Statistics
| Branch: | Tag: | Revision:

root / NEWS @ 7165b450

History | View | Annotate | Download (107.5 kB)

1
News
2
====
3

    
4

    
5
Version 2.9.0 beta1
6
-------------------
7

    
8
*(unreleased)*
9

    
10
- DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
11
  the correct command syntax. It is possible to use different DRBD versions
12
  on different nodes as long as they are compatible to each other. This
13
  enables rolling upgrades of DRBD with no downtime. As permanent operation
14
  of different DRBD versions within a node group is discouraged,
15
  ``gnt-cluster verify`` will emit a warning if it detects such a situation.
16
- hroller now also plans for capacity to move non-redundant instances off
17
  any node to be rebooted; the old behavior of completely ignoring any
18
  non-redundant instances can be restored by adding the --ignore-non-redundant
19
  option.
20
- The cluster option '--no-lvm-storage' was removed in favor of the new option
21
  '--enabled-disk-templates'.
22
- On instance creation, disk templates no longer need to be specified
23
  with '-t'. The default disk template will be taken from the list of
24
  enabled disk templates.
25
- New "inst-status-xen" data collector for the monitoring daemon, providing
26
  information about the state of the xen instances on the nodes.
27
- New "lv" data collector for the monitoring daemon, collecting data about the
28
  logical volumes on the nodes, and pairing them with the name of the instances
29
  they belong to.
30
- New "diskstats" data collector, collecting the data from /proc/diskstats and
31
  presenting them over the monitoring daemon interface.
32
- The monitoring daemon is now running as root, in order to be able to collect
33
  information only available to root (such as the state of Xen instances).
34
- The ConfD client is now IPv6 compatible.
35
- File storage is no longer dis/enabled at configure time, but using the
36
  option '--enabled-disk-templates' at cluster initialization and
37
  modification.
38
- The default directory for file storage is not anymore specified at
39
  configure time, but taken from the cluster's configuration. It can be
40
  set at cluster initialization and modification with '--file-storage-dir'.
41
- Cluster verification now includes stricter checks regarding the
42
  default file storage dir. It now checks that the file storage
43
  directory is explicitely allowed in the 'file-storage-paths' file and
44
  that the directory exists on all nodes.
45

    
46

    
47
Version 2.8.0 beta1
48
-------------------
49

    
50
*(Released Mon, 24 Jun 2013)*
51

    
52
Incompatible/important changes
53
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
54

    
55
- Instance policy can contain multiple instance specs, as described in
56
  the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
57
  <design-partitioned>`. As a consequence, it's not possible to partially change
58
  or override instance specs. Bounding specs (min and max) can be specified as a
59
  whole using the new option ``--ipolicy-bounds-specs``, while standard
60
  specs use the new option ``--ipolicy-std-specs``.
61
- The output of the info command of gnt-cluster, gnt-group, gnt-node,
62
  gnt-instance is a valid YAML object.
63
- hail now honors network restrictions when allocating nodes. This led to an
64
  update of the IAllocator protocol. See the IAllocator documentation for
65
  details.
66

    
67
New features
68
~~~~~~~~~~~~
69

    
70
- The :doc:`Remote API <rapi>` daemon now supports a command line flag
71
  to always require authentication, ``--require-authentication``. It can
72
  be specified in ``$sysconfdir/default/ganeti``.
73
- A new cluster attribute 'enabled_disk_templates' is introduced. It will
74
  be used to manage the disk templates to be used by instances in the cluster.
75
  Initially, it will be set to a list that includes plain, drbd, if they were
76
  enabled by specifying a volume group name, and file and sharedfile, if those
77
  were enabled at configure time. Additionally, it will include all disk
78
  templates that are currently used by instances. The order of disk templates
79
  will be based on Ganeti's history of supporting them. In the future, the
80
  first entry of the list will be used as a default disk template on instance
81
  creation.
82
- ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
83
  configuration back to the previous stable version.
84
- Disk templates in group ipolicy can be restored to the default value.
85
- Initial support for diskless instances and virtual clusters in QA.
86
- More QA and unit tests for instance policies.
87
- Every opcode now contains a reason trail (visible through ``gnt-job info``)
88
  describing why the opcode itself was executed.
89
- The monitoring daemon is now available. It allows users to query the cluster
90
  for obtaining information about the status of the system. The daemon is only
91
  responsible for providing the information over the network: the actual data
92
  gathering is performed by data collectors (currently, only the DRBD status
93
  collector is available).
94
- In order to help developers work on Ganeti, a new script
95
  (``devel/build_chroot``) is provided, for building a chroot that contains all
96
  the required development libraries and tools for compiling Ganeti on a Debian
97
  Squeeze system.
98
- A new tool, ``harep``, for performing self-repair and recreation of instances
99
  in Ganeti has been added.
100
- Split queries are enabled for tags, network, exports, cluster info, groups,
101
  jobs, nodes.
102
- New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
103
  It prints the command line to set the current policies, to ease
104
  changing them.
105

    
106
New dependencies
107
~~~~~~~~~~~~~~~~
108
The following new dependencies have been added:
109

    
110
For Haskell:
111
- The ``curl`` library is not optional anymore for compiling the Haskell code.
112
- ``snap-server`` library (if monitoring is enabled).
113

    
114
For Python:
115
- The minimum Python version needed to run Ganeti is now 2.6.
116
- ``yaml`` library (only for running the QA).
117

    
118

    
119
Version 2.7.0
120
-------------
121

    
122
*(Released Thu, 04 Jul 2013)*
123

    
124
Incompatible/important changes
125
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
126

    
127
- Instance policies for disk size were documented to be on a per-disk
128
  basis, but hail applied them to the sum of all disks. This has been
129
  fixed.
130
- ``hbal`` will now exit with status 0 if, during job execution over
131
  LUXI, early exit has been requested and all jobs are successful;
132
  before, exit status 1 was used, which cannot be differentiated from
133
  "job error" case
134
- Compatibility with newer versions of rbd has been fixed
135
- ``gnt-instance batch-create`` has been changed to use the bulk create
136
  opcode from Ganeti. This lead to incompatible changes in the format of
137
  the JSON file. It's now not a custom dict anymore but a dict
138
  compatible with the ``OpInstanceCreate`` opcode.
139
- Parent directories for file storage need to be listed in
140
  ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
141
  write the file automatically based on old configuration values, but it
142
  can not distribute it across all nodes and the file contents should be
143
  verified. Use ``gnt-cluster copyfile
144
  $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
145
  upgraded. The reason for requiring this list of paths now is that
146
  before it would have been possible to inject new paths via RPC,
147
  allowing files to be created in arbitrary locations. The RPC protocol
148
  is protected using SSL/X.509 certificates, but as a design principle
149
  Ganeti does not permit arbitrary paths to be passed.
150
- The parsing of the variants file for OSes (see
151
  :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
152
  lines and comment lines (starting with ``#``) are ignored for better
153
  readability.
154
- The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
155
  longer available. ``gnt-node add`` now invokes a new tool on the
156
  destination node, named ``prepare-node-join``, to configure the SSH
157
  daemon. Paramiko is no longer necessary to configure nodes' SSH
158
  daemons via ``gnt-node add``.
159
- Draining (``gnt-cluster queue drain``) and un-draining the job queue
160
  (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
161
  the flag is not reset after a master failover.
162
- Python 2.4 has *not* been tested with this release. Using 2.6 or above
163
  is recommended. 2.6 will be mandatory from the 2.8 series.
164

    
165

    
166
New features
167
~~~~~~~~~~~~
168

    
169
- New network management functionality to support automatic allocation
170
  of IP addresses and managing of network parameters. See
171
  :manpage:`gnt-network(8)` for more details.
172
- New external storage backend, to allow managing arbitrary storage
173
  systems external to the cluster. See
174
  :manpage:`ganeti-extstorage-interface(7)`.
175
- New ``exclusive-storage`` node parameter added, restricted to
176
  nodegroup level. When it's set to true, physical disks are assigned in
177
  an exclusive fashion to instances, as documented in :doc:`Partitioned
178
  Ganeti <design-partitioned>`.  Currently, only instances using the
179
  ``plain`` disk template are supported.
180
- The KVM hypervisor has been updated with many new hypervisor
181
  parameters, including a generic one for passing arbitrary command line
182
  values. See a complete list in :manpage:`gnt-instance(8)`. It is now
183
  compatible up to qemu 1.4.
184
- A new tool, called ``mon-collector``, is the stand-alone executor of
185
  the data collectors for a monitoring system. As of this version, it
186
  just includes the DRBD data collector, that can be executed by calling
187
  ``mon-collector`` using the ``drbd`` parameter. See
188
  :manpage:`mon-collector(7)`.
189
- A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
190
  for RAPI users. It allows granting permissions to query for
191
  information to a specific user without giving
192
  :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
193
- A new tool named ``node-cleanup`` has been added. It cleans remains of
194
  a cluster from a machine by stopping all daemons, removing
195
  certificates and ssconf files. Unless the ``--no-backup`` option is
196
  given, copies of the certificates are made.
197
- Instance creations now support the use of opportunistic locking,
198
  potentially speeding up the (parallel) creation of multiple instances.
199
  This feature is currently only available via the :doc:`RAPI
200
  <rapi>` interface and when an instance allocator is used. If the
201
  ``opportunistic_locking`` parameter is set the opcode will try to
202
  acquire as many locks as possible, but will not wait for any locks
203
  held by other opcodes. If not enough resources can be found to
204
  allocate the instance, the temporary error code
205
  :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
206
  retried thereafter, with or without opportunistic locking.
207
- New experimental linux-ha resource scripts.
208
- Restricted-commands support: ganeti can now be asked (via command line
209
  or rapi) to perform commands on a node. These are passed via ganeti
210
  RPC rather than ssh. This functionality is restricted to commands
211
  specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
212
  reasons. The file is not copied automatically.
213

    
214

    
215
Misc changes
216
~~~~~~~~~~~~
217

    
218
- Diskless instances are now externally mirrored (Issue 237). This for
219
  now has only been tested in conjunction with explicit target nodes for
220
  migration/failover.
221
- Queries not needing locks or RPC access to the node can now be
222
  performed by the confd daemon, making them independent from jobs, and
223
  thus faster to execute. This is selectable at configure time.
224
- The functionality for allocating multiple instances at once has been
225
  overhauled and is now also available through :doc:`RAPI <rapi>`.
226

    
227
There are no significant changes from version 2.7.0~rc3.
228

    
229

    
230
Version 2.7.0 rc3
231
-----------------
232

    
233
*(Released Tue, 25 Jun 2013)*
234

    
235
- Fix permissions on the confd query socket (Issue 477)
236
- Fix permissions on the job archive dir (Issue 498)
237
- Fix handling of an internal exception in replace-disks (Issue 472)
238
- Fix gnt-node info handling of shortened names (Issue 497)
239
- Fix gnt-instance grow-disk when wiping is enabled
240
- Documentation improvements, and support for newer pandoc
241
- Fix hspace honoring ipolicy for disks (Issue 484)
242
- Improve handling of the ``kvm_extra`` HV parameter
243

    
244

    
245
Version 2.7.0 rc2
246
-----------------
247

    
248
*(Released Fri, 24 May 2013)*
249

    
250
- ``devel/upload`` now works when ``/var/run`` on the target nodes is a
251
  symlink.
252
- Disks added through ``gnt-instance modify`` or created through
253
  ``gnt-instance recreate-disks`` are wiped, if the
254
  ``prealloc_wipe_disks`` flag is set.
255
- If wiping newly created disks fails, the disks are removed. Also,
256
  partial failures in creating disks through ``gnt-instance modify``
257
  triggers a cleanup of the partially-created disks.
258
- Removing the master IP address doesn't fail if the address has been
259
  already removed.
260
- Fix ownership of the OS log dir
261
- Workaround missing SO_PEERCRED constant (Issue 191)
262

    
263

    
264
Version 2.7.0 rc1
265
-----------------
266

    
267
*(Released Fri, 3 May 2013)*
268

    
269
This was the first release candidate of the 2.7 series. Since beta3:
270

    
271
- Fix kvm compatibility with qemu 1.4 (Issue 389)
272
- Documentation updates (admin guide, upgrade notes, install
273
  instructions) (Issue 372)
274
- Fix gnt-group list nodes and instances count (Issue 436)
275
- Fix compilation without non-mandatory libraries (Issue 441)
276
- Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
277
- Make confd logging more verbose at INFO level (Issue 435)
278
- Improve "networks" documentation in :manpage:`gnt-instance(8)`
279
- Fix failure path for instance storage type conversion (Issue 229)
280
- Update htools text backend documentation
281
- Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
282
- Disable inter-cluster instance move for file-based instances, because
283
  it is dependant on instance export, which is not supported for
284
  file-based instances. (Issue 414)
285
- Fix gnt-job crashes on non-ascii characters (Issue 427)
286
- Fix volume group checks on non-vm-capable nodes (Issue 432)
287

    
288

    
289
Version 2.7.0 beta3
290
-------------------
291

    
292
*(Released Mon, 22 Apr 2013)*
293

    
294
This was the third beta release of the 2.7 series. Since beta2:
295

    
296
- Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
297
- Fix data loss on wrong usage of ``gnt-instance move``
298
- Properly export errors in confd-based job queries
299
- Add ``users-setup`` tool
300
- Fix iallocator protocol to report 0 as a disk size for diskless
301
  instances. This avoids hail breaking when a diskless instance is
302
  present.
303
- Fix job queue directory permission problem that made confd job queries
304
  fail. This requires running an ``ensure-dirs --full-run`` on upgrade
305
  for access to archived jobs (Issue 406).
306
- Limit the sizes of networks supported by ``gnt-network`` to something
307
  between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
308
- Fix bugs in instance disk template conversion
309
- Fix GHC 7 compatibility
310
- Fix ``burnin`` install path (Issue 426).
311
- Allow very small disk grows (Issue 347).
312
- Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
313
  that noded doesn't import masterd code (Issue 419).
314
- Make sure the default metavg at cluster init is the same as the vg, if
315
  unspecified (Issue 358).
316
- Fix cleanup of partially created disks (part of Issue 416)
317

    
318

    
319
Version 2.7.0 beta2
320
-------------------
321

    
322
*(Released Tue, 2 Apr 2013)*
323

    
324
This was the second beta release of the 2.7 series. Since beta1:
325

    
326
- Networks no longer have a "type" slot, since this information was
327
  unused in Ganeti: instead of it tags should be used.
328
- The rapi client now has a ``target_node`` option to MigrateInstance.
329
- Fix early exit return code for hbal (Issue 386).
330
- Fix ``gnt-instance migrate/failover -n`` (Issue 396).
331
- Fix ``rbd showmapped`` output parsing (Issue 312).
332
- Networks are now referenced indexed by UUID, rather than name. This
333
  will require running cfgupgrade, from 2.7.0beta1, if networks are in
334
  use.
335
- The OS environment now includes network information.
336
- Deleting of a network is now disallowed if any instance nic is using
337
  it, to prevent dangling references.
338
- External storage is now documented in man pages.
339
- The exclusive_storage flag can now only be set at nodegroup level.
340
- Hbal can now submit an explicit priority with its jobs.
341
- Many network related locking fixes.
342
- Bump up the required pylint version to 0.25.1.
343
- Fix the ``no_remember`` option in RAPI client.
344
- Many ipolicy related tests, qa, and fixes.
345
- Many documentation improvements and fixes.
346
- Fix building with ``--disable-file-storage``.
347
- Fix ``-q`` option in htools, which was broken if passed more than
348
  once.
349
- Some haskell/python interaction improvements and fixes.
350
- Fix iallocator in case of missing LVM storage.
351
- Fix confd config load in case of ``--no-lvm-storage``.
352
- The confd/query functionality is now mentioned in the security
353
  documentation.
354

    
355

    
356
Version 2.7.0 beta1
357
-------------------
358

    
359
*(Released Wed, 6 Feb 2013)*
360

    
361
This was the first beta release of the 2.7 series. All important changes
362
are listed in the latest 2.7 entry.
363

    
364

    
365
Version 2.6.2
366
-------------
367

    
368
*(Released Fri, 21 Dec 2012)*
369

    
370
Important behaviour change: hbal won't rebalance anymore instances which
371
have the ``auto_balance`` attribute set to false. This was the intention
372
all along, but until now it only skipped those from the N+1 memory
373
reservation (DRBD-specific).
374

    
375
A significant number of bug fixes in this release:
376

    
377
- Fixed disk adoption interaction with ipolicy checks.
378
- Fixed networking issues when instances are started, stopped or
379
  migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
380
- Fixed the warning in cluster verify for shared storage instances not
381
  being redundant.
382
- Fixed removal of storage directory on shared file storage (issue 262).
383
- Fixed validation of LVM volume group name in OpClusterSetParams
384
  (``gnt-cluster modify``) (issue 285).
385
- Fixed runtime memory increases (``gnt-instance modify -m``).
386
- Fixed live migration under Xen's ``xl`` mode.
387
- Fixed ``gnt-instance console`` with ``xl``.
388
- Fixed building with newer Haskell compiler/libraries.
389
- Fixed PID file writing in Haskell daemons (confd); this prevents
390
  restart issues if confd was launched manually (outside of
391
  ``daemon-util``) while another copy of it was running
392
- Fixed a type error when doing live migrations with KVM (issue 297) and
393
  the error messages for failing migrations have been improved.
394
- Fixed opcode validation for the out-of-band commands (``gnt-node
395
  power``).
396
- Fixed a type error when unsetting OS hypervisor parameters (issue
397
  311); now it's possible to unset all OS-specific hypervisor
398
  parameters.
399
- Fixed the ``dry-run`` mode for many operations: verification of
400
  results was over-zealous but didn't take into account the ``dry-run``
401
  operation, resulting in "wrong" failures.
402
- Fixed bash completion in ``gnt-job list`` when the job queue has
403
  hundreds of entries; especially with older ``bash`` versions, this
404
  results in significant CPU usage.
405

    
406
And lastly, a few other improvements have been made:
407

    
408
- Added option to force master-failover without voting (issue 282).
409
- Clarified error message on lock conflict (issue 287).
410
- Logging of newly submitted jobs has been improved (issue 290).
411
- Hostname checks have been made uniform between instance rename and
412
  create (issue 291).
413
- The ``--submit`` option is now supported by ``gnt-debug delay``.
414
- Shutting down the master daemon by sending SIGTERM now stops it from
415
  processing jobs waiting for locks; instead, those jobs will be started
416
  once again after the master daemon is started the next time (issue
417
  296).
418
- Support for Xen's ``xl`` program has been improved (besides the fixes
419
  above).
420
- Reduced logging noise in the Haskell confd daemon (only show one log
421
  entry for each config reload, instead of two).
422
- Several man page updates and typo fixes.
423

    
424

    
425
Version 2.6.1
426
-------------
427

    
428
*(Released Fri, 12 Oct 2012)*
429

    
430
A small bugfix release. Among the bugs fixed:
431

    
432
- Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
433
  made the command unusable.
434
- Commands that issue many jobs don't fail anymore just because some jobs
435
  take so long that other jobs are archived.
436
- Failures during ``gnt-instance reinstall`` are reflected by the exit
437
  status.
438
- Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
439
  DRBD is enabled.
440
- When ``always_failover`` is set, ``--allow-failover`` is not required
441
  in migrate commands anymore.
442
- ``bash_completion`` works even if extglob is disabled.
443
- Fixed bug with locks that made failover for RDB-based instances fail.
444
- Fixed bug in non-mirrored instance allocation that made Ganeti choose
445
  a random node instead of one based on the allocator metric.
446
- Support for newer versions of pylint and pep8.
447
- Hail doesn't fail anymore when trying to add an instance of type
448
  ``file``, ``sharedfile`` or ``rbd``.
449
- Added new Makefile target to rebuild the whole distribution, so that
450
  all files are included.
451

    
452

    
453
Version 2.6.0
454
-------------
455

    
456
*(Released Fri, 27 Jul 2012)*
457

    
458

    
459
.. attention:: The ``LUXI`` protocol has been made more consistent
460
   regarding its handling of command arguments. This, however, leads to
461
   incompatibility issues with previous versions. Please ensure that you
462
   restart Ganeti daemons soon after the upgrade, otherwise most
463
   ``LUXI`` calls (job submission, setting/resetting the drain flag,
464
   pausing/resuming the watcher, cancelling and archiving jobs, querying
465
   the cluster configuration) will fail.
466

    
467

    
468
New features
469
~~~~~~~~~~~~
470

    
471
Instance run status
472
+++++++++++++++++++
473

    
474
The current ``admin_up`` field, which used to denote whether an instance
475
should be running or not, has been removed. Instead, ``admin_state`` is
476
introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
477

    
478
The rational behind this is that an instance being “down” can have
479
different meanings:
480

    
481
- it could be down during a reboot
482
- it could be temporarily be down for a reinstall
483
- or it could be down because it is deprecated and kept just for its
484
  disk
485

    
486
The previous Boolean state was making it difficult to do capacity
487
calculations: should Ganeti reserve memory for a down instance? Now, the
488
tri-state field makes it clear:
489

    
490
- in ``up`` and ``down`` state, all resources are reserved for the
491
  instance, and it can be at any time brought up if it is down
492
- in ``offline`` state, only disk space is reserved for it, but not
493
  memory or CPUs
494

    
495
The field can have an extra use: since the transition between ``up`` and
496
``down`` and vice-versus is done via ``gnt-instance start/stop``, but
497
transition between ``offline`` and ``down`` is done via ``gnt-instance
498
modify``, it is possible to given different rights to users. For
499
example, owners of an instance could be allowed to start/stop it, but
500
not transition it out of the offline state.
501

    
502
Instance policies and specs
503
+++++++++++++++++++++++++++
504

    
505
In previous Ganeti versions, an instance creation request was not
506
limited on the minimum size and on the maximum size just by the cluster
507
resources. As such, any policy could be implemented only in third-party
508
clients (RAPI clients, or shell wrappers over ``gnt-*``
509
tools). Furthermore, calculating cluster capacity via ``hspace`` again
510
required external input with regards to instance sizes.
511

    
512
In order to improve these workflows and to allow for example better
513
per-node group differentiation, we introduced instance specs, which
514
allow declaring:
515

    
516
- minimum instance disk size, disk count, memory size, cpu count
517
- maximum values for the above metrics
518
- and “standard” values (used in ``hspace`` to calculate the standard
519
  sized instances)
520

    
521
The minimum/maximum values can be also customised at node-group level,
522
for example allowing more powerful hardware to support bigger instance
523
memory sizes.
524

    
525
Beside the instance specs, there are a few other settings belonging to
526
the instance policy framework. It is possible now to customise, per
527
cluster and node-group:
528

    
529
- the list of allowed disk templates
530
- the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
531
- the maximum ratio of instance to spindles (see below for more
532
  information) for local storage
533

    
534
All these together should allow all tools that talk to Ganeti to know
535
what are the ranges of allowed values for instances and the
536
over-subscription that is allowed.
537

    
538
For the VCPU/PCPU ratio, we already have the VCPU configuration from the
539
instance configuration, and the physical CPU configuration from the
540
node. For the spindle ratios however, we didn't track before these
541
values, so new parameters have been added:
542

    
543
- a new node parameter ``spindle_count``, defaults to 1, customisable at
544
  node group or node level
545
- at new backend parameter (for instances), ``spindle_use`` defaults to 1
546

    
547
Note that spindles in this context doesn't need to mean actual
548
mechanical hard-drives; it's just a relative number for both the node
549
I/O capacity and instance I/O consumption.
550

    
551
Instance migration behaviour
552
++++++++++++++++++++++++++++
553

    
554
While live-migration is in general desirable over failover, it is
555
possible that for some workloads it is actually worse, due to the
556
variable time of the “suspend” phase during live migration.
557

    
558
To allow the tools to work consistently over such instances (without
559
having to hard-code instance names), a new backend parameter
560
``always_failover`` has been added to control the migration/failover
561
behaviour. When set to True, all migration requests for an instance will
562
instead fall-back to failover.
563

    
564
Instance memory ballooning
565
++++++++++++++++++++++++++
566

    
567
Initial support for memory ballooning has been added. The memory for an
568
instance is no longer fixed (backend parameter ``memory``), but instead
569
can vary between minimum and maximum values (backend parameters
570
``minmem`` and ``maxmem``). Currently we only change an instance's
571
memory when:
572

    
573
- live migrating or failing over and instance and the target node
574
  doesn't have enough memory
575
- user requests changing the memory via ``gnt-instance modify
576
  --runtime-memory``
577

    
578
Instance CPU pinning
579
++++++++++++++++++++
580

    
581
In order to control the use of specific CPUs by instance, support for
582
controlling CPU pinning has been added for the Xen, HVM and LXC
583
hypervisors. This is controlled by a new hypervisor parameter
584
``cpu_mask``; details about possible values for this are in the
585
:manpage:`gnt-instance(8)`. Note that use of the most specific (precise
586
VCPU-to-CPU mapping) form will work well only when all nodes in your
587
cluster have the same amount of CPUs.
588

    
589
Disk parameters
590
+++++++++++++++
591

    
592
Another area in which Ganeti was not customisable were the parameters
593
used for storage configuration, e.g. how many stripes to use for LVM,
594
DRBD resync configuration, etc.
595

    
596
To improve this area, we've added disks parameters, which are
597
customisable at cluster and node group level, and which allow to
598
specify various parameters for disks (DRBD has the most parameters
599
currently), for example:
600

    
601
- DRBD resync algorithm and parameters (e.g. speed)
602
- the default VG for meta-data volumes for DRBD
603
- number of stripes for LVM (plain disk template)
604
- the RBD pool
605

    
606
These parameters can be modified via ``gnt-cluster modify -D …`` and
607
``gnt-group modify -D …``, and are used at either instance creation (in
608
case of LVM stripes, for example) or at disk “activation” time
609
(e.g. resync speed).
610

    
611
Rados block device support
612
++++++++++++++++++++++++++
613

    
614
A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
615
denoted by the ``rbd`` disk template type. This is considered
616
experimental, feedback is welcome. For details on configuring it, see
617
the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
618

    
619
Master IP setup
620
+++++++++++++++
621

    
622
The existing master IP functionality works well only in simple setups (a
623
single network shared by all nodes); however, if nodes belong to
624
different networks, then the ``/32`` setup and lack of routing
625
information is not enough.
626

    
627
To allow the master IP to function well in more complex cases, the
628
system was reworked as follows:
629

    
630
- a master IP netmask setting has been added
631
- the master IP activation/turn-down code was moved from the node daemon
632
  to a separate script
633
- whether to run the Ganeti-supplied master IP script or a user-supplied
634
  on is a ``gnt-cluster init`` setting
635

    
636
Details about the location of the standard and custom setup scripts are
637
in the man page :manpage:`gnt-cluster(8)`; for information about the
638
setup script protocol, look at the Ganeti-supplied script.
639

    
640
SPICE support
641
+++++++++++++
642

    
643
The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
644
improved.
645

    
646
It is now possible to use TLS-protected connections, and when renewing
647
or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
648
it is now possible to specify spice or spice CA certificates. Also, it
649
is possible to configure a password for SPICE sessions via the
650
hypervisor parameter ``spice_password_file``.
651

    
652
There are also new parameters to control the compression and streaming
653
options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
654
etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
655
for the spice parameters.
656

    
657
Lastly, it is now possible to see the SPICE connection information via
658
``gnt-instance console``.
659

    
660
OVF converter
661
+++++++++++++
662

    
663
A new tool (``tools/ovfconverter``) has been added that supports
664
conversion between Ganeti and the `Open Virtualization Format
665
<http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
666
from).
667

    
668
This relies on the ``qemu-img`` tool to convert the disk formats, so the
669
actual compatibility with other virtualization solutions depends on it.
670

    
671
Confd daemon changes
672
++++++++++++++++++++
673

    
674
The configuration query daemon (``ganeti-confd``) is now optional, and
675
has been rewritten in Haskell; whether to use the daemon at all, use the
676
Python (default) or the Haskell version is selectable at configure time
677
via the ``--enable-confd`` parameter, which can take one of the
678
``haskell``, ``python`` or ``no`` values. If not used, disabling the
679
daemon will result in a smaller footprint; for larger systems, we
680
welcome feedback on the Haskell version which might become the default
681
in future versions.
682

    
683
If you want to use ``gnt-node list-drbd`` you need to have the Haskell
684
daemon running. The Python version doesn't implement the new call.
685

    
686

    
687
User interface changes
688
~~~~~~~~~~~~~~~~~~~~~~
689

    
690
We have replaced the ``--disks`` option of ``gnt-instance
691
replace-disks`` with a more flexible ``--disk`` option, which allows
692
adding and removing disks at arbitrary indices (Issue 188). Furthermore,
693
disk size and mode can be changed upon recreation (via ``gnt-instance
694
recreate-disks``, which accepts the same ``--disk`` option).
695

    
696
As many people are used to a ``show`` command, we have added that as an
697
alias to ``info`` on all ``gnt-*`` commands.
698

    
699
The ``gnt-instance grow-disk`` command has a new mode in which it can
700
accept the target size of the disk, instead of the delta; this can be
701
more safe since two runs in absolute mode will be idempotent, and
702
sometimes it's also easier to specify the desired size directly.
703

    
704
Also the handling of instances with regard to offline secondaries has
705
been improved. Instance operations should not fail because one of it's
706
secondary nodes is offline, even though it's safe to proceed.
707

    
708
A new command ``list-drbd`` has been added to the ``gnt-node`` script to
709
support debugging of DRBD issues on nodes. It provides a mapping of DRBD
710
minors to instance name.
711

    
712
API changes
713
~~~~~~~~~~~
714

    
715
RAPI coverage has improved, with (for example) new resources for
716
recreate-disks, node power-cycle, etc.
717

    
718
Compatibility
719
~~~~~~~~~~~~~
720

    
721
There is partial support for ``xl`` in the Xen hypervisor; feedback is
722
welcome.
723

    
724
Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
725
whether to still support Python 2.4 or move to Python 2.6 as minimum
726
required version.
727

    
728
Support for Fedora has been slightly improved; the provided example
729
init.d script should work better on it and the INSTALL file should
730
document the needed dependencies.
731

    
732
Internal changes
733
~~~~~~~~~~~~~~~~
734

    
735
The deprecated ``QueryLocks`` LUXI request has been removed. Use
736
``Query(what=QR_LOCK, ...)`` instead.
737

    
738
The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
739
:pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
740
:pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
741
:pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
742
future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
743

    
744
RAPI client: ``CertificateError`` now derives from
745
``GanetiApiError``. This should make it more easy to handle Ganeti
746
errors.
747

    
748
Deprecation warnings due to PyCrypto/paramiko import in
749
``tools/setup-ssh`` have been silenced, as usually they are safe; please
750
make sure to run an up-to-date paramiko version, if you use this tool.
751

    
752
The QA scripts now depend on Python 2.5 or above (the main code base
753
still works with Python 2.4).
754

    
755
The configuration file (``config.data``) is now written without
756
indentation for performance reasons; if you want to edit it, it can be
757
re-formatted via ``tools/fmtjson``.
758

    
759
A number of bugs has been fixed in the cluster merge tool.
760

    
761
``x509`` certification verification (used in import-export) has been
762
changed to allow the same clock skew as permitted by the cluster
763
verification. This will remove some rare but hard to diagnose errors in
764
import-export.
765

    
766

    
767
Version 2.6.0 rc4
768
-----------------
769

    
770
*(Released Thu, 19 Jul 2012)*
771

    
772
Very few changes from rc4 to the final release, only bugfixes:
773

    
774
- integrated fixes from release 2.5.2 (fix general boot flag for KVM
775
  instance, fix CDROM booting for KVM instances)
776
- fixed node group modification of node parameters
777
- fixed issue in LUClusterVerifyGroup with multi-group clusters
778
- fixed generation of bash completion to ensure a stable ordering
779
- fixed a few typos
780

    
781

    
782
Version 2.6.0 rc3
783
-----------------
784

    
785
*(Released Fri, 13 Jul 2012)*
786

    
787
Third release candidate for 2.6. The following changes were done from
788
rc3 to rc4:
789

    
790
- Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
791
- Fixed an inconsistency in the LUXI protocol with the provided
792
  arguments (NOT backwards compatible)
793
- Fixed a bug with node groups ipolicy where ``min`` was greater than
794
  the cluster ``std`` value
795
- Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
796
  easier instance debugging on nodes (requires ``hconfd`` to work)
797

    
798

    
799
Version 2.6.0 rc2
800
-----------------
801

    
802
*(Released Tue, 03 Jul 2012)*
803

    
804
Second release candidate for 2.6. The following changes were done from
805
rc2 to rc3:
806

    
807
- Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
808
  master candidates
809
- Fixed a RAPI regression on missing beparams/memory
810
- Fixed redistribution of files on offline nodes
811
- Added possibility to run activate-disks even though secondaries are
812
  offline. With this change it relaxes also the strictness on some other
813
  commands which use activate disks internally:
814
  * ``gnt-instance start|reboot|rename|backup|export``
815
- Made it possible to remove safely an instance if its secondaries are
816
  offline
817
- Made it possible to reinstall even though secondaries are offline
818

    
819

    
820
Version 2.6.0 rc1
821
-----------------
822

    
823
*(Released Mon, 25 Jun 2012)*
824

    
825
First release candidate for 2.6. The following changes were done from
826
rc1 to rc2:
827

    
828
- Fixed bugs with disk parameters and ``rbd`` templates as well as
829
  ``instance_os_add``
830
- Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
831
  behaviour. It supports now the modify operation
832
- ``hcheck`` implemented to analyze cluster health and possibility of
833
  improving health by rebalance
834
- ``hbal`` has been improved in dealing with split instances
835

    
836

    
837
Version 2.6.0 beta2
838
-------------------
839

    
840
*(Released Mon, 11 Jun 2012)*
841

    
842
Second beta release of 2.6. The following changes were done from beta2
843
to rc1:
844

    
845
- Fixed ``daemon-util`` with non-root user models
846
- Fixed creation of plain instances with ``--no-wait-for-sync``
847
- Fix wrong iv_names when running ``cfgupgrade``
848
- Export more information in RAPI group queries
849
- Fixed bug when changing instance network interfaces
850
- Extended burnin to do NIC changes
851
- query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
852
- Changed default for DRBD barriers
853
- Fixed DRBD error reporting for syncer rate
854
- Verify the options on disk parameters
855

    
856
And of course various fixes to documentation and improved unittests and
857
QA.
858

    
859

    
860
Version 2.6.0 beta1
861
-------------------
862

    
863
*(Released Wed, 23 May 2012)*
864

    
865
First beta release of 2.6. The following changes were done from beta1 to
866
beta2:
867

    
868
- integrated patch for distributions without ``start-stop-daemon``
869
- adapted example init.d script to work on Fedora
870
- fixed log handling in Haskell daemons
871
- adapted checks in the watcher for pycurl linked against libnss
872
- add partial support for ``xl`` instead of ``xm`` for Xen
873
- fixed a type issue in cluster verification
874
- fixed ssconf handling in the Haskell code (was breaking confd in IPv6
875
  clusters)
876

    
877
Plus integrated fixes from the 2.5 branch:
878

    
879
- fixed ``kvm-ifup`` to use ``/bin/bash``
880
- fixed parallel build failures
881
- KVM live migration when using a custom keymap
882

    
883

    
884
Version 2.5.2
885
-------------
886

    
887
*(Released Tue, 24 Jul 2012)*
888

    
889
A small bugfix release, with no new features:
890

    
891
- fixed bash-isms in kvm-ifup, for compatibility with systems which use a
892
  different default shell (e.g. Debian, Ubuntu)
893
- fixed KVM startup and live migration with a custom keymap (fixes Issue
894
  243 and Debian bug #650664)
895
- fixed compatibility with KVM versions that don't support multiple boot
896
  devices (fixes Issue 230 and Debian bug #624256)
897

    
898
Additionally, a few fixes were done to the build system (fixed parallel
899
build failures) and to the unittests (fixed race condition in test for
900
FileID functions, and the default enable/disable mode for QA test is now
901
customisable).
902

    
903

    
904
Version 2.5.1
905
-------------
906

    
907
*(Released Fri, 11 May 2012)*
908

    
909
A small bugfix release.
910

    
911
The main issues solved are on the topic of compatibility with newer LVM
912
releases:
913

    
914
- fixed parsing of ``lv_attr`` field
915
- adapted to new ``vgreduce --removemissing`` behaviour where sometimes
916
  the ``--force`` flag is needed
917

    
918
Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
919
to accept kernel 3.x too (was hardcoded to 2.6.*).
920

    
921
A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
922
of hook results and that also made display of other errors suboptimal
923
was fixed; the code behaves now like 2.4 and earlier.
924

    
925
Another change in 2.5, the cleanup of the OS scripts environment, is too
926
aggressive: it removed even the ``PATH`` variable, which requires the OS
927
scripts to *always* need to export it. Since this is a bit too strict,
928
we now export a minimal PATH, the same that we export for hooks.
929

    
930
The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
931
integrated into this release.
932

    
933
Finally, a few other miscellaneous changes were done (no new features,
934
just small improvements):
935

    
936
- Fix ``gnt-group --help`` display
937
- Fix hardcoded Xen kernel path
938
- Fix grow-disk handling of invalid units
939
- Update synopsis for ``gnt-cluster repair-disk-sizes``
940
- Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
941

    
942

    
943
Version 2.5.0
944
-------------
945

    
946
*(Released Thu, 12 Apr 2012)*
947

    
948
Incompatible/important changes and bugfixes
949
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
950

    
951
- The default of the ``/2/instances/[instance_name]/rename`` RAPI
952
  resource's ``ip_check`` parameter changed from ``True`` to ``False``
953
  to match the underlying LUXI interface.
954
- The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
955
  body parameters, see :doc:`RAPI documentation <rapi>`. The server does
956
  not maintain backwards-compatibility as the underlying operation
957
  changed in an incompatible way. The RAPI client can talk to old
958
  servers, but it needs to be told so as the return value changed.
959
- When creating file-based instances via RAPI, the ``file_driver``
960
  parameter no longer defaults to ``loop`` and must be specified.
961
- The deprecated ``bridge`` NIC parameter is no longer supported. Use
962
  ``link`` instead.
963
- Support for the undocumented and deprecated RAPI instance creation
964
  request format version 0 has been dropped. Use version 1, supported
965
  since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
966
- Pyparsing 1.4.6 or above is required, see :doc:`installation
967
  documentation <install>`.
968
- The "cluster-verify" hooks are now executed per group by the
969
  ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
970
  if you just run ``gnt-cluster verify``, which generates one opcode per
971
  group.
972
- The environment as passed to the OS scripts is cleared, and thus no
973
  environment variables defined in the node daemon's environment will be
974
  inherited by the scripts.
975
- The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
976
  deprecated.
977
- :doc:`New iallocator modes <design-multi-reloc>` have been added to
978
  support operations involving multiple node groups.
979
- Offline nodes are ignored when failing over an instance.
980
- Support for KVM version 1.0, which changed the version reporting format
981
  from 3 to 2 digits.
982
- TCP/IP ports used by DRBD disks are returned to a pool upon instance
983
  removal.
984
- ``Makefile`` is now compatible with Automake 1.11.2
985
- Includes all bugfixes made in the 2.4 series
986

    
987
New features
988
~~~~~~~~~~~~
989

    
990
- The ganeti-htools project has been merged into the ganeti-core source
991
  tree and will be built as part of Ganeti (see :doc:`install-quick`).
992
- Implemented support for :doc:`shared storage <design-shared-storage>`.
993
- Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
994
  GPT-style partition tables (requires `parted
995
  <http://www.gnu.org/s/parted/>`_).
996
- Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
997
- Allowed adding tags on instance creation.
998
- Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
999
- Allow instances to be started in a paused state, enabling the user to
1000
  see the complete console output on boot using the console.
1001
- Added new hypervisor flag to control default reboot behaviour
1002
  (``reboot_behavior``).
1003
- Added support for KVM keymaps (hypervisor parameter ``keymap``).
1004
- Improved out-of-band management support:
1005

    
1006
  - Added ``gnt-node health`` command reporting the health status of
1007
    nodes.
1008
  - Added ``gnt-node power`` command to manage power status of nodes.
1009
  - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1010

    
1011
- Instance migration can fall back to failover if instance is not
1012
  running.
1013
- Filters can be used when listing nodes, instances, groups and locks;
1014
  see :manpage:`ganeti(7)` manpage.
1015
- Added post-execution status as variables to :doc:`hooks <hooks>`
1016
  environment.
1017
- Instance tags are exported/imported together with the instance.
1018
- When given an explicit job ID, ``gnt-job info`` will work for archived
1019
  jobs.
1020
- Jobs can define dependencies on other jobs (not yet supported via
1021
  RAPI or command line, but used by internal commands and usable via
1022
  LUXI).
1023

    
1024
  - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1025
    dependencies.
1026

    
1027
- Instance failover is now available as a RAPI resource
1028
  (``/2/instances/[instance_name]/failover``).
1029
- ``gnt-instance info`` defaults to static information if primary node
1030
  is offline.
1031
- Opcodes have a new ``comment`` attribute.
1032
- Added basic SPICE support to KVM hypervisor.
1033
- ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1034

    
1035
Node group improvements
1036
~~~~~~~~~~~~~~~~~~~~~~~
1037

    
1038
- ``gnt-cluster verify`` has been modified to check groups separately,
1039
  thereby improving performance.
1040
- Node group support has been added to ``gnt-cluster verify-disks``,
1041
  which now operates per node group.
1042
- Watcher has been changed to work better with node groups.
1043

    
1044
  - One process and state file per node group.
1045
  - Slow watcher in one group doesn't block other group's watcher.
1046

    
1047
- Added new command, ``gnt-group evacuate``, to move all instances in a
1048
  node group to other groups.
1049
- Added ``gnt-instance change-group`` to move an instance to another
1050
  node group.
1051
- ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1052
  per-group operations.
1053
- Node groups can be tagged.
1054
- Some operations switch from an exclusive to a shared lock as soon as
1055
  possible.
1056
- Instance's primary and secondary nodes' groups are now available as
1057
  query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1058
  and ``snodes.group.uuid``).
1059

    
1060
Misc
1061
~~~~
1062

    
1063
- Numerous updates to documentation and manpages.
1064

    
1065
  - :doc:`RAPI <rapi>` documentation now has detailed parameter
1066
    descriptions.
1067
  - Some opcode/job results are now also documented, see :doc:`RAPI
1068
    <rapi>`.
1069

    
1070
- A lockset's internal lock is now also visible in lock monitor.
1071
- Log messages from job queue workers now contain information about the
1072
  opcode they're processing.
1073
- ``gnt-instance console`` no longer requires the instance lock.
1074
- A short delay when waiting for job changes reduces the number of LUXI
1075
  requests significantly.
1076
- DRBD metadata volumes are overwritten with zeros during disk creation.
1077
- Out-of-band commands no longer acquire the cluster lock in exclusive
1078
  mode.
1079
- ``devel/upload`` now uses correct permissions for directories.
1080

    
1081

    
1082
Version 2.5.0 rc6
1083
-----------------
1084

    
1085
*(Released Fri, 23 Mar 2012)*
1086

    
1087
This was the sixth release candidate of the 2.5 series.
1088

    
1089

    
1090
Version 2.5.0 rc5
1091
-----------------
1092

    
1093
*(Released Mon, 9 Jan 2012)*
1094

    
1095
This was the fifth release candidate of the 2.5 series.
1096

    
1097

    
1098
Version 2.5.0 rc4
1099
-----------------
1100

    
1101
*(Released Thu, 27 Oct 2011)*
1102

    
1103
This was the fourth release candidate of the 2.5 series.
1104

    
1105

    
1106
Version 2.5.0 rc3
1107
-----------------
1108

    
1109
*(Released Wed, 26 Oct 2011)*
1110

    
1111
This was the third release candidate of the 2.5 series.
1112

    
1113

    
1114
Version 2.5.0 rc2
1115
-----------------
1116

    
1117
*(Released Tue, 18 Oct 2011)*
1118

    
1119
This was the second release candidate of the 2.5 series.
1120

    
1121

    
1122
Version 2.5.0 rc1
1123
-----------------
1124

    
1125
*(Released Tue, 4 Oct 2011)*
1126

    
1127
This was the first release candidate of the 2.5 series.
1128

    
1129

    
1130
Version 2.5.0 beta3
1131
-------------------
1132

    
1133
*(Released Wed, 31 Aug 2011)*
1134

    
1135
This was the third beta release of the 2.5 series.
1136

    
1137

    
1138
Version 2.5.0 beta2
1139
-------------------
1140

    
1141
*(Released Mon, 22 Aug 2011)*
1142

    
1143
This was the second beta release of the 2.5 series.
1144

    
1145

    
1146
Version 2.5.0 beta1
1147
-------------------
1148

    
1149
*(Released Fri, 12 Aug 2011)*
1150

    
1151
This was the first beta release of the 2.5 series.
1152

    
1153

    
1154
Version 2.4.5
1155
-------------
1156

    
1157
*(Released Thu, 27 Oct 2011)*
1158

    
1159
- Fixed bug when parsing command line parameter values ending in
1160
  backslash
1161
- Fixed assertion error after unclean master shutdown
1162
- Disable HTTP client pool for RPC, significantly reducing memory usage
1163
  of master daemon
1164
- Fixed queue archive creation with wrong permissions
1165

    
1166

    
1167
Version 2.4.4
1168
-------------
1169

    
1170
*(Released Tue, 23 Aug 2011)*
1171

    
1172
Small bug-fixes:
1173

    
1174
- Fixed documentation for importing with ``--src-dir`` option
1175
- Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1176
- Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1177

    
1178

    
1179
Version 2.4.3
1180
-------------
1181

    
1182
*(Released Fri, 5 Aug 2011)*
1183

    
1184
Many bug-fixes and a few small features:
1185

    
1186
- Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1187
  issues when you tried to add an instance with two MAC addresses in one
1188
  request
1189
- KVM: fixed per-instance stored UID value
1190
- KVM: configure bridged NICs at migration start
1191
- KVM: Fix a bug where instance will not start with never KVM versions
1192
  (>= 0.14)
1193
- Added OS search path to ``gnt-cluster info``
1194
- Fixed an issue with ``file_storage_dir`` where you were forced to
1195
  provide an absolute path, but the documentation states it is a
1196
  relative path, the documentation was right
1197
- Added a new parameter to instance stop/start called ``--no-remember``
1198
  that will make the state change to not be remembered
1199
- Implemented ``no_remember`` at RAPI level
1200
- Improved the documentation
1201
- Node evacuation: don't call IAllocator if node is already empty
1202
- Fixed bug in DRBD8 replace disks on current nodes
1203
- Fixed bug in recreate-disks for DRBD instances
1204
- Moved assertion checking locks in ``gnt-instance replace-disks``
1205
  causing it to abort with not owning the right locks for some situation
1206
- Job queue: Fixed potential race condition when cancelling queued jobs
1207
- Fixed off-by-one bug in job serial generation
1208
- ``gnt-node volumes``: Fix instance names
1209
- Fixed aliases in bash completion
1210
- Fixed a bug in reopening log files after being sent a SIGHUP
1211
- Added a flag to burnin to allow specifying VCPU count
1212
- Bugfixes to non-root Ganeti configuration
1213

    
1214

    
1215
Version 2.4.2
1216
-------------
1217

    
1218
*(Released Thu, 12 May 2011)*
1219

    
1220
Many bug-fixes and a few new small features:
1221

    
1222
- Fixed a bug related to log opening failures
1223
- Fixed a bug in instance listing with orphan instances
1224
- Fixed a bug which prevented resetting the cluster-level node parameter
1225
  ``oob_program`` to the default
1226
- Many fixes related to the ``cluster-merge`` tool
1227
- Fixed a race condition in the lock monitor, which caused failures
1228
  during (at least) creation of many instances in parallel
1229
- Improved output for gnt-job info
1230
- Removed the quiet flag on some ssh calls which prevented debugging
1231
  failures
1232
- Improved the N+1 failure messages in cluster verify by actually
1233
  showing the memory values (needed and available)
1234
- Increased lock attempt timeouts so that when executing long operations
1235
  (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1236
  too early and thus prevent the use of the 'fair' mechanism
1237
- Changed instance query data (``gnt-instance info``) to not acquire
1238
  locks unless needed, thus allowing its use on locked instance if only
1239
  static information is asked for
1240
- Improved behaviour with filesystems that do not support rename on an
1241
  opened file
1242
- Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1243
  kept locks on all nodes during the wipe, which is unneeded
1244
- Fixed ``gnt-watcher`` handling of errors during hooks execution
1245
- Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1246
  10GiB) which caused the wipe to fail right at the end in some cases
1247
- Fixed master IP activation when doing master failover with no-voting
1248
- Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1249
  the master node itself
1250
- Fixed potential data-loss in under disk full conditions, where Ganeti
1251
  wouldn't check correctly the return code and would consider
1252
  partially-written files 'correct'
1253
- Fixed bug related to multiple VGs and DRBD disk replacing
1254
- Added new disk parameter ``metavg`` that allows placement of the meta
1255
  device for DRBD in a different volume group
1256
- Fixed error handling in the node daemon when the system libc doesn't
1257
  have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1258
- Fixed lock release during replace-disks, which kept cluster-wide locks
1259
  when doing disk replaces with an iallocator script
1260
- Added check for missing bridges in cluster verify
1261
- Handle EPIPE errors while writing to the terminal better, so that
1262
  piping the output to e.g. ``less`` doesn't cause a backtrace
1263
- Fixed rare case where a ^C during Luxi calls could have been
1264
  interpreted as server errors, instead of simply terminating
1265
- Fixed a race condition in LUGroupAssignNodes (``gnt-group
1266
  assign-nodes``)
1267
- Added a few more parameters to the KVM hypervisor, allowing a second
1268
  CDROM, custom disk type for CDROMs and a floppy image
1269
- Removed redundant message in instance rename when the name is given
1270
  already as a FQDN
1271
- Added option to ``gnt-instance recreate-disks`` to allow creating the
1272
  disks on new nodes, allowing recreation when the original instance
1273
  nodes are completely gone
1274
- Added option when converting disk templates to DRBD to skip waiting
1275
  for the resync, in order to make the instance available sooner
1276
- Added two new variables to the OS scripts environment (containing the
1277
  instance's nodes)
1278
- Made the root_path and optional parameter for the xen-pvm hypervisor,
1279
  to allow use of ``pvgrub`` as bootloader
1280
- Changed the instance memory modifications to only check out-of-memory
1281
  conditions on memory increases, and turned the secondary node warnings
1282
  into errors (they can still be overridden via ``--force``)
1283
- Fixed the handling of a corner case when the Python installation gets
1284
  corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1285
  to execute a command that doesn't exist
1286
- Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1287
  node of the instance returned failures during instance shutdown; this
1288
  adds the option ``--ignore-consistency`` to gnt-instance move
1289

    
1290
And as usual, various improvements to the error messages, documentation
1291
and man pages.
1292

    
1293

    
1294
Version 2.4.1
1295
-------------
1296

    
1297
*(Released Wed, 09 Mar 2011)*
1298

    
1299
Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1300
the RAPI users file if run twice (even with ``--dry-run``).
1301

    
1302
The release fixes that bug (nothing else changed).
1303

    
1304

    
1305
Version 2.4.0
1306
-------------
1307

    
1308
*(Released Mon, 07 Mar 2011)*
1309

    
1310
Final 2.4.0 release. Just a few small fixes:
1311

    
1312
- Fixed RAPI node evacuate
1313
- Fixed the kvm-ifup script
1314
- Fixed internal error handling for special job cases
1315
- Updated man page to specify the escaping feature for options
1316

    
1317

    
1318
Version 2.4.0 rc3
1319
-----------------
1320

    
1321
*(Released Mon, 28 Feb 2011)*
1322

    
1323
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1324
that this feature wiped the disks of the wrong instance, leading to loss
1325
of data.
1326

    
1327
Other changes:
1328

    
1329
- Fixed title of query field containing instance name
1330
- Expanded the glossary in the documentation
1331
- Fixed one unittest (internal issue)
1332

    
1333

    
1334
Version 2.4.0 rc2
1335
-----------------
1336

    
1337
*(Released Mon, 21 Feb 2011)*
1338

    
1339
A number of bug fixes plus just a couple functionality changes.
1340

    
1341
On the user-visible side, the ``gnt-* list`` command output has changed
1342
with respect to "special" field states. The current rc1 style of display
1343
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1344
the default output mode special fields states are displayed as follows:
1345

    
1346
- Offline resource: ``*``
1347
- Unavailable/not applicable: ``-``
1348
- Data missing (RPC failure): ``?``
1349
- Unknown field: ``??``
1350

    
1351
Another user-visible change is the addition of ``--force-join`` to
1352
``gnt-node add``.
1353

    
1354
As for bug fixes:
1355

    
1356
- ``tools/cluster-merge`` has seen many fixes and is now enabled again
1357
- Fixed regression in RAPI/instance reinstall where all parameters were
1358
  required (instead of optional)
1359
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1360
- Fixed iallocator usage (offline nodes were not considered offline)
1361
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1362
- Fixed hypervisor and OS parameter validation with respect to
1363
  non-vm_capable nodes
1364
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1365
  cosmetic)
1366
- Fixed ``tools/listrunner`` with respect to agent-based usage
1367

    
1368

    
1369
Version 2.4.0 rc1
1370
-----------------
1371

    
1372
*(Released Fri,  4 Feb 2011)*
1373

    
1374
Many changes and fixes since the beta1 release. While there were some
1375
internal changes, the code has been mostly stabilised for the RC
1376
release.
1377

    
1378
Note: the dumb allocator was removed in this release, as it was not kept
1379
up-to-date with the IAllocator protocol changes. It is recommended to
1380
use the ``hail`` command from the ganeti-htools package.
1381

    
1382
Note: the 2.4 and up versions of Ganeti are not compatible with the
1383
0.2.x branch of ganeti-htools. You need to upgrade to
1384
ganeti-htools-0.3.0 (or later).
1385

    
1386
Regressions fixed from 2.3
1387
~~~~~~~~~~~~~~~~~~~~~~~~~~
1388

    
1389
- Fixed the ``gnt-cluster verify-disks`` command
1390
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1391
  serially on nodes)
1392
- Fixed disk adoption breakage
1393
- Fixed wrong headers in instance listing for field aliases
1394

    
1395
Other bugs fixed
1396
~~~~~~~~~~~~~~~~
1397

    
1398
- Fixed corner case in KVM handling of NICs
1399
- Fixed many cases of wrong handling of non-vm_capable nodes
1400
- Fixed a bug where a missing instance symlink was not possible to
1401
  recreate with any ``gnt-*`` command (now ``gnt-instance
1402
  activate-disks`` does it)
1403
- Fixed the volume group name as reported by ``gnt-cluster
1404
  verify-disks``
1405
- Increased timeouts for the import-export code, hopefully leading to
1406
  fewer aborts due network or instance timeouts
1407
- Fixed bug in ``gnt-node list-storage``
1408
- Fixed bug where not all daemons were started on cluster
1409
  initialisation, but only at the first watcher run
1410
- Fixed many bugs in the OOB implementation
1411
- Fixed watcher behaviour in presence of instances with offline
1412
  secondaries
1413
- Fixed instance list output for instances running on the wrong node
1414
- a few fixes to the cluster-merge tool, but it still cannot merge
1415
  multi-node groups (currently it is not recommended to use this tool)
1416

    
1417

    
1418
Improvements
1419
~~~~~~~~~~~~
1420

    
1421
- Improved network configuration for the KVM hypervisor
1422
- Added e1000 as a supported NIC for Xen-HVM
1423
- Improved the lvmstrap tool to also be able to use partitions, as
1424
  opposed to full disks
1425
- Improved speed of disk wiping (the cluster parameter
1426
  ``prealloc_wipe_disks``, so that it has a low impact on the total time
1427
  of instance creations
1428
- Added documentation for the OS parameters
1429
- Changed ``gnt-instance deactivate-disks`` so that it can work if the
1430
  hypervisor is not responding
1431
- Added display of blacklisted and hidden OS information in
1432
  ``gnt-cluster info``
1433
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1434
  NIC and node parameters, which might create problems with currently
1435
  invalid (but undetected) configuration files, but prevents validation
1436
  failures when unrelated parameters are modified
1437
- Changed cluster initialisation to wait for the master daemon to become
1438
  available
1439
- Expanded the RAPI interface:
1440

    
1441
  - Added config redistribution resource
1442
  - Added activation/deactivation of instance disks
1443
  - Added export of console information
1444

    
1445
- Implemented log file reopening on SIGHUP, which allows using
1446
  logrotate(8) for the Ganeti log files
1447
- Added a basic OOB helper script as an example
1448

    
1449

    
1450
Version 2.4.0 beta1
1451
-------------------
1452

    
1453
*(Released Fri, 14 Jan 2011)*
1454

    
1455
User-visible
1456
~~~~~~~~~~~~
1457

    
1458
- Fixed timezone issues when formatting timestamps
1459
- Added support for node groups, available via ``gnt-group`` and other
1460
  commands
1461
- Added out-of-band framework and management, see :doc:`design
1462
  document <design-oob>`
1463
- Removed support for roman numbers from ``gnt-node list`` and
1464
  ``gnt-instance list``.
1465
- Allowed modification of master network interface via ``gnt-cluster
1466
  modify --master-netdev``
1467
- Accept offline secondaries while shutting down instance disks
1468
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1469
- Added support for multiple LVM volume groups
1470
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1471
  requested
1472
- Added commands to list available fields:
1473

    
1474
  - ``gnt-node list-fields``
1475
  - ``gnt-group list-fields``
1476
  - ``gnt-instance list-fields``
1477

    
1478
- Updated documentation and man pages
1479

    
1480
Integration
1481
~~~~~~~~~~~
1482

    
1483
- Moved ``rapi_users`` file into separate directory, now named
1484
  ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1485
  symlink
1486
- Added new tool for running commands on many machines,
1487
  ``tools/ganeti-listrunner``
1488
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1489
  improving the ``gnt-instance console`` output
1490
- Allowed customisation of disk index separator at ``configure`` time
1491
- Export node group allocation policy to :doc:`iallocator <iallocator>`
1492
- Added support for non-partitioned md disks in ``lvmstrap``
1493
- Added script to gracefully power off KVM instances
1494
- Split ``utils`` module into smaller parts
1495
- Changed query operations to return more detailed information, e.g.
1496
  whether an information is unavailable due to an offline node. To use
1497
  this new functionality, the LUXI call ``Query`` must be used. Field
1498
  information is now stored by the master daemon and can be retrieved
1499
  using ``QueryFields``. Instances, nodes and groups can also be queried
1500
  using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1501
  exposed via RAPI). The following commands make use of this
1502
  infrastructure change:
1503

    
1504
  - ``gnt-group list``
1505
  - ``gnt-group list-fields``
1506
  - ``gnt-node list``
1507
  - ``gnt-node list-fields``
1508
  - ``gnt-instance list``
1509
  - ``gnt-instance list-fields``
1510
  - ``gnt-debug locks``
1511

    
1512
Remote API
1513
~~~~~~~~~~
1514

    
1515
- New RAPI resources (see :doc:`rapi`):
1516

    
1517
  - ``/2/modify``
1518
  - ``/2/groups``
1519
  - ``/2/groups/[group_name]``
1520
  - ``/2/groups/[group_name]/assign-nodes``
1521
  - ``/2/groups/[group_name]/modify``
1522
  - ``/2/groups/[group_name]/rename``
1523
  - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1524

    
1525
- RAPI changes:
1526

    
1527
  - Implemented ``no_install`` for instance creation
1528
  - Implemented OS parameters for instance reinstallation, allowing
1529
    use of special settings on reinstallation (e.g. for preserving data)
1530

    
1531
Misc
1532
~~~~
1533

    
1534
- Added IPv6 support in import/export
1535
- Pause DRBD synchronization while wiping disks on instance creation
1536
- Updated unittests and QA scripts
1537
- Improved network parameters passed to KVM
1538
- Converted man pages from docbook to reStructuredText
1539

    
1540

    
1541
Version 2.3.1
1542
-------------
1543

    
1544
*(Released Mon, 20 Dec 2010)*
1545

    
1546
Released version 2.3.1~rc1 without any changes.
1547

    
1548

    
1549
Version 2.3.1 rc1
1550
-----------------
1551

    
1552
*(Released Wed, 1 Dec 2010)*
1553

    
1554
- impexpd: Disable OpenSSL compression in socat if possible (backport
1555
  from master, commit e90739d625b, see :doc:`installation guide
1556
  <install-quick>` for details)
1557
- Changed unittest coverage report to exclude test scripts
1558
- Added script to check version format
1559

    
1560

    
1561
Version 2.3.0
1562
-------------
1563

    
1564
*(Released Wed, 1 Dec 2010)*
1565

    
1566
Released version 2.3.0~rc1 without any changes.
1567

    
1568

    
1569
Version 2.3.0 rc1
1570
-----------------
1571

    
1572
*(Released Fri, 19 Nov 2010)*
1573

    
1574
A number of bugfixes and documentation updates:
1575

    
1576
- Update ganeti-os-interface documentation
1577
- Fixed a bug related to duplicate MACs or similar items which should be
1578
  unique
1579
- Fix breakage in OS state modify
1580
- Reinstall instance: disallow offline secondaries (fixes bug related to
1581
  OS changing but reinstall failing)
1582
- plus all the other fixes between 2.2.1 and 2.2.2
1583

    
1584

    
1585
Version 2.3.0 rc0
1586
-----------------
1587

    
1588
*(Released Tue, 2 Nov 2010)*
1589

    
1590
- Fixed clearing of the default iallocator using ``gnt-cluster modify``
1591
- Fixed master failover race with watcher
1592
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1593
  configuration
1594
- Accept previously stopped instance for export with instance removal
1595
- Simplify and extend the environment variables for instance OS scripts
1596
- Added new node flags, ``master_capable`` and ``vm_capable``
1597
- Added optional instance disk wiping prior during allocation. This is a
1598
  cluster-wide option and can be set/modified using
1599
  ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1600
- Added IPv6 support, see :doc:`design document <design-2.3>` and
1601
  :doc:`install-quick`
1602
- Added a new watcher option (``--ignore-pause``)
1603
- Added option to ignore offline node on instance start/stop
1604
  (``--ignore-offline``)
1605
- Allow overriding OS parameters with ``gnt-instance reinstall``
1606
- Added ability to change node's secondary IP address using ``gnt-node
1607
  modify``
1608
- Implemented privilege separation for all daemons except
1609
  ``ganeti-noded``, see ``configure`` options
1610
- Complain if an instance's disk is marked faulty in ``gnt-cluster
1611
  verify``
1612
- Implemented job priorities (see ``ganeti(7)`` manpage)
1613
- Ignore failures while shutting down instances during failover from
1614
  offline node
1615
- Exit daemon's bootstrap process only once daemon is ready
1616
- Export more information via ``LUInstanceQuery``/remote API
1617
- Improved documentation, QA and unittests
1618
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1619
  restart if the file was created or changed
1620
- Added LUXI protocol version sent with each request and response,
1621
  allowing detection of server/client mismatches
1622
- Moved the Python scripts among gnt-* and ganeti-* into modules
1623
- Moved all code related to setting up SSH to an external script,
1624
  ``setup-ssh``
1625
- Infrastructure changes for node group support in future versions
1626

    
1627

    
1628
Version 2.2.2
1629
-------------
1630

    
1631
*(Released Fri, 19 Nov 2010)*
1632

    
1633
A few small bugs fixed, and some improvements to the build system:
1634

    
1635
- Fix documentation regarding conversion to drbd
1636
- Fix validation of parameters in cluster modify (``gnt-cluster modify
1637
  -B``)
1638
- Fix error handling in node modify with multiple changes
1639
- Allow remote imports without checked names
1640

    
1641

    
1642
Version 2.2.1
1643
-------------
1644

    
1645
*(Released Tue, 19 Oct 2010)*
1646

    
1647
- Disable SSL session ID cache in RPC client
1648

    
1649

    
1650
Version 2.2.1 rc1
1651
-----------------
1652

    
1653
*(Released Thu, 14 Oct 2010)*
1654

    
1655
- Fix interaction between Curl/GnuTLS and the Python's HTTP server
1656
  (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1657
  with GnuTLS
1658
- Fix problems with interaction between Curl and Python's HTTP server,
1659
  resulting in increased speed in many RPC calls
1660
- Improve our release script to prevent breakage with older aclocal and
1661
  Python 2.6
1662

    
1663

    
1664
Version 2.2.1 rc0
1665
-----------------
1666

    
1667
*(Released Thu, 7 Oct 2010)*
1668

    
1669
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1670
  value retrieved from master
1671
- Added support for blacklisted or hidden OS definitions
1672
- Added simple lock monitor (accessible via (``gnt-debug locks``)
1673
- Added support for -mem-path in KVM hypervisor abstraction layer
1674
- Allow overriding instance parameters in tool for inter-cluster
1675
  instance moves (``tools/move-instance``)
1676
- Improved opcode summaries (e.g. in ``gnt-job list``)
1677
- Improve consistency of OS listing by sorting it
1678
- Documentation updates
1679

    
1680

    
1681
Version 2.2.0.1
1682
---------------
1683

    
1684
*(Released Fri, 8 Oct 2010)*
1685

    
1686
- Rebuild with a newer autotools version, to fix python 2.6 compatibility
1687

    
1688

    
1689
Version 2.2.0
1690
-------------
1691

    
1692
*(Released Mon, 4 Oct 2010)*
1693

    
1694
- Fixed regression in ``gnt-instance rename``
1695

    
1696

    
1697
Version 2.2.0 rc2
1698
-----------------
1699

    
1700
*(Released Wed, 22 Sep 2010)*
1701

    
1702
- Fixed OS_VARIANT variable for OS scripts
1703
- Fixed cluster tag operations via RAPI
1704
- Made ``setup-ssh`` exit with non-zero code if an error occurred
1705
- Disabled RAPI CA checks in watcher
1706

    
1707

    
1708
Version 2.2.0 rc1
1709
-----------------
1710

    
1711
*(Released Mon, 23 Aug 2010)*
1712

    
1713
- Support DRBD versions of the format "a.b.c.d"
1714
- Updated manpages
1715
- Re-introduce support for usage from multiple threads in RAPI client
1716
- Instance renames and modify via RAPI
1717
- Work around race condition between processing and archival in job
1718
  queue
1719
- Mark opcodes following failed one as failed, too
1720
- Job field ``lock_status`` was removed due to difficulties making it
1721
  work with the changed job queue in Ganeti 2.2; a better way to monitor
1722
  locks is expected for a later 2.2.x release
1723
- Fixed dry-run behaviour with many commands
1724
- Support ``ssh-agent`` again when adding nodes
1725
- Many additional bugfixes
1726

    
1727

    
1728
Version 2.2.0 rc0
1729
-----------------
1730

    
1731
*(Released Fri, 30 Jul 2010)*
1732

    
1733
Important change: the internal RPC mechanism between Ganeti nodes has
1734
changed from using a home-grown http library (based on the Python base
1735
libraries) to use the PycURL library. This requires that PycURL is
1736
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1737
against GnuTLS by default. cURL's support for GnuTLS had known issues
1738
before cURL 7.21.0 and we recommend using the latest cURL release or
1739
linking against OpenSSL. Most other distributions already link PycURL
1740
and cURL against OpenSSL. The command::
1741

    
1742
  python -c 'import pycurl; print pycurl.version'
1743

    
1744
can be used to determine the libraries PycURL and cURL are linked
1745
against.
1746

    
1747
Other significant changes:
1748

    
1749
- Rewrote much of the internals of the job queue, in order to achieve
1750
  better parallelism; this decouples job query operations from the job
1751
  processing, and it should allow much nicer behaviour of the master
1752
  daemon under load, and it also has uncovered some long-standing bugs
1753
  related to the job serialisation (now fixed)
1754
- Added a default iallocator setting to the cluster parameters,
1755
  eliminating the need to always pass nodes or an iallocator for
1756
  operations that require selection of new node(s)
1757
- Added experimental support for the LXC virtualization method
1758
- Added support for OS parameters, which allows the installation of
1759
  instances to pass parameter to OS scripts in order to customise the
1760
  instance
1761
- Added a hypervisor parameter controlling the migration type (live or
1762
  non-live), since hypervisors have various levels of reliability; this
1763
  has renamed the 'live' parameter to 'mode'
1764
- Added a cluster parameter ``reserved_lvs`` that denotes reserved
1765
  logical volumes, meaning that cluster verify will ignore them and not
1766
  flag their presence as errors
1767
- The watcher will now reset the error count for failed instances after
1768
  8 hours, thus allowing self-healing if the problem that caused the
1769
  instances to be down/fail to start has cleared in the meantime
1770
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1771
  check for, and warn, if the drbd module parameter ``usermode_helper``
1772
  is not consistent with the cluster-wide setting; this is needed to
1773
  make diagnose easier of failed drbd creations
1774
- Started adding base IPv6 support, but this is not yet
1775
  enabled/available for use
1776
- Rename operations (cluster, instance) will now return the new name,
1777
  which is especially useful if a short name was passed in
1778
- Added support for instance migration in RAPI
1779
- Added a tool to pre-configure nodes for the SSH setup, before joining
1780
  them to the cluster; this will allow in the future a simplified model
1781
  for node joining (but not yet fully enabled in 2.2); this needs the
1782
  paramiko python library
1783
- Fixed handling of name-resolving errors
1784
- Fixed consistency of job results on the error path
1785
- Fixed master-failover race condition when executed multiple times in
1786
  sequence
1787
- Fixed many bugs related to the job queue (mostly introduced during the
1788
  2.2 development cycle, so not all are impacting 2.1)
1789
- Fixed instance migration with missing disk symlinks
1790
- Fixed handling of unknown jobs in ``gnt-job archive``
1791
- And many other small fixes/improvements
1792

    
1793
Internal changes:
1794

    
1795
- Enhanced both the unittest and the QA coverage
1796
- Switched the opcode validation to a generic model, and extended the
1797
  validation to all opcode parameters
1798
- Changed more parts of the code that write shell scripts to use the
1799
  same class for this
1800
- Switched the master daemon to use the asyncore library for the Luxi
1801
  server endpoint
1802

    
1803

    
1804
Version 2.2.0 beta0
1805
-------------------
1806

    
1807
*(Released Thu, 17 Jun 2010)*
1808

    
1809
- Added tool (``move-instance``) and infrastructure to move instances
1810
  between separate clusters (see :doc:`separate documentation
1811
  <move-instance>` and :doc:`design document <design-2.2>`)
1812
- Added per-request RPC timeout
1813
- RAPI now requires a Content-Type header for requests with a body (e.g.
1814
  ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1815
  :rfc:`2616` (HTTP/1.1), section 7.2.1)
1816
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1817
  reachable
1818
- Implemented initial support for running Ganeti daemons as separate
1819
  users, see configure-time flags ``--with-user-prefix`` and
1820
  ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1821
  time)
1822
- Instances can be removed after export (``gnt-backup export
1823
  --remove-instance``)
1824
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1825
  key (instead of 1024 bit)
1826
- Added new cluster configuration file for cluster domain secret
1827
- Import/export now use SSL instead of SSH
1828
- Added support for showing estimated time when exporting an instance,
1829
  see the ``ganeti-os-interface(7)`` manpage and look for
1830
  ``EXP_SIZE_FD``
1831

    
1832

    
1833
Version 2.1.8
1834
-------------
1835

    
1836
*(Released Tue, 16 Nov 2010)*
1837

    
1838
Some more bugfixes. Unless critical bugs occur, this will be the last
1839
2.1 release:
1840

    
1841
- Fix case of MAC special-values
1842
- Fix mac checker regex
1843
- backend: Fix typo causing "out of range" error
1844
- Add missing --units in gnt-instance list man page
1845

    
1846

    
1847
Version 2.1.7
1848
-------------
1849

    
1850
*(Released Tue, 24 Aug 2010)*
1851

    
1852
Bugfixes only:
1853
  - Don't ignore secondary node silently on non-mirrored disk templates
1854
    (issue 113)
1855
  - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1856
  - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1857
  - Properly document the usb_mouse parameter
1858
  - Fix path in ganeti-rapi(8) (issue 116)
1859
  - Adjust error message when the ganeti user's .ssh directory is
1860
    missing
1861
  - Add same-node-check when changing the disk template to drbd
1862

    
1863

    
1864
Version 2.1.6
1865
-------------
1866

    
1867
*(Released Fri, 16 Jul 2010)*
1868

    
1869
Bugfixes only:
1870
  - Add an option to only select some reboot types during qa/burnin.
1871
    (on some hypervisors consequent reboots are not supported)
1872
  - Fix infrequent race condition in master failover. Sometimes the old
1873
    master ip address would be still detected as up for a short time
1874
    after it was removed, causing failover to fail.
1875
  - Decrease mlockall warnings when the ctypes module is missing. On
1876
    Python 2.4 we support running even if no ctypes module is installed,
1877
    but we were too verbose about this issue.
1878
  - Fix building on old distributions, on which man doesn't have a
1879
    --warnings option.
1880
  - Fix RAPI not to ignore the MAC address on instance creation
1881
  - Implement the old instance creation format in the RAPI client.
1882

    
1883

    
1884
Version 2.1.5
1885
-------------
1886

    
1887
*(Released Thu, 01 Jul 2010)*
1888

    
1889
A small bugfix release:
1890
  - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1891
  - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1892
    a non-existing option
1893
  - Fix instance create: the --force-variant option was ignored
1894
  - Improve pylint 0.21 compatibility and warnings with Python 2.6
1895
  - Fix modify node storage with non-FQDN arguments
1896
  - Fix RAPI client to authenticate under Python 2.6 when used
1897
    for more than 5 requests needing authentication
1898
  - Fix gnt-instance modify -t (storage) giving a wrong error message
1899
    when converting a non-shutdown drbd instance to plain
1900

    
1901

    
1902
Version 2.1.4
1903
-------------
1904

    
1905
*(Released Fri, 18 Jun 2010)*
1906

    
1907
A small bugfix release:
1908

    
1909
  - Fix live migration of KVM instances started with older Ganeti
1910
    versions which had fewer hypervisor parameters
1911
  - Fix gnt-instance grow-disk on down instances
1912
  - Fix an error-reporting bug during instance migration
1913
  - Better checking of the ``--net`` and ``--disk`` values, to avoid
1914
    silently ignoring broken ones
1915
  - Fix an RPC error reporting bug affecting, for example, RAPI client
1916
    users
1917
  - Fix bug triggered by different API version os-es on different nodes
1918
  - Fix a bug in instance startup with custom hvparams: OS level
1919
    parameters would fail to be applied.
1920
  - Fix the RAPI client under Python 2.6 (but more work is needed to
1921
    make it work completely well with OpenSSL)
1922
  - Fix handling of errors when resolving names from DNS
1923

    
1924

    
1925
Version 2.1.3
1926
-------------
1927

    
1928
*(Released Thu, 3 Jun 2010)*
1929

    
1930
A medium sized development cycle. Some new features, and some
1931
fixes/small improvements/cleanups.
1932

    
1933
Significant features
1934
~~~~~~~~~~~~~~~~~~~~
1935

    
1936
The node deamon now tries to mlock itself into memory, unless the
1937
``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1938
its logs, and falls back to console logging. This allows emergency
1939
features such as ``gnt-node powercycle`` to work even in the event of a
1940
broken node disk (tested offlining the disk hosting the node's
1941
filesystem and dropping its memory caches; don't try this at home)
1942

    
1943
KVM: add vhost-net acceleration support. It can be tested with a new
1944
enough version of the kernel and of qemu-kvm.
1945

    
1946
KVM: Add instance chrooting feature. If you use privilege dropping for
1947
your VMs you can also now force them to chroot to an empty directory,
1948
before starting the emulated guest.
1949

    
1950
KVM: Add maximum migration bandwith and maximum downtime tweaking
1951
support (requires a new-enough version of qemu-kvm).
1952

    
1953
Cluster verify will now warn if the master node doesn't have the master
1954
ip configured on it.
1955

    
1956
Add a new (incompatible) instance creation request format to RAPI which
1957
supports all parameters (previously only a subset was supported, and it
1958
wasn't possible to extend the old format to accomodate all the new
1959
features. The old format is still supported, and a client can check for
1960
this feature, before using it, by checking for its presence in the
1961
``features`` RAPI resource.
1962

    
1963
Now with ancient latin support. Try it passing the ``--roman`` option to
1964
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1965
(requires the python-roman module to be installed, in order to work).
1966

    
1967
Other changes
1968
~~~~~~~~~~~~~
1969

    
1970
As usual many internal code refactorings, documentation updates, and
1971
such. Among others:
1972

    
1973
  - Lots of improvements and cleanups to the experimental Remote API
1974
    (RAPI) client library.
1975
  - A new unit test suite for the core daemon libraries.
1976
  - A fix to creating missing directories makes sure the umask is not
1977
    applied anymore. This enforces the same directory permissions
1978
    everywhere.
1979
  - Better handling terminating daemons with ctrl+c (used when running
1980
    them in debugging mode).
1981
  - Fix a race condition in live migrating a KVM instance, when stat()
1982
    on the old proc status file returned EINVAL, which is an unexpected
1983
    value.
1984
  - Fixed manpage checking with newer man and utf-8 charachters. But now
1985
    you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1986

    
1987

    
1988
Version 2.1.2.1
1989
---------------
1990

    
1991
*(Released Fri, 7 May 2010)*
1992

    
1993
Fix a bug which prevented untagged KVM instances from starting.
1994

    
1995

    
1996
Version 2.1.2
1997
-------------
1998

    
1999
*(Released Fri, 7 May 2010)*
2000

    
2001
Another release with a long development cycle, during which many
2002
different features were added.
2003

    
2004
Significant features
2005
~~~~~~~~~~~~~~~~~~~~
2006

    
2007
The KVM hypervisor now can run the individual instances as non-root, to
2008
reduce the impact of a VM being hijacked due to bugs in the
2009
hypervisor. It is possible to run all instances as a single (non-root)
2010
user, to manually specify a user for each instance, or to dynamically
2011
allocate a user out of a cluster-wide pool to each instance, with the
2012
guarantee that no two instances will run under the same user ID on any
2013
given node.
2014

    
2015
An experimental RAPI client library, that can be used standalone
2016
(without the other Ganeti libraries), is provided in the source tree as
2017
``lib/rapi/client.py``. Note this client might change its interface in
2018
the future, as we iterate on its capabilities.
2019

    
2020
A new command, ``gnt-cluster renew-crypto`` has been added to easily
2021
replace the cluster's certificates and crypto keys. This might help in
2022
case they have been compromised, or have simply expired.
2023

    
2024
A new disk option for instance creation has been added that allows one
2025
to "adopt" currently existing logical volumes, with data
2026
preservation. This should allow easier migration to Ganeti from
2027
unmanaged (or managed via other software) instances.
2028

    
2029
Another disk improvement is the possibility to convert between redundant
2030
(DRBD) and plain (LVM) disk configuration for an instance. This should
2031
allow better scalability (starting with one node and growing the
2032
cluster, or shrinking a two-node cluster to one node).
2033

    
2034
A new feature that could help with automated node failovers has been
2035
implemented: if a node sees itself as offline (by querying the master
2036
candidates), it will try to shutdown (hard) all instances and any active
2037
DRBD devices. This reduces the risk of duplicate instances if an
2038
external script automatically failovers the instances on such nodes. To
2039
enable this, the cluster parameter ``maintain_node_health`` should be
2040
enabled; in the future this option (per the name) will enable other
2041
automatic maintenance features.
2042

    
2043
Instance export/import now will reuse the original instance
2044
specifications for all parameters; that means exporting an instance,
2045
deleting it and the importing it back should give an almost identical
2046
instance. Note that the default import behaviour has changed from
2047
before, where it created only one NIC; now it recreates the original
2048
number of NICs.
2049

    
2050
Cluster verify has added a few new checks: SSL certificates validity,
2051
/etc/hosts consistency across the cluster, etc.
2052

    
2053
Other changes
2054
~~~~~~~~~~~~~
2055

    
2056
As usual, many internal changes were done, documentation fixes,
2057
etc. Among others:
2058

    
2059
- Fixed cluster initialization with disabled cluster storage (regression
2060
  introduced in 2.1.1)
2061
- File-based storage supports growing the disks
2062
- Fixed behaviour of node role changes
2063
- Fixed cluster verify for some corner cases, plus a general rewrite of
2064
  cluster verify to allow future extension with more checks
2065
- Fixed log spamming by watcher and node daemon (regression introduced
2066
  in 2.1.1)
2067
- Fixed possible validation issues when changing the list of enabled
2068
  hypervisors
2069
- Fixed cleanup of /etc/hosts during node removal
2070
- Fixed RAPI response for invalid methods
2071
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2072
- Multiple small improvements to the KVM hypervisor (VNC usage, booting
2073
  from ide disks, etc.)
2074
- Allow OS changes without re-installation (to record a changed OS
2075
  outside of Ganeti, or to allow OS renames)
2076
- Allow instance creation without OS installation (useful for example if
2077
  the OS will be installed manually, or restored from a backup not in
2078
  Ganeti format)
2079
- Implemented option to make cluster ``copyfile`` use the replication
2080
  network
2081
- Added list of enabled hypervisors to ssconf (possibly useful for
2082
  external scripts)
2083
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2084
  1.2 clusters
2085
- A partial form of node re-IP is possible via node readd, which now
2086
  allows changed node primary IP
2087
- Command line utilities now show an informational message if the job is
2088
  waiting for a lock
2089
- The logs of the master daemon now show the PID/UID/GID of the
2090
  connected client
2091

    
2092

    
2093
Version 2.1.1
2094
-------------
2095

    
2096
*(Released Fri, 12 Mar 2010)*
2097

    
2098
During the 2.1.0 long release candidate cycle, a lot of improvements and
2099
changes have accumulated with were released later as 2.1.1.
2100

    
2101
Major changes
2102
~~~~~~~~~~~~~
2103

    
2104
The node evacuate command (``gnt-node evacuate``) was significantly
2105
rewritten, and as such the IAllocator protocol was changed - a new
2106
request type has been added. This unfortunate change during a stable
2107
series is designed to improve performance of node evacuations; on
2108
clusters with more than about five nodes and which are well-balanced,
2109
evacuation should proceed in parallel for all instances of the node
2110
being evacuated. As such, any existing IAllocator scripts need to be
2111
updated, otherwise the above command will fail due to the unknown
2112
request. The provided "dumb" allocator has not been updated; but the
2113
ganeti-htools package supports the new protocol since version 0.2.4.
2114

    
2115
Another important change is increased validation of node and instance
2116
names. This might create problems in special cases, if invalid host
2117
names are being used.
2118

    
2119
Also, a new layer of hypervisor parameters has been added, that sits at
2120
OS level between the cluster defaults and the instance ones. This allows
2121
customisation of virtualization parameters depending on the installed
2122
OS. For example instances with OS 'X' may have a different KVM kernel
2123
(or any other parameter) than the cluster defaults. This is intended to
2124
help managing a multiple OSes on the same cluster, without manual
2125
modification of each instance's parameters.
2126

    
2127
A tool for merging clusters, ``cluster-merge``, has been added in the
2128
tools sub-directory.
2129

    
2130
Bug fixes
2131
~~~~~~~~~
2132

    
2133
- Improved the int/float conversions that should make the code more
2134
  robust in face of errors from the node daemons
2135
- Fixed the remove node code in case of internal configuration errors
2136
- Fixed the node daemon behaviour in face of inconsistent queue
2137
  directory (e.g. read-only file-system where we can't open the files
2138
  read-write, etc.)
2139
- Fixed the behaviour of gnt-node modify for master candidate demotion;
2140
  now it either aborts cleanly or, if given the new "auto_promote"
2141
  parameter, will automatically promote other nodes as needed
2142
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2143
  completely prevent Ganeti from working
2144
- Fixed bug for instance export when not all disks were successfully
2145
  exported
2146
- Fixed behaviour of node add when the new node is slow in starting up
2147
  the node daemon
2148
- Fixed handling of signals in the LUXI client, which should improve
2149
  behaviour of command-line scripts
2150
- Added checks for invalid node/instance names in the configuration (now
2151
  flagged during cluster verify)
2152
- Fixed watcher behaviour for disk activation errors
2153
- Fixed two potentially endless loops in http library, which led to the
2154
  RAPI daemon hanging and consuming 100% CPU in some cases
2155
- Fixed bug in RAPI daemon related to hashed passwords
2156
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2157
  instances
2158
- Enhanced compatibility with non-Debian OSes, but not using absolute
2159
  path in some commands and allowing customisation of the ssh
2160
  configuration directory
2161
- Fixed possible future issue with new Python versions by abiding to the
2162
  proper use of ``__slots__`` attribute on classes
2163
- Added checks that should prevent directory traversal attacks
2164
- Many documentation fixes based on feedback from users
2165

    
2166
New features
2167
~~~~~~~~~~~~
2168

    
2169
- Added an "early_release" more for instance replace disks and node
2170
  evacuate, where we release locks earlier and thus allow higher
2171
  parallelism within the cluster
2172
- Added watcher hooks, intended to allow the watcher to restart other
2173
  daemons (e.g. from the ganeti-nbma project), but they can be used of
2174
  course for any other purpose
2175
- Added a compile-time disable for DRBD barriers, to increase
2176
  performance if the administrator trusts the power supply or the
2177
  storage system to not lose writes
2178
- Added the option of using syslog for logging instead of, or in
2179
  addition to, Ganeti's own log files
2180
- Removed boot restriction for paravirtual NICs for KVM, recent versions
2181
  can indeed boot from a paravirtual NIC
2182
- Added a generic debug level for many operations; while this is not
2183
  used widely yet, it allows one to pass the debug value all the way to
2184
  the OS scripts
2185
- Enhanced the hooks environment for instance moves (failovers,
2186
  migrations) where the primary/secondary nodes changed during the
2187
  operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2188
- Enhanced data validations for many user-supplied values; one important
2189
  item is the restrictions imposed on instance and node names, which
2190
  might reject some (invalid) host names
2191
- Add a configure-time option to disable file-based storage, if it's not
2192
  needed; this allows greater security separation between the master
2193
  node and the other nodes from the point of view of the inter-node RPC
2194
  protocol
2195
- Added user notification in interactive tools if job is waiting in the
2196
  job queue or trying to acquire locks
2197
- Added log messages when a job is waiting for locks
2198
- Added filtering by node tags in instance operations which admit
2199
  multiple instances (start, stop, reboot, reinstall)
2200
- Added a new tool for cluster mergers, ``cluster-merge``
2201
- Parameters from command line which are of the form ``a=b,c=d`` can now
2202
  use backslash escapes to pass in values which contain commas,
2203
  e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2204
  ``b,c``
2205
- For KVM, the instance name is the first parameter passed to KVM, so
2206
  that it's more visible in the process list
2207

    
2208

    
2209
Version 2.1.0
2210
-------------
2211

    
2212
*(Released Tue, 2 Mar 2010)*
2213

    
2214
Ganeti 2.1 brings many improvements with it. Major changes:
2215

    
2216
- Added infrastructure to ease automated disk repairs
2217
- Added new daemon to export configuration data in a cheaper way than
2218
  using the remote API
2219
- Instance NICs can now be routed instead of being associated with a
2220
  networking bridge
2221
- Improved job locking logic to reduce impact of jobs acquiring multiple
2222
  locks waiting for other long-running jobs
2223

    
2224
In-depth implementation details can be found in the Ganeti 2.1 design
2225
document.
2226

    
2227
Details
2228
~~~~~~~
2229

    
2230
- Added chroot hypervisor
2231
- Added more options to xen-hvm hypervisor (``kernel_path`` and
2232
  ``device_model``)
2233
- Added more options to xen-pvm hypervisor (``use_bootloader``,
2234
  ``bootloader_path`` and ``bootloader_args``)
2235
- Added the ``use_localtime`` option for the xen-hvm and kvm
2236
  hypervisors, and the default value for this has changed to false (in
2237
  2.0 xen-hvm always enabled it)
2238
- Added luxi call to submit multiple jobs in one go
2239
- Added cluster initialization option to not modify ``/etc/hosts``
2240
  file on nodes
2241
- Added network interface parameters
2242
- Added dry run mode to some LUs
2243
- Added RAPI resources:
2244

    
2245
  - ``/2/instances/[instance_name]/info``
2246
  - ``/2/instances/[instance_name]/replace-disks``
2247
  - ``/2/nodes/[node_name]/evacuate``
2248
  - ``/2/nodes/[node_name]/migrate``
2249
  - ``/2/nodes/[node_name]/role``
2250
  - ``/2/nodes/[node_name]/storage``
2251
  - ``/2/nodes/[node_name]/storage/modify``
2252
  - ``/2/nodes/[node_name]/storage/repair``
2253

    
2254
- Added OpCodes to evacuate or migrate all instances on a node
2255
- Added new command to list storage elements on nodes (``gnt-node
2256
  list-storage``) and modify them (``gnt-node modify-storage``)
2257
- Added new ssconf files with master candidate IP address
2258
  (``ssconf_master_candidates_ips``), node primary IP address
2259
  (``ssconf_node_primary_ips``) and node secondary IP address
2260
  (``ssconf_node_secondary_ips``)
2261
- Added ``ganeti-confd`` and a client library to query the Ganeti
2262
  configuration via UDP
2263
- Added ability to run hooks after cluster initialization and before
2264
  cluster destruction
2265
- Added automatic mode for disk replace (``gnt-instance replace-disks
2266
  --auto``)
2267
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2268
  after catastrophic data-loss
2269
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2270
  groups
2271
- Added ``gnt-instance move`` command to move instances
2272
- Added ``gnt-cluster watcher`` command to control watcher
2273
- Added ``gnt-node powercycle`` command to powercycle nodes
2274
- Added new job status field ``lock_status``
2275
- Added parseable error codes to cluster verification (``gnt-cluster
2276
  verify --error-codes``) and made output less verbose (use
2277
  ``--verbose`` to restore previous behaviour)
2278
- Added UUIDs to the main config entities (cluster, nodes, instances)
2279
- Added support for OS variants
2280
- Added support for hashed passwords in the Ganeti remote API users file
2281
  (``rapi_users``)
2282
- Added option to specify maximum timeout on instance shutdown
2283
- Added ``--no-ssh-init`` option to ``gnt-cluster init``
2284
- Added new helper script to start and stop Ganeti daemons
2285
  (``daemon-util``), with the intent to reduce the work necessary to
2286
  adjust Ganeti for non-Debian distributions and to start/stop daemons
2287
  from one place
2288
- Added more unittests
2289
- Fixed critical bug in ganeti-masterd startup
2290
- Removed the configure-time ``kvm-migration-port`` parameter, this is
2291
  now customisable at the cluster level for both the KVM and Xen
2292
  hypervisors using the new ``migration_port`` parameter
2293
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2294
  reinstalling an instance
2295
- Allowed ``@`` in tag names
2296
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2297
- Many documentation updates
2298
- Distribute hypervisor files on ``gnt-cluster redist-conf``
2299
- ``gnt-instance reinstall`` can now reinstall multiple instances
2300
- Updated many command line parameters
2301
- Introduced new OS API version 15
2302
- No longer support a default hypervisor
2303
- Treat virtual LVs as inexistent
2304
- Improved job locking logic to reduce lock contention
2305
- Match instance and node names case insensitively
2306
- Reimplemented bash completion script to be more complete
2307
- Improved burnin
2308

    
2309

    
2310
Version 2.0.6
2311
-------------
2312

    
2313
*(Released Thu, 4 Feb 2010)*
2314

    
2315
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2316
- Fix a string formatting bug
2317
- Improve safety of the code in some error paths
2318
- Improve data validation in the master of values returned from nodes
2319

    
2320

    
2321
Version 2.0.5
2322
-------------
2323

    
2324
*(Released Thu, 17 Dec 2009)*
2325

    
2326
- Fix security issue due to missing validation of iallocator names; this
2327
  allows local and remote execution of arbitrary executables
2328
- Fix failure of gnt-node list during instance removal
2329
- Ship the RAPI documentation in the archive
2330

    
2331

    
2332
Version 2.0.4
2333
-------------
2334

    
2335
*(Released Wed, 30 Sep 2009)*
2336

    
2337
- Fixed many wrong messages
2338
- Fixed a few bugs related to the locking library
2339
- Fixed MAC checking at instance creation time
2340
- Fixed a DRBD parsing bug related to gaps in /proc/drbd
2341
- Fixed a few issues related to signal handling in both daemons and
2342
  scripts
2343
- Fixed the example startup script provided
2344
- Fixed insserv dependencies in the example startup script (patch from
2345
  Debian)
2346
- Fixed handling of drained nodes in the iallocator framework
2347
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2348
  #528618)
2349
- Fixed error related to invalid job IDs in job polling
2350
- Fixed job/opcode persistence on unclean master shutdown
2351
- Fixed handling of partial job processing after unclean master
2352
  shutdown
2353
- Fixed error reporting from LUs, previously all errors were converted
2354
  into execution errors
2355
- Fixed error reporting from burnin
2356
- Decreased significantly the memory usage of the job queue
2357
- Optimised slightly multi-job submission
2358
- Optimised slightly opcode loading
2359
- Backported the multi-job submit framework from the development
2360
  branch; multi-instance start and stop should be faster
2361
- Added script to clean archived jobs after 21 days; this will reduce
2362
  the size of the queue directory
2363
- Added some extra checks in disk size tracking
2364
- Added an example ethers hook script
2365
- Added a cluster parameter that prevents Ganeti from modifying of
2366
  /etc/hosts
2367
- Added more node information to RAPI responses
2368
- Added a ``gnt-job watch`` command that allows following the ouput of a
2369
  job
2370
- Added a bind-address option to ganeti-rapi
2371
- Added more checks to the configuration verify
2372
- Enhanced the burnin script such that some operations can be retried
2373
  automatically
2374
- Converted instance reinstall to multi-instance model
2375

    
2376

    
2377
Version 2.0.3
2378
-------------
2379

    
2380
*(Released Fri, 7 Aug 2009)*
2381

    
2382
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2383
  to allow using the pre-2.0.2 behaviour in activation, if any existing
2384
  instances have mismatched disk sizes in the configuration
2385
- Added ``gnt-cluster repair-disk-sizes`` command to check and update
2386
  any configuration mismatches for disk sizes
2387
- Added ``gnt-master cluste-failover --no-voting`` to allow master
2388
  failover to work on two-node clusters
2389
- Fixed the ``--net`` option of ``gnt-backup import``, which was
2390
  unusable
2391
- Fixed detection of OS script errors in ``gnt-backup export``
2392
- Fixed exit code of ``gnt-backup export``
2393

    
2394

    
2395
Version 2.0.2
2396
-------------
2397

    
2398
*(Released Fri, 17 Jul 2009)*
2399

    
2400
- Added experimental support for stripped logical volumes; this should
2401
  enhance performance but comes with a higher complexity in the block
2402
  device handling; stripping is only enabled when passing
2403
  ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2404
  affected even in the non-stripped mode
2405
- Improved resiliency against transient failures at the end of DRBD
2406
  resyncs, and in general of DRBD resync checks
2407
- Fixed a couple of issues with exports and snapshot errors
2408
- Fixed a couple of issues in instance listing
2409
- Added display of the disk size in ``gnt-instance info``
2410
- Fixed checking for valid OSes in instance creation
2411
- Fixed handling of the "vcpus" parameter in instance listing and in
2412
  general of invalid parameters
2413
- Fixed http server library, and thus RAPI, to handle invalid
2414
  username/password combinations correctly; this means that now they
2415
  report unauthorized for queries too, not only for modifications,
2416
  allowing earlier detect of configuration problems
2417
- Added a new "role" node list field, equivalent to the master/master
2418
  candidate/drained/offline flags combinations
2419
- Fixed cluster modify and changes of candidate pool size
2420
- Fixed cluster verify error messages for wrong files on regular nodes
2421
- Fixed a couple of issues with node demotion from master candidate role
2422
- Fixed node readd issues
2423
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2424
- Added a new ``--no-voting`` option for masterfailover to fix failover
2425
  on two-nodes clusters when the former master node is unreachable
2426
- Added instance reinstall over RAPI
2427

    
2428

    
2429
Version 2.0.1
2430
-------------
2431

    
2432
*(Released Tue, 16 Jun 2009)*
2433

    
2434
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2435
  allow re-adding the start in single-user option (regression from 1.2)
2436
- the watcher writes the instance status to a file, to allow monitoring
2437
  to report the instance status (from the master) based on cached
2438
  results of the watcher's queries; while this can get stale if the
2439
  watcher is being locked due to other work on the cluster, this is
2440
  still an improvement
2441
- the watcher now also restarts the node daemon and the rapi daemon if
2442
  they died
2443
- fixed the watcher to handle full and drained queue cases
2444
- hooks export more instance data in the environment, which helps if
2445
  hook scripts need to take action based on the instance's properties
2446
  (no longer need to query back into ganeti)
2447
- instance failovers when the instance is stopped do not check for free
2448
  RAM, so that failing over a stopped instance is possible in low memory
2449
  situations
2450
- rapi uses queries for tags instead of jobs (for less job traffic), and
2451
  for cluster tags it won't talk to masterd at all but read them from
2452
  ssconf
2453
- a couple of error handling fixes in RAPI
2454
- drbd handling: improved the error handling of inconsistent disks after
2455
  resync to reduce the frequency of "there are some degraded disks for
2456
  this instance" messages
2457
- fixed a bug in live migration when DRBD doesn't want to reconnect (the
2458
  error handling path called a wrong function name)
2459

    
2460

    
2461
Version 2.0.0
2462
-------------
2463

    
2464
*(Released Wed, 27 May 2009)*
2465

    
2466
- no changes from rc5
2467

    
2468

    
2469
Version 2.0 rc5
2470
---------------
2471

    
2472
*(Released Wed, 20 May 2009)*
2473

    
2474
- fix a couple of bugs (validation, argument checks)
2475
- fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2476
- some small improvements to RAPI and IAllocator
2477
- make watcher automatically start the master daemon if down
2478

    
2479

    
2480
Version 2.0 rc4
2481
---------------
2482

    
2483
*(Released Mon, 27 Apr 2009)*
2484

    
2485
- change the OS list to not require locks; this helps with big clusters
2486
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2487
  volume group is broken
2488
- ``gnt-instance info``, without any arguments, doesn't run for all
2489
  instances anymore; either pass ``--all`` or pass the desired
2490
  instances; this helps against mistakes on big clusters where listing
2491
  the information for all instances takes a long time
2492
- miscellaneous doc and man pages fixes
2493

    
2494

    
2495
Version 2.0 rc3
2496
---------------
2497

    
2498
*(Released Wed, 8 Apr 2009)*
2499

    
2500
- Change the internal locking model of some ``gnt-node`` commands, in
2501
  order to reduce contention (and blocking of master daemon) when
2502
  batching many creation/reinstall jobs
2503
- Fixes to Xen soft reboot
2504
- No longer build documentation at build time, instead distribute it in
2505
  the archive, in order to reduce the need for the whole docbook/rst
2506
  toolchains
2507

    
2508

    
2509
Version 2.0 rc2
2510
---------------
2511

    
2512
*(Released Fri, 27 Mar 2009)*
2513

    
2514
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2515
- Fix watcher startup sequence, improves the behaviour of busy clusters
2516
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2517
  replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2518
  bind address and other places
2519
- Some documentation fixes and updates
2520

    
2521

    
2522
Version 2.0 rc1
2523
---------------
2524

    
2525
*(Released Mon, 2 Mar 2009)*
2526

    
2527
- More documentation updates, now all docs should be more-or-less
2528
  up-to-date
2529
- A couple of small fixes (mixed hypervisor clusters, offline nodes,
2530
  etc.)
2531
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2532
  and KVM)
2533
- Fix an issue related to $libdir/run/ganeti and cluster creation
2534

    
2535

    
2536
Version 2.0 beta2
2537
-----------------
2538

    
2539
*(Released Thu, 19 Feb 2009)*
2540

    
2541
- Xen PVM and KVM have switched the default value for the instance root
2542
  disk to the first partition on the first drive, instead of the whole
2543
  drive; this means that the OS installation scripts must be changed
2544
  accordingly
2545
- Man pages have been updated
2546
- RAPI has been switched by default to HTTPS, and the exported functions
2547
  should all work correctly
2548
- RAPI v1 has been removed
2549
- Many improvements to the KVM hypervisor
2550
- Block device errors are now better reported
2551
- Many other bugfixes and small improvements
2552

    
2553

    
2554
Version 2.0 beta1
2555
-----------------
2556

    
2557
*(Released Mon, 26 Jan 2009)*
2558

    
2559
- Version 2 is a general rewrite of the code and therefore the
2560
  differences are too many to list, see the design document for 2.0 in
2561
  the ``doc/`` subdirectory for more details
2562
- In this beta version there is not yet a migration path from 1.2 (there
2563
  will be one in the final 2.0 release)
2564
- A few significant changes are:
2565

    
2566
  - all commands are executed by a daemon (``ganeti-masterd``) and the
2567
    various ``gnt-*`` commands are just front-ends to it
2568
  - all the commands are entered into, and executed from a job queue,
2569
    see the ``gnt-job(8)`` manpage
2570
  - the RAPI daemon supports read-write operations, secured by basic
2571
    HTTP authentication on top of HTTPS
2572
  - DRBD version 0.7 support has been removed, DRBD 8 is the only
2573
    supported version (when migrating from Ganeti 1.2 to 2.0, you need
2574
    to migrate to DRBD 8 first while still running Ganeti 1.2)
2575
  - DRBD devices are using statically allocated minor numbers, which
2576
    will be assigned to existing instances during the migration process
2577
  - there is support for both Xen PVM and Xen HVM instances running on
2578
    the same cluster
2579
  - KVM virtualization is supported too
2580
  - file-based storage has been implemented, which means that it is
2581
    possible to run the cluster without LVM and DRBD storage, for
2582
    example using a shared filesystem exported from shared storage (and
2583
    still have live migration)
2584

    
2585

    
2586
Version 1.2.7
2587
-------------
2588

    
2589
*(Released Tue, 13 Jan 2009)*
2590

    
2591
- Change the default reboot type in ``gnt-instance reboot`` to "hard"
2592
- Reuse the old instance mac address by default on instance import, if
2593
  the instance name is the same.
2594
- Handle situations in which the node info rpc returns incomplete
2595
  results (issue 46)
2596
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2597
- Improved version of batcher:
2598

    
2599
  - state file support
2600
  - instance mac address support
2601
  - support for HVM clusters/instances
2602

    
2603
- Add an option to show the number of cpu sockets and nodes in
2604
  ``gnt-node list``
2605
- Support OSes that handle more than one version of the OS api (but do
2606
  not change the current API in any other way)
2607
- Fix ``gnt-node migrate``
2608
- ``gnt-debug`` man page
2609
- Fixes various more typos and small issues
2610
- Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2611

    
2612

    
2613
Version 1.2.6
2614
-------------
2615

    
2616
*(Released Wed, 24 Sep 2008)*
2617

    
2618
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2619
  type of disk exported to fully virtualized instances.
2620
- provide access to the serial console of HVM instances
2621
- instance auto_balance flag, set by default. If turned off it will
2622
  avoid warnings on cluster verify if there is not enough memory to fail
2623
  over an instance. in the future it will prevent automatically failing
2624
  it over when we will support that.
2625
- batcher tool for instance creation, see ``tools/README.batcher``
2626
- ``gnt-instance reinstall --select-os`` to interactively select a new
2627
  operating system when reinstalling an instance.
2628
- when changing the memory amount on instance modify a check has been
2629
  added that the instance will be able to start. also warnings are
2630
  emitted if the instance will not be able to fail over, if auto_balance
2631
  is true.
2632
- documentation fixes
2633
- sync fields between ``gnt-instance list/modify/add/import``
2634
- fix a race condition in drbd when the sync speed was set after giving
2635
  the device a remote peer.
2636

    
2637

    
2638
Version 1.2.5
2639
-------------
2640

    
2641
*(Released Tue, 22 Jul 2008)*
2642

    
2643
- note: the allowed size and number of tags per object were reduced
2644
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2645
- fixed twisted 8.x compatibility
2646
- fixed ``gnt-instance replace-disks`` with iallocator
2647
- add TCP keepalives on twisted connections to detect restarted nodes
2648
- disk increase support, see ``gnt-instance grow-disk``
2649
- implement bulk node/instance query for RAPI
2650
- add tags in node/instance listing (optional)
2651
- experimental migration (and live migration) support, read the man page
2652
  for ``gnt-instance migrate``
2653
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2654
  has some small improvements in handling its state file
2655

    
2656

    
2657
Version 1.2.4
2658
-------------
2659

    
2660
*(Released Fri, 13 Jun 2008)*
2661

    
2662
- Experimental readonly, REST-based remote API implementation;
2663
  automatically started on master node, TCP port 5080, if enabled by
2664
  ``--enable-rapi`` parameter to configure script.
2665
- Instance allocator support. Add and import instance accept a
2666
  ``--iallocator`` parameter, and call that instance allocator to decide
2667
  which node to use for the instance. The iallocator document describes
2668
  what's expected from an allocator script.
2669
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2670
  ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2671
  node is lost there is still enough memory to fail over the instances
2672
  that reside on it.
2673
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2674
  ``gnt-cluster verify``, to check for site-specific compliance. All the
2675
  hooks will run, and their output, if any, will be displayed. Any
2676
  failing hook will make the verification return an error value.
2677
- ``gnt-cluster verify`` now checks that its peers are reachable on the
2678
  primary and secondary interfaces
2679
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2680
  that is still declared as part of the cluster and has failed.
2681
- ``gnt-* list`` commands now accept a new ``-o +field`` way of
2682
  specifying output fields, that just adds the chosen fields to the
2683
  default ones.
2684
- ``gnt-backup`` now has a new ``remove`` command to delete an existing
2685
  export from the filesystem.
2686
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2687
  have been added. Using them you can enable/disable acpi and pae
2688
  support, and specify a path for a cd image to be exported to the
2689
  instance. These parameters as the name suggest only work on HVM
2690
  clusters.
2691
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2692
  PAE support will be set to the previously hardcoded values, but the
2693
  (previously hardcoded) path to the CDROM ISO image will be unset and
2694
  if required, needs to be set manually with ``gnt-instance modify``
2695
  after the upgrade.
2696
- The address to which an instance's VNC console is bound is now
2697
  selectable per-instance, rather than being cluster wide. Of course
2698
  this only applies to instances controlled via VNC, so currently just
2699
  applies to HVM clusters.
2700

    
2701

    
2702
Version 1.2.3
2703
-------------
2704

    
2705
*(Released Mon, 18 Feb 2008)*
2706

    
2707
- more tweaks to the disk activation code (especially helpful for DRBD)
2708
- change the default ``gnt-instance list`` output format, now there is
2709
  one combined status field (see the manpage for the exact values this
2710
  field will have)
2711
- some more fixes for the mac export to hooks change
2712
- make Ganeti not break with DRBD 8.2.x (which changed the version
2713
  format in ``/proc/drbd``) (issue 24)
2714
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2715
  template, allowing migration from DRBD0.7+MD to DRBD8
2716

    
2717

    
2718
Version 1.2.2
2719
-------------
2720

    
2721
*(Released Wed, 30 Jan 2008)*
2722

    
2723
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2724
  support (issue 23)
2725
- add command aliases infrastructure and a few aliases
2726
- allow listing of VCPUs in the ``gnt-instance list`` and improve the
2727
  man pages and the ``--help`` option of ``gnt-node
2728
  list``/``gnt-instance list``
2729
- fix ``gnt-backup list`` with down nodes (issue 21)
2730
- change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2731
- fix the dist archive and add a check for including svn/git files in
2732
  the future
2733
- some developer-related changes: improve the burnin and the QA suite,
2734
  add an upload script for testing during development
2735

    
2736

    
2737
Version 1.2.1
2738
-------------
2739

    
2740
*(Released Wed, 16 Jan 2008)*
2741

    
2742
- experimental HVM support, read the install document, section
2743
  "Initializing the cluster"
2744
- allow for the PVM hypervisor per-instance kernel and initrd paths
2745
- add a new command ``gnt-cluster verify-disks`` which uses a new
2746
  algorithm to improve the reconnection of the DRBD pairs if the device
2747
  on the secondary node has gone away
2748
- make logical volume code auto-activate LVs at disk activation time
2749
- slightly improve the speed of activating disks
2750
- allow specification of the MAC address at instance creation time, and
2751
  changing it later via ``gnt-instance modify``
2752
- fix handling of external commands that generate lots of output on
2753
  stderr
2754
- update documentation with regard to minimum version of DRBD8 supported
2755

    
2756

    
2757
Version 1.2.0
2758
-------------
2759

    
2760
*(Released Tue, 4 Dec 2007)*
2761

    
2762
- Log the ``xm create`` output to the node daemon log on failure (to
2763
  help diagnosing the error)
2764
- In debug mode, log all external commands output if failed to the logs
2765
- Change parsing of lvm commands to ignore stderr
2766

    
2767

    
2768
Version 1.2 beta3
2769
-----------------
2770

    
2771
*(Released Wed, 28 Nov 2007)*
2772

    
2773
- Another round of updates to the DRBD 8 code to deal with more failures
2774
  in the replace secondary node operation
2775
- Some more logging of failures in disk operations (lvm, drbd)
2776
- A few documentation updates
2777
- QA updates
2778

    
2779

    
2780
Version 1.2 beta2
2781
-----------------
2782

    
2783
*(Released Tue, 13 Nov 2007)*
2784

    
2785
- Change configuration file format from Python's Pickle to JSON.
2786
  Upgrading is possible using the cfgupgrade utility.
2787
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2788
  faster replace disks and is more stable (DRBD 8 has many improvements
2789
  compared to DRBD 0.7)
2790
- Added command line tags support (see man pages for ``gnt-instance``,
2791
  ``gnt-node``, ``gnt-cluster``)
2792
- Added instance rename support
2793
- Added multi-instance startup/shutdown
2794
- Added cluster rename support
2795
- Added ``gnt-node evacuate`` to simplify some node operations
2796
- Added instance reboot operation that can speedup reboot as compared to
2797
  stop and start
2798
- Soften the requirement that hostnames are in FQDN format
2799
- The ``ganeti-watcher`` now activates drbd pairs after secondary node
2800
  reboots
2801
- Removed dependency on debian's patched fping that uses the
2802
  non-standard ``-S`` option
2803
- Now the OS definitions are searched for in multiple, configurable
2804
  paths (easier for distros to package)
2805
- Some changes to the hooks infrastructure (especially the new
2806
  post-configuration update hook)
2807
- Other small bugfixes
2808

    
2809
.. vim: set textwidth=72 syntax=rst :
2810
.. Local Variables:
2811
.. mode: rst
2812
.. fill-column: 72
2813
.. End: