Compatibility with newer versions of rbd has been fixed
165
- ``gnt-instance batch-create`` has been changed to use the bulk create
166
  opcode from Ganeti. This lead to incompatible changes in the format of
167
  the JSON file. It's now not a custom dict anymore but a dict
168
  compatible with the ``OpInstanceCreate`` opcode.
169
- Parent directories for file storage need to be listed in
170
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
171
  write the file automatically based on old configuration values, but it
172
  can not distribute it across all nodes and the file contents should be
173
  verified. Use ``gnt-cluster copyfile
174
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
175
  upgraded. The reason for requiring this list of paths now is that
176
  before it would have been possible to inject new paths via RPC,
177
  allowing files to be created in arbitrary locations. The RPC protocol
178
  is protected using SSL/X.509 certificates, but as a design principle
179
  Ganeti does not permit arbitrary paths to be passed.
180
- The parsing of the variants file for OSes (see
181
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
182
  lines and comment lines (starting with ``#``) are ignored for better
183
  readability.
184
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
185
  longer available. ``gnt-node add`` now invokes a new tool on the
186
  destination node, named ``prepare-node-join``, to configure the SSH
187
  daemon. Paramiko is no longer necessary to configure nodes' SSH
188
  daemons via ``gnt-node add``.
189
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
190
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
191
  the flag is not reset after a master failover.
192
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
193
  is recommended. 2.6 will be mandatory from the 2.8 series.
194

    
195

    
196
New features
197
~~~~~~~~~~~~
198

    
199
- New network management functionality to support automatic allocation
200
  of IP addresses and managing of network parameters. See
201
  :manpage:`gnt-network(8)` for more details.
202
- New external storage backend, to allow managing arbitrary storage
203
  systems external to the cluster. See
204
  :manpage:`ganeti-extstorage-interface(7)`.
205
- New ``exclusive-storage`` node parameter added, restricted to
206
  nodegroup level. When it's set to true, physical disks are assigned in
207
  an exclusive fashion to instances, as documented in :doc:`Partitioned
208
  Ganeti <design-partitioned>`.  Currently, only instances using the
209
  ``plain`` disk template are supported.
210
- The KVM hypervisor has been updated with many new hypervisor
211
  parameters, including a generic one for passing arbitrary command line
212
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
213
  compatible up to qemu 1.4.
214
- A new tool, called ``mon-collector``, is the stand-alone executor of
215
  the data collectors for a monitoring system. As of this version, it
216
  just includes the DRBD data collector, that can be executed by calling
217
  ``mon-collector`` using the ``drbd`` parameter. See
218
  :manpage:`mon-collector(7)`.
219
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
220
  for RAPI users. It allows granting permissions to query for
221
  information to a specific user without giving
222
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
223
- A new tool named ``node-cleanup`` has been added. It cleans remains of
224
  a cluster from a machine by stopping all daemons, removing
225
  certificates and ssconf files. Unless the ``--no-backup`` option is
226
  given, copies of the certificates are made.
227
- Instance creations now support the use of opportunistic locking,
228
  potentially speeding up the (parallel) creation of multiple instances.
229
  This feature is currently only available via the :doc:`RAPI
230
  <rapi>` interface and when an instance allocator is used. If the
231
  ``opportunistic_locking`` parameter is set the opcode will try to
232
  acquire as many locks as possible, but will not wait for any locks
233
  held by other opcodes. If not enough resources can be found to
234
  allocate the instance, the temporary error code
235
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
236
  retried thereafter, with or without opportunistic locking.
237
- New experimental linux-ha resource scripts.
238
- Restricted-commands support: ganeti can now be asked (via command line
239
  or rapi) to perform commands on a node. These are passed via ganeti
240
  RPC rather than ssh. This functionality is restricted to commands
241
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
242
  reasons. The file is not copied automatically.
243

    
244

    
245
Misc changes
246
~~~~~~~~~~~~
247

    
248
- Diskless instances are now externally mirrored (Issue 237). This for
249
  now has only been tested in conjunction with explicit target nodes for
250
  migration/failover.
251
- Queries not needing locks or RPC access to the node can now be
252
  performed by the confd daemon, making them independent from jobs, and
253
  thus faster to execute. This is selectable at configure time.
254
- The functionality for allocating multiple instances at once has been
255
  overhauled and is now also available through :doc:`RAPI <rapi>`.
256

    
257
There are no significant changes from version 2.7.0~rc3.
258

    
259

    
260
Version 2.7.0 rc3
261
-----------------
262

    
263
*(Released Tue, 25 Jun 2013)*
264

    
265
- Fix permissions on the confd query socket (Issue 477)
266
- Fix permissions on the job archive dir (Issue 498)
267
- Fix handling of an internal exception in replace-disks (Issue 472)
268
- Fix gnt-node info handling of shortened names (Issue 497)
269
- Fix gnt-instance grow-disk when wiping is enabled
270
- Documentation improvements, and support for newer pandoc
271
- Fix hspace honoring ipolicy for disks (Issue 484)
272
- Improve handling of the ``kvm_extra`` HV parameter
273

    
274

    
275
Version 2.7.0 rc2
276
-----------------
277

    
278
*(Released Fri, 24 May 2013)*
279

    
280
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
281
  symlink.
282
- Disks added through ``gnt-instance modify`` or created through
283
  ``gnt-instance recreate-disks`` are wiped, if the
284
  ``prealloc_wipe_disks`` flag is set.
285
- If wiping newly created disks fails, the disks are removed. Also,
286
  partial failures in creating disks through ``gnt-instance modify``
287
  triggers a cleanup of the partially-created disks.
288
- Removing the master IP address doesn't fail if the address has been
289
  already removed.
290
- Fix ownership of the OS log dir
291
- Workaround missing SO_PEERCRED constant (Issue 191)
292

    
293

    
294
Version 2.7.0 rc1
295
-----------------
296

    
297
*(Released Fri, 3 May 2013)*
298

    
299
This was the first release candidate of the 2.7 series. Since beta3:
300

    
301
- Fix kvm compatibility with qemu 1.4 (Issue 389)
302
- Documentation updates (admin guide, upgrade notes, install
303
  instructions) (Issue 372)
304
- Fix gnt-group list nodes and instances count (Issue 436)
305
- Fix compilation without non-mandatory libraries (Issue 441)
306
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
307
- Make confd logging more verbose at INFO level (Issue 435)
308
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
309
- Fix failure path for instance storage type conversion (Issue 229)
310
- Update htools text backend documentation
311
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
312
- Disable inter-cluster instance move for file-based instances, because
313
  it is dependant on instance export, which is not supported for
314
  file-based instances. (Issue 414)
315
- Fix gnt-job crashes on non-ascii characters (Issue 427)
316
- Fix volume group checks on non-vm-capable nodes (Issue 432)
317

    
318

    
319
Version 2.7.0 beta3
320
-------------------
321

    
322
*(Released Mon, 22 Apr 2013)*
323

    
324
This was the third beta release of the 2.7 series. Since beta2:
325

    
326
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
327
- Fix data loss on wrong usage of ``gnt-instance move``
328
- Properly export errors in confd-based job queries
329
- Add ``users-setup`` tool
330
- Fix iallocator protocol to report 0 as a disk size for diskless
331
  instances. This avoids hail breaking when a diskless instance is
332
  present.
333
- Fix job queue directory permission problem that made confd job queries
334
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
335
  for access to archived jobs (Issue 406).
336
- Limit the sizes of networks supported by ``gnt-network`` to something
337
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
338
- Fix bugs in instance disk template conversion
339
- Fix GHC 7 compatibility
340
- Fix ``burnin`` install path (Issue 426).
341
- Allow very small disk grows (Issue 347).
342
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
343
  that noded doesn't import masterd code (Issue 419).
344
- Make sure the default metavg at cluster init is the same as the vg, if
345
  unspecified (Issue 358).
346
- Fix cleanup of partially created disks (part of Issue 416)
347

    
348

    
349
Version 2.7.0 beta2
350
-------------------
351

    
352
*(Released Tue, 2 Apr 2013)*
353

    
354
This was the second beta release of the 2.7 series. Since beta1:
355

    
356
- Networks no longer have a "type" slot, since this information was
357
  unused in Ganeti: instead of it tags should be used.
358
- The rapi client now has a ``target_node`` option to MigrateInstance.
359
- Fix early exit return code for hbal (Issue 386).
360
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
361
- Fix ``rbd showmapped`` output parsing (Issue 312).
362
- Networks are now referenced indexed by UUID, rather than name. This
363
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
364
  use.
365
- The OS environment now includes network information.
366
- Deleting of a network is now disallowed if any instance nic is using
367
  it, to prevent dangling references.
368
- External storage is now documented in man pages.
369
- The exclusive_storage flag can now only be set at nodegroup level.
370
- Hbal can now submit an explicit priority with its jobs.
371
- Many network related locking fixes.
372
- Bump up the required pylint version to 0.25.1.
373
- Fix the ``no_remember`` option in RAPI client.
374
- Many ipolicy related tests, qa, and fixes.
375
- Many documentation improvements and fixes.
376
- Fix building with ``--disable-file-storage``.
377
- Fix ``-q`` option in htools, which was broken if passed more than
378
  once.
379
- Some haskell/python interaction improvements and fixes.
380
- Fix iallocator in case of missing LVM storage.
381
- Fix confd config load in case of ``--no-lvm-storage``.
382
- The confd/query functionality is now mentioned in the security
383
  documentation.
384

    
385

    
386
Version 2.7.0 beta1
387
-------------------
388

    
389
*(Released Wed, 6 Feb 2013)*
390

    
391
This was the first beta release of the 2.7 series. All important changes
392
are listed in the latest 2.7 entry.
393

    
394

    
395
Version 2.6.2
396
-------------
397

    
398
*(Released Fri, 21 Dec 2012)*
399

    
400
Important behaviour change: hbal won't rebalance anymore instances which
401
have the ``auto_balance`` attribute set to false. This was the intention
402
all along, but until now it only skipped those from the N+1 memory
403
reservation (DRBD-specific).
404

    
405
A significant number of bug fixes in this release:
406

    
407
- Fixed disk adoption interaction with ipolicy checks.
408
- Fixed networking issues when instances are started, stopped or
409
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
410
- Fixed the warning in cluster verify for shared storage instances not
411
  being redundant.
412
- Fixed removal of storage directory on shared file storage (issue 262).
413
- Fixed validation of LVM volume group name in OpClusterSetParams
414
  (``gnt-cluster modify``) (issue 285).
415
- Fixed runtime memory increases (``gnt-instance modify -m``).
416
- Fixed live migration under Xen's ``xl`` mode.
417
- Fixed ``gnt-instance console`` with ``xl``.
418
- Fixed building with newer Haskell compiler/libraries.
419
- Fixed PID file writing in Haskell daemons (confd); this prevents
420
  restart issues if confd was launched manually (outside of
421
  ``daemon-util``) while another copy of it was running
422
- Fixed a type error when doing live migrations with KVM (issue 297) and
423
  the error messages for failing migrations have been improved.
424
- Fixed opcode validation for the out-of-band commands (``gnt-node
425
  power``).
426
- Fixed a type error when unsetting OS hypervisor parameters (issue
427
  311); now it's possible to unset all OS-specific hypervisor
428
  parameters.
429
- Fixed the ``dry-run`` mode for many operations: verification of
430
  results was over-zealous but didn't take into account the ``dry-run``
431
  operation, resulting in "wrong" failures.
432
- Fixed bash completion in ``gnt-job list`` when the job queue has
433
  hundreds of entries; especially with older ``bash`` versions, this
434
  results in significant CPU usage.
435

    
436
And lastly, a few other improvements have been made:
437

    
438
- Added option to force master-failover without voting (issue 282).
439
- Clarified error message on lock conflict (issue 287).
440
- Logging of newly submitted jobs has been improved (issue 290).
441
- Hostname checks have been made uniform between instance rename and
442
  create (issue 291).
443
- The ``--submit`` option is now supported by ``gnt-debug delay``.
444
- Shutting down the master daemon by sending SIGTERM now stops it from
445
  processing jobs waiting for locks; instead, those jobs will be started
446
  once again after the master daemon is started the next time (issue
447
  296).
448
- Support for Xen's ``xl`` program has been improved (besides the fixes
449
  above).
450
- Reduced logging noise in the Haskell confd daemon (only show one log
451
  entry for each config reload, instead of two).
452
- Several man page updates and typo fixes.
453

