root / NEWS @ e41a1c0c
History | View | Annotate | Download (55.8 kB)
1 |
News |
---|---|
2 |
==== |
3 |
|
4 |
Version 2.4.0 rc2 |
5 |
----------------- |
6 |
|
7 |
*(Released Mon, 21 Feb 2011)* |
8 |
|
9 |
A number of bug fixes plus just a couple functionality changes. |
10 |
|
11 |
On the user-visible side, the ``gnt-* list`` command output has changed |
12 |
with respect to “special” field states. The current rc1 style of display |
13 |
can be re-enabled by passing a new ``-v, --verbose`` flag, but in the |
14 |
default output mode special fields are displayed as follows: |
15 |
|
16 |
- offline field: ``*`` |
17 |
- unavailable/not applicable: ``-`` |
18 |
- data missing (RPC failure): ``?`` |
19 |
- unknown field: ``??`` |
20 |
|
21 |
Another user-visible change is the addition of ``--force-join`` to |
22 |
``gnt-node add``. |
23 |
|
24 |
As for bug fixes: |
25 |
|
26 |
- ``tools/cluster-merge`` has seen many fixes and is now enabled again |
27 |
- fixed regression in RAPI/instance reinstall where all parameters were |
28 |
required (instead of optional) |
29 |
- fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2 |
30 |
- fixed iallocator usage (offline nodes were not considered offline) |
31 |
- fixed ``gnt-node list`` with respect to non-vm_capable nodes |
32 |
- fixed hypervisor and OS parameter validation with respect to |
33 |
non-vm_capable nodes |
34 |
- fixed ``gnt-cluster verify`` with respect to offline nodes (mostly |
35 |
cosmetic) |
36 |
- fixed ``tools/listrunner`` with respect to agent-based usage |
37 |
|
38 |
|
39 |
Version 2.4.0 rc1 |
40 |
----------------- |
41 |
|
42 |
*(Released Fri, 4 Feb 2011)* |
43 |
|
44 |
Many changes and fixes since the beta1 release. While there were some |
45 |
internal changes, the code has been mostly stabilised for the RC |
46 |
release. |
47 |
|
48 |
Note: the dumb allocator was removed in this release, as it was not kept |
49 |
up-to-date with the IAllocator protocol changes. It is recommended to |
50 |
use the ``hail`` command from the ganeti-htools package. |
51 |
|
52 |
Note: the 2.4 and up versions of Ganeti are not compatible with the |
53 |
0.2.x branch of ganeti-htools. You need to upgrade to |
54 |
ganeti-htools-0.3.0 (or later). |
55 |
|
56 |
Regressions fixed from 2.3 |
57 |
~~~~~~~~~~~~~~~~~~~~~~~~~~ |
58 |
|
59 |
- Fixed the ``gnt-cluster verify-disks`` command |
60 |
- Made ``gnt-cluster verify-disks`` work in parallel (as opposed to |
61 |
serially on nodes) |
62 |
- Fixed disk adoption breakage |
63 |
- Fixed wrong headers in instance listing for field aliases |
64 |
|
65 |
Other bugs fixed |
66 |
~~~~~~~~~~~~~~~~ |
67 |
|
68 |
- Fixed corner case in KVM handling of NICs |
69 |
- Fixed many cases of wrong handling of non-vm_capable nodes |
70 |
- Fixed a bug where a missing instance symlink was not possible to |
71 |
recreate with any ``gnt-*`` command (now ``gnt-instance |
72 |
activate-disks`` does it) |
73 |
- Fixed the volume group name as reported by ``gnt-cluster |
74 |
verify-disks`` |
75 |
- Increased timeouts for the import-export code, hopefully leading to |
76 |
fewer aborts due network or instance timeouts |
77 |
- Fixed bug in ``gnt-node list-storage`` |
78 |
- Fixed bug where not all daemons were started on cluster |
79 |
initialisation, but only at the first watcher run |
80 |
- Fixed many bugs in the OOB implementation |
81 |
- Fixed watcher behaviour in presence of instances with offline |
82 |
secondaries |
83 |
- Fixed instance list output for instances running on the wrong node |
84 |
- a few fixes to the cluster-merge tool, but it still cannot merge |
85 |
multi-node groups (currently it is not recommended to use this tool) |
86 |
|
87 |
|
88 |
Improvements |
89 |
~~~~~~~~~~~~ |
90 |
|
91 |
- Improved network configuration for the KVM hypervisor |
92 |
- Added e1000 as a supported NIC for Xen-HVM |
93 |
- Improved the lvmstrap tool to also be able to use partitions, as |
94 |
opposed to full disks |
95 |
- Improved speed of disk wiping (the cluster parameter |
96 |
``prealloc_wipe_disks``, so that it has a low impact on the total time |
97 |
of instance creations |
98 |
- Added documentation for the OS parameters |
99 |
- Changed ``gnt-instance deactivate-disks`` so that it can work if the |
100 |
hypervisor is not responding |
101 |
- Added display of blacklisted and hidden OS information in |
102 |
``gnt-cluster info`` |
103 |
- Extended ``gnt-cluster verify`` to also validate hypervisor, backend, |
104 |
NIC and node parameters, which might create problems with currently |
105 |
invalid (but undetected) configuration files, but prevents validation |
106 |
failures when unrelated parameters are modified |
107 |
- Changed cluster initialisation to wait for the master daemon to become |
108 |
available |
109 |
- Expanded the RAPI interface: |
110 |
|
111 |
- Added config redistribution resource |
112 |
- Added activation/deactivation of instance disks |
113 |
- Added export of console information |
114 |
|
115 |
- Implemented log file reopening on SIGHUP, which allows using |
116 |
logrotate(8) for the Ganeti log files |
117 |
- Added a basic OOB helper script as an example |
118 |
|
119 |
|
120 |
Version 2.4.0 beta1 |
121 |
------------------- |
122 |
|
123 |
*(Released Fri, 14 Jan 2011)* |
124 |
|
125 |
User-visible |
126 |
~~~~~~~~~~~~ |
127 |
|
128 |
- Fixed timezone issues when formatting timestamps |
129 |
- Added support for node groups, available via ``gnt-group`` and other |
130 |
commands |
131 |
- Added out-of-band framework and management, see :doc:`design |
132 |
document <design-oob>` |
133 |
- Removed support for roman numbers from ``gnt-node list`` and |
134 |
``gnt-instance list``. |
135 |
- Allowed modification of master network interface via ``gnt-cluster |
136 |
modify --master-netdev`` |
137 |
- Accept offline secondaries while shutting down instance disks |
138 |
- Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors |
139 |
- Added support for multiple LVM volume groups |
140 |
- Avoid sorting nodes for ``gnt-node list`` if specific nodes are |
141 |
requested |
142 |
- Added commands to list available fields: |
143 |
|
144 |
- ``gnt-node list-fields`` |
145 |
- ``gnt-group list-fields`` |
146 |
- ``gnt-instance list-fields`` |
147 |
|
148 |
- Updated documentation and man pages |
149 |
|
150 |
Integration |
151 |
~~~~~~~~~~~ |
152 |
|
153 |
- Moved ``rapi_users`` file into separate directory, now named |
154 |
``…/ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a |
155 |
symlink |
156 |
- Added new tool for running commands on many machines, |
157 |
``tools/ganeti-listrunner`` |
158 |
- Implemented more verbose result in ``OpInstanceConsole`` opcode, also |
159 |
improving the ``gnt-instance console`` output |
160 |
- Allowed customisation of disk index separator at ``configure`` time |
161 |
- Export node group allocation policy to :doc:`iallocator <iallocator>` |
162 |
- Added support for non-partitioned md disks in ``lvmstrap`` |
163 |
- Added script to gracefully power off KVM instances |
164 |
- Split ``utils`` module into smaller parts |
165 |
- Changed query operations to return more detailed information, e.g. |
166 |
whether an information is unavailable due to an offline node. To use |
167 |
this new functionality, the LUXI call ``Query`` must be used. Field |
168 |
information is now stored by the master daemon and can be retrieved |
169 |
using ``QueryFields``. Instances, nodes and groups can also be queried |
170 |
using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet |
171 |
exposed via RAPI). The following commands make use of this |
172 |
infrastructure change: |
173 |
|
174 |
- ``gnt-group list`` |
175 |
- ``gnt-group list-fields`` |
176 |
- ``gnt-node list`` |
177 |
- ``gnt-node list-fields`` |
178 |
- ``gnt-instance list`` |
179 |
- ``gnt-instance list-fields`` |
180 |
- ``gnt-debug locks`` |
181 |
|
182 |
Remote API |
183 |
~~~~~~~~~~ |
184 |
|
185 |
- New RAPI resources (see :doc:`rapi`): |
186 |
|
187 |
- ``/2/modify`` |
188 |
- ``/2/groups`` |
189 |
- ``/2/groups/[group_name]`` |
190 |
- ``/2/groups/[group_name]/assign-nodes`` |
191 |
- ``/2/groups/[group_name]/modify`` |
192 |
- ``/2/groups/[group_name]/rename`` |
193 |
- ``/2/instances/[instance_name]/disk/[disk_index]/grow`` |
194 |
|
195 |
- RAPI changes: |
196 |
|
197 |
- Implemented ``no_install`` for instance creation |
198 |
- Implemented OS parameters for instance reinstallation, allowing |
199 |
use of special settings on reinstallation (e.g. for preserving data) |
200 |
|
201 |
Misc |
202 |
~~~~ |
203 |
|
204 |
- Added IPv6 support in import/export |
205 |
- Pause DRBD synchronization while wiping disks on instance creation |
206 |
- Updated unittests and QA scripts |
207 |
- Improved network parameters passed to KVM |
208 |
- Converted man pages from docbook to reStructuredText |
209 |
|
210 |
|
211 |
Version 2.3.1 |
212 |
------------- |
213 |
|
214 |
*(Released Mon, 20 Dec 2010)* |
215 |
|
216 |
Released version 2.3.1~rc1 without any changes. |
217 |
|
218 |
|
219 |
Version 2.3.1 rc1 |
220 |
----------------- |
221 |
|
222 |
*(Released Wed, 1 Dec 2010)* |
223 |
|
224 |
- impexpd: Disable OpenSSL compression in socat if possible (backport |
225 |
from master, commit e90739d625b, see :doc:`installation guide |
226 |
<install-quick>` for details) |
227 |
- Changed unittest coverage report to exclude test scripts |
228 |
- Added script to check version format |
229 |
|
230 |
|
231 |
Version 2.3.0 |
232 |
------------- |
233 |
|
234 |
*(Released Wed, 1 Dec 2010)* |
235 |
|
236 |
Released version 2.3.0~rc1 without any changes. |
237 |
|
238 |
|
239 |
Version 2.3.0 rc1 |
240 |
----------------- |
241 |
|
242 |
*(Released Fri, 19 Nov 2010)* |
243 |
|
244 |
A number of bugfixes and documentation updates: |
245 |
|
246 |
- Update ganeti-os-interface documentation |
247 |
- Fixed a bug related to duplicate MACs or similar items which should be |
248 |
unique |
249 |
- Fix breakage in OS state modify |
250 |
- Reinstall instance: disallow offline secondaries (fixes bug related to |
251 |
OS changing but reinstall failing) |
252 |
- plus all the other fixes between 2.2.1 and 2.2.2 |
253 |
|
254 |
|
255 |
Version 2.3.0 rc0 |
256 |
----------------- |
257 |
|
258 |
*(Released Tue, 2 Nov 2010)* |
259 |
|
260 |
- Fixed clearing of the default iallocator using ``gnt-cluster modify`` |
261 |
- Fixed master failover race with watcher |
262 |
- Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent |
263 |
configuration |
264 |
- Accept previously stopped instance for export with instance removal |
265 |
- Simplify and extend the environment variables for instance OS scripts |
266 |
- Added new node flags, ``master_capable`` and ``vm_capable`` |
267 |
- Added optional instance disk wiping prior during allocation. This is a |
268 |
cluster-wide option and can be set/modified using |
269 |
``gnt-cluster {init,modify} --prealloc-wipe-disks``. |
270 |
- Added IPv6 support, see :doc:`design document <design-2.3>` and |
271 |
:doc:`install-quick` |
272 |
- Added a new watcher option (``--ignore-pause``) |
273 |
- Added option to ignore offline node on instance start/stop |
274 |
(``--ignore-offline``) |
275 |
- Allow overriding OS parameters with ``gnt-instance reinstall`` |
276 |
- Added ability to change node's secondary IP address using ``gnt-node |
277 |
modify`` |
278 |
- Implemented privilege separation for all daemons except |
279 |
``ganeti-noded``, see ``configure`` options |
280 |
- Complain if an instance's disk is marked faulty in ``gnt-cluster |
281 |
verify`` |
282 |
- Implemented job priorities (see ``ganeti(7)`` manpage) |
283 |
- Ignore failures while shutting down instances during failover from |
284 |
offline node |
285 |
- Exit daemon's bootstrap process only once daemon is ready |
286 |
- Export more information via ``LUInstanceQuery``/remote API |
287 |
- Improved documentation, QA and unittests |
288 |
- RAPI daemon now watches ``rapi_users`` all the time and doesn't need a |
289 |
restart if the file was created or changed |
290 |
- Added LUXI protocol version sent with each request and response, |
291 |
allowing detection of server/client mismatches |
292 |
- Moved the Python scripts among gnt-* and ganeti-* into modules |
293 |
- Moved all code related to setting up SSH to an external script, |
294 |
``setup-ssh`` |
295 |
- Infrastructure changes for node group support in future versions |
296 |
|
297 |
|
298 |
Version 2.2.2 |
299 |
------------- |
300 |
|
301 |
*(Released Fri, 19 Nov 2010)* |
302 |
|
303 |
A few small bugs fixed, and some improvements to the build system: |
304 |
|
305 |
- Fix documentation regarding conversion to drbd |
306 |
- Fix validation of parameters in cluster modify (``gnt-cluster modify |
307 |
-B``) |
308 |
- Fix error handling in node modify with multiple changes |
309 |
- Allow remote imports without checked names |
310 |
|
311 |
|
312 |
Version 2.2.1 |
313 |
------------- |
314 |
|
315 |
*(Released Tue, 19 Oct 2010)* |
316 |
|
317 |
- Disable SSL session ID cache in RPC client |
318 |
|
319 |
|
320 |
Version 2.2.1 rc1 |
321 |
----------------- |
322 |
|
323 |
*(Released Thu, 14 Oct 2010)* |
324 |
|
325 |
- Fix interaction between Curl/GnuTLS and the Python's HTTP server |
326 |
(thanks Apollon Oikonomopoulos!), finally allowing the use of Curl |
327 |
with GnuTLS |
328 |
- Fix problems with interaction between Curl and Python's HTTP server, |
329 |
resulting in increased speed in many RPC calls |
330 |
- Improve our release script to prevent breakage with older aclocal and |
331 |
Python 2.6 |
332 |
|
333 |
|
334 |
Version 2.2.1 rc0 |
335 |
----------------- |
336 |
|
337 |
*(Released Thu, 7 Oct 2010)* |
338 |
|
339 |
- Fixed issue 125, replace hardcoded “xenvg” in ``gnt-cluster`` with |
340 |
value retrieved from master |
341 |
- Added support for blacklisted or hidden OS definitions |
342 |
- Added simple lock monitor (accessible via (``gnt-debug locks``) |
343 |
- Added support for -mem-path in KVM hypervisor abstraction layer |
344 |
- Allow overriding instance parameters in tool for inter-cluster |
345 |
instance moves (``tools/move-instance``) |
346 |
- Improved opcode summaries (e.g. in ``gnt-job list``) |
347 |
- Improve consistency of OS listing by sorting it |
348 |
- Documentation updates |
349 |
|
350 |
|
351 |
Version 2.2.0.1 |
352 |
--------------- |
353 |
|
354 |
*(Released Fri, 8 Oct 2010)* |
355 |
|
356 |
- Rebuild with a newer autotools version, to fix python 2.6 compatibility |
357 |
|
358 |
|
359 |
Version 2.2.0 |
360 |
------------- |
361 |
|
362 |
*(Released Mon, 4 Oct 2010)* |
363 |
|
364 |
- Fixed regression in ``gnt-instance rename`` |
365 |
|
366 |
|
367 |
Version 2.2.0 rc2 |
368 |
----------------- |
369 |
|
370 |
*(Released Wed, 22 Sep 2010)* |
371 |
|
372 |
- Fixed OS_VARIANT variable for OS scripts |
373 |
- Fixed cluster tag operations via RAPI |
374 |
- Made ``setup-ssh`` exit with non-zero code if an error occurred |
375 |
- Disabled RAPI CA checks in watcher |
376 |
|
377 |
|
378 |
Version 2.2.0 rc1 |
379 |
----------------- |
380 |
|
381 |
*(Released Mon, 23 Aug 2010)* |
382 |
|
383 |
- Support DRBD versions of the format "a.b.c.d" |
384 |
- Updated manpages |
385 |
- Re-introduce support for usage from multiple threads in RAPI client |
386 |
- Instance renames and modify via RAPI |
387 |
- Work around race condition between processing and archival in job |
388 |
queue |
389 |
- Mark opcodes following failed one as failed, too |
390 |
- Job field ``lock_status`` was removed due to difficulties making it |
391 |
work with the changed job queue in Ganeti 2.2; a better way to monitor |
392 |
locks is expected for a later 2.2.x release |
393 |
- Fixed dry-run behaviour with many commands |
394 |
- Support ``ssh-agent`` again when adding nodes |
395 |
- Many additional bugfixes |
396 |
|
397 |
|
398 |
Version 2.2.0 rc0 |
399 |
----------------- |
400 |
|
401 |
*(Released Fri, 30 Jul 2010)* |
402 |
|
403 |
Important change: the internal RPC mechanism between Ganeti nodes has |
404 |
changed from using a home-grown http library (based on the Python base |
405 |
libraries) to use the PycURL library. This requires that PycURL is |
406 |
installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked |
407 |
against GnuTLS by default. cURL's support for GnuTLS had known issues |
408 |
before cURL 7.21.0 and we recommend using the latest cURL release or |
409 |
linking against OpenSSL. Most other distributions already link PycURL |
410 |
and cURL against OpenSSL. The command:: |
411 |
|
412 |
python -c 'import pycurl; print pycurl.version' |
413 |
|
414 |
can be used to determine the libraries PycURL and cURL are linked |
415 |
against. |
416 |
|
417 |
Other significant changes: |
418 |
|
419 |
- Rewrote much of the internals of the job queue, in order to achieve |
420 |
better parallelism; this decouples job query operations from the job |
421 |
processing, and it should allow much nicer behaviour of the master |
422 |
daemon under load, and it also has uncovered some long-standing bugs |
423 |
related to the job serialisation (now fixed) |
424 |
- Added a default iallocator setting to the cluster parameters, |
425 |
eliminating the need to always pass nodes or an iallocator for |
426 |
operations that require selection of new node(s) |
427 |
- Added experimental support for the LXC virtualization method |
428 |
- Added support for OS parameters, which allows the installation of |
429 |
instances to pass parameter to OS scripts in order to customise the |
430 |
instance |
431 |
- Added a hypervisor parameter controlling the migration type (live or |
432 |
non-live), since hypervisors have various levels of reliability; this |
433 |
has renamed the 'live' parameter to 'mode' |
434 |
- Added a cluster parameter ``reserved_lvs`` that denotes reserved |
435 |
logical volumes, meaning that cluster verify will ignore them and not |
436 |
flag their presence as errors |
437 |
- The watcher will now reset the error count for failed instances after |
438 |
8 hours, thus allowing self-healing if the problem that caused the |
439 |
instances to be down/fail to start has cleared in the meantime |
440 |
- Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti |
441 |
check for, and warn, if the drbd module parameter ``usermode_helper`` |
442 |
is not consistent with the cluster-wide setting; this is needed to |
443 |
make diagnose easier of failed drbd creations |
444 |
- Started adding base IPv6 support, but this is not yet |
445 |
enabled/available for use |
446 |
- Rename operations (cluster, instance) will now return the new name, |
447 |
which is especially useful if a short name was passed in |
448 |
- Added support for instance migration in RAPI |
449 |
- Added a tool to pre-configure nodes for the SSH setup, before joining |
450 |
them to the cluster; this will allow in the future a simplified model |
451 |
for node joining (but not yet fully enabled in 2.2); this needs the |
452 |
paramiko python library |
453 |
- Fixed handling of name-resolving errors |
454 |
- Fixed consistency of job results on the error path |
455 |
- Fixed master-failover race condition when executed multiple times in |
456 |
sequence |
457 |
- Fixed many bugs related to the job queue (mostly introduced during the |
458 |
2.2 development cycle, so not all are impacting 2.1) |
459 |
- Fixed instance migration with missing disk symlinks |
460 |
- Fixed handling of unknown jobs in ``gnt-job archive`` |
461 |
- And many other small fixes/improvements |
462 |
|
463 |
Internal changes: |
464 |
|
465 |
- Enhanced both the unittest and the QA coverage |
466 |
- Switched the opcode validation to a generic model, and extended the |
467 |
validation to all opcode parameters |
468 |
- Changed more parts of the code that write shell scripts to use the |
469 |
same class for this |
470 |
- Switched the master daemon to use the asyncore library for the Luxi |
471 |
server endpoint |
472 |
|
473 |
|
474 |
Version 2.2.0 beta 0 |
475 |
-------------------- |
476 |
|
477 |
*(Released Thu, 17 Jun 2010)* |
478 |
|
479 |
- Added tool (``move-instance``) and infrastructure to move instances |
480 |
between separate clusters (see :doc:`separate documentation |
481 |
<move-instance>` and :doc:`design document <design-2.2>`) |
482 |
- Added per-request RPC timeout |
483 |
- RAPI now requires a Content-Type header for requests with a body (e.g. |
484 |
``PUT`` or ``POST``) which must be set to ``application/json`` (see |
485 |
:rfc:`2616` (HTTP/1.1), section 7.2.1) |
486 |
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not |
487 |
reachable |
488 |
- Implemented initial support for running Ganeti daemons as separate |
489 |
users, see configure-time flags ``--with-user-prefix`` and |
490 |
``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this |
491 |
time) |
492 |
- Instances can be removed after export (``gnt-backup export |
493 |
--remove-instance``) |
494 |
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA |
495 |
key (instead of 1024 bit) |
496 |
- Added new cluster configuration file for cluster domain secret |
497 |
- Import/export now use SSL instead of SSH |
498 |
- Added support for showing estimated time when exporting an instance, |
499 |
see the ``ganeti-os-interface(7)`` manpage and look for |
500 |
``EXP_SIZE_FD`` |
501 |
|
502 |
|
503 |
Version 2.1.8 |
504 |
------------- |
505 |
|
506 |
*(Released Tue, 16 Nov 2010)* |
507 |
|
508 |
Some more bugfixes. Unless critical bugs occur, this will be the last |
509 |
2.1 release: |
510 |
|
511 |
- Fix case of MAC special-values |
512 |
- Fix mac checker regex |
513 |
- backend: Fix typo causing “out of range” error |
514 |
- Add missing --units in gnt-instance list man page |
515 |
|
516 |
|
517 |
Version 2.1.7 |
518 |
------------- |
519 |
|
520 |
*(Released Tue, 24 Aug 2010)* |
521 |
|
522 |
Bugfixes only: |
523 |
- Don't ignore secondary node silently on non-mirrored disk templates |
524 |
(issue 113) |
525 |
- Fix --master-netdev arg name in gnt-cluster(8) (issue 114) |
526 |
- Fix usb_mouse parameter breaking with vnc_console (issue 109) |
527 |
- Properly document the usb_mouse parameter |
528 |
- Fix path in ganeti-rapi(8) (issue 116) |
529 |
- Adjust error message when the ganeti user's .ssh directory is |
530 |
missing |
531 |
- Add same-node-check when changing the disk template to drbd |
532 |
|
533 |
|
534 |
Version 2.1.6 |
535 |
------------- |
536 |
|
537 |
*(Released Fri, 16 Jul 2010)* |
538 |
|
539 |
Bugfixes only: |
540 |
- Add an option to only select some reboot types during qa/burnin. |
541 |
(on some hypervisors consequent reboots are not supported) |
542 |
- Fix infrequent race condition in master failover. Sometimes the old |
543 |
master ip address would be still detected as up for a short time |
544 |
after it was removed, causing failover to fail. |
545 |
- Decrease mlockall warnings when the ctypes module is missing. On |
546 |
Python 2.4 we support running even if no ctypes module is installed, |
547 |
but we were too verbose about this issue. |
548 |
- Fix building on old distributions, on which man doesn't have a |
549 |
--warnings option. |
550 |
- Fix RAPI not to ignore the MAC address on instance creation |
551 |
- Implement the old instance creation format in the RAPI client. |
552 |
|
553 |
|
554 |
Version 2.1.5 |
555 |
------------- |
556 |
|
557 |
*(Released Thu, 01 Jul 2010)* |
558 |
|
559 |
A small bugfix release: |
560 |
- Fix disk adoption: broken by strict --disk option checking in 2.1.4 |
561 |
- Fix batch-create: broken in the whole 2.1 series due to a lookup on |
562 |
a non-existing option |
563 |
- Fix instance create: the --force-variant option was ignored |
564 |
- Improve pylint 0.21 compatibility and warnings with Python 2.6 |
565 |
- Fix modify node storage with non-FQDN arguments |
566 |
- Fix RAPI client to authenticate under Python 2.6 when used |
567 |
for more than 5 requests needing authentication |
568 |
- Fix gnt-instance modify -t (storage) giving a wrong error message |
569 |
when converting a non-shutdown drbd instance to plain |
570 |
|
571 |
|
572 |
Version 2.1.4 |
573 |
------------- |
574 |
|
575 |
*(Released Fri, 18 Jun 2010)* |
576 |
|
577 |
A small bugfix release: |
578 |
|
579 |
- Fix live migration of KVM instances started with older Ganeti |
580 |
versions which had fewer hypervisor parameters |
581 |
- Fix gnt-instance grow-disk on down instances |
582 |
- Fix an error-reporting bug during instance migration |
583 |
- Better checking of the ``--net`` and ``--disk`` values, to avoid |
584 |
silently ignoring broken ones |
585 |
- Fix an RPC error reporting bug affecting, for example, RAPI client |
586 |
users |
587 |
- Fix bug triggered by different API version os-es on different nodes |
588 |
- Fix a bug in instance startup with custom hvparams: OS level |
589 |
parameters would fail to be applied. |
590 |
- Fix the RAPI client under Python 2.6 (but more work is needed to |
591 |
make it work completely well with OpenSSL) |
592 |
- Fix handling of errors when resolving names from DNS |
593 |
|
594 |
|
595 |
Version 2.1.3 |
596 |
------------- |
597 |
|
598 |
*(Released Thu, 3 Jun 2010)* |
599 |
|
600 |
A medium sized development cycle. Some new features, and some |
601 |
fixes/small improvements/cleanups. |
602 |
|
603 |
Significant features |
604 |
~~~~~~~~~~~~~~~~~~~~ |
605 |
|
606 |
The node deamon now tries to mlock itself into memory, unless the |
607 |
``--no-mlock`` flag is passed. It also doesn't fail if it can't write |
608 |
its logs, and falls back to console logging. This allows emergency |
609 |
features such as ``gnt-node powercycle`` to work even in the event of a |
610 |
broken node disk (tested offlining the disk hosting the node's |
611 |
filesystem and dropping its memory caches; don't try this at home) |
612 |
|
613 |
KVM: add vhost-net acceleration support. It can be tested with a new |
614 |
enough version of the kernel and of qemu-kvm. |
615 |
|
616 |
KVM: Add instance chrooting feature. If you use privilege dropping for |
617 |
your VMs you can also now force them to chroot to an empty directory, |
618 |
before starting the emulated guest. |
619 |
|
620 |
KVM: Add maximum migration bandwith and maximum downtime tweaking |
621 |
support (requires a new-enough version of qemu-kvm). |
622 |
|
623 |
Cluster verify will now warn if the master node doesn't have the master |
624 |
ip configured on it. |
625 |
|
626 |
Add a new (incompatible) instance creation request format to RAPI which |
627 |
supports all parameters (previously only a subset was supported, and it |
628 |
wasn't possible to extend the old format to accomodate all the new |
629 |
features. The old format is still supported, and a client can check for |
630 |
this feature, before using it, by checking for its presence in the |
631 |
``features`` RAPI resource. |
632 |
|
633 |
Now with ancient latin support. Try it passing the ``--roman`` option to |
634 |
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list`` |
635 |
(requires the python-roman module to be installed, in order to work). |
636 |
|
637 |
Other changes |
638 |
~~~~~~~~~~~~~ |
639 |
|
640 |
As usual many internal code refactorings, documentation updates, and |
641 |
such. Among others: |
642 |
|
643 |
- Lots of improvements and cleanups to the experimental Remote API |
644 |
(RAPI) client library. |
645 |
- A new unit test suite for the core daemon libraries. |
646 |
- A fix to creating missing directories makes sure the umask is not |
647 |
applied anymore. This enforces the same directory permissions |
648 |
everywhere. |
649 |
- Better handling terminating daemons with ctrl+c (used when running |
650 |
them in debugging mode). |
651 |
- Fix a race condition in live migrating a KVM instance, when stat() |
652 |
on the old proc status file returned EINVAL, which is an unexpected |
653 |
value. |
654 |
- Fixed manpage checking with newer man and utf-8 charachters. But now |
655 |
you need the en_US.UTF-8 locale enabled to build Ganeti from git. |
656 |
|
657 |
|
658 |
Version 2.1.2.1 |
659 |
--------------- |
660 |
|
661 |
*(Released Fri, 7 May 2010)* |
662 |
|
663 |
Fix a bug which prevented untagged KVM instances from starting. |
664 |
|
665 |
|
666 |
Version 2.1.2 |
667 |
------------- |
668 |
|
669 |
*(Released Fri, 7 May 2010)* |
670 |
|
671 |
Another release with a long development cycle, during which many |
672 |
different features were added. |
673 |
|
674 |
Significant features |
675 |
~~~~~~~~~~~~~~~~~~~~ |
676 |
|
677 |
The KVM hypervisor now can run the individual instances as non-root, to |
678 |
reduce the impact of a VM being hijacked due to bugs in the |
679 |
hypervisor. It is possible to run all instances as a single (non-root) |
680 |
user, to manually specify a user for each instance, or to dynamically |
681 |
allocate a user out of a cluster-wide pool to each instance, with the |
682 |
guarantee that no two instances will run under the same user ID on any |
683 |
given node. |
684 |
|
685 |
An experimental RAPI client library, that can be used standalone |
686 |
(without the other Ganeti libraries), is provided in the source tree as |
687 |
``lib/rapi/client.py``. Note this client might change its interface in |
688 |
the future, as we iterate on its capabilities. |
689 |
|
690 |
A new command, ``gnt-cluster renew-crypto`` has been added to easily |
691 |
replace the cluster's certificates and crypto keys. This might help in |
692 |
case they have been compromised, or have simply expired. |
693 |
|
694 |
A new disk option for instance creation has been added that allows one |
695 |
to "adopt" currently existing logical volumes, with data |
696 |
preservation. This should allow easier migration to Ganeti from |
697 |
unmanaged (or managed via other software) instances. |
698 |
|
699 |
Another disk improvement is the possibility to convert between redundant |
700 |
(DRBD) and plain (LVM) disk configuration for an instance. This should |
701 |
allow better scalability (starting with one node and growing the |
702 |
cluster, or shrinking a two-node cluster to one node). |
703 |
|
704 |
A new feature that could help with automated node failovers has been |
705 |
implemented: if a node sees itself as offline (by querying the master |
706 |
candidates), it will try to shutdown (hard) all instances and any active |
707 |
DRBD devices. This reduces the risk of duplicate instances if an |
708 |
external script automatically failovers the instances on such nodes. To |
709 |
enable this, the cluster parameter ``maintain_node_health`` should be |
710 |
enabled; in the future this option (per the name) will enable other |
711 |
automatic maintenance features. |
712 |
|
713 |
Instance export/import now will reuse the original instance |
714 |
specifications for all parameters; that means exporting an instance, |
715 |
deleting it and the importing it back should give an almost identical |
716 |
instance. Note that the default import behaviour has changed from |
717 |
before, where it created only one NIC; now it recreates the original |
718 |
number of NICs. |
719 |
|
720 |
Cluster verify has added a few new checks: SSL certificates validity, |
721 |
/etc/hosts consistency across the cluster, etc. |
722 |
|
723 |
Other changes |
724 |
~~~~~~~~~~~~~ |
725 |
|
726 |
As usual, many internal changes were done, documentation fixes, |
727 |
etc. Among others: |
728 |
|
729 |
- Fixed cluster initialization with disabled cluster storage (regression |
730 |
introduced in 2.1.1) |
731 |
- File-based storage supports growing the disks |
732 |
- Fixed behaviour of node role changes |
733 |
- Fixed cluster verify for some corner cases, plus a general rewrite of |
734 |
cluster verify to allow future extension with more checks |
735 |
- Fixed log spamming by watcher and node daemon (regression introduced |
736 |
in 2.1.1) |
737 |
- Fixed possible validation issues when changing the list of enabled |
738 |
hypervisors |
739 |
- Fixed cleanup of /etc/hosts during node removal |
740 |
- Fixed RAPI response for invalid methods |
741 |
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon |
742 |
- Multiple small improvements to the KVM hypervisor (VNC usage, booting |
743 |
from ide disks, etc.) |
744 |
- Allow OS changes without re-installation (to record a changed OS |
745 |
outside of Ganeti, or to allow OS renames) |
746 |
- Allow instance creation without OS installation (useful for example if |
747 |
the OS will be installed manually, or restored from a backup not in |
748 |
Ganeti format) |
749 |
- Implemented option to make cluster ``copyfile`` use the replication |
750 |
network |
751 |
- Added list of enabled hypervisors to ssconf (possibly useful for |
752 |
external scripts) |
753 |
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from |
754 |
1.2 clusters |
755 |
- A partial form of node re-IP is possible via node readd, which now |
756 |
allows changed node primary IP |
757 |
- Command line utilities now show an informational message if the job is |
758 |
waiting for a lock |
759 |
- The logs of the master daemon now show the PID/UID/GID of the |
760 |
connected client |
761 |
|
762 |
|
763 |
Version 2.1.1 |
764 |
------------- |
765 |
|
766 |
*(Released Fri, 12 Mar 2010)* |
767 |
|
768 |
During the 2.1.0 long release candidate cycle, a lot of improvements and |
769 |
changes have accumulated with were released later as 2.1.1. |
770 |
|
771 |
Major changes |
772 |
~~~~~~~~~~~~~ |
773 |
|
774 |
The node evacuate command (``gnt-node evacuate``) was significantly |
775 |
rewritten, and as such the IAllocator protocol was changed - a new |
776 |
request type has been added. This unfortunate change during a stable |
777 |
series is designed to improve performance of node evacuations; on |
778 |
clusters with more than about five nodes and which are well-balanced, |
779 |
evacuation should proceed in parallel for all instances of the node |
780 |
being evacuated. As such, any existing IAllocator scripts need to be |
781 |
updated, otherwise the above command will fail due to the unknown |
782 |
request. The provided "dumb" allocator has not been updated; but the |
783 |
ganeti-htools package supports the new protocol since version 0.2.4. |
784 |
|
785 |
Another important change is increased validation of node and instance |
786 |
names. This might create problems in special cases, if invalid host |
787 |
names are being used. |
788 |
|
789 |
Also, a new layer of hypervisor parameters has been added, that sits at |
790 |
OS level between the cluster defaults and the instance ones. This allows |
791 |
customisation of virtualization parameters depending on the installed |
792 |
OS. For example instances with OS 'X' may have a different KVM kernel |
793 |
(or any other parameter) than the cluster defaults. This is intended to |
794 |
help managing a multiple OSes on the same cluster, without manual |
795 |
modification of each instance's parameters. |
796 |
|
797 |
A tool for merging clusters, ``cluster-merge``, has been added in the |
798 |
tools sub-directory. |
799 |
|
800 |
Bug fixes |
801 |
~~~~~~~~~ |
802 |
|
803 |
- Improved the int/float conversions that should make the code more |
804 |
robust in face of errors from the node daemons |
805 |
- Fixed the remove node code in case of internal configuration errors |
806 |
- Fixed the node daemon behaviour in face of inconsistent queue |
807 |
directory (e.g. read-only file-system where we can't open the files |
808 |
read-write, etc.) |
809 |
- Fixed the behaviour of gnt-node modify for master candidate demotion; |
810 |
now it either aborts cleanly or, if given the new “auto_promote” |
811 |
parameter, will automatically promote other nodes as needed |
812 |
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would |
813 |
completely prevent Ganeti from working |
814 |
- Fixed bug for instance export when not all disks were successfully |
815 |
exported |
816 |
- Fixed behaviour of node add when the new node is slow in starting up |
817 |
the node daemon |
818 |
- Fixed handling of signals in the LUXI client, which should improve |
819 |
behaviour of command-line scripts |
820 |
- Added checks for invalid node/instance names in the configuration (now |
821 |
flagged during cluster verify) |
822 |
- Fixed watcher behaviour for disk activation errors |
823 |
- Fixed two potentially endless loops in http library, which led to the |
824 |
RAPI daemon hanging and consuming 100% CPU in some cases |
825 |
- Fixed bug in RAPI daemon related to hashed passwords |
826 |
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM |
827 |
instances |
828 |
- Enhanced compatibility with non-Debian OSes, but not using absolute |
829 |
path in some commands and allowing customisation of the ssh |
830 |
configuration directory |
831 |
- Fixed possible future issue with new Python versions by abiding to the |
832 |
proper use of ``__slots__`` attribute on classes |
833 |
- Added checks that should prevent directory traversal attacks |
834 |
- Many documentation fixes based on feedback from users |
835 |
|
836 |
New features |
837 |
~~~~~~~~~~~~ |
838 |
|
839 |
- Added an “early_release” more for instance replace disks and node |
840 |
evacuate, where we release locks earlier and thus allow higher |
841 |
parallelism within the cluster |
842 |
- Added watcher hooks, intended to allow the watcher to restart other |
843 |
daemons (e.g. from the ganeti-nbma project), but they can be used of |
844 |
course for any other purpose |
845 |
- Added a compile-time disable for DRBD barriers, to increase |
846 |
performance if the administrator trusts the power supply or the |
847 |
storage system to not lose writes |
848 |
- Added the option of using syslog for logging instead of, or in |
849 |
addition to, Ganeti's own log files |
850 |
- Removed boot restriction for paravirtual NICs for KVM, recent versions |
851 |
can indeed boot from a paravirtual NIC |
852 |
- Added a generic debug level for many operations; while this is not |
853 |
used widely yet, it allows one to pass the debug value all the way to |
854 |
the OS scripts |
855 |
- Enhanced the hooks environment for instance moves (failovers, |
856 |
migrations) where the primary/secondary nodes changed during the |
857 |
operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars |
858 |
- Enhanced data validations for many user-supplied values; one important |
859 |
item is the restrictions imposed on instance and node names, which |
860 |
might reject some (invalid) host names |
861 |
- Add a configure-time option to disable file-based storage, if it's not |
862 |
needed; this allows greater security separation between the master |
863 |
node and the other nodes from the point of view of the inter-node RPC |
864 |
protocol |
865 |
- Added user notification in interactive tools if job is waiting in the |
866 |
job queue or trying to acquire locks |
867 |
- Added log messages when a job is waiting for locks |
868 |
- Added filtering by node tags in instance operations which admit |
869 |
multiple instances (start, stop, reboot, reinstall) |
870 |
- Added a new tool for cluster mergers, ``cluster-merge`` |
871 |
- Parameters from command line which are of the form ``a=b,c=d`` can now |
872 |
use backslash escapes to pass in values which contain commas, |
873 |
e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value |
874 |
``b,c`` |
875 |
- For KVM, the instance name is the first parameter passed to KVM, so |
876 |
that it's more visible in the process list |
877 |
|
878 |
|
879 |
Version 2.1.0 |
880 |
------------- |
881 |
|
882 |
*(Released Tue, 2 Mar 2010)* |
883 |
|
884 |
Ganeti 2.1 brings many improvements with it. Major changes: |
885 |
|
886 |
- Added infrastructure to ease automated disk repairs |
887 |
- Added new daemon to export configuration data in a cheaper way than |
888 |
using the remote API |
889 |
- Instance NICs can now be routed instead of being associated with a |
890 |
networking bridge |
891 |
- Improved job locking logic to reduce impact of jobs acquiring multiple |
892 |
locks waiting for other long-running jobs |
893 |
|
894 |
In-depth implementation details can be found in the Ganeti 2.1 design |
895 |
document. |
896 |
|
897 |
Details |
898 |
~~~~~~~ |
899 |
|
900 |
- Added chroot hypervisor |
901 |
- Added more options to xen-hvm hypervisor (``kernel_path`` and |
902 |
``device_model``) |
903 |
- Added more options to xen-pvm hypervisor (``use_bootloader``, |
904 |
``bootloader_path`` and ``bootloader_args``) |
905 |
- Added the ``use_localtime`` option for the xen-hvm and kvm |
906 |
hypervisors, and the default value for this has changed to false (in |
907 |
2.0 xen-hvm always enabled it) |
908 |
- Added luxi call to submit multiple jobs in one go |
909 |
- Added cluster initialization option to not modify ``/etc/hosts`` |
910 |
file on nodes |
911 |
- Added network interface parameters |
912 |
- Added dry run mode to some LUs |
913 |
- Added RAPI resources: |
914 |
|
915 |
- ``/2/instances/[instance_name]/info`` |
916 |
- ``/2/instances/[instance_name]/replace-disks`` |
917 |
- ``/2/nodes/[node_name]/evacuate`` |
918 |
- ``/2/nodes/[node_name]/migrate`` |
919 |
- ``/2/nodes/[node_name]/role`` |
920 |
- ``/2/nodes/[node_name]/storage`` |
921 |
- ``/2/nodes/[node_name]/storage/modify`` |
922 |
- ``/2/nodes/[node_name]/storage/repair`` |
923 |
|
924 |
- Added OpCodes to evacuate or migrate all instances on a node |
925 |
- Added new command to list storage elements on nodes (``gnt-node |
926 |
list-storage``) and modify them (``gnt-node modify-storage``) |
927 |
- Added new ssconf files with master candidate IP address |
928 |
(``ssconf_master_candidates_ips``), node primary IP address |
929 |
(``ssconf_node_primary_ips``) and node secondary IP address |
930 |
(``ssconf_node_secondary_ips``) |
931 |
- Added ``ganeti-confd`` and a client library to query the Ganeti |
932 |
configuration via UDP |
933 |
- Added ability to run hooks after cluster initialization and before |
934 |
cluster destruction |
935 |
- Added automatic mode for disk replace (``gnt-instance replace-disks |
936 |
--auto``) |
937 |
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks |
938 |
after catastrophic data-loss |
939 |
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume |
940 |
groups |
941 |
- Added ``gnt-instance move`` command to move instances |
942 |
- Added ``gnt-cluster watcher`` command to control watcher |
943 |
- Added ``gnt-node powercycle`` command to powercycle nodes |
944 |
- Added new job status field ``lock_status`` |
945 |
- Added parseable error codes to cluster verification (``gnt-cluster |
946 |
verify --error-codes``) and made output less verbose (use |
947 |
``--verbose`` to restore previous behaviour) |
948 |
- Added UUIDs to the main config entities (cluster, nodes, instances) |
949 |
- Added support for OS variants |
950 |
- Added support for hashed passwords in the Ganeti remote API users file |
951 |
(``rapi_users``) |
952 |
- Added option to specify maximum timeout on instance shutdown |
953 |
- Added ``--no-ssh-init`` option to ``gnt-cluster init`` |
954 |
- Added new helper script to start and stop Ganeti daemons |
955 |
(``daemon-util``), with the intent to reduce the work necessary to |
956 |
adjust Ganeti for non-Debian distributions and to start/stop daemons |
957 |
from one place |
958 |
- Added more unittests |
959 |
- Fixed critical bug in ganeti-masterd startup |
960 |
- Removed the configure-time ``kvm-migration-port`` parameter, this is |
961 |
now customisable at the cluster level for both the KVM and Xen |
962 |
hypervisors using the new ``migration_port`` parameter |
963 |
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when |
964 |
reinstalling an instance |
965 |
- Allowed ``@`` in tag names |
966 |
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation |
967 |
- Many documentation updates |
968 |
- Distribute hypervisor files on ``gnt-cluster redist-conf`` |
969 |
- ``gnt-instance reinstall`` can now reinstall multiple instances |
970 |
- Updated many command line parameters |
971 |
- Introduced new OS API version 15 |
972 |
- No longer support a default hypervisor |
973 |
- Treat virtual LVs as inexistent |
974 |
- Improved job locking logic to reduce lock contention |
975 |
- Match instance and node names case insensitively |
976 |
- Reimplemented bash completion script to be more complete |
977 |
- Improved burnin |
978 |
|
979 |
|
980 |
Version 2.0.6 |
981 |
------------- |
982 |
|
983 |
*(Released Thu, 4 Feb 2010)* |
984 |
|
985 |
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105) |
986 |
- Fix a string formatting bug |
987 |
- Improve safety of the code in some error paths |
988 |
- Improve data validation in the master of values returned from nodes |
989 |
|
990 |
|
991 |
Version 2.0.5 |
992 |
------------- |
993 |
|
994 |
*(Released Thu, 17 Dec 2009)* |
995 |
|
996 |
- Fix security issue due to missing validation of iallocator names; this |
997 |
allows local and remote execution of arbitrary executables |
998 |
- Fix failure of gnt-node list during instance removal |
999 |
- Ship the RAPI documentation in the archive |
1000 |
|
1001 |
|
1002 |
Version 2.0.4 |
1003 |
------------- |
1004 |
|
1005 |
*(Released Wed, 30 Sep 2009)* |
1006 |
|
1007 |
- Fixed many wrong messages |
1008 |
- Fixed a few bugs related to the locking library |
1009 |
- Fixed MAC checking at instance creation time |
1010 |
- Fixed a DRBD parsing bug related to gaps in /proc/drbd |
1011 |
- Fixed a few issues related to signal handling in both daemons and |
1012 |
scripts |
1013 |
- Fixed the example startup script provided |
1014 |
- Fixed insserv dependencies in the example startup script (patch from |
1015 |
Debian) |
1016 |
- Fixed handling of drained nodes in the iallocator framework |
1017 |
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug |
1018 |
#528618) |
1019 |
- Fixed error related to invalid job IDs in job polling |
1020 |
- Fixed job/opcode persistence on unclean master shutdown |
1021 |
- Fixed handling of partial job processing after unclean master |
1022 |
shutdown |
1023 |
- Fixed error reporting from LUs, previously all errors were converted |
1024 |
into execution errors |
1025 |
- Fixed error reporting from burnin |
1026 |
- Decreased significantly the memory usage of the job queue |
1027 |
- Optimised slightly multi-job submission |
1028 |
- Optimised slightly opcode loading |
1029 |
- Backported the multi-job submit framework from the development |
1030 |
branch; multi-instance start and stop should be faster |
1031 |
- Added script to clean archived jobs after 21 days; this will reduce |
1032 |
the size of the queue directory |
1033 |
- Added some extra checks in disk size tracking |
1034 |
- Added an example ethers hook script |
1035 |
- Added a cluster parameter that prevents Ganeti from modifying of |
1036 |
/etc/hosts |
1037 |
- Added more node information to RAPI responses |
1038 |
- Added a ``gnt-job watch`` command that allows following the ouput of a |
1039 |
job |
1040 |
- Added a bind-address option to ganeti-rapi |
1041 |
- Added more checks to the configuration verify |
1042 |
- Enhanced the burnin script such that some operations can be retried |
1043 |
automatically |
1044 |
- Converted instance reinstall to multi-instance model |
1045 |
|
1046 |
|
1047 |
Version 2.0.3 |
1048 |
------------- |
1049 |
|
1050 |
*(Released Fri, 7 Aug 2009)* |
1051 |
|
1052 |
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command |
1053 |
to allow using the pre-2.0.2 behaviour in activation, if any existing |
1054 |
instances have mismatched disk sizes in the configuration |
1055 |
- Added ``gnt-cluster repair-disk-sizes`` command to check and update |
1056 |
any configuration mismatches for disk sizes |
1057 |
- Added ``gnt-master cluste-failover --no-voting`` to allow master |
1058 |
failover to work on two-node clusters |
1059 |
- Fixed the ``--net`` option of ``gnt-backup import``, which was |
1060 |
unusable |
1061 |
- Fixed detection of OS script errors in ``gnt-backup export`` |
1062 |
- Fixed exit code of ``gnt-backup export`` |
1063 |
|
1064 |
|
1065 |
Version 2.0.2 |
1066 |
------------- |
1067 |
|
1068 |
*(Released Fri, 17 Jul 2009)* |
1069 |
|
1070 |
- Added experimental support for stripped logical volumes; this should |
1071 |
enhance performance but comes with a higher complexity in the block |
1072 |
device handling; stripping is only enabled when passing |
1073 |
``--with-lvm-stripecount=N`` to ``configure``, but codepaths are |
1074 |
affected even in the non-stripped mode |
1075 |
- Improved resiliency against transient failures at the end of DRBD |
1076 |
resyncs, and in general of DRBD resync checks |
1077 |
- Fixed a couple of issues with exports and snapshot errors |
1078 |
- Fixed a couple of issues in instance listing |
1079 |
- Added display of the disk size in ``gnt-instance info`` |
1080 |
- Fixed checking for valid OSes in instance creation |
1081 |
- Fixed handling of the "vcpus" parameter in instance listing and in |
1082 |
general of invalid parameters |
1083 |
- Fixed http server library, and thus RAPI, to handle invalid |
1084 |
username/password combinations correctly; this means that now they |
1085 |
report unauthorized for queries too, not only for modifications, |
1086 |
allowing earlier detect of configuration problems |
1087 |
- Added a new "role" node list field, equivalent to the master/master |
1088 |
candidate/drained/offline flags combinations |
1089 |
- Fixed cluster modify and changes of candidate pool size |
1090 |
- Fixed cluster verify error messages for wrong files on regular nodes |
1091 |
- Fixed a couple of issues with node demotion from master candidate role |
1092 |
- Fixed node readd issues |
1093 |
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup |
1094 |
- Added a new ``--no-voting`` option for masterfailover to fix failover |
1095 |
on two-nodes clusters when the former master node is unreachable |
1096 |
- Added instance reinstall over RAPI |
1097 |
|
1098 |
|
1099 |
Version 2.0.1 |
1100 |
------------- |
1101 |
|
1102 |
*(Released Tue, 16 Jun 2009)* |
1103 |
|
1104 |
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will |
1105 |
allow re-adding the start in single-user option (regression from 1.2) |
1106 |
- the watcher writes the instance status to a file, to allow monitoring |
1107 |
to report the instance status (from the master) based on cached |
1108 |
results of the watcher's queries; while this can get stale if the |
1109 |
watcher is being locked due to other work on the cluster, this is |
1110 |
still an improvement |
1111 |
- the watcher now also restarts the node daemon and the rapi daemon if |
1112 |
they died |
1113 |
- fixed the watcher to handle full and drained queue cases |
1114 |
- hooks export more instance data in the environment, which helps if |
1115 |
hook scripts need to take action based on the instance's properties |
1116 |
(no longer need to query back into ganeti) |
1117 |
- instance failovers when the instance is stopped do not check for free |
1118 |
RAM, so that failing over a stopped instance is possible in low memory |
1119 |
situations |
1120 |
- rapi uses queries for tags instead of jobs (for less job traffic), and |
1121 |
for cluster tags it won't talk to masterd at all but read them from |
1122 |
ssconf |
1123 |
- a couple of error handling fixes in RAPI |
1124 |
- drbd handling: improved the error handling of inconsistent disks after |
1125 |
resync to reduce the frequency of "there are some degraded disks for |
1126 |
this instance" messages |
1127 |
- fixed a bug in live migration when DRBD doesn't want to reconnect (the |
1128 |
error handling path called a wrong function name) |
1129 |
|
1130 |
|
1131 |
Version 2.0.0 final |
1132 |
------------------- |
1133 |
|
1134 |
*(Released Wed, 27 May 2009)* |
1135 |
|
1136 |
- no changes from rc5 |
1137 |
|
1138 |
|
1139 |
Version 2.0 release candidate 5 |
1140 |
------------------------------- |
1141 |
|
1142 |
*(Released Wed, 20 May 2009)* |
1143 |
|
1144 |
- fix a couple of bugs (validation, argument checks) |
1145 |
- fix ``gnt-cluster getmaster`` on non-master nodes (regression) |
1146 |
- some small improvements to RAPI and IAllocator |
1147 |
- make watcher automatically start the master daemon if down |
1148 |
|
1149 |
|
1150 |
Version 2.0 release candidate 4 |
1151 |
------------------------------- |
1152 |
|
1153 |
*(Released Mon, 27 Apr 2009)* |
1154 |
|
1155 |
- change the OS list to not require locks; this helps with big clusters |
1156 |
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the |
1157 |
volume group is broken |
1158 |
- ``gnt-instance info``, without any arguments, doesn't run for all |
1159 |
instances anymore; either pass ``--all`` or pass the desired |
1160 |
instances; this helps against mistakes on big clusters where listing |
1161 |
the information for all instances takes a long time |
1162 |
- miscellaneous doc and man pages fixes |
1163 |
|
1164 |
|
1165 |
Version 2.0 release candidate 3 |
1166 |
------------------------------- |
1167 |
|
1168 |
*(Released Wed, 8 Apr 2009)* |
1169 |
|
1170 |
- Change the internal locking model of some ``gnt-node`` commands, in |
1171 |
order to reduce contention (and blocking of master daemon) when |
1172 |
batching many creation/reinstall jobs |
1173 |
- Fixes to Xen soft reboot |
1174 |
- No longer build documentation at build time, instead distribute it in |
1175 |
the archive, in order to reduce the need for the whole docbook/rst |
1176 |
toolchains |
1177 |
|
1178 |
|
1179 |
Version 2.0 release candidate 2 |
1180 |
------------------------------- |
1181 |
|
1182 |
*(Released Fri, 27 Mar 2009)* |
1183 |
|
1184 |
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0 |
1185 |
- Fix watcher startup sequence, improves the behaviour of busy clusters |
1186 |
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance |
1187 |
replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC |
1188 |
bind address and other places |
1189 |
- Some documentation fixes and updates |
1190 |
|
1191 |
|
1192 |
Version 2.0 release candidate 1 |
1193 |
------------------------------- |
1194 |
|
1195 |
*(Released Mon, 2 Mar 2009)* |
1196 |
|
1197 |
- More documentation updates, now all docs should be more-or-less |
1198 |
up-to-date |
1199 |
- A couple of small fixes (mixed hypervisor clusters, offline nodes, |
1200 |
etc.) |
1201 |
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM |
1202 |
and KVM) |
1203 |
- Fix an issue related to $libdir/run/ganeti and cluster creation |
1204 |
|
1205 |
|
1206 |
Version 2.0 beta 2 |
1207 |
------------------ |
1208 |
|
1209 |
*(Released Thu, 19 Feb 2009)* |
1210 |
|
1211 |
- Xen PVM and KVM have switched the default value for the instance root |
1212 |
disk to the first partition on the first drive, instead of the whole |
1213 |
drive; this means that the OS installation scripts must be changed |
1214 |
accordingly |
1215 |
- Man pages have been updated |
1216 |
- RAPI has been switched by default to HTTPS, and the exported functions |
1217 |
should all work correctly |
1218 |
- RAPI v1 has been removed |
1219 |
- Many improvements to the KVM hypervisor |
1220 |
- Block device errors are now better reported |
1221 |
- Many other bugfixes and small improvements |
1222 |
|
1223 |
|
1224 |
Version 2.0 beta 1 |
1225 |
------------------ |
1226 |
|
1227 |
*(Released Mon, 26 Jan 2009)* |
1228 |
|
1229 |
- Version 2 is a general rewrite of the code and therefore the |
1230 |
differences are too many to list, see the design document for 2.0 in |
1231 |
the ``doc/`` subdirectory for more details |
1232 |
- In this beta version there is not yet a migration path from 1.2 (there |
1233 |
will be one in the final 2.0 release) |
1234 |
- A few significant changes are: |
1235 |
|
1236 |
- all commands are executed by a daemon (``ganeti-masterd``) and the |
1237 |
various ``gnt-*`` commands are just front-ends to it |
1238 |
- all the commands are entered into, and executed from a job queue, |
1239 |
see the ``gnt-job(8)`` manpage |
1240 |
- the RAPI daemon supports read-write operations, secured by basic |
1241 |
HTTP authentication on top of HTTPS |
1242 |
- DRBD version 0.7 support has been removed, DRBD 8 is the only |
1243 |
supported version (when migrating from Ganeti 1.2 to 2.0, you need |
1244 |
to migrate to DRBD 8 first while still running Ganeti 1.2) |
1245 |
- DRBD devices are using statically allocated minor numbers, which |
1246 |
will be assigned to existing instances during the migration process |
1247 |
- there is support for both Xen PVM and Xen HVM instances running on |
1248 |
the same cluster |
1249 |
- KVM virtualization is supported too |
1250 |
- file-based storage has been implemented, which means that it is |
1251 |
possible to run the cluster without LVM and DRBD storage, for |
1252 |
example using a shared filesystem exported from shared storage (and |
1253 |
still have live migration) |
1254 |
|
1255 |
|
1256 |
Version 1.2.7 |
1257 |
------------- |
1258 |
|
1259 |
*(Released Tue, 13 Jan 2009)* |
1260 |
|
1261 |
- Change the default reboot type in ``gnt-instance reboot`` to "hard" |
1262 |
- Reuse the old instance mac address by default on instance import, if |
1263 |
the instance name is the same. |
1264 |
- Handle situations in which the node info rpc returns incomplete |
1265 |
results (issue 46) |
1266 |
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify`` |
1267 |
- Improved version of batcher: |
1268 |
|
1269 |
- state file support |
1270 |
- instance mac address support |
1271 |
- support for HVM clusters/instances |
1272 |
|
1273 |
- Add an option to show the number of cpu sockets and nodes in |
1274 |
``gnt-node list`` |
1275 |
- Support OSes that handle more than one version of the OS api (but do |
1276 |
not change the current API in any other way) |
1277 |
- Fix ``gnt-node migrate`` |
1278 |
- ``gnt-debug`` man page |
1279 |
- Fixes various more typos and small issues |
1280 |
- Increase disk resync maximum speed to 60MB/s (from 30MB/s) |
1281 |
|
1282 |
|
1283 |
Version 1.2.6 |
1284 |
------------- |
1285 |
|
1286 |
*(Released Wed, 24 Sep 2008)* |
1287 |
|
1288 |
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the |
1289 |
type of disk exported to fully virtualized instances. |
1290 |
- provide access to the serial console of HVM instances |
1291 |
- instance auto_balance flag, set by default. If turned off it will |
1292 |
avoid warnings on cluster verify if there is not enough memory to fail |
1293 |
over an instance. in the future it will prevent automatically failing |
1294 |
it over when we will support that. |
1295 |
- batcher tool for instance creation, see ``tools/README.batcher`` |
1296 |
- ``gnt-instance reinstall --select-os`` to interactively select a new |
1297 |
operating system when reinstalling an instance. |
1298 |
- when changing the memory amount on instance modify a check has been |
1299 |
added that the instance will be able to start. also warnings are |
1300 |
emitted if the instance will not be able to fail over, if auto_balance |
1301 |
is true. |
1302 |
- documentation fixes |
1303 |
- sync fields between ``gnt-instance list/modify/add/import`` |
1304 |
- fix a race condition in drbd when the sync speed was set after giving |
1305 |
the device a remote peer. |
1306 |
|
1307 |
|
1308 |
Version 1.2.5 |
1309 |
------------- |
1310 |
|
1311 |
*(Released Tue, 22 Jul 2008)* |
1312 |
|
1313 |
- note: the allowed size and number of tags per object were reduced |
1314 |
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups |
1315 |
- fixed twisted 8.x compatibility |
1316 |
- fixed ``gnt-instance replace-disks`` with iallocator |
1317 |
- add TCP keepalives on twisted connections to detect restarted nodes |
1318 |
- disk increase support, see ``gnt-instance grow-disk`` |
1319 |
- implement bulk node/instance query for RAPI |
1320 |
- add tags in node/instance listing (optional) |
1321 |
- experimental migration (and live migration) support, read the man page |
1322 |
for ``gnt-instance migrate`` |
1323 |
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also |
1324 |
has some small improvements in handling its state file |
1325 |
|
1326 |
|
1327 |
Version 1.2.4 |
1328 |
------------- |
1329 |
|
1330 |
*(Released Fri, 13 Jun 2008)* |
1331 |
|
1332 |
- Experimental readonly, REST-based remote API implementation; |
1333 |
automatically started on master node, TCP port 5080, if enabled by |
1334 |
``--enable-rapi`` parameter to configure script. |
1335 |
- Instance allocator support. Add and import instance accept a |
1336 |
``--iallocator`` parameter, and call that instance allocator to decide |
1337 |
which node to use for the instance. The iallocator document describes |
1338 |
what's expected from an allocator script. |
1339 |
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the |
1340 |
``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a |
1341 |
node is lost there is still enough memory to fail over the instances |
1342 |
that reside on it. |
1343 |
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to |
1344 |
``gnt-cluster verify``, to check for site-specific compliance. All the |
1345 |
hooks will run, and their output, if any, will be displayed. Any |
1346 |
failing hook will make the verification return an error value. |
1347 |
- ``gnt-cluster verify`` now checks that its peers are reachable on the |
1348 |
primary and secondary interfaces |
1349 |
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node |
1350 |
that is still declared as part of the cluster and has failed. |
1351 |
- ``gnt-* list`` commands now accept a new ``-o +field`` way of |
1352 |
specifying output fields, that just adds the chosen fields to the |
1353 |
default ones. |
1354 |
- ``gnt-backup`` now has a new ``remove`` command to delete an existing |
1355 |
export from the filesystem. |
1356 |
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path |
1357 |
have been added. Using them you can enable/disable acpi and pae |
1358 |
support, and specify a path for a cd image to be exported to the |
1359 |
instance. These parameters as the name suggest only work on HVM |
1360 |
clusters. |
1361 |
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and |
1362 |
PAE support will be set to the previously hardcoded values, but the |
1363 |
(previously hardcoded) path to the CDROM ISO image will be unset and |
1364 |
if required, needs to be set manually with ``gnt-instance modify`` |
1365 |
after the upgrade. |
1366 |
- The address to which an instance's VNC console is bound is now |
1367 |
selectable per-instance, rather than being cluster wide. Of course |
1368 |
this only applies to instances controlled via VNC, so currently just |
1369 |
applies to HVM clusters. |
1370 |
|
1371 |
|
1372 |
Version 1.2.3 |
1373 |
------------- |
1374 |
|
1375 |
*(Released Mon, 18 Feb 2008)* |
1376 |
|
1377 |
- more tweaks to the disk activation code (especially helpful for DRBD) |
1378 |
- change the default ``gnt-instance list`` output format, now there is |
1379 |
one combined status field (see the manpage for the exact values this |
1380 |
field will have) |
1381 |
- some more fixes for the mac export to hooks change |
1382 |
- make Ganeti not break with DRBD 8.2.x (which changed the version |
1383 |
format in ``/proc/drbd``) (issue 24) |
1384 |
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk |
1385 |
template, allowing migration from DRBD0.7+MD to DRBD8 |
1386 |
|
1387 |
|
1388 |
Version 1.2.2 |
1389 |
------------- |
1390 |
|
1391 |
*(Released Wed, 30 Jan 2008)* |
1392 |
|
1393 |
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM |
1394 |
support (issue 23) |
1395 |
- add command aliases infrastructure and a few aliases |
1396 |
- allow listing of VCPUs in the ``gnt-instance list`` and improve the |
1397 |
man pages and the ``--help`` option of ``gnt-node |
1398 |
list``/``gnt-instance list`` |
1399 |
- fix ``gnt-backup list`` with down nodes (issue 21) |
1400 |
- change the tools location (move from $pkgdatadir to $pkglibdir/tools) |
1401 |
- fix the dist archive and add a check for including svn/git files in |
1402 |
the future |
1403 |
- some developer-related changes: improve the burnin and the QA suite, |
1404 |
add an upload script for testing during development |
1405 |
|
1406 |
|
1407 |
Version 1.2.1 |
1408 |
------------- |
1409 |
|
1410 |
*(Released Wed, 16 Jan 2008)* |
1411 |
|
1412 |
- experimental HVM support, read the install document, section |
1413 |
"Initializing the cluster" |
1414 |
- allow for the PVM hypervisor per-instance kernel and initrd paths |
1415 |
- add a new command ``gnt-cluster verify-disks`` which uses a new |
1416 |
algorithm to improve the reconnection of the DRBD pairs if the device |
1417 |
on the secondary node has gone away |
1418 |
- make logical volume code auto-activate LVs at disk activation time |
1419 |
- slightly improve the speed of activating disks |
1420 |
- allow specification of the MAC address at instance creation time, and |
1421 |
changing it later via ``gnt-instance modify`` |
1422 |
- fix handling of external commands that generate lots of output on |
1423 |
stderr |
1424 |
- update documentation with regard to minimum version of DRBD8 supported |
1425 |
|
1426 |
|
1427 |
Version 1.2.0 |
1428 |
------------- |
1429 |
|
1430 |
*(Released Tue, 4 Dec 2007)* |
1431 |
|
1432 |
- Log the ``xm create`` output to the node daemon log on failure (to |
1433 |
help diagnosing the error) |
1434 |
- In debug mode, log all external commands output if failed to the logs |
1435 |
- Change parsing of lvm commands to ignore stderr |
1436 |
|
1437 |
|
1438 |
Version 1.2b3 |
1439 |
------------- |
1440 |
|
1441 |
*(Released Wed, 28 Nov 2007)* |
1442 |
|
1443 |
- Another round of updates to the DRBD 8 code to deal with more failures |
1444 |
in the replace secondary node operation |
1445 |
- Some more logging of failures in disk operations (lvm, drbd) |
1446 |
- A few documentation updates |
1447 |
- QA updates |
1448 |
|
1449 |
|
1450 |
Version 1.2b2 |
1451 |
------------- |
1452 |
|
1453 |
*(Released Tue, 13 Nov 2007)* |
1454 |
|
1455 |
- Change configuration file format from Python's Pickle to JSON. |
1456 |
Upgrading is possible using the cfgupgrade utility. |
1457 |
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for |
1458 |
faster replace disks and is more stable (DRBD 8 has many improvements |
1459 |
compared to DRBD 0.7) |
1460 |
- Added command line tags support (see man pages for ``gnt-instance``, |
1461 |
``gnt-node``, ``gnt-cluster``) |
1462 |
- Added instance rename support |
1463 |
- Added multi-instance startup/shutdown |
1464 |
- Added cluster rename support |
1465 |
- Added ``gnt-node evacuate`` to simplify some node operations |
1466 |
- Added instance reboot operation that can speedup reboot as compared to |
1467 |
stop and start |
1468 |
- Soften the requirement that hostnames are in FQDN format |
1469 |
- The ``ganeti-watcher`` now activates drbd pairs after secondary node |
1470 |
reboots |
1471 |
- Removed dependency on debian's patched fping that uses the |
1472 |
non-standard ``-S`` option |
1473 |
- Now the OS definitions are searched for in multiple, configurable |
1474 |
paths (easier for distros to package) |
1475 |
- Some changes to the hooks infrastructure (especially the new |
1476 |
post-configuration update hook) |
1477 |
- Other small bugfixes |
1478 |
|
1479 |
.. vim: set textwidth=72 syntax=rst : |
1480 |
.. Local Variables: |
1481 |
.. mode: rst |
1482 |
.. fill-column: 72 |
1483 |
.. End: |