root / NEWS @ 20203756
History | View | Annotate | Download (56.5 kB)
1 |
News |
---|---|
2 |
==== |
3 |
|
4 |
Version 2.4.0 |
5 |
------------- |
6 |
|
7 |
*(Released Mon, 07 Mar 2011)* |
8 |
|
9 |
Final 2.4.0 release. Just a few small fixes: |
10 |
|
11 |
- Fixed RAPI node evacuate |
12 |
- Fixed the kvm-ifup script |
13 |
- Fixed internal error handling for special job cases |
14 |
- Updated man page to specify the escaping feature for options |
15 |
|
16 |
|
17 |
Version 2.4.0 rc3 |
18 |
----------------- |
19 |
|
20 |
*(Released Mon, 28 Feb 2011)* |
21 |
|
22 |
A critical fix for the ``prealloc_wipe_disks`` feature: it is possible |
23 |
that this feature wiped the disks of the wrong instance, leading to loss |
24 |
of data. |
25 |
|
26 |
Other changes: |
27 |
|
28 |
- Fixed title of query field containing instance name |
29 |
- Expanded the glossary in the documentation |
30 |
- Fixed one unittest (internal issue) |
31 |
|
32 |
|
33 |
Version 2.4.0 rc2 |
34 |
----------------- |
35 |
|
36 |
*(Released Mon, 21 Feb 2011)* |
37 |
|
38 |
A number of bug fixes plus just a couple functionality changes. |
39 |
|
40 |
On the user-visible side, the ``gnt-* list`` command output has changed |
41 |
with respect to "special" field states. The current rc1 style of display |
42 |
can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in |
43 |
the default output mode special fields states are displayed as follows: |
44 |
|
45 |
- Offline resource: ``*`` |
46 |
- Unavailable/not applicable: ``-`` |
47 |
- Data missing (RPC failure): ``?`` |
48 |
- Unknown field: ``??`` |
49 |
|
50 |
Another user-visible change is the addition of ``--force-join`` to |
51 |
``gnt-node add``. |
52 |
|
53 |
As for bug fixes: |
54 |
|
55 |
- ``tools/cluster-merge`` has seen many fixes and is now enabled again |
56 |
- Fixed regression in RAPI/instance reinstall where all parameters were |
57 |
required (instead of optional) |
58 |
- Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2 |
59 |
- Fixed iallocator usage (offline nodes were not considered offline) |
60 |
- Fixed ``gnt-node list`` with respect to non-vm_capable nodes |
61 |
- Fixed hypervisor and OS parameter validation with respect to |
62 |
non-vm_capable nodes |
63 |
- Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly |
64 |
cosmetic) |
65 |
- Fixed ``tools/listrunner`` with respect to agent-based usage |
66 |
|
67 |
|
68 |
Version 2.4.0 rc1 |
69 |
----------------- |
70 |
|
71 |
*(Released Fri, 4 Feb 2011)* |
72 |
|
73 |
Many changes and fixes since the beta1 release. While there were some |
74 |
internal changes, the code has been mostly stabilised for the RC |
75 |
release. |
76 |
|
77 |
Note: the dumb allocator was removed in this release, as it was not kept |
78 |
up-to-date with the IAllocator protocol changes. It is recommended to |
79 |
use the ``hail`` command from the ganeti-htools package. |
80 |
|
81 |
Note: the 2.4 and up versions of Ganeti are not compatible with the |
82 |
0.2.x branch of ganeti-htools. You need to upgrade to |
83 |
ganeti-htools-0.3.0 (or later). |
84 |
|
85 |
Regressions fixed from 2.3 |
86 |
~~~~~~~~~~~~~~~~~~~~~~~~~~ |
87 |
|
88 |
- Fixed the ``gnt-cluster verify-disks`` command |
89 |
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to |
90 |
serially on nodes) |
91 |
- Fixed disk adoption breakage |
92 |
- Fixed wrong headers in instance listing for field aliases |
93 |
|
94 |
Other bugs fixed |
95 |
~~~~~~~~~~~~~~~~ |
96 |
|
97 |
- Fixed corner case in KVM handling of NICs |
98 |
- Fixed many cases of wrong handling of non-vm_capable nodes |
99 |
- Fixed a bug where a missing instance symlink was not possible to |
100 |
recreate with any ``gnt-*`` command (now ``gnt-instance |
101 |
activate-disks`` does it) |
102 |
- Fixed the volume group name as reported by ``gnt-cluster |
103 |
verify-disks`` |
104 |
- Increased timeouts for the import-export code, hopefully leading to |
105 |
fewer aborts due network or instance timeouts |
106 |
- Fixed bug in ``gnt-node list-storage`` |
107 |
- Fixed bug where not all daemons were started on cluster |
108 |
initialisation, but only at the first watcher run |
109 |
- Fixed many bugs in the OOB implementation |
110 |
- Fixed watcher behaviour in presence of instances with offline |
111 |
secondaries |
112 |
- Fixed instance list output for instances running on the wrong node |
113 |
- a few fixes to the cluster-merge tool, but it still cannot merge |
114 |
multi-node groups (currently it is not recommended to use this tool) |
115 |
|
116 |
|
117 |
Improvements |
118 |
~~~~~~~~~~~~ |
119 |
|
120 |
- Improved network configuration for the KVM hypervisor |
121 |
- Added e1000 as a supported NIC for Xen-HVM |
122 |
- Improved the lvmstrap tool to also be able to use partitions, as |
123 |
opposed to full disks |
124 |
- Improved speed of disk wiping (the cluster parameter |
125 |
``prealloc_wipe_disks``, so that it has a low impact on the total time |
126 |
of instance creations |
127 |
- Added documentation for the OS parameters |
128 |
- Changed ``gnt-instance deactivate-disks`` so that it can work if the |
129 |
hypervisor is not responding |
130 |
- Added display of blacklisted and hidden OS information in |
131 |
``gnt-cluster info`` |
132 |
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend, |
133 |
NIC and node parameters, which might create problems with currently |
134 |
invalid (but undetected) configuration files, but prevents validation |
135 |
failures when unrelated parameters are modified |
136 |
- Changed cluster initialisation to wait for the master daemon to become |
137 |
available |
138 |
- Expanded the RAPI interface: |
139 |
|
140 |
- Added config redistribution resource |
141 |
- Added activation/deactivation of instance disks |
142 |
- Added export of console information |
143 |
|
144 |
- Implemented log file reopening on SIGHUP, which allows using |
145 |
logrotate(8) for the Ganeti log files |
146 |
- Added a basic OOB helper script as an example |
147 |
|
148 |
|
149 |
Version 2.4.0 beta1 |
150 |
------------------- |
151 |
|
152 |
*(Released Fri, 14 Jan 2011)* |
153 |
|
154 |
User-visible |
155 |
~~~~~~~~~~~~ |
156 |
|
157 |
- Fixed timezone issues when formatting timestamps |
158 |
- Added support for node groups, available via ``gnt-group`` and other |
159 |
commands |
160 |
- Added out-of-band framework and management, see :doc:`design |
161 |
document <design-oob>` |
162 |
- Removed support for roman numbers from ``gnt-node list`` and |
163 |
``gnt-instance list``. |
164 |
- Allowed modification of master network interface via ``gnt-cluster |
165 |
modify --master-netdev`` |
166 |
- Accept offline secondaries while shutting down instance disks |
167 |
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors |
168 |
- Added support for multiple LVM volume groups |
169 |
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are |
170 |
requested |
171 |
- Added commands to list available fields: |
172 |
|
173 |
- ``gnt-node list-fields`` |
174 |
- ``gnt-group list-fields`` |
175 |
- ``gnt-instance list-fields`` |
176 |
|
177 |
- Updated documentation and man pages |
178 |
|
179 |
Integration |
180 |
~~~~~~~~~~~ |
181 |
|
182 |
- Moved ``rapi_users`` file into separate directory, now named |
183 |
``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a |
184 |
symlink |
185 |
- Added new tool for running commands on many machines, |
186 |
``tools/ganeti-listrunner`` |
187 |
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also |
188 |
improving the ``gnt-instance console`` output |
189 |
- Allowed customisation of disk index separator at ``configure`` time |
190 |
- Export node group allocation policy to :doc:`iallocator <iallocator>` |
191 |
- Added support for non-partitioned md disks in ``lvmstrap`` |
192 |
- Added script to gracefully power off KVM instances |
193 |
- Split ``utils`` module into smaller parts |
194 |
- Changed query operations to return more detailed information, e.g. |
195 |
whether an information is unavailable due to an offline node. To use |
196 |
this new functionality, the LUXI call ``Query`` must be used. Field |
197 |
information is now stored by the master daemon and can be retrieved |
198 |
using ``QueryFields``. Instances, nodes and groups can also be queried |
199 |
using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet |
200 |
exposed via RAPI). The following commands make use of this |
201 |
infrastructure change: |
202 |
|
203 |
- ``gnt-group list`` |
204 |
- ``gnt-group list-fields`` |
205 |
- ``gnt-node list`` |
206 |
- ``gnt-node list-fields`` |
207 |
- ``gnt-instance list`` |
208 |
- ``gnt-instance list-fields`` |
209 |
- ``gnt-debug locks`` |
210 |
|
211 |
Remote API |
212 |
~~~~~~~~~~ |
213 |
|
214 |
- New RAPI resources (see :doc:`rapi`): |
215 |
|
216 |
- ``/2/modify`` |
217 |
- ``/2/groups`` |
218 |
- ``/2/groups/[group_name]`` |
219 |
- ``/2/groups/[group_name]/assign-nodes`` |
220 |
- ``/2/groups/[group_name]/modify`` |
221 |
- ``/2/groups/[group_name]/rename`` |
222 |
- ``/2/instances/[instance_name]/disk/[disk_index]/grow`` |
223 |
|
224 |
- RAPI changes: |
225 |
|
226 |
- Implemented ``no_install`` for instance creation |
227 |
- Implemented OS parameters for instance reinstallation, allowing |
228 |
use of special settings on reinstallation (e.g. for preserving data) |
229 |
|
230 |
Misc |
231 |
~~~~ |
232 |
|
233 |
- Added IPv6 support in import/export |
234 |
- Pause DRBD synchronization while wiping disks on instance creation |
235 |
- Updated unittests and QA scripts |
236 |
- Improved network parameters passed to KVM |
237 |
- Converted man pages from docbook to reStructuredText |
238 |
|
239 |
|
240 |
Version 2.3.1 |
241 |
------------- |
242 |
|
243 |
*(Released Mon, 20 Dec 2010)* |
244 |
|
245 |
Released version 2.3.1~rc1 without any changes. |
246 |
|
247 |
|
248 |
Version 2.3.1 rc1 |
249 |
----------------- |
250 |
|
251 |
*(Released Wed, 1 Dec 2010)* |
252 |
|
253 |
- impexpd: Disable OpenSSL compression in socat if possible (backport |
254 |
from master, commit e90739d625b, see :doc:`installation guide |
255 |
<install-quick>` for details) |
256 |
- Changed unittest coverage report to exclude test scripts |
257 |
- Added script to check version format |
258 |
|
259 |
|
260 |
Version 2.3.0 |
261 |
------------- |
262 |
|
263 |
*(Released Wed, 1 Dec 2010)* |
264 |
|
265 |
Released version 2.3.0~rc1 without any changes. |
266 |
|
267 |
|
268 |
Version 2.3.0 rc1 |
269 |
----------------- |
270 |
|
271 |
*(Released Fri, 19 Nov 2010)* |
272 |
|
273 |
A number of bugfixes and documentation updates: |
274 |
|
275 |
- Update ganeti-os-interface documentation |
276 |
- Fixed a bug related to duplicate MACs or similar items which should be |
277 |
unique |
278 |
- Fix breakage in OS state modify |
279 |
- Reinstall instance: disallow offline secondaries (fixes bug related to |
280 |
OS changing but reinstall failing) |
281 |
- plus all the other fixes between 2.2.1 and 2.2.2 |
282 |
|
283 |
|
284 |
Version 2.3.0 rc0 |
285 |
----------------- |
286 |
|
287 |
*(Released Tue, 2 Nov 2010)* |
288 |
|
289 |
- Fixed clearing of the default iallocator using ``gnt-cluster modify`` |
290 |
- Fixed master failover race with watcher |
291 |
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent |
292 |
configuration |
293 |
- Accept previously stopped instance for export with instance removal |
294 |
- Simplify and extend the environment variables for instance OS scripts |
295 |
- Added new node flags, ``master_capable`` and ``vm_capable`` |
296 |
- Added optional instance disk wiping prior during allocation. This is a |
297 |
cluster-wide option and can be set/modified using |
298 |
``gnt-cluster {init,modify} --prealloc-wipe-disks``. |
299 |
- Added IPv6 support, see :doc:`design document <design-2.3>` and |
300 |
:doc:`install-quick` |
301 |
- Added a new watcher option (``--ignore-pause``) |
302 |
- Added option to ignore offline node on instance start/stop |
303 |
(``--ignore-offline``) |
304 |
- Allow overriding OS parameters with ``gnt-instance reinstall`` |
305 |
- Added ability to change node's secondary IP address using ``gnt-node |
306 |
modify`` |
307 |
- Implemented privilege separation for all daemons except |
308 |
``ganeti-noded``, see ``configure`` options |
309 |
- Complain if an instance's disk is marked faulty in ``gnt-cluster |
310 |
verify`` |
311 |
- Implemented job priorities (see ``ganeti(7)`` manpage) |
312 |
- Ignore failures while shutting down instances during failover from |
313 |
offline node |
314 |
- Exit daemon's bootstrap process only once daemon is ready |
315 |
- Export more information via ``LUInstanceQuery``/remote API |
316 |
- Improved documentation, QA and unittests |
317 |
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a |
318 |
restart if the file was created or changed |
319 |
- Added LUXI protocol version sent with each request and response, |
320 |
allowing detection of server/client mismatches |
321 |
- Moved the Python scripts among gnt-* and ganeti-* into modules |
322 |
- Moved all code related to setting up SSH to an external script, |
323 |
``setup-ssh`` |
324 |
- Infrastructure changes for node group support in future versions |
325 |
|
326 |
|
327 |
Version 2.2.2 |
328 |
------------- |
329 |
|
330 |
*(Released Fri, 19 Nov 2010)* |
331 |
|
332 |
A few small bugs fixed, and some improvements to the build system: |
333 |
|
334 |
- Fix documentation regarding conversion to drbd |
335 |
- Fix validation of parameters in cluster modify (``gnt-cluster modify |
336 |
-B``) |
337 |
- Fix error handling in node modify with multiple changes |
338 |
- Allow remote imports without checked names |
339 |
|
340 |
|
341 |
Version 2.2.1 |
342 |
------------- |
343 |
|
344 |
*(Released Tue, 19 Oct 2010)* |
345 |
|
346 |
- Disable SSL session ID cache in RPC client |
347 |
|
348 |
|
349 |
Version 2.2.1 rc1 |
350 |
----------------- |
351 |
|
352 |
*(Released Thu, 14 Oct 2010)* |
353 |
|
354 |
- Fix interaction between Curl/GnuTLS and the Python's HTTP server |
355 |
(thanks Apollon Oikonomopoulos!), finally allowing the use of Curl |
356 |
with GnuTLS |
357 |
- Fix problems with interaction between Curl and Python's HTTP server, |
358 |
resulting in increased speed in many RPC calls |
359 |
- Improve our release script to prevent breakage with older aclocal and |
360 |
Python 2.6 |
361 |
|
362 |
|
363 |
Version 2.2.1 rc0 |
364 |
----------------- |
365 |
|
366 |
*(Released Thu, 7 Oct 2010)* |
367 |
|
368 |
- Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with |
369 |
value retrieved from master |
370 |
- Added support for blacklisted or hidden OS definitions |
371 |
- Added simple lock monitor (accessible via (``gnt-debug locks``) |
372 |
- Added support for -mem-path in KVM hypervisor abstraction layer |
373 |
- Allow overriding instance parameters in tool for inter-cluster |
374 |
instance moves (``tools/move-instance``) |
375 |
- Improved opcode summaries (e.g. in ``gnt-job list``) |
376 |
- Improve consistency of OS listing by sorting it |
377 |
- Documentation updates |
378 |
|
379 |
|
380 |
Version 2.2.0.1 |
381 |
--------------- |
382 |
|
383 |
*(Released Fri, 8 Oct 2010)* |
384 |
|
385 |
- Rebuild with a newer autotools version, to fix python 2.6 compatibility |
386 |
|
387 |
|
388 |
Version 2.2.0 |
389 |
------------- |
390 |
|
391 |
*(Released Mon, 4 Oct 2010)* |
392 |
|
393 |
- Fixed regression in ``gnt-instance rename`` |
394 |
|
395 |
|
396 |
Version 2.2.0 rc2 |
397 |
----------------- |
398 |
|
399 |
*(Released Wed, 22 Sep 2010)* |
400 |
|
401 |
- Fixed OS_VARIANT variable for OS scripts |
402 |
- Fixed cluster tag operations via RAPI |
403 |
- Made ``setup-ssh`` exit with non-zero code if an error occurred |
404 |
- Disabled RAPI CA checks in watcher |
405 |
|
406 |
|
407 |
Version 2.2.0 rc1 |
408 |
----------------- |
409 |
|
410 |
*(Released Mon, 23 Aug 2010)* |
411 |
|
412 |
- Support DRBD versions of the format "a.b.c.d" |
413 |
- Updated manpages |
414 |
- Re-introduce support for usage from multiple threads in RAPI client |
415 |
- Instance renames and modify via RAPI |
416 |
- Work around race condition between processing and archival in job |
417 |
queue |
418 |
- Mark opcodes following failed one as failed, too |
419 |
- Job field ``lock_status`` was removed due to difficulties making it |
420 |
work with the changed job queue in Ganeti 2.2; a better way to monitor |
421 |
locks is expected for a later 2.2.x release |
422 |
- Fixed dry-run behaviour with many commands |
423 |
- Support ``ssh-agent`` again when adding nodes |
424 |
- Many additional bugfixes |
425 |
|
426 |
|
427 |
Version 2.2.0 rc0 |
428 |
----------------- |
429 |
|
430 |
*(Released Fri, 30 Jul 2010)* |
431 |
|
432 |
Important change: the internal RPC mechanism between Ganeti nodes has |
433 |
changed from using a home-grown http library (based on the Python base |
434 |
libraries) to use the PycURL library. This requires that PycURL is |
435 |
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked |
436 |
against GnuTLS by default. cURL's support for GnuTLS had known issues |
437 |
before cURL 7.21.0 and we recommend using the latest cURL release or |
438 |
linking against OpenSSL. Most other distributions already link PycURL |
439 |
and cURL against OpenSSL. The command:: |
440 |
|
441 |
python -c 'import pycurl; print pycurl.version' |
442 |
|
443 |
can be used to determine the libraries PycURL and cURL are linked |
444 |
against. |
445 |
|
446 |
Other significant changes: |
447 |
|
448 |
- Rewrote much of the internals of the job queue, in order to achieve |
449 |
better parallelism; this decouples job query operations from the job |
450 |
processing, and it should allow much nicer behaviour of the master |
451 |
daemon under load, and it also has uncovered some long-standing bugs |
452 |
related to the job serialisation (now fixed) |
453 |
- Added a default iallocator setting to the cluster parameters, |
454 |
eliminating the need to always pass nodes or an iallocator for |
455 |
operations that require selection of new node(s) |
456 |
- Added experimental support for the LXC virtualization method |
457 |
- Added support for OS parameters, which allows the installation of |
458 |
instances to pass parameter to OS scripts in order to customise the |
459 |
instance |
460 |
- Added a hypervisor parameter controlling the migration type (live or |
461 |
non-live), since hypervisors have various levels of reliability; this |
462 |
has renamed the 'live' parameter to 'mode' |
463 |
- Added a cluster parameter ``reserved_lvs`` that denotes reserved |
464 |
logical volumes, meaning that cluster verify will ignore them and not |
465 |
flag their presence as errors |
466 |
- The watcher will now reset the error count for failed instances after |
467 |
8 hours, thus allowing self-healing if the problem that caused the |
468 |
instances to be down/fail to start has cleared in the meantime |
469 |
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti |
470 |
check for, and warn, if the drbd module parameter ``usermode_helper`` |
471 |
is not consistent with the cluster-wide setting; this is needed to |
472 |
make diagnose easier of failed drbd creations |
473 |
- Started adding base IPv6 support, but this is not yet |
474 |
enabled/available for use |
475 |
- Rename operations (cluster, instance) will now return the new name, |
476 |
which is especially useful if a short name was passed in |
477 |
- Added support for instance migration in RAPI |
478 |
- Added a tool to pre-configure nodes for the SSH setup, before joining |
479 |
them to the cluster; this will allow in the future a simplified model |
480 |
for node joining (but not yet fully enabled in 2.2); this needs the |
481 |
paramiko python library |
482 |
- Fixed handling of name-resolving errors |
483 |
- Fixed consistency of job results on the error path |
484 |
- Fixed master-failover race condition when executed multiple times in |
485 |
sequence |
486 |
- Fixed many bugs related to the job queue (mostly introduced during the |
487 |
2.2 development cycle, so not all are impacting 2.1) |
488 |
- Fixed instance migration with missing disk symlinks |
489 |
- Fixed handling of unknown jobs in ``gnt-job archive`` |
490 |
- And many other small fixes/improvements |
491 |
|
492 |
Internal changes: |
493 |
|
494 |
- Enhanced both the unittest and the QA coverage |
495 |
- Switched the opcode validation to a generic model, and extended the |
496 |
validation to all opcode parameters |
497 |
- Changed more parts of the code that write shell scripts to use the |
498 |
same class for this |
499 |
- Switched the master daemon to use the asyncore library for the Luxi |
500 |
server endpoint |
501 |
|
502 |
|
503 |
Version 2.2.0 beta 0 |
504 |
-------------------- |
505 |
|
506 |
*(Released Thu, 17 Jun 2010)* |
507 |
|
508 |
- Added tool (``move-instance``) and infrastructure to move instances |
509 |
between separate clusters (see :doc:`separate documentation |
510 |
<move-instance>` and :doc:`design document <design-2.2>`) |
511 |
- Added per-request RPC timeout |
512 |
- RAPI now requires a Content-Type header for requests with a body (e.g. |
513 |
``PUT`` or ``POST``) which must be set to ``application/json`` (see |
514 |
:rfc:`2616` (HTTP/1.1), section 7.2.1) |
515 |
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not |
516 |
reachable |
517 |
- Implemented initial support for running Ganeti daemons as separate |
518 |
users, see configure-time flags ``--with-user-prefix`` and |
519 |
``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this |
520 |
time) |
521 |
- Instances can be removed after export (``gnt-backup export |
522 |
--remove-instance``) |
523 |
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA |
524 |
key (instead of 1024 bit) |
525 |
- Added new cluster configuration file for cluster domain secret |
526 |
- Import/export now use SSL instead of SSH |
527 |
- Added support for showing estimated time when exporting an instance, |
528 |
see the ``ganeti-os-interface(7)`` manpage and look for |
529 |
``EXP_SIZE_FD`` |
530 |
|
531 |
|
532 |
Version 2.1.8 |
533 |
------------- |
534 |
|
535 |
*(Released Tue, 16 Nov 2010)* |
536 |
|
537 |
Some more bugfixes. Unless critical bugs occur, this will be the last |
538 |
2.1 release: |
539 |
|
540 |
- Fix case of MAC special-values |
541 |
- Fix mac checker regex |
542 |
- backend: Fix typo causing "out of range" error |
543 |
- Add missing --units in gnt-instance list man page |
544 |
|
545 |
|
546 |
Version 2.1.7 |
547 |
------------- |
548 |
|
549 |
*(Released Tue, 24 Aug 2010)* |
550 |
|
551 |
Bugfixes only: |
552 |
- Don't ignore secondary node silently on non-mirrored disk templates |
553 |
(issue 113) |
554 |
- Fix --master-netdev arg name in gnt-cluster(8) (issue 114) |
555 |
- Fix usb_mouse parameter breaking with vnc_console (issue 109) |
556 |
- Properly document the usb_mouse parameter |
557 |
- Fix path in ganeti-rapi(8) (issue 116) |
558 |
- Adjust error message when the ganeti user's .ssh directory is |
559 |
missing |
560 |
- Add same-node-check when changing the disk template to drbd |
561 |
|
562 |
|
563 |
Version 2.1.6 |
564 |
------------- |
565 |
|
566 |
*(Released Fri, 16 Jul 2010)* |
567 |
|
568 |
Bugfixes only: |
569 |
- Add an option to only select some reboot types during qa/burnin. |
570 |
(on some hypervisors consequent reboots are not supported) |
571 |
- Fix infrequent race condition in master failover. Sometimes the old |
572 |
master ip address would be still detected as up for a short time |
573 |
after it was removed, causing failover to fail. |
574 |
- Decrease mlockall warnings when the ctypes module is missing. On |
575 |
Python 2.4 we support running even if no ctypes module is installed, |
576 |
but we were too verbose about this issue. |
577 |
- Fix building on old distributions, on which man doesn't have a |
578 |
--warnings option. |
579 |
- Fix RAPI not to ignore the MAC address on instance creation |
580 |
- Implement the old instance creation format in the RAPI client. |
581 |
|
582 |
|
583 |
Version 2.1.5 |
584 |
------------- |
585 |
|
586 |
*(Released Thu, 01 Jul 2010)* |
587 |
|
588 |
A small bugfix release: |
589 |
- Fix disk adoption: broken by strict --disk option checking in 2.1.4 |
590 |
- Fix batch-create: broken in the whole 2.1 series due to a lookup on |
591 |
a non-existing option |
592 |
- Fix instance create: the --force-variant option was ignored |
593 |
- Improve pylint 0.21 compatibility and warnings with Python 2.6 |
594 |
- Fix modify node storage with non-FQDN arguments |
595 |
- Fix RAPI client to authenticate under Python 2.6 when used |
596 |
for more than 5 requests needing authentication |
597 |
- Fix gnt-instance modify -t (storage) giving a wrong error message |
598 |
when converting a non-shutdown drbd instance to plain |
599 |
|
600 |
|
601 |
Version 2.1.4 |
602 |
------------- |
603 |
|
604 |
*(Released Fri, 18 Jun 2010)* |
605 |
|
606 |
A small bugfix release: |
607 |
|
608 |
- Fix live migration of KVM instances started with older Ganeti |
609 |
versions which had fewer hypervisor parameters |
610 |
- Fix gnt-instance grow-disk on down instances |
611 |
- Fix an error-reporting bug during instance migration |
612 |
- Better checking of the ``--net`` and ``--disk`` values, to avoid |
613 |
silently ignoring broken ones |
614 |
- Fix an RPC error reporting bug affecting, for example, RAPI client |
615 |
users |
616 |
- Fix bug triggered by different API version os-es on different nodes |
617 |
- Fix a bug in instance startup with custom hvparams: OS level |
618 |
parameters would fail to be applied. |
619 |
- Fix the RAPI client under Python 2.6 (but more work is needed to |
620 |
make it work completely well with OpenSSL) |
621 |
- Fix handling of errors when resolving names from DNS |
622 |
|
623 |
|
624 |
Version 2.1.3 |
625 |
------------- |
626 |
|
627 |
*(Released Thu, 3 Jun 2010)* |
628 |
|
629 |
A medium sized development cycle. Some new features, and some |
630 |
fixes/small improvements/cleanups. |
631 |
|
632 |
Significant features |
633 |
~~~~~~~~~~~~~~~~~~~~ |
634 |
|
635 |
The node deamon now tries to mlock itself into memory, unless the |
636 |
``--no-mlock`` flag is passed. It also doesn't fail if it can't write |
637 |
its logs, and falls back to console logging. This allows emergency |
638 |
features such as ``gnt-node powercycle`` to work even in the event of a |
639 |
broken node disk (tested offlining the disk hosting the node's |
640 |
filesystem and dropping its memory caches; don't try this at home) |
641 |
|
642 |
KVM: add vhost-net acceleration support. It can be tested with a new |
643 |
enough version of the kernel and of qemu-kvm. |
644 |
|
645 |
KVM: Add instance chrooting feature. If you use privilege dropping for |
646 |
your VMs you can also now force them to chroot to an empty directory, |
647 |
before starting the emulated guest. |
648 |
|
649 |
KVM: Add maximum migration bandwith and maximum downtime tweaking |
650 |
support (requires a new-enough version of qemu-kvm). |
651 |
|
652 |
Cluster verify will now warn if the master node doesn't have the master |
653 |
ip configured on it. |
654 |
|
655 |
Add a new (incompatible) instance creation request format to RAPI which |
656 |
supports all parameters (previously only a subset was supported, and it |
657 |
wasn't possible to extend the old format to accomodate all the new |
658 |
features. The old format is still supported, and a client can check for |
659 |
this feature, before using it, by checking for its presence in the |
660 |
``features`` RAPI resource. |
661 |
|
662 |
Now with ancient latin support. Try it passing the ``--roman`` option to |
663 |
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list`` |
664 |
(requires the python-roman module to be installed, in order to work). |
665 |
|
666 |
Other changes |
667 |
~~~~~~~~~~~~~ |
668 |
|
669 |
As usual many internal code refactorings, documentation updates, and |
670 |
such. Among others: |
671 |
|
672 |
- Lots of improvements and cleanups to the experimental Remote API |
673 |
(RAPI) client library. |
674 |
- A new unit test suite for the core daemon libraries. |
675 |
- A fix to creating missing directories makes sure the umask is not |
676 |
applied anymore. This enforces the same directory permissions |
677 |
everywhere. |
678 |
- Better handling terminating daemons with ctrl+c (used when running |
679 |
them in debugging mode). |
680 |
- Fix a race condition in live migrating a KVM instance, when stat() |
681 |
on the old proc status file returned EINVAL, which is an unexpected |
682 |
value. |
683 |
- Fixed manpage checking with newer man and utf-8 charachters. But now |
684 |
you need the en_US.UTF-8 locale enabled to build Ganeti from git. |
685 |
|
686 |
|
687 |
Version 2.1.2.1 |
688 |
--------------- |
689 |
|
690 |
*(Released Fri, 7 May 2010)* |
691 |
|
692 |
Fix a bug which prevented untagged KVM instances from starting. |
693 |
|
694 |
|
695 |
Version 2.1.2 |
696 |
------------- |
697 |
|
698 |
*(Released Fri, 7 May 2010)* |
699 |
|
700 |
Another release with a long development cycle, during which many |
701 |
different features were added. |
702 |
|
703 |
Significant features |
704 |
~~~~~~~~~~~~~~~~~~~~ |
705 |
|
706 |
The KVM hypervisor now can run the individual instances as non-root, to |
707 |
reduce the impact of a VM being hijacked due to bugs in the |
708 |
hypervisor. It is possible to run all instances as a single (non-root) |
709 |
user, to manually specify a user for each instance, or to dynamically |
710 |
allocate a user out of a cluster-wide pool to each instance, with the |
711 |
guarantee that no two instances will run under the same user ID on any |
712 |
given node. |
713 |
|
714 |
An experimental RAPI client library, that can be used standalone |
715 |
(without the other Ganeti libraries), is provided in the source tree as |
716 |
``lib/rapi/client.py``. Note this client might change its interface in |
717 |
the future, as we iterate on its capabilities. |
718 |
|
719 |
A new command, ``gnt-cluster renew-crypto`` has been added to easily |
720 |
replace the cluster's certificates and crypto keys. This might help in |
721 |
case they have been compromised, or have simply expired. |
722 |
|
723 |
A new disk option for instance creation has been added that allows one |
724 |
to "adopt" currently existing logical volumes, with data |
725 |
preservation. This should allow easier migration to Ganeti from |
726 |
unmanaged (or managed via other software) instances. |
727 |
|
728 |
Another disk improvement is the possibility to convert between redundant |
729 |
(DRBD) and plain (LVM) disk configuration for an instance. This should |
730 |
allow better scalability (starting with one node and growing the |
731 |
cluster, or shrinking a two-node cluster to one node). |
732 |
|
733 |
A new feature that could help with automated node failovers has been |
734 |
implemented: if a node sees itself as offline (by querying the master |
735 |
candidates), it will try to shutdown (hard) all instances and any active |
736 |
DRBD devices. This reduces the risk of duplicate instances if an |
737 |
external script automatically failovers the instances on such nodes. To |
738 |
enable this, the cluster parameter ``maintain_node_health`` should be |
739 |
enabled; in the future this option (per the name) will enable other |
740 |
automatic maintenance features. |
741 |
|
742 |
Instance export/import now will reuse the original instance |
743 |
specifications for all parameters; that means exporting an instance, |
744 |
deleting it and the importing it back should give an almost identical |
745 |
instance. Note that the default import behaviour has changed from |
746 |
before, where it created only one NIC; now it recreates the original |
747 |
number of NICs. |
748 |
|
749 |
Cluster verify has added a few new checks: SSL certificates validity, |
750 |
/etc/hosts consistency across the cluster, etc. |
751 |
|
752 |
Other changes |
753 |
~~~~~~~~~~~~~ |
754 |
|
755 |
As usual, many internal changes were done, documentation fixes, |
756 |
etc. Among others: |
757 |
|
758 |
- Fixed cluster initialization with disabled cluster storage (regression |
759 |
introduced in 2.1.1) |
760 |
- File-based storage supports growing the disks |
761 |
- Fixed behaviour of node role changes |
762 |
- Fixed cluster verify for some corner cases, plus a general rewrite of |
763 |
cluster verify to allow future extension with more checks |
764 |
- Fixed log spamming by watcher and node daemon (regression introduced |
765 |
in 2.1.1) |
766 |
- Fixed possible validation issues when changing the list of enabled |
767 |
hypervisors |
768 |
- Fixed cleanup of /etc/hosts during node removal |
769 |
- Fixed RAPI response for invalid methods |
770 |
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon |
771 |
- Multiple small improvements to the KVM hypervisor (VNC usage, booting |
772 |
from ide disks, etc.) |
773 |
- Allow OS changes without re-installation (to record a changed OS |
774 |
outside of Ganeti, or to allow OS renames) |
775 |
- Allow instance creation without OS installation (useful for example if |
776 |
the OS will be installed manually, or restored from a backup not in |
777 |
Ganeti format) |
778 |
- Implemented option to make cluster ``copyfile`` use the replication |
779 |
network |
780 |
- Added list of enabled hypervisors to ssconf (possibly useful for |
781 |
external scripts) |
782 |
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from |
783 |
1.2 clusters |
784 |
- A partial form of node re-IP is possible via node readd, which now |
785 |
allows changed node primary IP |
786 |
- Command line utilities now show an informational message if the job is |
787 |
waiting for a lock |
788 |
- The logs of the master daemon now show the PID/UID/GID of the |
789 |
connected client |
790 |
|
791 |
|
792 |
Version 2.1.1 |
793 |
------------- |
794 |
|
795 |
*(Released Fri, 12 Mar 2010)* |
796 |
|
797 |
During the 2.1.0 long release candidate cycle, a lot of improvements and |
798 |
changes have accumulated with were released later as 2.1.1. |
799 |
|
800 |
Major changes |
801 |
~~~~~~~~~~~~~ |
802 |
|
803 |
The node evacuate command (``gnt-node evacuate``) was significantly |
804 |
rewritten, and as such the IAllocator protocol was changed - a new |
805 |
request type has been added. This unfortunate change during a stable |
806 |
series is designed to improve performance of node evacuations; on |
807 |
clusters with more than about five nodes and which are well-balanced, |
808 |
evacuation should proceed in parallel for all instances of the node |
809 |
being evacuated. As such, any existing IAllocator scripts need to be |
810 |
updated, otherwise the above command will fail due to the unknown |
811 |
request. The provided "dumb" allocator has not been updated; but the |
812 |
ganeti-htools package supports the new protocol since version 0.2.4. |
813 |
|
814 |
Another important change is increased validation of node and instance |
815 |
names. This might create problems in special cases, if invalid host |
816 |
names are being used. |
817 |
|
818 |
Also, a new layer of hypervisor parameters has been added, that sits at |
819 |
OS level between the cluster defaults and the instance ones. This allows |
820 |
customisation of virtualization parameters depending on the installed |
821 |
OS. For example instances with OS 'X' may have a different KVM kernel |
822 |
(or any other parameter) than the cluster defaults. This is intended to |
823 |
help managing a multiple OSes on the same cluster, without manual |
824 |
modification of each instance's parameters. |
825 |
|
826 |
A tool for merging clusters, ``cluster-merge``, has been added in the |
827 |
tools sub-directory. |
828 |
|
829 |
Bug fixes |
830 |
~~~~~~~~~ |
831 |
|
832 |
- Improved the int/float conversions that should make the code more |
833 |
robust in face of errors from the node daemons |
834 |
- Fixed the remove node code in case of internal configuration errors |
835 |
- Fixed the node daemon behaviour in face of inconsistent queue |
836 |
directory (e.g. read-only file-system where we can't open the files |
837 |
read-write, etc.) |
838 |
- Fixed the behaviour of gnt-node modify for master candidate demotion; |
839 |
now it either aborts cleanly or, if given the new "auto_promote" |
840 |
parameter, will automatically promote other nodes as needed |
841 |
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would |
842 |
completely prevent Ganeti from working |
843 |
- Fixed bug for instance export when not all disks were successfully |
844 |
exported |
845 |
- Fixed behaviour of node add when the new node is slow in starting up |
846 |
the node daemon |
847 |
- Fixed handling of signals in the LUXI client, which should improve |
848 |
behaviour of command-line scripts |
849 |
- Added checks for invalid node/instance names in the configuration (now |
850 |
flagged during cluster verify) |
851 |
- Fixed watcher behaviour for disk activation errors |
852 |
- Fixed two potentially endless loops in http library, which led to the |
853 |
RAPI daemon hanging and consuming 100% CPU in some cases |
854 |
- Fixed bug in RAPI daemon related to hashed passwords |
855 |
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM |
856 |
instances |
857 |
- Enhanced compatibility with non-Debian OSes, but not using absolute |
858 |
path in some commands and allowing customisation of the ssh |
859 |
configuration directory |
860 |
- Fixed possible future issue with new Python versions by abiding to the |
861 |
proper use of ``__slots__`` attribute on classes |
862 |
- Added checks that should prevent directory traversal attacks |
863 |
- Many documentation fixes based on feedback from users |
864 |
|
865 |
New features |
866 |
~~~~~~~~~~~~ |
867 |
|
868 |
- Added an "early_release" more for instance replace disks and node |
869 |
evacuate, where we release locks earlier and thus allow higher |
870 |
parallelism within the cluster |
871 |
- Added watcher hooks, intended to allow the watcher to restart other |
872 |
daemons (e.g. from the ganeti-nbma project), but they can be used of |
873 |
course for any other purpose |
874 |
- Added a compile-time disable for DRBD barriers, to increase |
875 |
performance if the administrator trusts the power supply or the |
876 |
storage system to not lose writes |
877 |
- Added the option of using syslog for logging instead of, or in |
878 |
addition to, Ganeti's own log files |
879 |
- Removed boot restriction for paravirtual NICs for KVM, recent versions |
880 |
can indeed boot from a paravirtual NIC |
881 |
- Added a generic debug level for many operations; while this is not |
882 |
used widely yet, it allows one to pass the debug value all the way to |
883 |
the OS scripts |
884 |
- Enhanced the hooks environment for instance moves (failovers, |
885 |
migrations) where the primary/secondary nodes changed during the |
886 |
operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars |
887 |
- Enhanced data validations for many user-supplied values; one important |
888 |
item is the restrictions imposed on instance and node names, which |
889 |
might reject some (invalid) host names |
890 |
- Add a configure-time option to disable file-based storage, if it's not |
891 |
needed; this allows greater security separation between the master |
892 |
node and the other nodes from the point of view of the inter-node RPC |
893 |
protocol |
894 |
- Added user notification in interactive tools if job is waiting in the |
895 |
job queue or trying to acquire locks |
896 |
- Added log messages when a job is waiting for locks |
897 |
- Added filtering by node tags in instance operations which admit |
898 |
multiple instances (start, stop, reboot, reinstall) |
899 |
- Added a new tool for cluster mergers, ``cluster-merge`` |
900 |
- Parameters from command line which are of the form ``a=b,c=d`` can now |
901 |
use backslash escapes to pass in values which contain commas, |
902 |
e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value |
903 |
``b,c`` |
904 |
- For KVM, the instance name is the first parameter passed to KVM, so |
905 |
that it's more visible in the process list |
906 |
|
907 |
|
908 |
Version 2.1.0 |
909 |
------------- |
910 |
|
911 |
*(Released Tue, 2 Mar 2010)* |
912 |
|
913 |
Ganeti 2.1 brings many improvements with it. Major changes: |
914 |
|
915 |
- Added infrastructure to ease automated disk repairs |
916 |
- Added new daemon to export configuration data in a cheaper way than |
917 |
using the remote API |
918 |
- Instance NICs can now be routed instead of being associated with a |
919 |
networking bridge |
920 |
- Improved job locking logic to reduce impact of jobs acquiring multiple |
921 |
locks waiting for other long-running jobs |
922 |
|
923 |
In-depth implementation details can be found in the Ganeti 2.1 design |
924 |
document. |
925 |
|
926 |
Details |
927 |
~~~~~~~ |
928 |
|
929 |
- Added chroot hypervisor |
930 |
- Added more options to xen-hvm hypervisor (``kernel_path`` and |
931 |
``device_model``) |
932 |
- Added more options to xen-pvm hypervisor (``use_bootloader``, |
933 |
``bootloader_path`` and ``bootloader_args``) |
934 |
- Added the ``use_localtime`` option for the xen-hvm and kvm |
935 |
hypervisors, and the default value for this has changed to false (in |
936 |
2.0 xen-hvm always enabled it) |
937 |
- Added luxi call to submit multiple jobs in one go |
938 |
- Added cluster initialization option to not modify ``/etc/hosts`` |
939 |
file on nodes |
940 |
- Added network interface parameters |
941 |
- Added dry run mode to some LUs |
942 |
- Added RAPI resources: |
943 |
|
944 |
- ``/2/instances/[instance_name]/info`` |
945 |
- ``/2/instances/[instance_name]/replace-disks`` |
946 |
- ``/2/nodes/[node_name]/evacuate`` |
947 |
- ``/2/nodes/[node_name]/migrate`` |
948 |
- ``/2/nodes/[node_name]/role`` |
949 |
- ``/2/nodes/[node_name]/storage`` |
950 |
- ``/2/nodes/[node_name]/storage/modify`` |
951 |
- ``/2/nodes/[node_name]/storage/repair`` |
952 |
|
953 |
- Added OpCodes to evacuate or migrate all instances on a node |
954 |
- Added new command to list storage elements on nodes (``gnt-node |
955 |
list-storage``) and modify them (``gnt-node modify-storage``) |
956 |
- Added new ssconf files with master candidate IP address |
957 |
(``ssconf_master_candidates_ips``), node primary IP address |
958 |
(``ssconf_node_primary_ips``) and node secondary IP address |
959 |
(``ssconf_node_secondary_ips``) |
960 |
- Added ``ganeti-confd`` and a client library to query the Ganeti |
961 |
configuration via UDP |
962 |
- Added ability to run hooks after cluster initialization and before |
963 |
cluster destruction |
964 |
- Added automatic mode for disk replace (``gnt-instance replace-disks |
965 |
--auto``) |
966 |
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks |
967 |
after catastrophic data-loss |
968 |
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume |
969 |
groups |
970 |
- Added ``gnt-instance move`` command to move instances |
971 |
- Added ``gnt-cluster watcher`` command to control watcher |
972 |
- Added ``gnt-node powercycle`` command to powercycle nodes |
973 |
- Added new job status field ``lock_status`` |
974 |
- Added parseable error codes to cluster verification (``gnt-cluster |
975 |
verify --error-codes``) and made output less verbose (use |
976 |
``--verbose`` to restore previous behaviour) |
977 |
- Added UUIDs to the main config entities (cluster, nodes, instances) |
978 |
- Added support for OS variants |
979 |
- Added support for hashed passwords in the Ganeti remote API users file |
980 |
(``rapi_users``) |
981 |
- Added option to specify maximum timeout on instance shutdown |
982 |
- Added ``--no-ssh-init`` option to ``gnt-cluster init`` |
983 |
- Added new helper script to start and stop Ganeti daemons |
984 |
(``daemon-util``), with the intent to reduce the work necessary to |
985 |
adjust Ganeti for non-Debian distributions and to start/stop daemons |
986 |
from one place |
987 |
- Added more unittests |
988 |
- Fixed critical bug in ganeti-masterd startup |
989 |
- Removed the configure-time ``kvm-migration-port`` parameter, this is |
990 |
now customisable at the cluster level for both the KVM and Xen |
991 |
hypervisors using the new ``migration_port`` parameter |
992 |
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when |
993 |
reinstalling an instance |
994 |
- Allowed ``@`` in tag names |
995 |
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation |
996 |
- Many documentation updates |
997 |
- Distribute hypervisor files on ``gnt-cluster redist-conf`` |
998 |
- ``gnt-instance reinstall`` can now reinstall multiple instances |
999 |
- Updated many command line parameters |
1000 |
- Introduced new OS API version 15 |
1001 |
- No longer support a default hypervisor |
1002 |
- Treat virtual LVs as inexistent |
1003 |
- Improved job locking logic to reduce lock contention |
1004 |
- Match instance and node names case insensitively |
1005 |
- Reimplemented bash completion script to be more complete |
1006 |
- Improved burnin |
1007 |
|
1008 |
|
1009 |
Version 2.0.6 |
1010 |
------------- |
1011 |
|
1012 |
*(Released Thu, 4 Feb 2010)* |
1013 |
|
1014 |
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105) |
1015 |
- Fix a string formatting bug |
1016 |
- Improve safety of the code in some error paths |
1017 |
- Improve data validation in the master of values returned from nodes |
1018 |
|
1019 |
|
1020 |
Version 2.0.5 |
1021 |
------------- |
1022 |
|
1023 |
*(Released Thu, 17 Dec 2009)* |
1024 |
|
1025 |
- Fix security issue due to missing validation of iallocator names; this |
1026 |
allows local and remote execution of arbitrary executables |
1027 |
- Fix failure of gnt-node list during instance removal |
1028 |
- Ship the RAPI documentation in the archive |
1029 |
|
1030 |
|
1031 |
Version 2.0.4 |
1032 |
------------- |
1033 |
|
1034 |
*(Released Wed, 30 Sep 2009)* |
1035 |
|
1036 |
- Fixed many wrong messages |
1037 |
- Fixed a few bugs related to the locking library |
1038 |
- Fixed MAC checking at instance creation time |
1039 |
- Fixed a DRBD parsing bug related to gaps in /proc/drbd |
1040 |
- Fixed a few issues related to signal handling in both daemons and |
1041 |
scripts |
1042 |
- Fixed the example startup script provided |
1043 |
- Fixed insserv dependencies in the example startup script (patch from |
1044 |
Debian) |
1045 |
- Fixed handling of drained nodes in the iallocator framework |
1046 |
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug |
1047 |
#528618) |
1048 |
- Fixed error related to invalid job IDs in job polling |
1049 |
- Fixed job/opcode persistence on unclean master shutdown |
1050 |
- Fixed handling of partial job processing after unclean master |
1051 |
shutdown |
1052 |
- Fixed error reporting from LUs, previously all errors were converted |
1053 |
into execution errors |
1054 |
- Fixed error reporting from burnin |
1055 |
- Decreased significantly the memory usage of the job queue |
1056 |
- Optimised slightly multi-job submission |
1057 |
- Optimised slightly opcode loading |
1058 |
- Backported the multi-job submit framework from the development |
1059 |
branch; multi-instance start and stop should be faster |
1060 |
- Added script to clean archived jobs after 21 days; this will reduce |
1061 |
the size of the queue directory |
1062 |
- Added some extra checks in disk size tracking |
1063 |
- Added an example ethers hook script |
1064 |
- Added a cluster parameter that prevents Ganeti from modifying of |
1065 |
/etc/hosts |
1066 |
- Added more node information to RAPI responses |
1067 |
- Added a ``gnt-job watch`` command that allows following the ouput of a |
1068 |
job |
1069 |
- Added a bind-address option to ganeti-rapi |
1070 |
- Added more checks to the configuration verify |
1071 |
- Enhanced the burnin script such that some operations can be retried |
1072 |
automatically |
1073 |
- Converted instance reinstall to multi-instance model |
1074 |
|
1075 |
|
1076 |
Version 2.0.3 |
1077 |
------------- |
1078 |
|
1079 |
*(Released Fri, 7 Aug 2009)* |
1080 |
|
1081 |
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command |
1082 |
to allow using the pre-2.0.2 behaviour in activation, if any existing |
1083 |
instances have mismatched disk sizes in the configuration |
1084 |
- Added ``gnt-cluster repair-disk-sizes`` command to check and update |
1085 |
any configuration mismatches for disk sizes |
1086 |
- Added ``gnt-master cluste-failover --no-voting`` to allow master |
1087 |
failover to work on two-node clusters |
1088 |
- Fixed the ``--net`` option of ``gnt-backup import``, which was |
1089 |
unusable |
1090 |
- Fixed detection of OS script errors in ``gnt-backup export`` |
1091 |
- Fixed exit code of ``gnt-backup export`` |
1092 |
|
1093 |
|
1094 |
Version 2.0.2 |
1095 |
------------- |
1096 |
|
1097 |
*(Released Fri, 17 Jul 2009)* |
1098 |
|
1099 |
- Added experimental support for stripped logical volumes; this should |
1100 |
enhance performance but comes with a higher complexity in the block |
1101 |
device handling; stripping is only enabled when passing |
1102 |
``--with-lvm-stripecount=N`` to ``configure``, but codepaths are |
1103 |
affected even in the non-stripped mode |
1104 |
- Improved resiliency against transient failures at the end of DRBD |
1105 |
resyncs, and in general of DRBD resync checks |
1106 |
- Fixed a couple of issues with exports and snapshot errors |
1107 |
- Fixed a couple of issues in instance listing |
1108 |
- Added display of the disk size in ``gnt-instance info`` |
1109 |
- Fixed checking for valid OSes in instance creation |
1110 |
- Fixed handling of the "vcpus" parameter in instance listing and in |
1111 |
general of invalid parameters |
1112 |
- Fixed http server library, and thus RAPI, to handle invalid |
1113 |
username/password combinations correctly; this means that now they |
1114 |
report unauthorized for queries too, not only for modifications, |
1115 |
allowing earlier detect of configuration problems |
1116 |
- Added a new "role" node list field, equivalent to the master/master |
1117 |
candidate/drained/offline flags combinations |
1118 |
- Fixed cluster modify and changes of candidate pool size |
1119 |
- Fixed cluster verify error messages for wrong files on regular nodes |
1120 |
- Fixed a couple of issues with node demotion from master candidate role |
1121 |
- Fixed node readd issues |
1122 |
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup |
1123 |
- Added a new ``--no-voting`` option for masterfailover to fix failover |
1124 |
on two-nodes clusters when the former master node is unreachable |
1125 |
- Added instance reinstall over RAPI |
1126 |
|
1127 |
|
1128 |
Version 2.0.1 |
1129 |
------------- |
1130 |
|
1131 |
*(Released Tue, 16 Jun 2009)* |
1132 |
|
1133 |
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will |
1134 |
allow re-adding the start in single-user option (regression from 1.2) |
1135 |
- the watcher writes the instance status to a file, to allow monitoring |
1136 |
to report the instance status (from the master) based on cached |
1137 |
results of the watcher's queries; while this can get stale if the |
1138 |
watcher is being locked due to other work on the cluster, this is |
1139 |
still an improvement |
1140 |
- the watcher now also restarts the node daemon and the rapi daemon if |
1141 |
they died |
1142 |
- fixed the watcher to handle full and drained queue cases |
1143 |
- hooks export more instance data in the environment, which helps if |
1144 |
hook scripts need to take action based on the instance's properties |
1145 |
(no longer need to query back into ganeti) |
1146 |
- instance failovers when the instance is stopped do not check for free |
1147 |
RAM, so that failing over a stopped instance is possible in low memory |
1148 |
situations |
1149 |
- rapi uses queries for tags instead of jobs (for less job traffic), and |
1150 |
for cluster tags it won't talk to masterd at all but read them from |
1151 |
ssconf |
1152 |
- a couple of error handling fixes in RAPI |
1153 |
- drbd handling: improved the error handling of inconsistent disks after |
1154 |
resync to reduce the frequency of "there are some degraded disks for |
1155 |
this instance" messages |
1156 |
- fixed a bug in live migration when DRBD doesn't want to reconnect (the |
1157 |
error handling path called a wrong function name) |
1158 |
|
1159 |
|
1160 |
Version 2.0.0 final |
1161 |
------------------- |
1162 |
|
1163 |
*(Released Wed, 27 May 2009)* |
1164 |
|
1165 |
- no changes from rc5 |
1166 |
|
1167 |
|
1168 |
Version 2.0 release candidate 5 |
1169 |
------------------------------- |
1170 |
|
1171 |
*(Released Wed, 20 May 2009)* |
1172 |
|
1173 |
- fix a couple of bugs (validation, argument checks) |
1174 |
- fix ``gnt-cluster getmaster`` on non-master nodes (regression) |
1175 |
- some small improvements to RAPI and IAllocator |
1176 |
- make watcher automatically start the master daemon if down |
1177 |
|
1178 |
|
1179 |
Version 2.0 release candidate 4 |
1180 |
------------------------------- |
1181 |
|
1182 |
*(Released Mon, 27 Apr 2009)* |
1183 |
|
1184 |
- change the OS list to not require locks; this helps with big clusters |
1185 |
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the |
1186 |
volume group is broken |
1187 |
- ``gnt-instance info``, without any arguments, doesn't run for all |
1188 |
instances anymore; either pass ``--all`` or pass the desired |
1189 |
instances; this helps against mistakes on big clusters where listing |
1190 |
the information for all instances takes a long time |
1191 |
- miscellaneous doc and man pages fixes |
1192 |
|
1193 |
|
1194 |
Version 2.0 release candidate 3 |
1195 |
------------------------------- |
1196 |
|
1197 |
*(Released Wed, 8 Apr 2009)* |
1198 |
|
1199 |
- Change the internal locking model of some ``gnt-node`` commands, in |
1200 |
order to reduce contention (and blocking of master daemon) when |
1201 |
batching many creation/reinstall jobs |
1202 |
- Fixes to Xen soft reboot |
1203 |
- No longer build documentation at build time, instead distribute it in |
1204 |
the archive, in order to reduce the need for the whole docbook/rst |
1205 |
toolchains |
1206 |
|
1207 |
|
1208 |
Version 2.0 release candidate 2 |
1209 |
------------------------------- |
1210 |
|
1211 |
*(Released Fri, 27 Mar 2009)* |
1212 |
|
1213 |
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0 |
1214 |
- Fix watcher startup sequence, improves the behaviour of busy clusters |
1215 |
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance |
1216 |
replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC |
1217 |
bind address and other places |
1218 |
- Some documentation fixes and updates |
1219 |
|
1220 |
|
1221 |
Version 2.0 release candidate 1 |
1222 |
------------------------------- |
1223 |
|
1224 |
*(Released Mon, 2 Mar 2009)* |
1225 |
|
1226 |
- More documentation updates, now all docs should be more-or-less |
1227 |
up-to-date |
1228 |
- A couple of small fixes (mixed hypervisor clusters, offline nodes, |
1229 |
etc.) |
1230 |
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM |
1231 |
and KVM) |
1232 |
- Fix an issue related to $libdir/run/ganeti and cluster creation |
1233 |
|
1234 |
|
1235 |
Version 2.0 beta 2 |
1236 |
------------------ |
1237 |
|
1238 |
*(Released Thu, 19 Feb 2009)* |
1239 |
|
1240 |
- Xen PVM and KVM have switched the default value for the instance root |
1241 |
disk to the first partition on the first drive, instead of the whole |
1242 |
drive; this means that the OS installation scripts must be changed |
1243 |
accordingly |
1244 |
- Man pages have been updated |
1245 |
- RAPI has been switched by default to HTTPS, and the exported functions |
1246 |
should all work correctly |
1247 |
- RAPI v1 has been removed |
1248 |
- Many improvements to the KVM hypervisor |
1249 |
- Block device errors are now better reported |
1250 |
- Many other bugfixes and small improvements |
1251 |
|
1252 |
|
1253 |
Version 2.0 beta 1 |
1254 |
------------------ |
1255 |
|
1256 |
*(Released Mon, 26 Jan 2009)* |
1257 |
|
1258 |
- Version 2 is a general rewrite of the code and therefore the |
1259 |
differences are too many to list, see the design document for 2.0 in |
1260 |
the ``doc/`` subdirectory for more details |
1261 |
- In this beta version there is not yet a migration path from 1.2 (there |
1262 |
will be one in the final 2.0 release) |
1263 |
- A few significant changes are: |
1264 |
|
1265 |
- all commands are executed by a daemon (``ganeti-masterd``) and the |
1266 |
various ``gnt-*`` commands are just front-ends to it |
1267 |
- all the commands are entered into, and executed from a job queue, |
1268 |
see the ``gnt-job(8)`` manpage |
1269 |
- the RAPI daemon supports read-write operations, secured by basic |
1270 |
HTTP authentication on top of HTTPS |
1271 |
- DRBD version 0.7 support has been removed, DRBD 8 is the only |
1272 |
supported version (when migrating from Ganeti 1.2 to 2.0, you need |
1273 |
to migrate to DRBD 8 first while still running Ganeti 1.2) |
1274 |
- DRBD devices are using statically allocated minor numbers, which |
1275 |
will be assigned to existing instances during the migration process |
1276 |
- there is support for both Xen PVM and Xen HVM instances running on |
1277 |
the same cluster |
1278 |
- KVM virtualization is supported too |
1279 |
- file-based storage has been implemented, which means that it is |
1280 |
possible to run the cluster without LVM and DRBD storage, for |
1281 |
example using a shared filesystem exported from shared storage (and |
1282 |
still have live migration) |
1283 |
|
1284 |
|
1285 |
Version 1.2.7 |
1286 |
------------- |
1287 |
|
1288 |
*(Released Tue, 13 Jan 2009)* |
1289 |
|
1290 |
- Change the default reboot type in ``gnt-instance reboot`` to "hard" |
1291 |
- Reuse the old instance mac address by default on instance import, if |
1292 |
the instance name is the same. |
1293 |
- Handle situations in which the node info rpc returns incomplete |
1294 |
results (issue 46) |
1295 |
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify`` |
1296 |
- Improved version of batcher: |
1297 |
|
1298 |
- state file support |
1299 |
- instance mac address support |
1300 |
- support for HVM clusters/instances |
1301 |
|
1302 |
- Add an option to show the number of cpu sockets and nodes in |
1303 |
``gnt-node list`` |
1304 |
- Support OSes that handle more than one version of the OS api (but do |
1305 |
not change the current API in any other way) |
1306 |
- Fix ``gnt-node migrate`` |
1307 |
- ``gnt-debug`` man page |
1308 |
- Fixes various more typos and small issues |
1309 |
- Increase disk resync maximum speed to 60MB/s (from 30MB/s) |
1310 |
|
1311 |
|
1312 |
Version 1.2.6 |
1313 |
------------- |
1314 |
|
1315 |
*(Released Wed, 24 Sep 2008)* |
1316 |
|
1317 |
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the |
1318 |
type of disk exported to fully virtualized instances. |
1319 |
- provide access to the serial console of HVM instances |
1320 |
- instance auto_balance flag, set by default. If turned off it will |
1321 |
avoid warnings on cluster verify if there is not enough memory to fail |
1322 |
over an instance. in the future it will prevent automatically failing |
1323 |
it over when we will support that. |
1324 |
- batcher tool for instance creation, see ``tools/README.batcher`` |
1325 |
- ``gnt-instance reinstall --select-os`` to interactively select a new |
1326 |
operating system when reinstalling an instance. |
1327 |
- when changing the memory amount on instance modify a check has been |
1328 |
added that the instance will be able to start. also warnings are |
1329 |
emitted if the instance will not be able to fail over, if auto_balance |
1330 |
is true. |
1331 |
- documentation fixes |
1332 |
- sync fields between ``gnt-instance list/modify/add/import`` |
1333 |
- fix a race condition in drbd when the sync speed was set after giving |
1334 |
the device a remote peer. |
1335 |
|
1336 |
|
1337 |
Version 1.2.5 |
1338 |
------------- |
1339 |
|
1340 |
*(Released Tue, 22 Jul 2008)* |
1341 |
|
1342 |
- note: the allowed size and number of tags per object were reduced |
1343 |
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups |
1344 |
- fixed twisted 8.x compatibility |
1345 |
- fixed ``gnt-instance replace-disks`` with iallocator |
1346 |
- add TCP keepalives on twisted connections to detect restarted nodes |
1347 |
- disk increase support, see ``gnt-instance grow-disk`` |
1348 |
- implement bulk node/instance query for RAPI |
1349 |
- add tags in node/instance listing (optional) |
1350 |
- experimental migration (and live migration) support, read the man page |
1351 |
for ``gnt-instance migrate`` |
1352 |
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also |
1353 |
has some small improvements in handling its state file |
1354 |
|
1355 |
|
1356 |
Version 1.2.4 |
1357 |
------------- |
1358 |
|
1359 |
*(Released Fri, 13 Jun 2008)* |
1360 |
|
1361 |
- Experimental readonly, REST-based remote API implementation; |
1362 |
automatically started on master node, TCP port 5080, if enabled by |
1363 |
``--enable-rapi`` parameter to configure script. |
1364 |
- Instance allocator support. Add and import instance accept a |
1365 |
``--iallocator`` parameter, and call that instance allocator to decide |
1366 |
which node to use for the instance. The iallocator document describes |
1367 |
what's expected from an allocator script. |
1368 |
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the |
1369 |
``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a |
1370 |
node is lost there is still enough memory to fail over the instances |
1371 |
that reside on it. |
1372 |
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to |
1373 |
``gnt-cluster verify``, to check for site-specific compliance. All the |
1374 |
hooks will run, and their output, if any, will be displayed. Any |
1375 |
failing hook will make the verification return an error value. |
1376 |
- ``gnt-cluster verify`` now checks that its peers are reachable on the |
1377 |
primary and secondary interfaces |
1378 |
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node |
1379 |
that is still declared as part of the cluster and has failed. |
1380 |
- ``gnt-* list`` commands now accept a new ``-o +field`` way of |
1381 |
specifying output fields, that just adds the chosen fields to the |
1382 |
default ones. |
1383 |
- ``gnt-backup`` now has a new ``remove`` command to delete an existing |
1384 |
export from the filesystem. |
1385 |
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path |
1386 |
have been added. Using them you can enable/disable acpi and pae |
1387 |
support, and specify a path for a cd image to be exported to the |
1388 |
instance. These parameters as the name suggest only work on HVM |
1389 |
clusters. |
1390 |
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and |
1391 |
PAE support will be set to the previously hardcoded values, but the |
1392 |
(previously hardcoded) path to the CDROM ISO image will be unset and |
1393 |
if required, needs to be set manually with ``gnt-instance modify`` |
1394 |
after the upgrade. |
1395 |
- The address to which an instance's VNC console is bound is now |
1396 |
selectable per-instance, rather than being cluster wide. Of course |
1397 |
this only applies to instances controlled via VNC, so currently just |
1398 |
applies to HVM clusters. |
1399 |
|
1400 |
|
1401 |
Version 1.2.3 |
1402 |
------------- |
1403 |
|
1404 |
*(Released Mon, 18 Feb 2008)* |
1405 |
|
1406 |
- more tweaks to the disk activation code (especially helpful for DRBD) |
1407 |
- change the default ``gnt-instance list`` output format, now there is |
1408 |
one combined status field (see the manpage for the exact values this |
1409 |
field will have) |
1410 |
- some more fixes for the mac export to hooks change |
1411 |
- make Ganeti not break with DRBD 8.2.x (which changed the version |
1412 |
format in ``/proc/drbd``) (issue 24) |
1413 |
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk |
1414 |
template, allowing migration from DRBD0.7+MD to DRBD8 |
1415 |
|
1416 |
|
1417 |
Version 1.2.2 |
1418 |
------------- |
1419 |
|
1420 |
*(Released Wed, 30 Jan 2008)* |
1421 |
|
1422 |
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM |
1423 |
support (issue 23) |
1424 |
- add command aliases infrastructure and a few aliases |
1425 |
- allow listing of VCPUs in the ``gnt-instance list`` and improve the |
1426 |
man pages and the ``--help`` option of ``gnt-node |
1427 |
list``/``gnt-instance list`` |
1428 |
- fix ``gnt-backup list`` with down nodes (issue 21) |
1429 |
- change the tools location (move from $pkgdatadir to $pkglibdir/tools) |
1430 |
- fix the dist archive and add a check for including svn/git files in |
1431 |
the future |
1432 |
- some developer-related changes: improve the burnin and the QA suite, |
1433 |
add an upload script for testing during development |
1434 |
|
1435 |
|
1436 |
Version 1.2.1 |
1437 |
------------- |
1438 |
|
1439 |
*(Released Wed, 16 Jan 2008)* |
1440 |
|
1441 |
- experimental HVM support, read the install document, section |
1442 |
"Initializing the cluster" |
1443 |
- allow for the PVM hypervisor per-instance kernel and initrd paths |
1444 |
- add a new command ``gnt-cluster verify-disks`` which uses a new |
1445 |
algorithm to improve the reconnection of the DRBD pairs if the device |
1446 |
on the secondary node has gone away |
1447 |
- make logical volume code auto-activate LVs at disk activation time |
1448 |
- slightly improve the speed of activating disks |
1449 |
- allow specification of the MAC address at instance creation time, and |
1450 |
changing it later via ``gnt-instance modify`` |
1451 |
- fix handling of external commands that generate lots of output on |
1452 |
stderr |
1453 |
- update documentation with regard to minimum version of DRBD8 supported |
1454 |
|
1455 |
|
1456 |
Version 1.2.0 |
1457 |
------------- |
1458 |
|
1459 |
*(Released Tue, 4 Dec 2007)* |
1460 |
|
1461 |
- Log the ``xm create`` output to the node daemon log on failure (to |
1462 |
help diagnosing the error) |
1463 |
- In debug mode, log all external commands output if failed to the logs |
1464 |
- Change parsing of lvm commands to ignore stderr |
1465 |
|
1466 |
|
1467 |
Version 1.2b3 |
1468 |
------------- |
1469 |
|
1470 |
*(Released Wed, 28 Nov 2007)* |
1471 |
|
1472 |
- Another round of updates to the DRBD 8 code to deal with more failures |
1473 |
in the replace secondary node operation |
1474 |
- Some more logging of failures in disk operations (lvm, drbd) |
1475 |
- A few documentation updates |
1476 |
- QA updates |
1477 |
|
1478 |
|
1479 |
Version 1.2b2 |
1480 |
------------- |
1481 |
|
1482 |
*(Released Tue, 13 Nov 2007)* |
1483 |
|
1484 |
- Change configuration file format from Python's Pickle to JSON. |
1485 |
Upgrading is possible using the cfgupgrade utility. |
1486 |
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for |
1487 |
faster replace disks and is more stable (DRBD 8 has many improvements |
1488 |
compared to DRBD 0.7) |
1489 |
- Added command line tags support (see man pages for ``gnt-instance``, |
1490 |
``gnt-node``, ``gnt-cluster``) |
1491 |
- Added instance rename support |
1492 |
- Added multi-instance startup/shutdown |
1493 |
- Added cluster rename support |
1494 |
- Added ``gnt-node evacuate`` to simplify some node operations |
1495 |
- Added instance reboot operation that can speedup reboot as compared to |
1496 |
stop and start |
1497 |
- Soften the requirement that hostnames are in FQDN format |
1498 |
- The ``ganeti-watcher`` now activates drbd pairs after secondary node |
1499 |
reboots |
1500 |
- Removed dependency on debian's patched fping that uses the |
1501 |
non-standard ``-S`` option |
1502 |
- Now the OS definitions are searched for in multiple, configurable |
1503 |
paths (easier for distros to package) |
1504 |
- Some changes to the hooks infrastructure (especially the new |
1505 |
post-configuration update hook) |
1506 |
- Other small bugfixes |
1507 |
|
1508 |
.. vim: set textwidth=72 syntax=rst : |
1509 |
.. Local Variables: |
1510 |
.. mode: rst |
1511 |
.. fill-column: 72 |
1512 |
.. End: |