    
454

    
455
Version 2.6.1
456
-------------
457

    
458
*(Released Fri, 12 Oct 2012)*
459

    
460
A small bugfix release. Among the bugs fixed:
461

    
462
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
463
  made the command unusable.
464
- Commands that issue many jobs don't fail anymore just because some jobs
465
  take so long that other jobs are archived.
466
- Failures during ``gnt-instance reinstall`` are reflected by the exit
467
  status.
468
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
469
  DRBD is enabled.
470
- When ``always_failover`` is set, ``--allow-failover`` is not required
471
  in migrate commands anymore.
472
- ``bash_completion`` works even if extglob is disabled.
473
- Fixed bug with locks that made failover for RDB-based instances fail.
474
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
475
  a random node instead of one based on the allocator metric.
476
- Support for newer versions of pylint and pep8.
477
- Hail doesn't fail anymore when trying to add an instance of type
478
  ``file``, ``sharedfile`` or ``rbd``.
479
- Added new Makefile target to rebuild the whole distribution, so that
480
  all files are included.
481

    
482

    
483
Version 2.6.0
484
-------------
485

    
486
*(Released Fri, 27 Jul 2012)*
487

    
488

    
489
.. attention:: The ``LUXI`` protocol has been made more consistent
490
   regarding its handling of command arguments. This, however, leads to
491
   incompatibility issues with previous versions. Please ensure that you
492
   restart Ganeti daemons soon after the upgrade, otherwise most
493
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
494
   pausing/resuming the watcher, cancelling and archiving jobs, querying
495
   the cluster configuration) will fail.
496

    
497

    
498
New features
499
~~~~~~~~~~~~
500

    
501
Instance run status
502
+++++++++++++++++++
503

    
504
The current ``admin_up`` field, which used to denote whether an instance
505
should be running or not, has been removed. Instead, ``admin_state`` is
506
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
507

    
508
The rational behind this is that an instance being “down” can have
509
different meanings:
510

    
511
- it could be down during a reboot
512
- it could be temporarily be down for a reinstall
513
- or it could be down because it is deprecated and kept just for its
514
  disk
515

    
516
The previous Boolean state was making it difficult to do capacity
517
calculations: should Ganeti reserve memory for a down instance? Now, the
518
tri-state field makes it clear:
519

    
520
- in ``up`` and ``down`` state, all resources are reserved for the
521
  instance, and it can be at any time brought up if it is down
522
- in ``offline`` state, only disk space is reserved for it, but not
523
  memory or CPUs
524

    
525
The field can have an extra use: since the transition between ``up`` and
526
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
527
transition between ``offline`` and ``down`` is done via ``gnt-instance
528
modify``, it is possible to given different rights to users. For
529
example, owners of an instance could be allowed to start/stop it, but
530
not transition it out of the offline state.
531

    
532
Instance policies and specs
533
+++++++++++++++++++++++++++
534

    
535
In previous Ganeti versions, an instance creation request was not
536
limited on the minimum size and on the maximum size just by the cluster
537
resources. As such, any policy could be implemented only in third-party
538
clients (RAPI clients, or shell wrappers over ``gnt-*``
539
tools). Furthermore, calculating cluster capacity via ``hspace`` again
540
required external input with regards to instance sizes.
541

    
542
In order to improve these workflows and to allow for example better
543
per-node group differentiation, we introduced instance specs, which
544
allow declaring:
545

    
546
- minimum instance disk size, disk count, memory size, cpu count
547
- maximum values for the above metrics
548
- and “standard” values (used in ``hspace`` to calculate the standard
549
  sized instances)
550

    
551
The minimum/maximum values can be also customised at node-group level,
552
for example allowing more powerful hardware to support bigger instance
553
memory sizes.
554

    
555
Beside the instance specs, there are a few other settings belonging to
556
the instance policy framework. It is possible now to customise, per
557
cluster and node-group:
558

    
559
- the list of allowed disk templates
560
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
561
- the maximum ratio of instance to spindles (see below for more
562
  information) for local storage
563

    
564
All these together should allow all tools that talk to Ganeti to know
565
what are the ranges of allowed values for instances and the
566
over-subscription that is allowed.
567

    
568
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
569
instance configuration, and the physical CPU configuration from the
570
node. For the spindle ratios however, we didn't track before these
571
values, so new parameters have been added:
572

    
573
- a new node parameter ``spindle_count``, defaults to 1, customisable at
574
  node group or node level
575
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
576

    
577
Note that spindles in this context doesn't need to mean actual
578
mechanical hard-drives; it's just a relative number for both the node
579
I/O capacity and instance I/O consumption.
580

    
581
Instance migration behaviour
582
++++++++++++++++++++++++++++
583

    
584
While live-migration is in general desirable over failover, it is
585
possible that for some workloads it is actually worse, due to the
586
variable time of the “suspend” phase during live migration.
587

    
588
To allow the tools to work consistently over such instances (without
589
having to hard-code instance names), a new backend parameter
590
``always_failover`` has been added to control the migration/failover
591
behaviour. When set to True, all migration requests for an instance will
592
instead fall-back to failover.
593

    
594
Instance memory ballooning
595
++++++++++++++++++++++++++
596

    
597
Initial support for memory ballooning has been added. The memory for an
598
instance is no longer fixed (backend parameter ``memory``), but instead
599
can vary between minimum and maximum values (backend parameters
600
``minmem`` and ``maxmem``). Currently we only change an instance's
601
memory when:
602

    
603
- live migrating or failing over and instance and the target node
604
  doesn't have enough memory
605
- user requests changing the memory via ``gnt-instance modify
606
  --runtime-memory``
607

    
608
Instance CPU pinning
609
++++++++++++++++++++
610

    
611
In order to control the use of specific CPUs by instance, support for
612
controlling CPU pinning has been added for the Xen, HVM and LXC
613
hypervisors. This is controlled by a new hypervisor parameter
614
``cpu_mask``; details about possible values for this are in the
615
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
616
VCPU-to-CPU mapping) form will work well only when all nodes in your
617
cluster have the same amount of CPUs.
618

    
619
Disk parameters
620
+++++++++++++++
621

    
622
Another area in which Ganeti was not customisable were the parameters
623
used for storage configuration, e.g. how many stripes to use for LVM,
624
DRBD resync configuration, etc.
625

    
626
To improve this area, we've added disks parameters, which are
627
customisable at cluster and node group level, and which allow to
628
specify various parameters for disks (DRBD has the most parameters
629
currently), for example:
630

    
631
- DRBD resync algorithm and parameters (e.g. speed)
632
- the default VG for meta-data volumes for DRBD
633
- number of stripes for LVM (plain disk template)
634
- the RBD pool
635

    
636
These parameters can be modified via ``gnt-cluster modify -D …`` and
637
``gnt-group modify -D …``, and are used at either instance creation (in
638
case of LVM stripes, for example) or at disk “activation” time
639
(e.g. resync speed).
640

    
641
Rados block device support
642
++++++++++++++++++++++++++
643

    
644
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
645
denoted by the ``rbd`` disk template type. This is considered
646
experimental, feedback is welcome. For details on configuring it, see
647
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
648

    
649
Master IP setup
650
+++++++++++++++
651

    
652
The existing master IP functionality works well only in simple setups (a
653
single network shared by all nodes); however, if nodes belong to
654
different networks, then the ``/32`` setup and lack of routing
655
information is not enough.
656

    
657
To allow the master IP to function well in more complex cases, the
658
system was reworked as follows:
659

    
660
- a master IP netmask setting has been added
661
- the master IP activation/turn-down code was moved from the node daemon
662
  to a separate script
663
- whether to run the Ganeti-supplied master IP script or a user-supplied
664
  on is a ``gnt-cluster init`` setting
665

    
666
Details about the location of the standard and custom setup scripts are
667
in the man page :manpage:`gnt-cluster(8)`; for information about the
668
setup script protocol, look at the Ganeti-supplied script.
669

    
670
SPICE support
671
+++++++++++++
672

    
673
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
674
improved.
675

    
676
It is now possible to use TLS-protected connections, and when renewing
677
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
678
it is now possible to specify spice or spice CA certificates. Also, it
679
is possible to configure a password for SPICE sessions via the
680
hypervisor parameter ``spice_password_file``.
681

    
682
There are also new parameters to control the compression and streaming
683
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
684
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
685
for the spice parameters.
686

    
687
Lastly, it is now possible to see the SPICE connection information via
688
``gnt-instance console``.
689

    
690
OVF converter
691
+++++++++++++
692

    
693
A new tool (``tools/ovfconverter``) has been added that supports
694
conversion between Ganeti and the `Open Virtualization Format
695
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
696
from).
697

    
698
This relies on the ``qemu-img`` tool to convert the disk formats, so the
699
actual compatibility with other virtualization solutions depends on it.
700

    
701
Confd daemon changes
702
++++++++++++++++++++
703

    
704
The configuration query daemon (``ganeti-confd``) is now optional, and
705
has been rewritten in Haskell; whether to use the daemon at all, use the
706
Python (default) or the Haskell version is selectable at configure time
707
via the ``--enable-confd`` parameter, which can take one of the
708
``haskell``, ``python`` or ``no`` values. If not used, disabling the
709
daemon will result in a smaller footprint; for larger systems, we
710
welcome feedback on the Haskell version which might become the default
711
in future versions.
712

    
713
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
714
daemon running. The Python version doesn't implement the new call.
715

    
716

    
717
User interface changes
718
~~~~~~~~~~~~~~~~~~~~~~
719

    
720
We have replaced the ``--disks`` option of ``gnt-instance
721
replace-disks`` with a more flexible ``--disk`` option, which allows
722
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
723
disk size and mode can be changed upon recreation (via ``gnt-instance
724
recreate-disks``, which accepts the same ``--disk`` option).
725

    
726
As many people are used to a ``show`` command, we have added that as an
727
alias to ``info`` on all ``gnt-*`` commands.
728

    
729
The ``gnt-instance grow-disk`` command has a new mode in which it can
730
accept the target size of the disk, instead of the delta; this can be
731
more safe since two runs in absolute mode will be idempotent, and
732
sometimes it's also easier to specify the desired size directly.
733

    
734
Also the handling of instances with regard to offline secondaries has
735
been improved. Instance operations should not fail because one of it's
736
secondary nodes is offline, even though it's safe to proceed.
737

    
738
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
739
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
740
minors to instance name.
741

    
742
API changes
743
~~~~~~~~~~~
744

    
745
RAPI coverage has improved, with (for example) new resources for
746
recreate-disks, node power-cycle, etc.
747

    
748
Compatibility
749
~~~~~~~~~~~~~
750

    
751
There is partial support for ``xl`` in the Xen hypervisor; feedback is
752
welcome.
753

    
754
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
755
whether to still support Python 2.4 or move to Python 2.6 as minimum
756
required version.
757

    
758
Support for Fedora has been slightly improved; the provided example
759
init.d script should work better on it and the INSTALL file should
760
document the needed dependencies.
761

    
762
Internal changes
763
~~~~~~~~~~~~~~~~
764

    
765
The deprecated ``QueryLocks`` LUXI request has been removed. Use
766
``Query(what=QR_LOCK, ...)`` instead.
767

    
768
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
769
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
770
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
771
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
772
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
773

    
774
RAPI client: ``CertificateError`` now derives from
775
``GanetiApiError``. This should make it more easy to handle Ganeti
776
errors.
777

    
778
Deprecation warnings due to PyCrypto/paramiko import in
779
``tools/setup-ssh`` have been silenced, as usually they are safe; please
780
make sure to run an up-to-date paramiko version, if you use this tool.
781

    
782
The QA scripts now depend on Python 2.5 or above (the main code base
783
still works with Python 2.4).
784

    
785
The configuration file (``config.data``) is now written without
786
indentation for performance reasons; if you want to edit it, it can be
787
re-formatted via ``tools/fmtjson``.
788

    
789
A number of bugs has been fixed in the cluster merge tool.
790

    
791
``x509`` certification verification (used in import-export) has been
792
changed to allow the same clock skew as permitted by the cluster
793
verification. This will remove some rare but hard to diagnose errors in
794
import-export.
795

    
796

    
797
Version 2.6.0 rc4
798
-----------------
799

    
800
*(Released Thu, 19 Jul 2012)*
801

    
802
Very few changes from rc4 to the final release, only bugfixes:
803

    
804
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
805
  instance, fix CDROM booting for KVM instances)
806
- fixed node group modification of node parameters
807
- fixed issue in LUClusterVerifyGroup with multi-group clusters
808
- fixed generation of bash completion to ensure a stable ordering
809
- fixed a few typos
810

    
811

    
812
Version 2.6.0 rc3
813
-----------------
814

    
815
*(Released Fri, 13 Jul 2012)*
816

    
817
Third release candidate for 2.6. The following changes were done from
818
rc3 to rc4:
819

    
820
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
821
- Fixed an inconsistency in the LUXI protocol with the provided
822
  arguments (NOT backwards compatible)
823
- Fixed a bug with node groups ipolicy where ``min`` was greater than
824
  the cluster ``std`` value
825
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
826
  easier instance debugging on nodes (requires ``hconfd`` to work)
827

    
828

    
829
Version 2.6.0 rc2
830
-----------------
831

    
832
*(Released Tue, 03 Jul 2012)*
833

    
834
Second release candidate for 2.6. The following changes were done from
835
rc2 to rc3:
836

    
837
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
838
  master candidates
839
- Fixed a RAPI regression on missing beparams/memory
840
- Fixed redistribution of files on offline nodes
841
- Added possibility to run activate-disks even though secondaries are
842
  offline. With this change it relaxes also the strictness on some other
843
  commands which use activate disks internally:
844
  * ``gnt-instance start|reboot|rename|backup|export``
845
- Made it possible to remove safely an instance if its secondaries are
846
  offline
847
- Made it possible to reinstall even though secondaries are offline
848

    
849

    
850
Version 2.6.0 rc1
851
-----------------
852

    
853
*(Released Mon, 25 Jun 2012)*
854

    
855
First release candidate for 2.6. The following changes were done from
856
rc1 to rc2:
857

    
858
- Fixed bugs with disk parameters and ``rbd`` templates as well as
859
  ``instance_os_add``
860
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
861
  behaviour. It supports now the modify operation
862
- ``hcheck`` implemented to analyze cluster health and possibility of
863
  improving health by rebalance
864
- ``hbal`` has been improved in dealing with split instances
865

    
866

    
867
Version 2.6.0 beta2
868
-------------------
869

    
870
*(Released Mon, 11 Jun 2012)*
871

    
872
Second beta release of 2.6. The following changes were done from beta2
873
to rc1:
874

    
875
- Fixed ``daemon-util`` with non-root user models
876
- Fixed creation of plain instances with ``--no-wait-for-sync``
877
- Fix wrong iv_names when running ``cfgupgrade``
878
- Export more information in RAPI group queries
879
- Fixed bug when changing instance network interfaces
880
- Extended burnin to do NIC changes
881
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
882
- Changed default for DRBD barriers
883
- Fixed DRBD error reporting for syncer rate
884
- Verify the options on disk parameters
885

    
886
And of course various fixes to documentation and improved unittests and
887
QA.
888

    
889

    
890
Version 2.6.0 beta1
891
-------------------
892

    
893
*(Released Wed, 23 May 2012)*
894

    
895
First beta release of 2.6. The following changes were done from beta1 to
896
beta2:
897

    
898
- integrated patch for distributions without ``start-stop-daemon``
899
- adapted example init.d script to work on Fedora
900
- fixed log handling in Haskell daemons
901
- adapted checks in the watcher for pycurl linked against libnss
902
- add partial support for ``xl`` instead of ``xm`` for Xen
903
- fixed a type issue in cluster verification
904
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
905
  clusters)
906

    
907
Plus integrated fixes from the 2.5 branch:
908

    
909
- fixed ``kvm-ifup`` to use ``/bin/bash``
910
- fixed parallel build failures
911
- KVM live migration when using a custom keymap
912

    
913

    
914
Version 2.5.2
915
-------------
916

    
917
*(Released Tue, 24 Jul 2012)*
918

    
919
A small bugfix release, with no new features:
920

    
921
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
922
  different default shell (e.g. Debian, Ubuntu)
923
- fixed KVM startup and live migration with a custom keymap (fixes Issue
924
  243 and Debian bug #650664)
925
- fixed compatibility with KVM versions that don't support multiple boot
926
  devices (fixes Issue 230 and Debian bug #624256)
927

    
928
Additionally, a few fixes were done to the build system (fixed parallel
929
build failures) and to the unittests (fixed race condition in test for
930
FileID functions, and the default enable/disable mode for QA test is now
931
customisable).
932

    
933

    
934
Version 2.5.1
935
-------------
936

    
937
*(Released Fri, 11 May 2012)*
938

    
939
A small bugfix release.
940

    
941
The main issues solved are on the topic of compatibility with newer LVM
942
releases:
943

    
944
- fixed parsing of ``lv_attr`` field
945
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
946
  the ``--force`` flag is needed
947

    
948
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
949
to accept kernel 3.x too (was hardcoded to 2.6.*).
950

    
951
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
952
of hook results and that also made display of other errors suboptimal
953
was fixed; the code behaves now like 2.4 and earlier.
954

    
955
Another change in 2.5, the cleanup of the OS scripts environment, is too
956
aggressive: it removed even the ``PATH`` variable, which requires the OS
957
scripts to *always* need to export it. Since this is a bit too strict,
958
we now export a minimal PATH, the same that we export for hooks.
959

    
960
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
961
integrated into this release.
962

    
963
Finally, a few other miscellaneous changes were done (no new features,
964
just small improvements):
965

    
966
- Fix ``gnt-group --help`` display
967
- Fix hardcoded Xen kernel path
968
- Fix grow-disk handling of invalid units
969
- Update synopsis for ``gnt-cluster repair-disk-sizes``
970
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
971

    
972

    
973
Version 2.5.0
974
-------------
975

    
976
*(Released Thu, 12 Apr 2012)*
977

    
978
Incompatible/important changes and bugfixes
979
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
980

    
981
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
982
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
983
  to match the underlying LUXI interface.
984
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
985
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
986
  not maintain backwards-compatibility as the underlying operation
987
  changed in an incompatible way. The RAPI client can talk to old
988
  servers, but it needs to be told so as the return value changed.
989
- When creating file-based instances via RAPI, the ``file_driver``
990
  parameter no longer defaults to ``loop`` and must be specified.
991
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
992
  ``link`` instead.
993
- Support for the undocumented and deprecated RAPI instance creation
994
  request format version 0 has been dropped. Use version 1, supported
995
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
996
- Pyparsing 1.4.6 or above is required, see :doc:`installation
997
  documentation <install>`.
998
- The "cluster-verify" hooks are now executed per group by the
999
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1000
  if you just run ``gnt-cluster verify``, which generates one opcode per
1001
  group.
1002
- The environment as passed to the OS scripts is cleared, and thus no
1003
  environment variables defined in the node daemon's environment will be
1004
  inherited by the scripts.
1005
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1006
  deprecated.
1007
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
1008
  support operations involving multiple node groups.
1009
- Offline nodes are ignored when failing over an instance.
1010
- Support for KVM version 1.0, which changed the version reporting format
1011
  from 3 to 2 digits.
1012
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
1013
  removal.
1014
- ``Makefile`` is now compatible with Automake 1.11.2
1015
- Includes all bugfixes made in the 2.4 series
1016

    
1017
New features
1018
~~~~~~~~~~~~
1019

    
1020
- The ganeti-htools project has been merged into the ganeti-core source
1021
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
1022
- Implemented support for :doc:`shared storage <design-shared-storage>`.
1023
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1024
  GPT-style partition tables (requires `parted
1025
  <http://www.gnu.org/s/parted/>`_).
1026
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1027
- Allowed adding tags on instance creation.
1028
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1029
- Allow instances to be started in a paused state, enabling the user to
1030
  see the complete console output on boot using the console.
1031
- Added new hypervisor flag to control default reboot behaviour
1032
  (``reboot_behavior``).
1033
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1034
- Improved out-of-band management support:
1035

    
1036
  - Added ``gnt-node health`` command reporting the health status of
1037
    nodes.
1038
  - Added ``gnt-node power`` command to manage power status of nodes.
1039
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1040

    
1041
- Instance migration can fall back to failover if instance is not
1042
  running.
1043
- Filters can be used when listing nodes, instances, groups and locks;
1044
  see :manpage:`ganeti(7)` manpage.
1045
- Added post-execution status as variables to :doc:`hooks <hooks>`
1046
  environment.
1047
- Instance tags are exported/imported together with the instance.
1048
- When given an explicit job ID, ``gnt-job info`` will work for archived
1049
  jobs.
1050
- Jobs can define dependencies on other jobs (not yet supported via
1051
  RAPI or command line, but used by internal commands and usable via
1052
  LUXI).
1053

    
1054
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1055
    dependencies.
1056

    
1057
- Instance failover is now available as a RAPI resource
1058
  (``/2/instances/[instance_name]/failover``).
1059
- ``gnt-instance info`` defaults to static information if primary node
1060
  is offline.
1061
- Opcodes have a new ``comment`` attribute.
1062
- Added basic SPICE support to KVM hypervisor.
1063
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1064

    
1065
Node group improvements
1066
~~~~~~~~~~~~~~~~~~~~~~~
1067

    
1068
- ``gnt-cluster verify`` has been modified to check groups separately,
1069
  thereby improving performance.
1070
- Node group support has been added to ``gnt-cluster verify-disks``,
1071
  which now operates per node group.
1072
- Watcher has been changed to work better with node groups.
1073

    
1074
  - One process and state file per node group.
1075
  - Slow watcher in one group doesn't block other group's watcher.
1076

    
1077
- Added new command, ``gnt-group evacuate``, to move all instances in a
1078
  node group to other groups.
1079
- Added ``gnt-instance change-group`` to move an instance to another
1080
  node group.
1081
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1082
  per-group operations.
1083
- Node groups can be tagged.
1084
- Some operations switch from an exclusive to a shared lock as soon as
1085
  possible.
1086
- Instance's primary and secondary nodes' groups are now available as
1087
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1088
  and ``snodes.group.uuid``).
1089

    
1090
Misc
1091
~~~~
1092

    
1093
- Numerous updates to documentation and manpages.
1094

    
1095
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1096
    descriptions.
1097
  - Some opcode/job results are now also documented, see :doc:`RAPI
1098
    <rapi>`.
1099

    
1100
- A lockset's internal lock is now also visible in lock monitor.
1101
- Log messages from job queue workers now contain information about the
1102
  opcode they're processing.
1103
- ``gnt-instance console`` no longer requires the instance lock.
1104
- A short delay when waiting for job changes reduces the number of LUXI
1105
  requests significantly.
1106
- DRBD metadata volumes are overwritten with zeros during disk creation.
1107
- Out-of-band commands no longer acquire the cluster lock in exclusive
1108
  mode.
1109
- ``devel/upload`` now uses correct permissions for directories.
1110

    
1111

    
1112
Version 2.5.0 rc6
1113
-----------------
1114

    
1115
*(Released Fri, 23 Mar 2012)*
1116

    
1117
This was the sixth release candidate of the 2.5 series.
1118

    
1119

    
1120
Version 2.5.0 rc5
1121
-----------------
1122

    
1123
*(Released Mon, 9 Jan 2012)*
1124

    
1125
This was the fifth release candidate of the 2.5 series.
1126

    
1127

    
1128
Version 2.5.0 rc4
1129
-----------------
1130

    
1131
*(Released Thu, 27 Oct 2011)*
1132

    
1133
This was the fourth release candidate of the 2.5 series.
1134

    
1135

    
1136
Version 2.5.0 rc3
1137
-----------------
1138

    
1139
*(Released Wed, 26 Oct 2011)*
1140

    
1141
This was the third release candidate of the 2.5 series.
1142

    
1143

    
1144
Version 2.5.0 rc2
1145
-----------------
1146

    
1147
*(Released Tue, 18 Oct 2011)*
1148

    
1149
This was the second release candidate of the 2.5 series.
1150

    
1151

    
1152
Version 2.5.0 rc1
1153
-----------------
1154

    
1155
*(Released Tue, 4 Oct 2011)*
1156

    
1157
This was the first release candidate of the 2.5 series.
1158

    
1159

    
1160
Version 2.5.0 beta3
1161
-------------------
1162

    
1163
*(Released Wed, 31 Aug 2011)*
1164

    
1165
This was the third beta release of the 2.5 series.
1166

    
1167

    
1168
Version 2.5.0 beta2
1169
-------------------
1170

    
1171
*(Released Mon, 22 Aug 2011)*
1172

    
1173
This was the second beta release of the 2.5 series.
1174

    
1175

    
1176
Version 2.5.0 beta1
1177
-------------------
1178

    
1179
*(Released Fri, 12 Aug 2011)*
1180

    
1181
This was the first beta release of the 2.5 series.
1182

    
1183

    
1184
Version 2.4.5
1185
-------------
1186

    
1187
*(Released Thu, 27 Oct 2011)*
1188

    
1189
- Fixed bug when parsing command line parameter values ending in
1190
  backslash
1191
- Fixed assertion error after unclean master shutdown
1192
- Disable HTTP client pool for RPC, significantly reducing memory usage
1193
  of master daemon
1194
- Fixed queue archive creation with wrong permissions
1195

    
1196

    
1197
Version 2.4.4
1198
-------------
1199

    
1200
*(Released Tue, 23 Aug 2011)*
1201

    
1202
Small bug-fixes:
1203

    
1204
- Fixed documentation for importing with ``--src-dir`` option
1205
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1206
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1207

    
1208

    
1209
Version 2.4.3
1210
-------------
1211

    
1212
*(Released Fri, 5 Aug 2011)*
1213

    
1214
Many bug-fixes and a few small features:
1215

    
1216
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1217
  issues when you tried to add an instance with two MAC addresses in one
1218
  request
1219
- KVM: fixed per-instance stored UID value
1220
- KVM: configure bridged NICs at migration start
1221
- KVM: Fix a bug where instance will not start with never KVM versions
1222
  (>= 0.14)
1223
- Added OS search path to ``gnt-cluster info``
1224
- Fixed an issue with ``file_storage_dir`` where you were forced to
1225
  provide an absolute path, but the documentation states it is a
1226
  relative path, the documentation was right
1227
- Added a new parameter to instance stop/start called ``--no-remember``
1228
  that will make the state change to not be remembered
1229
- Implemented ``no_remember`` at RAPI level
1230
- Improved the documentation
1231
- Node evacuation: don't call IAllocator if node is already empty
1232
- Fixed bug in DRBD8 replace disks on current nodes
1233
- Fixed bug in recreate-disks for DRBD instances
1234
- Moved assertion checking locks in ``gnt-instance replace-disks``
1235
  causing it to abort with not owning the right locks for some situation
1236
- Job queue: Fixed potential race condition when cancelling queued jobs
1237
- Fixed off-by-one bug in job serial generation
1238
- ``gnt-node volumes``: Fix instance names
1239
- Fixed aliases in bash completion
1240
- Fixed a bug in reopening log files after being sent a SIGHUP
1241
- Added a flag to burnin to allow specifying VCPU count
1242
- Bugfixes to non-root Ganeti configuration
1243

    
1244

    
1245
Version 2.4.2
1246
-------------
1247

    
1248
*(Released Thu, 12 May 2011)*
1249

    
1250
Many bug-fixes and a few new small features:
1251

    
1252
- Fixed a bug related to log opening failures
1253
- Fixed a bug in instance listing with orphan instances
1254
- Fixed a bug which prevented resetting the cluster-level node parameter
1255
  ``oob_program`` to the default
1256
- Many fixes related to the ``cluster-merge`` tool
1257
- Fixed a race condition in the lock monitor, which caused failures
1258
  during (at least) creation of many instances in parallel
1259
- Improved output for gnt-job info
1260
- Removed the quiet flag on some ssh calls which prevented debugging
1261
  failures
1262
- Improved the N+1 failure messages in cluster verify by actually
1263
  showing the memory values (needed and available)
1264
- Increased lock attempt timeouts so that when executing long operations
1265
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1266
  too early and thus prevent the use of the 'fair' mechanism
1267
- Changed instance query data (``gnt-instance info``) to not acquire
1268
  locks unless needed, thus allowing its use on locked instance if only
1269
  static information is asked for
1270
- Improved behaviour with filesystems that do not support rename on an
1271
  opened file
1272
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1273
  kept locks on all nodes during the wipe, which is unneeded
1274
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1275
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1276
  10GiB) which caused the wipe to fail right at the end in some cases
1277
- Fixed master IP activation when doing master failover with no-voting
1278
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1279
  the master node itself
1280
- Fixed potential data-loss in under disk full conditions, where Ganeti
1281
  wouldn't check correctly the return code and would consider
1282
  partially-written files 'correct'
1283
- Fixed bug related to multiple VGs and DRBD disk replacing
1284
- Added new disk parameter ``metavg`` that allows placement of the meta
1285
  device for DRBD in a different volume group
1286
- Fixed error handling in the node daemon when the system libc doesn't
1287
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1288
- Fixed lock release during replace-disks, which kept cluster-wide locks
1289
  when doing disk replaces with an iallocator script
1290
- Added check for missing bridges in cluster verify
1291
- Handle EPIPE errors while writing to the terminal better, so that
1292
  piping the output to e.g. ``less`` doesn't cause a backtrace
1293
- Fixed rare case where a ^C during Luxi calls could have been
1294
  interpreted as server errors, instead of simply terminating
1295
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1296
  assign-nodes``)
1297
- Added a few more parameters to the KVM hypervisor, allowing a second
1298
  CDROM, custom disk type for CDROMs and a floppy image
1299
- Removed redundant message in instance rename when the name is given
1300
  already as a FQDN
1301
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1302
  disks on new nodes, allowing recreation when the original instance
1303
  nodes are completely gone
1304
- Added option when converting disk templates to DRBD to skip waiting
1305
  for the resync, in order to make the instance available sooner
1306
- Added two new variables to the OS scripts environment (containing the
1307
  instance's nodes)
1308
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1309
  to allow use of ``pvgrub`` as bootloader
1310
- Changed the instance memory modifications to only check out-of-memory
1311
  conditions on memory increases, and turned the secondary node warnings
1312
  into errors (they can still be overridden via ``--force``)
1313
- Fixed the handling of a corner case when the Python installation gets
1314
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1315
  to execute a command that doesn't exist
1316
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1317
  node of the instance returned failures during instance shutdown; this
1318
  adds the option ``--ignore-consistency`` to gnt-instance move
1319

    
1320
And as usual, various improvements to the error messages, documentation
1321
and man pages.
1322

    
1323

    
1324
Version 2.4.1
1325
-------------
1326

    
1327
*(Released Wed, 09 Mar 2011)*
1328

    
1329
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1330
the RAPI users file if run twice (even with ``--dry-run``).
1331

    
1332
The release fixes that bug (nothing else changed).
1333

    
1334

    
1335
Version 2.4.0
1336
-------------
1337

    
1338
*(Released Mon, 07 Mar 2011)*
1339

    
1340
Final 2.4.0 release. Just a few small fixes:
1341

    
1342
- Fixed RAPI node evacuate
1343
- Fixed the kvm-ifup script
1344
- Fixed internal error handling for special job cases
1345
- Updated man page to specify the escaping feature for options
1346

    
1347

    
1348
Version 2.4.0 rc3
1349
-----------------
1350

    
1351
*(Released Mon, 28 Feb 2011)*
1352

    
1353
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1354
that this feature wiped the disks of the wrong instance, leading to loss
1355
of data.
1356

    
1357
Other changes:
1358

    
1359
- Fixed title of query field containing instance name
1360
- Expanded the glossary in the documentation
1361
- Fixed one unittest (internal issue)
1362

    
1363

    
1364
Version 2.4.0 rc2
1365
-----------------
1366

    
1367
*(Released Mon, 21 Feb 2011)*
1368

    
1369
A number of bug fixes plus just a couple functionality changes.
1370

    
1371
On the user-visible side, the ``gnt-* list`` command output has changed
1372
with respect to "special" field states. The current rc1 style of display
1373
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1374
the default output mode special fields states are displayed as follows:
1375

    
1376
- Offline resource: ``*``
1377
- Unavailable/not applicable: ``-``
1378
- Data missing (RPC failure): ``?``
1379
- Unknown field: ``??``
1380

    
1381
Another user-visible change is the addition of ``--force-join`` to
1382
``gnt-node add``.
1383

    
1384
As for bug fixes:
1385

    
1386
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1387
- Fixed regression in RAPI/instance reinstall where all parameters were
1388
  required (instead of optional)
1389
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1390
- Fixed iallocator usage (offline nodes were not considered offline)
1391
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1392
- Fixed hypervisor and OS parameter validation with respect to
1393
  non-vm_capable nodes
1394
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1395
  cosmetic)
1396
- Fixed ``tools/listrunner`` with respect to agent-based usage
1397

    
1398

    
1399
Version 2.4.0 rc1
1400
-----------------
1401

    
1402
*(Released Fri,  4 Feb 2011)*
1403

    
1404
Many changes and fixes since the beta1 release. While there were some
1405
internal changes, the code has been mostly stabilised for the RC
1406
release.
1407

    
1408
Note: the dumb allocator was removed in this release, as it was not kept
1409
up-to-date with the IAllocator protocol changes. It is recommended to
1410
use the ``hail`` command from the ganeti-htools package.
1411

    
1412
Note: the 2.4 and up versions of Ganeti are not compatible with the
1413
0.2.x branch of ganeti-htools. You need to upgrade to
1414
ganeti-htools-0.3.0 (or later).
1415

    
1416
Regressions fixed from 2.3
1417
~~~~~~~~~~~~~~~~~~~~~~~~~~
1418

    
1419
- Fixed the ``gnt-cluster verify-disks`` command
1420
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1421
  serially on nodes)
1422
- Fixed disk adoption breakage
1423
- Fixed wrong headers in instance listing for field aliases
1424

    
1425
Other bugs fixed
1426
~~~~~~~~~~~~~~~~
1427

    
1428
- Fixed corner case in KVM handling of NICs
1429
- Fixed many cases of wrong handling of non-vm_capable nodes
1430
- Fixed a bug where a missing instance symlink was not possible to
1431
  recreate with any ``gnt-*`` command (now ``gnt-instance
1432
  activate-disks`` does it)
1433
- Fixed the volume group name as reported by ``gnt-cluster
1434
  verify-disks``
1435
- Increased timeouts for the import-export code, hopefully leading to
1436
  fewer aborts due network or instance timeouts
1437
- Fixed bug in ``gnt-node list-storage``
1438
- Fixed bug where not all daemons were started on cluster
1439
  initialisation, but only at the first watcher run
1440
- Fixed many bugs in the OOB implementation
1441
- Fixed watcher behaviour in presence of instances with offline
1442
  secondaries
1443
- Fixed instance list output for instances running on the wrong node
1444
- a few fixes to the cluster-merge tool, but it still cannot merge
1445
  multi-node groups (currently it is not recommended to use this tool)
1446

    
1447

    
1448
Improvements
1449
~~~~~~~~~~~~
1450

    
1451
- Improved network configuration for the KVM hypervisor
1452
- Added e1000 as a supported NIC for Xen-HVM
1453
- Improved the lvmstrap tool to also be able to use partitions, as
1454
  opposed to full disks
1455
- Improved speed of disk wiping (the cluster parameter
1456
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1457
  of instance creations
1458
- Added documentation for the OS parameters
1459
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1460
  hypervisor is not responding
1461
- Added display of blacklisted and hidden OS information in
1462
  ``gnt-cluster info``
1463
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1464
  NIC and node parameters, which might create problems with currently
1465
  invalid (but undetected) configuration files, but prevents validation
1466
  failures when unrelated parameters are modified
1467
- Changed cluster initialisation to wait for the master daemon to become
1468
  available
1469
- Expanded the RAPI interface:
1470

    
1471
  - Added config redistribution resource
1472
  - Added activation/deactivation of instance disks
1473
  - Added export of console information
1474

    
1475
- Implemented log file reopening on SIGHUP, which allows using
1476
  logrotate(8) for the Ganeti log files
1477
- Added a basic OOB helper script as an example
1478

    
1479

    
1480
Version 2.4.0 beta1
1481
-------------------
1482

    
1483
*(Released Fri, 14 Jan 2011)*
1484

    
1485
User-visible
1486
~~~~~~~~~~~~
1487

    
1488
- Fixed timezone issues when formatting timestamps
1489
- Added support for node groups, available via ``gnt-group`` and other
1490
  commands
1491
- Added out-of-band framework and management, see :doc:`design
1492
  document <design-oob>`
1493
- Removed support for roman numbers from ``gnt-node list`` and
1494
  ``gnt-instance list``.
1495
- Allowed modification of master network interface via ``gnt-cluster
1496
  modify --master-netdev``
1497
- Accept offline secondaries while shutting down instance disks
1498
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1499
- Added support for multiple LVM volume groups
1500
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1501
  requested
1502
- Added commands to list available fields:
1503

    
1504
  - ``gnt-node list-fields``
1505
  - ``gnt-group list-fields``
1506
  - ``gnt-instance list-fields``
1507

    
1508
- Updated documentation and man pages
1509

    
1510
Integration
1511
~~~~~~~~~~~
1512

    
1513
- Moved ``rapi_users`` file into separate directory, now named
1514
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1515
  symlink
1516
- Added new tool for running commands on many machines,
1517
  ``tools/ganeti-listrunner``
1518
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1519
  improving the ``gnt-instance console`` output
1520
- Allowed customisation of disk index separator at ``configure`` time
1521
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1522
- Added support for non-partitioned md disks in ``lvmstrap``
1523
- Added script to gracefully power off KVM instances
1524
- Split ``utils`` module into smaller parts
1525
- Changed query operations to return more detailed information, e.g.
1526
  whether an information is unavailable due to an offline node. To use
1527
  this new functionality, the LUXI call ``Query`` must be used. Field
1528
  information is now stored by the master daemon and can be retrieved
1529
  using ``QueryFields``. Instances, nodes and groups can also be queried
1530
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1531
  exposed via RAPI). The following commands make use of this
1532
  infrastructure change:
1533

    
1534
  - ``gnt-group list``
1535
  - ``gnt-group list-fields``
1536
  - ``gnt-node list``
1537
  - ``gnt-node list-fields``
1538
  - ``gnt-instance list``
1539
  - ``gnt-instance list-fields``
1540
  - ``gnt-debug locks``
1541

    
1542
Remote API
1543
~~~~~~~~~~
1544

    
1545
- New RAPI resources (see :doc:`rapi`):
1546

    
1547
  - ``/2/modify``
1548
  - ``/2/groups``
1549
  - ``/2/groups/[group_name]``
1550
  - ``/2/groups/[group_name]/assign-nodes``
1551
  - ``/2/groups/[group_name]/modify``
1552
  - ``/2/groups/[group_name]/rename``
1553
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1554

    
1555
- RAPI changes:
1556

    
1557
  - Implemented ``no_install`` for instance creation
1558
  - Implemented OS parameters for instance reinstallation, allowing
1559
    use of special settings on reinstallation (e.g. for preserving data)
1560

    
1561
Misc
1562
~~~~
1563

    
1564
- Added IPv6 support in import/export
1565
- Pause DRBD synchronization while wiping disks on instance creation
1566
- Updated unittests and QA scripts
1567
- Improved network parameters passed to KVM
1568
- Converted man pages from docbook to reStructuredText
1569

    
1570

    
1571
Version 2.3.1
1572
-------------
1573

    
1574
*(Released Mon, 20 Dec 2010)*
1575

    
1576
Released version 2.3.1~rc1 without any changes.
1577

    
1578

    
1579
Version 2.3.1 rc1
1580
-----------------
1581

    
1582
*(Released Wed, 1 Dec 2010)*
1583

    
1584
- impexpd: Disable OpenSSL compression in socat if possible (backport
1585
  from master, commit e90739d625b, see :doc:`installation guide
1586
  <install-quick>` for details)
1587
- Changed unittest coverage report to exclude test scripts
1588
- Added script to check version format
1589

    
1590

    
1591
Version 2.3.0
1592
-------------
1593

    
1594
*(Released Wed, 1 Dec 2010)*
1595

    
1596
Released version 2.3.0~rc1 without any changes.
1597

    
1598

    
1599
Version 2.3.0 rc1
1600
-----------------
1601

    
1602
*(Released Fri, 19 Nov 2010)*
1603

    
1604
A number of bugfixes and documentation updates:
1605

    
1606
- Update ganeti-os-interface documentation
1607
- Fixed a bug related to duplicate MACs or similar items which should be
1608
  unique
1609
- Fix breakage in OS state modify
1610
- Reinstall instance: disallow offline secondaries (fixes bug related to
1611
  OS changing but reinstall failing)
1612
- plus all the other fixes between 2.2.1 and 2.2.2
1613

    
1614

    
1615
Version 2.3.0 rc0
1616
-----------------
1617

    
1618
*(Released Tue, 2 Nov 2010)*
1619

    
1620
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1621
- Fixed master failover race with watcher
1622
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1623
  configuration
1624
- Accept previously stopped instance for export with instance removal
1625
- Simplify and extend the environment variables for instance OS scripts
1626
- Added new node flags, ``master_capable`` and ``vm_capable``
1627
- Added optional instance disk wiping prior during allocation. This is a
1628
  cluster-wide option and can be set/modified using
1629
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1630
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1631
  :doc:`install-quick`
1632
- Added a new watcher option (``--ignore-pause``)
1633
- Added option to ignore offline node on instance start/stop
1634
  (``--ignore-offline``)
1635
- Allow overriding OS parameters with ``gnt-instance reinstall``
1636
- Added ability to change node's secondary IP address using ``gnt-node
1637
  modify``
1638
- Implemented privilege separation for all daemons except
1639
  ``ganeti-noded``, see ``configure`` options
1640
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1641
  verify``
1642
- Implemented job priorities (see ``ganeti(7)`` manpage)
1643
- Ignore failures while shutting down instances during failover from
1644
  offline node
1645
- Exit daemon's bootstrap process only once daemon is ready
1646
- Export more information via ``LUInstanceQuery``/remote API
1647
- Improved documentation, QA and unittests
1648
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1649
  restart if the file was created or changed
1650
- Added LUXI protocol version sent with each request and response,
1651
  allowing detection of server/client mismatches
1652
- Moved the Python scripts among gnt-* and ganeti-* into modules
1653
- Moved all code related to setting up SSH to an external script,
1654
  ``setup-ssh``
1655
- Infrastructure changes for node group support in future versions
1656

    
1657

    
1658
Version 2.2.2
1659
-------------
1660

    
1661
*(Released Fri, 19 Nov 2010)*
1662

    
1663
A few small bugs fixed, and some improvements to the build system:
1664

    
1665
- Fix documentation regarding conversion to drbd
1666
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1667
  -B``)
1668
- Fix error handling in node modify with multiple changes
1669
- Allow remote imports without checked names
1670

    
1671

    
1672
Version 2.2.1
1673
-------------
1674

    
1675
*(Released Tue, 19 Oct 2010)*
1676

    
1677
- Disable SSL session ID cache in RPC client
1678

    
1679

    
1680
Version 2.2.1 rc1
1681
-----------------
1682

    
1683
*(Released Thu, 14 Oct 2010)*
1684

    
1685
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1686
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1687
  with GnuTLS
1688
- Fix problems with interaction between Curl and Python's HTTP server,
1689
  resulting in increased speed in many RPC calls
1690
- Improve our release script to prevent breakage with older aclocal and
1691
  Python 2.6
1692

    
1693

    
1694
Version 2.2.1 rc0
1695
-----------------
1696

    
1697
*(Released Thu, 7 Oct 2010)*
1698

    
1699
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1700
  value retrieved from master
1701
- Added support for blacklisted or hidden OS definitions
1702
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1703
- Added support for -mem-path in KVM hypervisor abstraction layer
1704
- Allow overriding instance parameters in tool for inter-cluster
1705
  instance moves (``tools/move-instance``)
1706
- Improved opcode summaries (e.g. in ``gnt-job list``)
1707
- Improve consistency of OS listing by sorting it
1708
- Documentation updates
1709

    
1710

    
1711
Version 2.2.0.1
1712
---------------
1713

    
1714
*(Released Fri, 8 Oct 2010)*
1715

    
1716
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1717

    
1718

    
1719
Version 2.2.0
1720
-------------
1721

    
1722
*(Released Mon, 4 Oct 2010)*
1723

    
1724
- Fixed regression in ``gnt-instance rename``
1725

    
1726

    
1727
Version 2.2.0 rc2
1728
-----------------
1729

    
1730
*(Released Wed, 22 Sep 2010)*
1731

    
1732
- Fixed OS_VARIANT variable for OS scripts
1733
- Fixed cluster tag operations via RAPI
1734
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1735
- Disabled RAPI CA checks in watcher
1736

    
1737

    
1738
Version 2.2.0 rc1
1739
-----------------
1740

    
1741
*(Released Mon, 23 Aug 2010)*
1742

    
1743
- Support DRBD versions of the format "a.b.c.d"
1744
- Updated manpages
1745
- Re-introduce support for usage from multiple threads in RAPI client
1746
- Instance renames and modify via RAPI
1747
- Work around race condition between processing and archival in job
1748
  queue
1749
- Mark opcodes following failed one as failed, too
1750
- Job field ``lock_status`` was removed due to difficulties making it
1751
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1752
  locks is expected for a later 2.2.x release
1753
- Fixed dry-run behaviour with many commands
1754
- Support ``ssh-agent`` again when adding nodes
1755
- Many additional bugfixes
1756

    
1757

    
1758
Version 2.2.0 rc0
1759
-----------------
1760

    
1761
*(Released Fri, 30 Jul 2010)*
1762

    
1763
Important change: the internal RPC mechanism between Ganeti nodes has
1764
changed from using a home-grown http library (based on the Python base
1765
libraries) to use the PycURL library. This requires that PycURL is
1766
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1767
against GnuTLS by default. cURL's support for GnuTLS had known issues
1768
before cURL 7.21.0 and we recommend using the latest cURL release or
1769
linking against OpenSSL. Most other distributions already link PycURL
1770
and cURL against OpenSSL. The command::
1771

    
1772
  python -c 'import pycurl; print pycurl.version'
1773

    
1774
can be used to determine the libraries PycURL and cURL are linked
1775
against.
1776

    
1777
Other significant changes:
1778

    
1779
- Rewrote much of the internals of the job queue, in order to achieve
1780
  better parallelism; this decouples job query operations from the job
1781
  processing, and it should allow much nicer behaviour of the master
1782
  daemon under load, and it also has uncovered some long-standing bugs
1783
  related to the job serialisation (now fixed)
1784
- Added a default iallocator setting to the cluster parameters,
1785
  eliminating the need to always pass nodes or an iallocator for
1786
  operations that require selection of new node(s)
1787
- Added experimental support for the LXC virtualization method
1788
- Added support for OS parameters, which allows the installation of
1789
  instances to pass parameter to OS scripts in order to customise the
1790
  instance
1791
- Added a hypervisor parameter controlling the migration type (live or
1792
  non-live), since hypervisors have various levels of reliability; this
1793
  has renamed the 'live' parameter to 'mode'
1794
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
1795
  logical volumes, meaning that cluster verify will ignore them and not
1796
  flag their presence as errors
1797
- The watcher will now reset the error count for failed instances after
1798
  8 hours, thus allowing self-healing if the problem that caused the
1799
  instances to be down/fail to start has cleared in the meantime
1800
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1801
  check for, and warn, if the drbd module parameter ``usermode_helper``
1802
  is not consistent with the cluster-wide setting; this is needed to
1803
  make diagnose easier of failed drbd creations
1804
- Started adding base IPv6 support, but this is not yet
1805
  enabled/available for use
1806
- Rename operations (cluster, instance) will now return the new name,
1807
  which is especially useful if a short name was passed in
1808
- Added support for instance migration in RAPI
1809
- Added a tool to pre-configure nodes for the SSH setup, before joining
1810
  them to the cluster; this will allow in the future a simplified model
1811
  for node joining (but not yet fully enabled in 2.2); this needs the
1812
  paramiko python library
1813
- Fixed handling of name-resolving errors
1814
- Fixed consistency of job results on the error path
1815
- Fixed master-failover race condition when executed multiple times in
1816
  sequence
1817
- Fixed many bugs related to the job queue (mostly introduced during the
1818
  2.2 development cycle, so not all are impacting 2.1)
1819
- Fixed instance migration with missing disk symlinks
1820
- Fixed handling of unknown jobs in ``gnt-job archive``
1821
- And many other small fixes/improvements
1822

    
1823
Internal changes:
1824

    
1825
- Enhanced both the unittest and the QA coverage
1826
- Switched the opcode validation to a generic model, and extended the
1827
  validation to all opcode parameters
1828
- Changed more parts of the code that write shell scripts to use the
1829
  same class for this
1830
- Switched the master daemon to use the asyncore library for the Luxi
1831
  server endpoint
1832

    
1833

    
1834
Version 2.2.0 beta0
1835
-------------------
1836

    
1837
*(Released Thu, 17 Jun 2010)*
1838

    
1839
- Added tool (``move-instance``) and infrastructure to move instances
1840
  between separate clusters (see :doc:`separate documentation
1841
  <move-instance>` and :doc:`design document <design-2.2>`)
1842
- Added per-request RPC timeout
1843
- RAPI now requires a Content-Type header for requests with a body (e.g.
1844
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1845
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
1846
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1847
  reachable
1848
- Implemented initial support for running Ganeti daemons as separate
1849
  users, see configure-time flags ``--with-user-prefix`` and
1850
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1851
  time)
1852
- Instances can be removed after export (``gnt-backup export
1853
  --remove-instance``)
1854
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1855
  key (instead of 1024 bit)
1856
- Added new cluster configuration file for cluster domain secret
1857
- Import/export now use SSL instead of SSH
1858
- Added support for showing estimated time when exporting an instance,
1859
  see the ``ganeti-os-interface(7)`` manpage and look for
1860
  ``EXP_SIZE_FD``
1861

    
1862

    
1863
Version 2.1.8
1864
-------------
1865

    
1866
*(Released Tue, 16 Nov 2010)*
1867

    
1868
Some more bugfixes. Unless critical bugs occur, this will be the last
1869
2.1 release:
1870

    
1871
- Fix case of MAC special-values
1872
- Fix mac checker regex
1873
- backend: Fix typo causing "out of range" error
1874
- Add missing --units in gnt-instance list man page
1875

    
1876

    
1877
Version 2.1.7
1878
-------------
1879

    
1880
*(Released Tue, 24 Aug 2010)*
1881

    
1882
Bugfixes only:
1883
  - Don't ignore secondary node silently on non-mirrored disk templates
1884
    (issue 113)
1885
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1886
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1887
  - Properly document the usb_mouse parameter
1888
  - Fix path in ganeti-rapi(8) (issue 116)
1889
  - Adjust error message when the ganeti user's .ssh directory is
1890
    missing
1891
  - Add same-node-check when changing the disk template to drbd
1892

    
1893

    
1894
Version 2.1.6
1895
-------------
1896

    
1897
*(Released Fri, 16 Jul 2010)*
1898

    
1899
Bugfixes only:
1900
  - Add an option to only select some reboot types during qa/burnin.
1901
    (on some hypervisors consequent reboots are not supported)
1902
  - Fix infrequent race condition in master failover. Sometimes the old
1903
    master ip address would be still detected as up for a short time
1904
    after it was removed, causing failover to fail.
1905
  - Decrease mlockall warnings when the ctypes module is missing. On
1906
    Python 2.4 we support running even if no ctypes module is installed,
1907
    but we were too verbose about this issue.
1908
  - Fix building on old distributions, on which man doesn't have a
1909
    --warnings option.
1910
  - Fix RAPI not to ignore the MAC address on instance creation
1911
  - Implement the old instance creation format in the RAPI client.
1912

    
1913

    
1914
Version 2.1.5
1915
-------------
1916

    
1917
*(Released Thu, 01 Jul 2010)*
1918

    
1919
A small bugfix release:
1920
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1921
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1922
    a non-existing option
1923
  - Fix instance create: the --force-variant option was ignored
1924
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
1925
  - Fix modify node storage with non-FQDN arguments
1926
  - Fix RAPI client to authenticate under Python 2.6 when used
1927
    for more than 5 requests needing authentication
1928
  - Fix gnt-instance modify -t (storage) giving a wrong error message
1929
    when converting a non-shutdown drbd instance to plain
1930

    
1931

    
1932
Version 2.1.4
1933
-------------
1934

    
1935
*(Released Fri, 18 Jun 2010)*
1936

    
1937
A small bugfix release:
1938

    
1939
  - Fix live migration of KVM instances started with older Ganeti
1940
    versions which had fewer hypervisor parameters
1941
  - Fix gnt-instance grow-disk on down instances
1942
  - Fix an error-reporting bug during instance migration
1943
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
1944
    silently ignoring broken ones
1945
  - Fix an RPC error reporting bug affecting, for example, RAPI client
1946
    users
1947
  - Fix bug triggered by different API version os-es on different nodes
1948
  - Fix a bug in instance startup with custom hvparams: OS level
1949
    parameters would fail to be applied.
1950
  - Fix the RAPI client under Python 2.6 (but more work is needed to
1951
    make it work completely well with OpenSSL)
1952
  - Fix handling of errors when resolving names from DNS
1953

    
1954

    
1955
Version 2.1.3
1956
-------------
1957

    
1958
*(Released Thu, 3 Jun 2010)*
1959

    
1960
A medium sized development cycle. Some new features, and some
1961
fixes/small improvements/cleanups.
1962

    
1963
Significant features
1964
~~~~~~~~~~~~~~~~~~~~
1965

    
1966
The node deamon now tries to mlock itself into memory, unless the
1967
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1968
its logs, and falls back to console logging. This allows emergency
1969
features such as ``gnt-node powercycle`` to work even in the event of a
1970
broken node disk (tested offlining the disk hosting the node's
1971
filesystem and dropping its memory caches; don't try this at home)
1972

    
1973
KVM: add vhost-net acceleration support. It can be tested with a new
1974
enough version of the kernel and of qemu-kvm.
1975

    
1976
KVM: Add instance chrooting feature. If you use privilege dropping for
1977
your VMs you can also now force them to chroot to an empty directory,
1978
before starting the emulated guest.
1979

    
1980
KVM: Add maximum migration bandwith and maximum downtime tweaking
1981
support (requires a new-enough version of qemu-kvm).
1982

    
1983
Cluster verify will now warn if the master node doesn't have the master
1984
ip configured on it.
1985

    
1986
Add a new (incompatible) instance creation request format to RAPI which
1987
supports all parameters (previously only a subset was supported, and it
1988
wasn't possible to extend the old format to accomodate all the new
1989
features. The old format is still supported, and a client can check for
1990
this feature, before using it, by checking for its presence in the
1991
``features`` RAPI resource.
1992

    
1993
Now with ancient latin support. Try it passing the ``--roman`` option to
1994
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1995
(requires the python-roman module to be installed, in order to work).
1996

    
1997
Other changes
1998
~~~~~~~~~~~~~
1999

    
2000
As usual many internal code refactorings, documentation updates, and
2001
such. Among others:
2002

    
2003
  - Lots of improvements and cleanups to the experimental Remote API
2004
    (RAPI) client library.
2005
  - A new unit test suite for the core daemon libraries.
2006
  - A fix to creating missing directories makes sure the umask is not
2007
    applied anymore. This enforces the same directory permissions
2008
    everywhere.
2009
  - Better handling terminating daemons with ctrl+c (used when running
2010
    them in debugging mode).
2011
  - Fix a race condition in live migrating a KVM instance, when stat()
2012
    on the old proc status file returned EINVAL, which is an unexpected
2013
    value.
2014
  - Fixed manpage checking with newer man and utf-8 charachters. But now
2015
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2016

    
2017

    
2018
Version 2.1.2.1
2019
---------------
2020

    
2021
*(Released Fri, 7 May 2010)*
2022

    
2023
Fix a bug which prevented untagged KVM instances from starting.
2024

    
2025

    
2026
Version 2.1.2
2027
-------------
2028

    
2029
*(Released Fri, 7 May 2010)*
2030

    
2031
Another release with a long development cycle, during which many
2032
different features were added.
2033

    
2034
Significant features
2035
~~~~~~~~~~~~~~~~~~~~
2036

    
2037
The KVM hypervisor now can run the individual instances as non-root, to
2038
reduce the impact of a VM being hijacked due to bugs in the
2039
hypervisor. It is possible to run all instances as a single (non-root)
2040
user, to manually specify a user for each instance, or to dynamically
2041
allocate a user out of a cluster-wide pool to each instance, with the
2042
guarantee that no two instances will run under the same user ID on any
2043
given node.
2044

    
2045
An experimental RAPI client library, that can be used standalone
2046
(without the other Ganeti libraries), is provided in the source tree as
2047
``lib/rapi/client.py``. Note this client might change its interface in
2048
the future, as we iterate on its capabilities.
2049

    
2050
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2051
replace the cluster's certificates and crypto keys. This might help in
2052
case they have been compromised, or have simply expired.
2053

    
2054
A new disk option for instance creation has been added that allows one
2055
to "adopt" currently existing logical volumes, with data
2056
preservation. This should allow easier migration to Ganeti from
2057
unmanaged (or managed via other software) instances.
2058

    
2059
Another disk improvement is the possibility to convert between redundant
2060
(DRBD) and plain (LVM) disk configuration for an instance. This should
2061
allow better scalability (starting with one node and growing the
2062
cluster, or shrinking a two-node cluster to one node).
2063

    
2064
A new feature that could help with automated node failovers has been
2065
implemented: if a node sees itself as offline (by querying the master
2066
candidates), it will try to shutdown (hard) all instances and any active
2067
DRBD devices. This reduces the risk of duplicate instances if an
2068
external script automatically failovers the instances on such nodes. To
2069
enable this, the cluster parameter ``maintain_node_health`` should be
2070
enabled; in the future this option (per the name) will enable other
2071
automatic maintenance features.
2072

    
2073
Instance export/import now will reuse the original instance
2074
specifications for all parameters; that means exporting an instance,
2075
deleting it and the importing it back should give an almost identical
2076
instance. Note that the default import behaviour has changed from
2077
before, where it created only one NIC; now it recreates the original
2078
number of NICs.
2079

    
2080
Cluster verify has added a few new checks: SSL certificates validity,
2081
/etc/hosts consistency across the cluster, etc.
2082

    
2083
Other changes
2084
~~~~~~~~~~~~~
2085

    
2086
As usual, many internal changes were done, documentation fixes,
2087
etc. Among others:
2088

    
2089
- Fixed cluster initialization with disabled cluster storage (regression
2090
  introduced in 2.1.1)
2091
- File-based storage supports growing the disks
2092
- Fixed behaviour of node role changes
2093
- Fixed cluster verify for some corner cases, plus a general rewrite of
2094
  cluster verify to allow future extension with more checks
2095
- Fixed log spamming by watcher and node daemon (regression introduced
2096
  in 2.1.1)
2097
- Fixed possible validation issues when changing the list of enabled
2098
  hypervisors
2099
- Fixed cleanup of /etc/hosts during node removal
2100
- Fixed RAPI response for invalid methods
2101
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2102
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2103
  from ide disks, etc.)
2104
- Allow OS changes without re-installation (to record a changed OS
2105
  outside of Ganeti, or to allow OS renames)
2106
- Allow instance creation without OS installation (useful for example if
2107
  the OS will be installed manually, or restored from a backup not in
2108
  Ganeti format)
2109
- Implemented option to make cluster ``copyfile`` use the replication
2110
  network
2111
- Added list of enabled hypervisors to ssconf (possibly useful for
2112
  external scripts)
2113
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2114
  1.2 clusters
2115
- A partial form of node re-IP is possible via node readd, which now
2116
  allows changed node primary IP
2117
- Command line utilities now show an informational message if the job is
2118
  waiting for a lock
2119
- The logs of the master daemon now show the PID/UID/GID of the
2120
  connected client
2121

    
2122

    
2123
Version 2.1.1
2124
-------------
2125

    
2126
*(Released Fri, 12 Mar 2010)*
2127

    
2128
During the 2.1.0 long release candidate cycle, a lot of improvements and
2129
changes have accumulated with were released later as 2.1.1.
2130

    
2131
Major changes
2132
~~~~~~~~~~~~~
2133

    
2134
The node evacuate command (``gnt-node evacuate``) was significantly
2135
rewritten, and as such the IAllocator protocol was changed - a new
2136
request type has been added. This unfortunate change during a stable
2137
series is designed to improve performance of node evacuations; on
2138
clusters with more than about five nodes and which are well-balanced,
2139
evacuation should proceed in parallel for all instances of the node
2140
being evacuated. As such, any existing IAllocator scripts need to be
2141
updated, otherwise the above command will fail due to the unknown
2142
request. The provided "dumb" allocator has not been updated; but the
2143
ganeti-htools package supports the new protocol since version 0.2.4.
2144

    
2145
Another important change is increased validation of node and instance
2146
names. This might create problems in special cases, if invalid host
2147
names are being used.
2148

    
2149
Also, a new layer of hypervisor parameters has been added, that sits at
2150
OS level between the cluster defaults and the instance ones. This allows
2151
customisation of virtualization parameters depending on the installed
2152
OS. For example instances with OS 'X' may have a different KVM kernel
2153
(or any other parameter) than the cluster defaults. This is intended to
2154
help managing a multiple OSes on the same cluster, without manual
2155
modification of each instance's parameters.
2156

    
2157
A tool for merging clusters, ``cluster-merge``, has been added in the
2158
tools sub-directory.
2159

    
2160
Bug fixes
2161
~~~~~~~~~
2162

    
2163
- Improved the int/float conversions that should make the code more
2164
  robust in face of errors from the node daemons
2165
- Fixed the remove node code in case of internal configuration errors
2166
- Fixed the node daemon behaviour in face of inconsistent queue
2167
  directory (e.g. read-only file-system where we can't open the files
2168
  read-write, etc.)
2169
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2170
  now it either aborts cleanly or, if given the new "auto_promote"
2171
  parameter, will automatically promote other nodes as needed
2172
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2173
  completely prevent Ganeti from working
2174
- Fixed bug for instance export when not all disks were successfully
2175
  exported
2176
- Fixed behaviour of node add when the new node is slow in starting up
2177
  the node daemon
2178
- Fixed handling of signals in the LUXI client, which should improve
2179
  behaviour of command-line scripts
2180
- Added checks for invalid node/instance names in the configuration (now
2181
  flagged during cluster verify)
2182
- Fixed watcher behaviour for disk activation errors
2183
- Fixed two potentially endless loops in http library, which led to the
2184
  RAPI daemon hanging and consuming 100% CPU in some cases
2185
- Fixed bug in RAPI daemon related to hashed passwords
2186
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2187
  instances
2188
- Enhanced compatibility with non-Debian OSes, but not using absolute
2189
  path in some commands and allowing customisation of the ssh
2190
  configuration directory
2191
- Fixed possible future issue with new Python versions by abiding to the
2192
  proper use of ``__slots__`` attribute on classes
2193
- Added checks that should prevent directory traversal attacks
2194
- Many documentation fixes based on feedback from users
2195

    
2196
New features
2197
~~~~~~~~~~~~
2198

    
2199
- Added an "early_release" more for instance replace disks and node
2200
  evacuate, where we release locks earlier and thus allow higher
2201
  parallelism within the cluster
2202
- Added watcher hooks, intended to allow the watcher to restart other
2203
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2204
  course for any other purpose
2205
- Added a compile-time disable for DRBD barriers, to increase
2206
  performance if the administrator trusts the power supply or the
2207
  storage system to not lose writes
2208
- Added the option of using syslog for logging instead of, or in
2209
  addition to, Ganeti's own log files
2210
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2211
  can indeed boot from a paravirtual NIC
2212
- Added a generic debug level for many operations; while this is not
2213
  used widely yet, it allows one to pass the debug value all the way to
2214
  the OS scripts
2215
- Enhanced the hooks environment for instance moves (failovers,
2216
  migrations) where the primary/secondary nodes changed during the
2217
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2218
- Enhanced data validations for many user-supplied values; one important
2219
  item is the restrictions imposed on instance and node names, which
2220
  might reject some (invalid) host names
2221
- Add a configure-time option to disable file-based storage, if it's not
2222
  needed; this allows greater security separation between the master
2223
  node and the other nodes from the point of view of the inter-node RPC
2224
  protocol
2225
- Added user notification in interactive tools if job is waiting in the
2226
  job queue or trying to acquire locks
2227
- Added log messages when a job is waiting for locks
2228
- Added filtering by node tags in instance operations which admit
2229
  multiple instances (start, stop, reboot, reinstall)
2230
- Added a new tool for cluster mergers, ``cluster-merge``
2231
- Parameters from command line which are of the form ``a=b,c=d`` can now
2232
  use backslash escapes to pass in values which contain commas,
2233
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2234
  ``b,c``
2235
- For KVM, the instance name is the first parameter passed to KVM, so
2236
  that it's more visible in the process list
2237

    
2238

    
2239
Version 2.1.0
2240
-------------
2241

    
2242
*(Released Tue, 2 Mar 2010)*
2243

    
2244
Ganeti 2.1 brings many improvements with it. Major changes:
2245

    
2246
- Added infrastructure to ease automated disk repairs
2247
- Added new daemon to export configuration data in a cheaper way than
2248
  using the remote API
2249
- Instance NICs can now be routed instead of being associated with a
2250
  networking bridge
2251
- Improved job locking logic to reduce impact of jobs acquiring multiple
2252
  locks waiting for other long-running jobs
2253

    
2254
In-depth implementation details can be found in the Ganeti 2.1 design
2255
document.
2256

    
2257
Details
2258
~~~~~~~
2259

    
2260
- Added chroot hypervisor
2261
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2262
  ``device_model``)
2263
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2264
  ``bootloader_path`` and ``bootloader_args``)
2265
- Added the ``use_localtime`` option for the xen-hvm and kvm
2266
  hypervisors, and the default value for this has changed to false (in
2267
  2.0 xen-hvm always enabled it)
2268
- Added luxi call to submit multiple jobs in one go
2269
- Added cluster initialization option to not modify ``/etc/hosts``
2270
  file on nodes
2271
- Added network interface parameters
2272
- Added dry run mode to some LUs
2273
- Added RAPI resources:
2274

    
2275
  - ``/2/instances/[instance_name]/info``
2276
  - ``/2/instances/[instance_name]/replace-disks``
2277
  - ``/2/nodes/[node_name]/evacuate``
2278
  - ``/2/nodes/[node_name]/migrate``
2279
  - ``/2/nodes/[node_name]/role``
2280
  - ``/2/nodes/[node_name]/storage``
2281
  - ``/2/nodes/[node_name]/storage/modify``
2282
  - ``/2/nodes/[node_name]/storage/repair``
2283

    
2284
- Added OpCodes to evacuate or migrate all instances on a node
2285
- Added new command to list storage elements on nodes (``gnt-node
2286
  list-storage``) and modify them (``gnt-node modify-storage``)
2287
- Added new ssconf files with master candidate IP address
2288
  (``ssconf_master_candidates_ips``), node primary IP address
2289
  (``ssconf_node_primary_ips``) and node secondary IP address
2290
  (``ssconf_node_secondary_ips``)
2291
- Added ``ganeti-confd`` and a client library to query the Ganeti
2292
  configuration via UDP
2293
- Added ability to run hooks after cluster initialization and before
2294
  cluster destruction
2295
- Added automatic mode for disk replace (``gnt-instance replace-disks
2296
  --auto``)
2297
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2298
  after catastrophic data-loss
2299
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2300
  groups
2301
- Added ``gnt-instance move`` command to move instances
2302
- Added ``gnt-cluster watcher`` command to control watcher
2303
- Added ``gnt-node powercycle`` command to powercycle nodes
2304
- Added new job status field ``lock_status``
2305
- Added parseable error codes to cluster verification (``gnt-cluster
2306
  verify --error-codes``) and made output less verbose (use
2307
  ``--verbose`` to restore previous behaviour)
2308
- Added UUIDs to the main config entities (cluster, nodes, instances)
2309
- Added support for OS variants
2310
- Added support for hashed passwords in the Ganeti remote API users file
2311
  (``rapi_users``)
2312
- Added option to specify maximum timeout on instance shutdown
2313
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2314
- Added new helper script to start and stop Ganeti daemons
2315
  (``daemon-util``), with the intent to reduce the work necessary to
2316
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2317
  from one place
2318
- Added more unittests
2319
- Fixed critical bug in ganeti-masterd startup
2320
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2321
  now customisable at the cluster level for both the KVM and Xen
2322
  hypervisors using the new ``migration_port`` parameter
2323
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2324
  reinstalling an instance
2325
- Allowed ``@`` in tag names
2326
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2327
- Many documentation updates
2328
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2329
- ``gnt-instance reinstall`` can now reinstall multiple instances
2330
- Updated many command line parameters
2331
- Introduced new OS API version 15
2332
- No longer support a default hypervisor
2333
- Treat virtual LVs as inexistent
2334
- Improved job locking logic to reduce lock contention
2335
- Match instance and node names case insensitively
2336
- Reimplemented bash completion script to be more complete
2337
- Improved burnin
2338

    
2339

    
2340
Version 2.0.6
2341
-------------
2342

    
2343
*(Released Thu, 4 Feb 2010)*
2344

    
2345
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2346
- Fix a string formatting bug
2347
- Improve safety of the code in some error paths
2348
- Improve data validation in the master of values returned from nodes
2349

    
2350

    
2351
Version 2.0.5
2352
-------------
2353

    
2354
*(Released Thu, 17 Dec 2009)*
2355

    
2356
- Fix security issue due to missing validation of iallocator names; this
2357
  allows local and remote execution of arbitrary executables
2358
- Fix failure of gnt-node list during instance removal
2359
- Ship the RAPI documentation in the archive
2360

    
2361

    
2362
Version 2.0.4
2363
-------------
2364

    
2365
*(Released Wed, 30 Sep 2009)*
2366

    
2367
- Fixed many wrong messages
2368
- Fixed a few bugs related to the locking library
2369
- Fixed MAC checking at instance creation time
2370
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2371
- Fixed a few issues related to signal handling in both daemons and
2372
  scripts
2373
- Fixed the example startup script provided
2374
- Fixed insserv dependencies in the example startup script (patch from
2375
  Debian)
2376
- Fixed handling of drained nodes in the iallocator framework
2377
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2378
  #528618)
2379
- Fixed error related to invalid job IDs in job polling
2380
- Fixed job/opcode persistence on unclean master shutdown
2381
- Fixed handling of partial job processing after unclean master
2382
  shutdown
2383
- Fixed error reporting from LUs, previously all errors were converted
2384
  into execution errors
2385
- Fixed error reporting from burnin
2386
- Decreased significantly the memory usage of the job queue
2387
- Optimised slightly multi-job submission
2388
- Optimised slightly opcode loading
2389
- Backported the multi-job submit framework from the development
2390
  branch; multi-instance start and stop should be faster
2391
- Added script to clean archived jobs after 21 days; this will reduce
2392
  the size of the queue directory
2393
- Added some extra checks in disk size tracking
2394
- Added an example ethers hook script
2395
- Added a cluster parameter that prevents Ganeti from modifying of
2396
  /etc/hosts
2397
- Added more node information to RAPI responses
2398
- Added a ``gnt-job watch`` command that allows following the ouput of a
2399
  job
2400
- Added a bind-address option to ganeti-rapi
2401
- Added more checks to the configuration verify
2402
- Enhanced the burnin script such that some operations can be retried
2403
  automatically
2404
- Converted instance reinstall to multi-instance model
2405

    
2406

    
2407
Version 2.0.3
2408
-------------
2409

    
2410
*(Released Fri, 7 Aug 2009)*
2411

    
2412
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2413
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2414
  instances have mismatched disk sizes in the configuration
2415
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2416
  any configuration mismatches for disk sizes
2417
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2418
  failover to work on two-node clusters
2419
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2420
  unusable
2421
- Fixed detection of OS script errors in ``gnt-backup export``
2422
- Fixed exit code of ``gnt-backup export``
2423

    
2424

    
2425
Version 2.0.2
2426
-------------
2427

    
2428
*(Released Fri, 17 Jul 2009)*
2429

    
2430
- Added experimental support for stripped logical volumes; this should
2431
  enhance performance but comes with a higher complexity in the block
2432
  device handling; stripping is only enabled when passing
2433
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2434
  affected even in the non-stripped mode
2435
- Improved resiliency against transient failures at the end of DRBD
2436
  resyncs, and in general of DRBD resync checks
2437
- Fixed a couple of issues with exports and snapshot errors
2438
- Fixed a couple of issues in instance listing
2439
- Added display of the disk size in ``gnt-instance info``
2440
- Fixed checking for valid OSes in instance creation
2441
- Fixed handling of the "vcpus" parameter in instance listing and in
2442
  general of invalid parameters
2443
- Fixed http server library, and thus RAPI, to handle invalid
2444
  username/password combinations correctly; this means that now they
2445
  report unauthorized for queries too, not only for modifications,
2446
  allowing earlier detect of configuration problems
2447
- Added a new "role" node list field, equivalent to the master/master
2448
  candidate/drained/offline flags combinations
2449
- Fixed cluster modify and changes of candidate pool size
2450
- Fixed cluster verify error messages for wrong files on regular nodes
2451
- Fixed a couple of issues with node demotion from master candidate role
2452
- Fixed node readd issues
2453
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2454
- Added a new ``--no-voting`` option for masterfailover to fix failover
2455
  on two-nodes clusters when the former master node is unreachable
2456
- Added instance reinstall over RAPI
2457

    
2458

    
2459
Version 2.0.1
2460
-------------
2461

    
2462
*(Released Tue, 16 Jun 2009)*
2463

    
2464
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2465
  allow re-adding the start in single-user option (regression from 1.2)
2466
- the watcher writes the instance status to a file, to allow monitoring
2467
  to report the instance status (from the master) based on cached
2468
  results of the watcher's queries; while this can get stale if the
2469
  watcher is being locked due to other work on the cluster, this is
2470
  still an improvement
2471
- the watcher now also restarts the node daemon and the rapi daemon if
2472
  they died
2473
- fixed the watcher to handle full and drained queue cases
2474
- hooks export more instance data in the environment, which helps if
2475
  hook scripts need to take action based on the instance's properties
2476
  (no longer need to query back into ganeti)
2477
- instance failovers when the instance is stopped do not check for free
2478
  RAM, so that failing over a stopped instance is possible in low memory
2479
  situations
2480
- rapi uses queries for tags instead of jobs (for less job traffic), and
2481
  for cluster tags it won't talk to masterd at all but read them from
2482
  ssconf
2483
- a couple of error handling fixes in RAPI
2484
- drbd handling: improved the error handling of inconsistent disks after
2485
  resync to reduce the frequency of "there are some degraded disks for
2486
  this instance" messages
2487
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2488
  error handling path called a wrong function name)
2489

    
2490

    
2491
Version 2.0.0
2492
-------------
2493

    
2494
*(Released Wed, 27 May 2009)*
2495

    
2496
- no changes from rc5
2497

    
2498

    
2499
Version 2.0 rc5
2500
---------------
2501

    
2502
*(Released Wed, 20 May 2009)*
2503

    
2504
- fix a couple of bugs (validation, argument checks)
2505
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2506
- some small improvements to RAPI and IAllocator
2507
- make watcher automatically start the master daemon if down
2508

    
2509

    
2510
Version 2.0 rc4
2511
---------------
2512

    
2513
*(Released Mon, 27 Apr 2009)*
2514

    
2515
- change the OS list to not require locks; this helps with big clusters
2516
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2517
  volume group is broken
2518
- ``gnt-instance info``, without any arguments, doesn't run for all
2519
  instances anymore; either pass ``--all`` or pass the desired
2520
  instances; this helps against mistakes on big clusters where listing
2521
  the information for all instances takes a long time
2522
- miscellaneous doc and man pages fixes
2523

    
2524

    
2525
Version 2.0 rc3
2526
---------------
2527

    
2528
*(Released Wed, 8 Apr 2009)*
2529

    
2530
- Change the internal locking model of some ``gnt-node`` commands, in
2531
  order to reduce contention (and blocking of master daemon) when
2532
  batching many creation/reinstall jobs
2533
- Fixes to Xen soft reboot
2534
- No longer build documentation at build time, instead distribute it in
2535
  the archive, in order to reduce the need for the whole docbook/rst
2536
  toolchains
2537

    
2538

    
2539
Version 2.0 rc2
2540
---------------
2541

    
2542
*(Released Fri, 27 Mar 2009)*
2543

    
2544
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2545
- Fix watcher startup sequence, improves the behaviour of busy clusters
2546
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2547
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2548
  bind address and other places
2549
- Some documentation fixes and updates
2550

    
2551

    
2552
Version 2.0 rc1
2553
---------------
2554

    
2555
*(Released Mon, 2 Mar 2009)*
2556

    
2557
- More documentation updates, now all docs should be more-or-less
2558
  up-to-date
2559
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2560
  etc.)
2561
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2562
  and KVM)
2563
- Fix an issue related to $libdir/run/ganeti and cluster creation
2564

    
2565

    
2566
Version 2.0 beta2
2567
-----------------
2568

    
2569
*(Released Thu, 19 Feb 2009)*
2570

    
2571
- Xen PVM and KVM have switched the default value for the instance root
2572
  disk to the first partition on the first drive, instead of the whole
2573
  drive; this means that the OS installation scripts must be changed
2574
  accordingly
2575
- Man pages have been updated
2576
- RAPI has been switched by default to HTTPS, and the exported functions
2577
  should all work correctly
2578
- RAPI v1 has been removed
2579
- Many improvements to the KVM hypervisor
2580
- Block device errors are now better reported
2581
- Many other bugfixes and small improvements
2582

    
2583

    
2584
Version 2.0 beta1
2585
-----------------
2586

    
2587
*(Released Mon, 26 Jan 2009)*
2588

    
2589
- Version 2 is a general rewrite of the code and therefore the
2590
  differences are too many to list, see the design document for 2.0 in
2591
  the ``doc/`` subdirectory for more details
2592
- In this beta version there is not yet a migration path from 1.2 (there
2593
  will be one in the final 2.0 release)
2594
- A few significant changes are:
2595

    
2596
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2597
    various ``gnt-*`` commands are just front-ends to it
2598
  - all the commands are entered into, and executed from a job queue,
2599
    see the ``gnt-job(8)`` manpage
2600
  - the RAPI daemon supports read-write operations, secured by basic
2601
    HTTP authentication on top of HTTPS
2602
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2603
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2604
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2605
  - DRBD devices are using statically allocated minor numbers, which
2606
    will be assigned to existing instances during the migration process
2607
  - there is support for both Xen PVM and Xen HVM instances running on
2608
    the same cluster
2609
  - KVM virtualization is supported too
2610
  - file-based storage has been implemented, which means that it is
2611
    possible to run the cluster without LVM and DRBD storage, for
2612
    example using a shared filesystem exported from shared storage (and
2613
    still have live migration)
2614

    
2615

    
2616
Version 1.2.7
2617
-------------
2618

    
2619
*(Released Tue, 13 Jan 2009)*
2620

    
2621
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2622
- Reuse the old instance mac address by default on instance import, if
2623
  the instance name is the same.
2624
- Handle situations in which the node info rpc returns incomplete
2625
  results (issue 46)
2626
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2627
- Improved version of batcher:
2628

    
2629
  - state file support
2630
  - instance mac address support
2631
  - support for HVM clusters/instances
2632

    
2633
- Add an option to show the number of cpu sockets and nodes in
2634
  ``gnt-node list``
2635
- Support OSes that handle more than one version of the OS api (but do
2636
  not change the current API in any other way)
2637
- Fix ``gnt-node migrate``
2638
- ``gnt-debug`` man page
2639
- Fixes various more typos and small issues
2640
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2641

    
2642

    
2643
Version 1.2.6
2644
-------------
2645

    
2646
*(Released Wed, 24 Sep 2008)*
2647

    
2648
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2649
  type of disk exported to fully virtualized instances.
2650
- provide access to the serial console of HVM instances
2651
- instance auto_balance flag, set by default. If turned off it will
2652
  avoid warnings on cluster verify if there is not enough memory to fail
2653
  over an instance. in the future it will prevent automatically failing
2654
  it over when we will support that.
2655
- batcher tool for instance creation, see ``tools/README.batcher``
2656
- ``gnt-instance reinstall --select-os`` to interactively select a new
2657
  operating system when reinstalling an instance.
2658
- when changing the memory amount on instance modify a check has been
2659
  added that the instance will be able to start. also warnings are
2660
  emitted if the instance will not be able to fail over, if auto_balance
2661
  is true.
2662
- documentation fixes
2663
- sync fields between ``gnt-instance list/modify/add/import``
2664
- fix a race condition in drbd when the sync speed was set after giving
2665
  the device a remote peer.
2666

    
2667

    
2668
Version 1.2.5
2669
-------------
2670

    
2671
*(Released Tue, 22 Jul 2008)*
2672

    
2673
- note: the allowed size and number of tags per object were reduced
2674
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2675
- fixed twisted 8.x compatibility
2676
- fixed ``gnt-instance replace-disks`` with iallocator
2677
- add TCP keepalives on twisted connections to detect restarted nodes
2678
- disk increase support, see ``gnt-instance grow-disk``
2679
- implement bulk node/instance query for RAPI
2680
- add tags in node/instance listing (optional)
2681
- experimental migration (and live migration) support, read the man page
2682
  for ``gnt-instance migrate``
2683
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2684
  has some small improvements in handling its state file
2685

    
2686

    
2687
Version 1.2.4
2688
-------------
2689

    
2690
*(Released Fri, 13 Jun 2008)*
2691

    
2692
- Experimental readonly, REST-based remote API implementation;
2693
  automatically started on master node, TCP port 5080, if enabled by
2694
  ``--enable-rapi`` parameter to configure script.
2695
- Instance allocator support. Add and import instance accept a
2696
  ``--iallocator`` parameter, and call that instance allocator to decide
2697
  which node to use for the instance. The iallocator document describes
2698
  what's expected from an allocator script.
2699
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2700
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2701
  node is lost there is still enough memory to fail over the instances
2702
  that reside on it.
2703
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2704
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2705
  hooks will run, and their output, if any, will be displayed. Any
2706
  failing hook will make the verification return an error value.
2707
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2708
  primary and secondary interfaces
2709
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2710
  that is still declared as part of the cluster and has failed.
2711
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2712
  specifying output fields, that just adds the chosen fields to the
2713
  default ones.
2714
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2715
  export from the filesystem.
2716
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2717
  have been added. Using them you can enable/disable acpi and pae
2718
  support, and specify a path for a cd image to be exported to the
2719
  instance. These parameters as the name suggest only work on HVM
2720
  clusters.
2721
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2722
  PAE support will be set to the previously hardcoded values, but the
2723
  (previously hardcoded) path to the CDROM ISO image will be unset and
2724
  if required, needs to be set manually with ``gnt-instance modify``
2725
  after the upgrade.
2726
- The address to which an instance's VNC console is bound is now
2727
  selectable per-instance, rather than being cluster wide. Of course
2728
  this only applies to instances controlled via VNC, so currently just
2729
  applies to HVM clusters.
2730

    
2731

    
2732
Version 1.2.3
2733
-------------
2734

    
2735
*(Released Mon, 18 Feb 2008)*
2736

    
2737
- more tweaks to the disk activation code (especially helpful for DRBD)
2738
- change the default ``gnt-instance list`` output format, now there is
2739
  one combined status field (see the manpage for the exact values this
2740
  field will have)
2741
- some more fixes for the mac export to hooks change
2742
- make Ganeti not break with DRBD 8.2.x (which changed the version
2743
  format in ``/proc/drbd``) (issue 24)
2744
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2745
  template, allowing migration from DRBD0.7+MD to DRBD8
2746

    
2747

    
2748
Version 1.2.2
2749
-------------
2750

    
2751
*(Released Wed, 30 Jan 2008)*
2752

    
2753
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2754
  support (issue 23)
2755
- add command aliases infrastructure and a few aliases
2756
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2757
  man pages and the ``--help`` option of ``gnt-node
2758
  list``/``gnt-instance list``
2759
- fix ``gnt-backup list`` with down nodes (issue 21)
2760
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2761
- fix the dist archive and add a check for including svn/git files in
2762
  the future
2763
- some developer-related changes: improve the burnin and the QA suite,
2764
  add an upload script for testing during development
2765

    
2766

    
2767
Version 1.2.1
2768
-------------
2769

    
2770
*(Released Wed, 16 Jan 2008)*
2771

    
2772
- experimental HVM support, read the install document, section
2773
  "Initializing the cluster"
2774
- allow for the PVM hypervisor per-instance kernel and initrd paths
2775
- add a new command ``gnt-cluster verify-disks`` which uses a new
2776
  algorithm to improve the reconnection of the DRBD pairs if the device
2777
  on the secondary node has gone away
2778
- make logical volume code auto-activate LVs at disk activation time
2779
- slightly improve the speed of activating disks
2780
- allow specification of the MAC address at instance creation time, and
2781
  changing it later via ``gnt-instance modify``
2782
- fix handling of external commands that generate lots of output on
2783
  stderr
2784
- update documentation with regard to minimum version of DRBD8 supported
2785

    
2786

    
2787
Version 1.2.0
2788
-------------
2789

    
2790
*(Released Tue, 4 Dec 2007)*
2791

    
2792
- Log the ``xm create`` output to the node daemon log on failure (to
2793
  help diagnosing the error)
2794
- In debug mode, log all external commands output if failed to the logs
2795
- Change parsing of lvm commands to ignore stderr
2796

    
2797

    
2798
Version 1.2 beta3
2799
-----------------
2800

    
2801
*(Released Wed, 28 Nov 2007)*
2802

    
2803
- Another round of updates to the DRBD 8 code to deal with more failures
2804
  in the replace secondary node operation
2805
- Some more logging of failures in disk operations (lvm, drbd)
2806
- A few documentation updates
2807
- QA updates
2808

    
2809

    
2810
Version 1.2 beta2
2811
-----------------
2812

    
2813
*(Released Tue, 13 Nov 2007)*
2814

    
2815
- Change configuration file format from Python's Pickle to JSON.
2816
  Upgrading is possible using the cfgupgrade utility.
2817
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2818
  faster replace disks and is more stable (DRBD 8 has many improvements
2819
  compared to DRBD 0.7)
2820
- Added command line tags support (see man pages for ``gnt-instance``,
2821
  ``gnt-node``, ``gnt-cluster``)
2822
- Added instance rename support
2823
- Added multi-instance startup/shutdown
2824
- Added cluster rename support
2825
- Added ``gnt-node evacuate`` to simplify some node operations
2826
- Added instance reboot operation that can speedup reboot as compared to
2827
  stop and start
2828
- Soften the requirement that hostnames are in FQDN format
2829
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
2830
  reboots
2831
- Removed dependency on debian's patched fping that uses the
2832
  non-standard ``-S`` option
2833
- Now the OS definitions are searched for in multiple, configurable
2834
  paths (easier for distros to package)
2835
- Some changes to the hooks infrastructure (especially the new
2836
  post-configuration update hook)
2837
- Other small bugfixes
2838

    
2839
.. vim: set textwidth=72 syntax=rst :
2840
.. Local Variables:
2841
.. mode: rst
2842
.. fill-column: 72
2843
.. End: