root / NEWS @ 70c81511
History | View | Annotate | Download (36.4 kB)
1 |
News |
---|---|
2 |
==== |
3 |
|
4 |
Version 2.2.0 beta 0 |
5 |
-------------------- |
6 |
|
7 |
*(Released Thu, 17 Jun 2010)* |
8 |
|
9 |
- Added tool (``move-instance``) and infrastructure to move instances |
10 |
between separate clusters (see :doc:`separate documentation |
11 |
<move-instance>` and :doc:`design document <design-2.2>`) |
12 |
- Added per-request RPC timeout |
13 |
- RAPI now requires a Content-Type header for requests with a body (e.g. |
14 |
``PUT`` or ``POST``) which must be set to ``application/json`` (see |
15 |
RFC2616 (HTTP/1.1), section 7.2.1) |
16 |
- ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not |
17 |
reachable |
18 |
- Implemented initial support for running Ganeti daemons as separate |
19 |
users, see configure-time flags ``--with-user-prefix`` and |
20 |
``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this |
21 |
time) |
22 |
- Instances can be removed after export (``gnt-backup export |
23 |
--remove-instance``) |
24 |
- Self-signed certificates generated by Ganeti now use a 2048 bit RSA |
25 |
key (instead of 1024 bit) |
26 |
- Added new cluster configuration file for cluster domain secret |
27 |
- Import/export now use SSL instead of SSH |
28 |
- Added support for showing estimated time when exporting an instance, |
29 |
see the ``ganeti-os-interface(7)`` manpage and look for |
30 |
``EXP_SIZE_FD`` |
31 |
|
32 |
|
33 |
Version 2.1.3 |
34 |
------------- |
35 |
|
36 |
*(Released Thu, 3 Jun 2010)* |
37 |
|
38 |
A medium sized development cycle. Some new features, and some |
39 |
fixes/small improvements/cleanups. |
40 |
|
41 |
Significant features |
42 |
~~~~~~~~~~~~~~~~~~~~ |
43 |
|
44 |
The node deamon now tries to mlock itself into memory, unless the |
45 |
``--no-mlock`` flag is passed. It also doesn't fail if it can't write |
46 |
its logs, and falls back to console logging. This allows emergency |
47 |
features such as ``gnt-node powercycle`` to work even in the event of a |
48 |
broken node disk (tested offlining the disk hosting the node's |
49 |
filesystem and dropping its memory caches; don't try this at home) |
50 |
|
51 |
KVM: add vhost-net acceleration support. It can be tested with a new |
52 |
enough version of the kernel and of qemu-kvm. |
53 |
|
54 |
KVM: Add instance chrooting feature. If you use privilege dropping for |
55 |
your VMs you can also now force them to chroot to an empty directory, |
56 |
before starting the emulated guest. |
57 |
|
58 |
KVM: Add maximum migration bandwith and maximum downtime tweaking |
59 |
support (requires a new-enough version of qemu-kvm). |
60 |
|
61 |
Cluster verify will now warn if the master node doesn't have the master |
62 |
ip configured on it. |
63 |
|
64 |
Add a new (incompatible) instance creation request format to RAPI which |
65 |
supports all parameters (previously only a subset was supported, and it |
66 |
wasn't possible to extend the old format to accomodate all the new |
67 |
features. The old format is still supported, and a client can check for |
68 |
this feature, before using it, by checking for its presence in the |
69 |
``features`` RAPI resource. |
70 |
|
71 |
Now with ancient latin support. Try it passing the ``--roman`` option to |
72 |
``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list`` |
73 |
(requires the python-roman module to be installed, in order to work). |
74 |
|
75 |
Other changes |
76 |
~~~~~~~~~~~~~ |
77 |
|
78 |
As usual many internal code refactorings, documentation updates, and |
79 |
such. Among others: |
80 |
|
81 |
- Lots of improvements and cleanups to the experimental Remote API |
82 |
(RAPI) client library. |
83 |
- A new unit test suite for the core daemon libraries. |
84 |
- A fix to creating missing directories makes sure the umask is not |
85 |
applied anymore. This enforces the same directory permissions |
86 |
everywhere. |
87 |
- Better handling terminating daemons with ctrl+c (used when running |
88 |
them in debugging mode). |
89 |
- Fix a race condition in live migrating a KVM instance, when stat() |
90 |
on the old proc status file returned EINVAL, which is an unexpected |
91 |
value. |
92 |
- Fixed manpage checking with newer man and utf-8 charachters. But now |
93 |
you need the en_US.UTF-8 locale enabled to build Ganeti from git. |
94 |
|
95 |
|
96 |
Version 2.1.2.1 |
97 |
--------------- |
98 |
|
99 |
*(Released Fri, 7 May 2010)* |
100 |
|
101 |
Fix a bug which prevented untagged KVM instances from starting. |
102 |
|
103 |
|
104 |
Version 2.1.2 |
105 |
------------- |
106 |
|
107 |
*(Released Fri, 7 May 2010)* |
108 |
|
109 |
Another release with a long development cycle, during which many |
110 |
different features were added. |
111 |
|
112 |
Significant features |
113 |
~~~~~~~~~~~~~~~~~~~~ |
114 |
|
115 |
The KVM hypervisor now can run the individual instances as non-root, to |
116 |
reduce the impact of a VM being hijacked due to bugs in the |
117 |
hypervisor. It is possible to run all instances as a single (non-root) |
118 |
user, to manually specify a user for each instance, or to dynamically |
119 |
allocate a user out of a cluster-wide pool to each instance, with the |
120 |
guarantee that no two instances will run under the same user ID on any |
121 |
given node. |
122 |
|
123 |
An experimental RAPI client library, that can be used standalone |
124 |
(without the other Ganeti libraries), is provided in the source tree as |
125 |
``lib/rapi/client.py``. Note this client might change its interface in |
126 |
the future, as we iterate on its capabilities. |
127 |
|
128 |
A new command, ``gnt-cluster renew-crypto`` has been added to easily |
129 |
replace the cluster's certificates and crypto keys. This might help in |
130 |
case they have been compromised, or have simply expired. |
131 |
|
132 |
A new disk option for instance creation has been added that allows one |
133 |
to "adopt" currently existing logical volumes, with data |
134 |
preservation. This should allow easier migration to Ganeti from |
135 |
unmanaged (or managed via other software) instances. |
136 |
|
137 |
Another disk improvement is the possibility to convert between redundant |
138 |
(DRBD) and plain (LVM) disk configuration for an instance. This should |
139 |
allow better scalability (starting with one node and growing the |
140 |
cluster, or shrinking a two-node cluster to one node). |
141 |
|
142 |
A new feature that could help with automated node failovers has been |
143 |
implemented: if a node sees itself as offline (by querying the master |
144 |
candidates), it will try to shutdown (hard) all instances and any active |
145 |
DRBD devices. This reduces the risk of duplicate instances if an |
146 |
external script automatically failovers the instances on such nodes. To |
147 |
enable this, the cluster parameter ``maintain_node_health`` should be |
148 |
enabled; in the future this option (per the name) will enable other |
149 |
automatic maintenance features. |
150 |
|
151 |
Instance export/import now will reuse the original instance |
152 |
specifications for all parameters; that means exporting an instance, |
153 |
deleting it and the importing it back should give an almost identical |
154 |
instance. Note that the default import behaviour has changed from |
155 |
before, where it created only one NIC; now it recreates the original |
156 |
number of NICs. |
157 |
|
158 |
Cluster verify has added a few new checks: SSL certificates validity, |
159 |
/etc/hosts consistency across the cluster, etc. |
160 |
|
161 |
Other changes |
162 |
~~~~~~~~~~~~~ |
163 |
|
164 |
As usual, many internal changes were done, documentation fixes, |
165 |
etc. Among others: |
166 |
|
167 |
- Fixed cluster initialization with disabled cluster storage (regression |
168 |
introduced in 2.1.1) |
169 |
- File-based storage supports growing the disks |
170 |
- Fixed behaviour of node role changes |
171 |
- Fixed cluster verify for some corner cases, plus a general rewrite of |
172 |
cluster verify to allow future extension with more checks |
173 |
- Fixed log spamming by watcher and node daemon (regression introduced |
174 |
in 2.1.1) |
175 |
- Fixed possible validation issues when changing the list of enabled |
176 |
hypervisors |
177 |
- Fixed cleanup of /etc/hosts during node removal |
178 |
- Fixed RAPI response for invalid methods |
179 |
- Fixed bug with hashed passwords in ``ganeti-rapi`` daemon |
180 |
- Multiple small improvements to the KVM hypervisor (VNC usage, booting |
181 |
from ide disks, etc.) |
182 |
- Allow OS changes without re-installation (to record a changed OS |
183 |
outside of Ganeti, or to allow OS renames) |
184 |
- Allow instance creation without OS installation (useful for example if |
185 |
the OS will be installed manually, or restored from a backup not in |
186 |
Ganeti format) |
187 |
- Implemented option to make cluster ``copyfile`` use the replication |
188 |
network |
189 |
- Added list of enabled hypervisors to ssconf (possibly useful for |
190 |
external scripts) |
191 |
- Added a new tool (``tools/cfgupgrade12``) that allows upgrading from |
192 |
1.2 clusters |
193 |
- A partial form of node re-IP is possible via node readd, which now |
194 |
allows changed node primary IP |
195 |
- Command line utilities now show an informational message if the job is |
196 |
waiting for a lock |
197 |
- The logs of the master daemon now show the PID/UID/GID of the |
198 |
connected client |
199 |
|
200 |
|
201 |
Version 2.1.1 |
202 |
------------- |
203 |
|
204 |
*(Released Fri, 12 Mar 2010)* |
205 |
|
206 |
During the 2.1.0 long release candidate cycle, a lot of improvements and |
207 |
changes have accumulated with were released later as 2.1.1. |
208 |
|
209 |
Major changes |
210 |
~~~~~~~~~~~~~ |
211 |
|
212 |
The node evacuate command (``gnt-node evacuate``) was significantly |
213 |
rewritten, and as such the IAllocator protocol was changed - a new |
214 |
request type has been added. This unfortunate change during a stable |
215 |
series is designed to improve performance of node evacuations; on |
216 |
clusters with more than about five nodes and which are well-balanced, |
217 |
evacuation should proceed in parallel for all instances of the node |
218 |
being evacuated. As such, any existing IAllocator scripts need to be |
219 |
updated, otherwise the above command will fail due to the unknown |
220 |
request. The provided "dumb" allocator has not been updated; but the |
221 |
ganeti-htools package supports the new protocol since version 0.2.4. |
222 |
|
223 |
Another important change is increased validation of node and instance |
224 |
names. This might create problems in special cases, if invalid host |
225 |
names are being used. |
226 |
|
227 |
Also, a new layer of hypervisor parameters has been added, that sits at |
228 |
OS level between the cluster defaults and the instance ones. This allows |
229 |
customisation of virtualization parameters depending on the installed |
230 |
OS. For example instances with OS 'X' may have a different KVM kernel |
231 |
(or any other parameter) than the cluster defaults. This is intended to |
232 |
help managing a multiple OSes on the same cluster, without manual |
233 |
modification of each instance's parameters. |
234 |
|
235 |
A tool for merging clusters, ``cluster-merge``, has been added in the |
236 |
tools sub-directory. |
237 |
|
238 |
Bug fixes |
239 |
~~~~~~~~~ |
240 |
|
241 |
- Improved the int/float conversions that should make the code more |
242 |
robust in face of errors from the node daemons |
243 |
- Fixed the remove node code in case of internal configuration errors |
244 |
- Fixed the node daemon behaviour in face of inconsistent queue |
245 |
directory (e.g. read-only file-system where we can't open the files |
246 |
read-write, etc.) |
247 |
- Fixed the behaviour of gnt-node modify for master candidate demotion; |
248 |
now it either aborts cleanly or, if given the new โauto_promoteโ |
249 |
parameter, will automatically promote other nodes as needed |
250 |
- Fixed compatibility with (unreleased yet) Python 2.6.5 that would |
251 |
completely prevent Ganeti from working |
252 |
- Fixed bug for instance export when not all disks were successfully |
253 |
exported |
254 |
- Fixed behaviour of node add when the new node is slow in starting up |
255 |
the node daemon |
256 |
- Fixed handling of signals in the LUXI client, which should improve |
257 |
behaviour of command-line scripts |
258 |
- Added checks for invalid node/instance names in the configuration (now |
259 |
flagged during cluster verify) |
260 |
- Fixed watcher behaviour for disk activation errors |
261 |
- Fixed two potentially endless loops in http library, which led to the |
262 |
RAPI daemon hanging and consuming 100% CPU in some cases |
263 |
- Fixed bug in RAPI daemon related to hashed passwords |
264 |
- Fixed bug for unintended qemu-level bridging of multi-NIC KVM |
265 |
instances |
266 |
- Enhanced compatibility with non-Debian OSes, but not using absolute |
267 |
path in some commands and allowing customisation of the ssh |
268 |
configuration directory |
269 |
- Fixed possible future issue with new Python versions by abiding to the |
270 |
proper use of ``__slots__`` attribute on classes |
271 |
- Added checks that should prevent directory traversal attacks |
272 |
- Many documentation fixes based on feedback from users |
273 |
|
274 |
New features |
275 |
~~~~~~~~~~~~ |
276 |
|
277 |
- Added an โearly_releaseโ more for instance replace disks and node |
278 |
evacuate, where we release locks earlier and thus allow higher |
279 |
parallelism within the cluster |
280 |
- Added watcher hooks, intended to allow the watcher to restart other |
281 |
daemons (e.g. from the ganeti-nbma project), but they can be used of |
282 |
course for any other purpose |
283 |
- Added a compile-time disable for DRBD barriers, to increase |
284 |
performance if the administrator trusts the power supply or the |
285 |
storage system to not lose writes |
286 |
- Added the option of using syslog for logging instead of, or in |
287 |
addition to, Ganeti's own log files |
288 |
- Removed boot restriction for paravirtual NICs for KVM, recent versions |
289 |
can indeed boot from a paravirtual NIC |
290 |
- Added a generic debug level for many operations; while this is not |
291 |
used widely yet, it allows one to pass the debug value all the way to |
292 |
the OS scripts |
293 |
- Enhanced the hooks environment for instance moves (failovers, |
294 |
migrations) where the primary/secondary nodes changed during the |
295 |
operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars |
296 |
- Enhanced data validations for many user-supplied values; one important |
297 |
item is the restrictions imposed on instance and node names, which |
298 |
might reject some (invalid) host names |
299 |
- Add a configure-time option to disable file-based storage, if it's not |
300 |
needed; this allows greater security separation between the master |
301 |
node and the other nodes from the point of view of the inter-node RPC |
302 |
protocol |
303 |
- Added user notification in interactive tools if job is waiting in the |
304 |
job queue or trying to acquire locks |
305 |
- Added log messages when a job is waiting for locks |
306 |
- Added filtering by node tags in instance operations which admit |
307 |
multiple instances (start, stop, reboot, reinstall) |
308 |
- Added a new tool for cluster mergers, ``cluster-merge`` |
309 |
- Parameters from command line which are of the form ``a=b,c=d`` can now |
310 |
use backslash escapes to pass in values which contain commas, |
311 |
e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value |
312 |
``b,c`` |
313 |
- For KVM, the instance name is the first parameter passed to KVM, so |
314 |
that it's more visible in the process list |
315 |
|
316 |
|
317 |
Version 2.1.0 |
318 |
------------- |
319 |
|
320 |
*(Released Tue, 2 Mar 2010)* |
321 |
|
322 |
Ganeti 2.1 brings many improvements with it. Major changes: |
323 |
|
324 |
- Added infrastructure to ease automated disk repairs |
325 |
- Added new daemon to export configuration data in a cheaper way than |
326 |
using the remote API |
327 |
- Instance NICs can now be routed instead of being associated with a |
328 |
networking bridge |
329 |
- Improved job locking logic to reduce impact of jobs acquiring multiple |
330 |
locks waiting for other long-running jobs |
331 |
|
332 |
In-depth implementation details can be found in the Ganeti 2.1 design |
333 |
document. |
334 |
|
335 |
Details |
336 |
~~~~~~~ |
337 |
|
338 |
- Added chroot hypervisor |
339 |
- Added more options to xen-hvm hypervisor (``kernel_path`` and |
340 |
``device_model``) |
341 |
- Added more options to xen-pvm hypervisor (``use_bootloader``, |
342 |
``bootloader_path`` and ``bootloader_args``) |
343 |
- Added the ``use_localtime`` option for the xen-hvm and kvm |
344 |
hypervisors, and the default value for this has changed to false (in |
345 |
2.0 xen-hvm always enabled it) |
346 |
- Added luxi call to submit multiple jobs in one go |
347 |
- Added cluster initialization option to not modify ``/etc/hosts`` |
348 |
file on nodes |
349 |
- Added network interface parameters |
350 |
- Added dry run mode to some LUs |
351 |
- Added RAPI resources: |
352 |
|
353 |
- ``/2/instances/[instance_name]/info`` |
354 |
- ``/2/instances/[instance_name]/replace-disks`` |
355 |
- ``/2/nodes/[node_name]/evacuate`` |
356 |
- ``/2/nodes/[node_name]/migrate`` |
357 |
- ``/2/nodes/[node_name]/role`` |
358 |
- ``/2/nodes/[node_name]/storage`` |
359 |
- ``/2/nodes/[node_name]/storage/modify`` |
360 |
- ``/2/nodes/[node_name]/storage/repair`` |
361 |
|
362 |
- Added OpCodes to evacuate or migrate all instances on a node |
363 |
- Added new command to list storage elements on nodes (``gnt-node |
364 |
list-storage``) and modify them (``gnt-node modify-storage``) |
365 |
- Added new ssconf files with master candidate IP address |
366 |
(``ssconf_master_candidates_ips``), node primary IP address |
367 |
(``ssconf_node_primary_ips``) and node secondary IP address |
368 |
(``ssconf_node_secondary_ips``) |
369 |
- Added ``ganeti-confd`` and a client library to query the Ganeti |
370 |
configuration via UDP |
371 |
- Added ability to run hooks after cluster initialization and before |
372 |
cluster destruction |
373 |
- Added automatic mode for disk replace (``gnt-instance replace-disks |
374 |
--auto``) |
375 |
- Added ``gnt-instance recreate-disks`` to re-create (empty) disks |
376 |
after catastrophic data-loss |
377 |
- Added ``gnt-node repair-storage`` command to repair damaged LVM volume |
378 |
groups |
379 |
- Added ``gnt-instance move`` command to move instances |
380 |
- Added ``gnt-cluster watcher`` command to control watcher |
381 |
- Added ``gnt-node powercycle`` command to powercycle nodes |
382 |
- Added new job status field ``lock_status`` |
383 |
- Added parseable error codes to cluster verification (``gnt-cluster |
384 |
verify --error-codes``) and made output less verbose (use |
385 |
``--verbose`` to restore previous behaviour) |
386 |
- Added UUIDs to the main config entities (cluster, nodes, instances) |
387 |
- Added support for OS variants |
388 |
- Added support for hashed passwords in the Ganeti remote API users file |
389 |
(``rapi_users``) |
390 |
- Added option to specify maximum timeout on instance shutdown |
391 |
- Added ``--no-ssh-init`` option to ``gnt-cluster init`` |
392 |
- Added new helper script to start and stop Ganeti daemons |
393 |
(``daemon-util``), with the intent to reduce the work necessary to |
394 |
adjust Ganeti for non-Debian distributions and to start/stop daemons |
395 |
from one place |
396 |
- Added more unittests |
397 |
- Fixed critical bug in ganeti-masterd startup |
398 |
- Removed the configure-time ``kvm-migration-port`` parameter, this is |
399 |
now customisable at the cluster level for both the KVM and Xen |
400 |
hypervisors using the new ``migration_port`` parameter |
401 |
- Pass ``INSTANCE_REINSTALL`` variable to OS installation script when |
402 |
reinstalling an instance |
403 |
- Allowed ``@`` in tag names |
404 |
- Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation |
405 |
- Many documentation updates |
406 |
- Distribute hypervisor files on ``gnt-cluster redist-conf`` |
407 |
- ``gnt-instance reinstall`` can now reinstall multiple instances |
408 |
- Updated many command line parameters |
409 |
- Introduced new OS API version 15 |
410 |
- No longer support a default hypervisor |
411 |
- Treat virtual LVs as inexistent |
412 |
- Improved job locking logic to reduce lock contention |
413 |
- Match instance and node names case insensitively |
414 |
- Reimplemented bash completion script to be more complete |
415 |
- Improved burnin |
416 |
|
417 |
|
418 |
Version 2.0.6 |
419 |
------------- |
420 |
|
421 |
*(Released Thu, 4 Feb 2010)* |
422 |
|
423 |
- Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105) |
424 |
- Fix a string formatting bug |
425 |
- Improve safety of the code in some error paths |
426 |
- Improve data validation in the master of values returned from nodes |
427 |
|
428 |
|
429 |
Version 2.0.5 |
430 |
------------- |
431 |
|
432 |
*(Released Thu, 17 Dec 2009)* |
433 |
|
434 |
- Fix security issue due to missing validation of iallocator names; this |
435 |
allows local and remote execution of arbitrary executables |
436 |
- Fix failure of gnt-node list during instance removal |
437 |
- Ship the RAPI documentation in the archive |
438 |
|
439 |
|
440 |
Version 2.0.4 |
441 |
------------- |
442 |
|
443 |
*(Released Wed, 30 Sep 2009)* |
444 |
|
445 |
- Fixed many wrong messages |
446 |
- Fixed a few bugs related to the locking library |
447 |
- Fixed MAC checking at instance creation time |
448 |
- Fixed a DRBD parsing bug related to gaps in /proc/drbd |
449 |
- Fixed a few issues related to signal handling in both daemons and |
450 |
scripts |
451 |
- Fixed the example startup script provided |
452 |
- Fixed insserv dependencies in the example startup script (patch from |
453 |
Debian) |
454 |
- Fixed handling of drained nodes in the iallocator framework |
455 |
- Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug |
456 |
#528618) |
457 |
- Fixed error related to invalid job IDs in job polling |
458 |
- Fixed job/opcode persistence on unclean master shutdown |
459 |
- Fixed handling of partial job processing after unclean master |
460 |
shutdown |
461 |
- Fixed error reporting from LUs, previously all errors were converted |
462 |
into execution errors |
463 |
- Fixed error reporting from burnin |
464 |
- Decreased significantly the memory usage of the job queue |
465 |
- Optimised slightly multi-job submission |
466 |
- Optimised slightly opcode loading |
467 |
- Backported the multi-job submit framework from the development |
468 |
branch; multi-instance start and stop should be faster |
469 |
- Added script to clean archived jobs after 21 days; this will reduce |
470 |
the size of the queue directory |
471 |
- Added some extra checks in disk size tracking |
472 |
- Added an example ethers hook script |
473 |
- Added a cluster parameter that prevents Ganeti from modifying of |
474 |
/etc/hosts |
475 |
- Added more node information to RAPI responses |
476 |
- Added a ``gnt-job watch`` command that allows following the ouput of a |
477 |
job |
478 |
- Added a bind-address option to ganeti-rapi |
479 |
- Added more checks to the configuration verify |
480 |
- Enhanced the burnin script such that some operations can be retried |
481 |
automatically |
482 |
- Converted instance reinstall to multi-instance model |
483 |
|
484 |
|
485 |
Version 2.0.3 |
486 |
------------- |
487 |
|
488 |
*(Released Fri, 7 Aug 2009)* |
489 |
|
490 |
- Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command |
491 |
to allow using the pre-2.0.2 behaviour in activation, if any existing |
492 |
instances have mismatched disk sizes in the configuration |
493 |
- Added ``gnt-cluster repair-disk-sizes`` command to check and update |
494 |
any configuration mismatches for disk sizes |
495 |
- Added ``gnt-master cluste-failover --no-voting`` to allow master |
496 |
failover to work on two-node clusters |
497 |
- Fixed the ``--net`` option of ``gnt-backup import``, which was |
498 |
unusable |
499 |
- Fixed detection of OS script errors in ``gnt-backup export`` |
500 |
- Fixed exit code of ``gnt-backup export`` |
501 |
|
502 |
|
503 |
Version 2.0.2 |
504 |
------------- |
505 |
|
506 |
*(Released Fri, 17 Jul 2009)* |
507 |
|
508 |
- Added experimental support for stripped logical volumes; this should |
509 |
enhance performance but comes with a higher complexity in the block |
510 |
device handling; stripping is only enabled when passing |
511 |
``--with-lvm-stripecount=N`` to ``configure``, but codepaths are |
512 |
affected even in the non-stripped mode |
513 |
- Improved resiliency against transient failures at the end of DRBD |
514 |
resyncs, and in general of DRBD resync checks |
515 |
- Fixed a couple of issues with exports and snapshot errors |
516 |
- Fixed a couple of issues in instance listing |
517 |
- Added display of the disk size in ``gnt-instance info`` |
518 |
- Fixed checking for valid OSes in instance creation |
519 |
- Fixed handling of the "vcpus" parameter in instance listing and in |
520 |
general of invalid parameters |
521 |
- Fixed http server library, and thus RAPI, to handle invalid |
522 |
username/password combinations correctly; this means that now they |
523 |
report unauthorized for queries too, not only for modifications, |
524 |
allowing earlier detect of configuration problems |
525 |
- Added a new "role" node list field, equivalent to the master/master |
526 |
candidate/drained/offline flags combinations |
527 |
- Fixed cluster modify and changes of candidate pool size |
528 |
- Fixed cluster verify error messages for wrong files on regular nodes |
529 |
- Fixed a couple of issues with node demotion from master candidate role |
530 |
- Fixed node readd issues |
531 |
- Added non-interactive mode for ``ganeti-masterd --no-voting`` startup |
532 |
- Added a new ``--no-voting`` option for masterfailover to fix failover |
533 |
on two-nodes clusters when the former master node is unreachable |
534 |
- Added instance reinstall over RAPI |
535 |
|
536 |
|
537 |
Version 2.0.1 |
538 |
------------- |
539 |
|
540 |
*(Released Tue, 16 Jun 2009)* |
541 |
|
542 |
- added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will |
543 |
allow re-adding the start in single-user option (regression from 1.2) |
544 |
- the watcher writes the instance status to a file, to allow monitoring |
545 |
to report the instance status (from the master) based on cached |
546 |
results of the watcher's queries; while this can get stale if the |
547 |
watcher is being locked due to other work on the cluster, this is |
548 |
still an improvement |
549 |
- the watcher now also restarts the node daemon and the rapi daemon if |
550 |
they died |
551 |
- fixed the watcher to handle full and drained queue cases |
552 |
- hooks export more instance data in the environment, which helps if |
553 |
hook scripts need to take action based on the instance's properties |
554 |
(no longer need to query back into ganeti) |
555 |
- instance failovers when the instance is stopped do not check for free |
556 |
RAM, so that failing over a stopped instance is possible in low memory |
557 |
situations |
558 |
- rapi uses queries for tags instead of jobs (for less job traffic), and |
559 |
for cluster tags it won't talk to masterd at all but read them from |
560 |
ssconf |
561 |
- a couple of error handling fixes in RAPI |
562 |
- drbd handling: improved the error handling of inconsistent disks after |
563 |
resync to reduce the frequency of "there are some degraded disks for |
564 |
this instance" messages |
565 |
- fixed a bug in live migration when DRBD doesn't want to reconnect (the |
566 |
error handling path called a wrong function name) |
567 |
|
568 |
|
569 |
Version 2.0.0 final |
570 |
------------------- |
571 |
|
572 |
*(Released Wed, 27 May 2009)* |
573 |
|
574 |
- no changes from rc5 |
575 |
|
576 |
|
577 |
Version 2.0 release candidate 5 |
578 |
------------------------------- |
579 |
|
580 |
*(Released Wed, 20 May 2009)* |
581 |
|
582 |
- fix a couple of bugs (validation, argument checks) |
583 |
- fix ``gnt-cluster getmaster`` on non-master nodes (regression) |
584 |
- some small improvements to RAPI and IAllocator |
585 |
- make watcher automatically start the master daemon if down |
586 |
|
587 |
|
588 |
Version 2.0 release candidate 4 |
589 |
------------------------------- |
590 |
|
591 |
*(Released Mon, 27 Apr 2009)* |
592 |
|
593 |
- change the OS list to not require locks; this helps with big clusters |
594 |
- fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the |
595 |
volume group is broken |
596 |
- ``gnt-instance info``, without any arguments, doesn't run for all |
597 |
instances anymore; either pass ``--all`` or pass the desired |
598 |
instances; this helps against mistakes on big clusters where listing |
599 |
the information for all instances takes a long time |
600 |
- miscellaneous doc and man pages fixes |
601 |
|
602 |
|
603 |
Version 2.0 release candidate 3 |
604 |
------------------------------- |
605 |
|
606 |
*(Released Wed, 8 Apr 2009)* |
607 |
|
608 |
- Change the internal locking model of some ``gnt-node`` commands, in |
609 |
order to reduce contention (and blocking of master daemon) when |
610 |
batching many creation/reinstall jobs |
611 |
- Fixes to Xen soft reboot |
612 |
- No longer build documentation at build time, instead distribute it in |
613 |
the archive, in order to reduce the need for the whole docbook/rst |
614 |
toolchains |
615 |
|
616 |
|
617 |
Version 2.0 release candidate 2 |
618 |
------------------------------- |
619 |
|
620 |
*(Released Fri, 27 Mar 2009)* |
621 |
|
622 |
- Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0 |
623 |
- Fix watcher startup sequence, improves the behaviour of busy clusters |
624 |
- Some other fixes in ``gnt-cluster verify``, ``gnt-instance |
625 |
replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC |
626 |
bind address and other places |
627 |
- Some documentation fixes and updates |
628 |
|
629 |
|
630 |
Version 2.0 release candidate 1 |
631 |
------------------------------- |
632 |
|
633 |
*(Released Mon, 2 Mar 2009)* |
634 |
|
635 |
- More documentation updates, now all docs should be more-or-less |
636 |
up-to-date |
637 |
- A couple of small fixes (mixed hypervisor clusters, offline nodes, |
638 |
etc.) |
639 |
- Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM |
640 |
and KVM) |
641 |
- Fix an issue related to $libdir/run/ganeti and cluster creation |
642 |
|
643 |
|
644 |
Version 2.0 beta 2 |
645 |
------------------ |
646 |
|
647 |
*(Released Thu, 19 Feb 2009)* |
648 |
|
649 |
- Xen PVM and KVM have switched the default value for the instance root |
650 |
disk to the first partition on the first drive, instead of the whole |
651 |
drive; this means that the OS installation scripts must be changed |
652 |
accordingly |
653 |
- Man pages have been updated |
654 |
- RAPI has been switched by default to HTTPS, and the exported functions |
655 |
should all work correctly |
656 |
- RAPI v1 has been removed |
657 |
- Many improvements to the KVM hypervisor |
658 |
- Block device errors are now better reported |
659 |
- Many other bugfixes and small improvements |
660 |
|
661 |
|
662 |
Version 2.0 beta 1 |
663 |
------------------ |
664 |
|
665 |
*(Released Mon, 26 Jan 2009)* |
666 |
|
667 |
- Version 2 is a general rewrite of the code and therefore the |
668 |
differences are too many to list, see the design document for 2.0 in |
669 |
the ``doc/`` subdirectory for more details |
670 |
- In this beta version there is not yet a migration path from 1.2 (there |
671 |
will be one in the final 2.0 release) |
672 |
- A few significant changes are: |
673 |
|
674 |
- all commands are executed by a daemon (``ganeti-masterd``) and the |
675 |
various ``gnt-*`` commands are just front-ends to it |
676 |
- all the commands are entered into, and executed from a job queue, |
677 |
see the ``gnt-job(8)`` manpage |
678 |
- the RAPI daemon supports read-write operations, secured by basic |
679 |
HTTP authentication on top of HTTPS |
680 |
- DRBD version 0.7 support has been removed, DRBD 8 is the only |
681 |
supported version (when migrating from Ganeti 1.2 to 2.0, you need |
682 |
to migrate to DRBD 8 first while still running Ganeti 1.2) |
683 |
- DRBD devices are using statically allocated minor numbers, which |
684 |
will be assigned to existing instances during the migration process |
685 |
- there is support for both Xen PVM and Xen HVM instances running on |
686 |
the same cluster |
687 |
- KVM virtualization is supported too |
688 |
- file-based storage has been implemented, which means that it is |
689 |
possible to run the cluster without LVM and DRBD storage, for |
690 |
example using a shared filesystem exported from shared storage (and |
691 |
still have live migration) |
692 |
|
693 |
|
694 |
Version 1.2.7 |
695 |
------------- |
696 |
|
697 |
*(Released Tue, 13 Jan 2009)* |
698 |
|
699 |
- Change the default reboot type in ``gnt-instance reboot`` to "hard" |
700 |
- Reuse the old instance mac address by default on instance import, if |
701 |
the instance name is the same. |
702 |
- Handle situations in which the node info rpc returns incomplete |
703 |
results (issue 46) |
704 |
- Add checks for tcp/udp ports collisions in ``gnt-cluster verify`` |
705 |
- Improved version of batcher: |
706 |
|
707 |
- state file support |
708 |
- instance mac address support |
709 |
- support for HVM clusters/instances |
710 |
|
711 |
- Add an option to show the number of cpu sockets and nodes in |
712 |
``gnt-node list`` |
713 |
- Support OSes that handle more than one version of the OS api (but do |
714 |
not change the current API in any other way) |
715 |
- Fix ``gnt-node migrate`` |
716 |
- ``gnt-debug`` man page |
717 |
- Fixes various more typos and small issues |
718 |
- Increase disk resync maximum speed to 60MB/s (from 30MB/s) |
719 |
|
720 |
|
721 |
Version 1.2.6 |
722 |
------------- |
723 |
|
724 |
*(Released Wed, 24 Sep 2008)* |
725 |
|
726 |
- new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the |
727 |
type of disk exported to fully virtualized instances. |
728 |
- provide access to the serial console of HVM instances |
729 |
- instance auto_balance flag, set by default. If turned off it will |
730 |
avoid warnings on cluster verify if there is not enough memory to fail |
731 |
over an instance. in the future it will prevent automatically failing |
732 |
it over when we will support that. |
733 |
- batcher tool for instance creation, see ``tools/README.batcher`` |
734 |
- ``gnt-instance reinstall --select-os`` to interactively select a new |
735 |
operating system when reinstalling an instance. |
736 |
- when changing the memory amount on instance modify a check has been |
737 |
added that the instance will be able to start. also warnings are |
738 |
emitted if the instance will not be able to fail over, if auto_balance |
739 |
is true. |
740 |
- documentation fixes |
741 |
- sync fields between ``gnt-instance list/modify/add/import`` |
742 |
- fix a race condition in drbd when the sync speed was set after giving |
743 |
the device a remote peer. |
744 |
|
745 |
|
746 |
Version 1.2.5 |
747 |
------------- |
748 |
|
749 |
*(Released Tue, 22 Jul 2008)* |
750 |
|
751 |
- note: the allowed size and number of tags per object were reduced |
752 |
- fix a bug in ``gnt-cluster verify`` with inconsistent volume groups |
753 |
- fixed twisted 8.x compatibility |
754 |
- fixed ``gnt-instance replace-disks`` with iallocator |
755 |
- add TCP keepalives on twisted connections to detect restarted nodes |
756 |
- disk increase support, see ``gnt-instance grow-disk`` |
757 |
- implement bulk node/instance query for RAPI |
758 |
- add tags in node/instance listing (optional) |
759 |
- experimental migration (and live migration) support, read the man page |
760 |
for ``gnt-instance migrate`` |
761 |
- the ``ganeti-watcher`` logs are now timestamped, and the watcher also |
762 |
has some small improvements in handling its state file |
763 |
|
764 |
|
765 |
Version 1.2.4 |
766 |
------------- |
767 |
|
768 |
*(Released Fri, 13 Jun 2008)* |
769 |
|
770 |
- Experimental readonly, REST-based remote API implementation; |
771 |
automatically started on master node, TCP port 5080, if enabled by |
772 |
``--enable-rapi`` parameter to configure script. |
773 |
- Instance allocator support. Add and import instance accept a |
774 |
``--iallocator`` parameter, and call that instance allocator to decide |
775 |
which node to use for the instance. The iallocator document describes |
776 |
what's expected from an allocator script. |
777 |
- ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the |
778 |
``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a |
779 |
node is lost there is still enough memory to fail over the instances |
780 |
that reside on it. |
781 |
- ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to |
782 |
``gnt-cluster verify``, to check for site-specific compliance. All the |
783 |
hooks will run, and their output, if any, will be displayed. Any |
784 |
failing hook will make the verification return an error value. |
785 |
- ``gnt-cluster verify`` now checks that its peers are reachable on the |
786 |
primary and secondary interfaces |
787 |
- ``gnt-node add`` now supports the ``--readd`` option, to readd a node |
788 |
that is still declared as part of the cluster and has failed. |
789 |
- ``gnt-* list`` commands now accept a new ``-o +field`` way of |
790 |
specifying output fields, that just adds the chosen fields to the |
791 |
default ones. |
792 |
- ``gnt-backup`` now has a new ``remove`` command to delete an existing |
793 |
export from the filesystem. |
794 |
- New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path |
795 |
have been added. Using them you can enable/disable acpi and pae |
796 |
support, and specify a path for a cd image to be exported to the |
797 |
instance. These parameters as the name suggest only work on HVM |
798 |
clusters. |
799 |
- When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and |
800 |
PAE support will be set to the previously hardcoded values, but the |
801 |
(previously hardcoded) path to the CDROM ISO image will be unset and |
802 |
if required, needs to be set manually with ``gnt-instance modify`` |
803 |
after the upgrade. |
804 |
- The address to which an instance's VNC console is bound is now |
805 |
selectable per-instance, rather than being cluster wide. Of course |
806 |
this only applies to instances controlled via VNC, so currently just |
807 |
applies to HVM clusters. |
808 |
|
809 |
|
810 |
Version 1.2.3 |
811 |
------------- |
812 |
|
813 |
*(Released Mon, 18 Feb 2008)* |
814 |
|
815 |
- more tweaks to the disk activation code (especially helpful for DRBD) |
816 |
- change the default ``gnt-instance list`` output format, now there is |
817 |
one combined status field (see the manpage for the exact values this |
818 |
field will have) |
819 |
- some more fixes for the mac export to hooks change |
820 |
- make Ganeti not break with DRBD 8.2.x (which changed the version |
821 |
format in ``/proc/drbd``) (issue 24) |
822 |
- add an upgrade tool from "remote_raid1" disk template to "drbd" disk |
823 |
template, allowing migration from DRBD0.7+MD to DRBD8 |
824 |
|
825 |
|
826 |
Version 1.2.2 |
827 |
------------- |
828 |
|
829 |
*(Released Wed, 30 Jan 2008)* |
830 |
|
831 |
- fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM |
832 |
support (issue 23) |
833 |
- add command aliases infrastructure and a few aliases |
834 |
- allow listing of VCPUs in the ``gnt-instance list`` and improve the |
835 |
man pages and the ``--help`` option of ``gnt-node |
836 |
list``/``gnt-instance list`` |
837 |
- fix ``gnt-backup list`` with down nodes (issue 21) |
838 |
- change the tools location (move from $pkgdatadir to $pkglibdir/tools) |
839 |
- fix the dist archive and add a check for including svn/git files in |
840 |
the future |
841 |
- some developer-related changes: improve the burnin and the QA suite, |
842 |
add an upload script for testing during development |
843 |
|
844 |
|
845 |
Version 1.2.1 |
846 |
------------- |
847 |
|
848 |
*(Released Wed, 16 Jan 2008)* |
849 |
|
850 |
- experimental HVM support, read the install document, section |
851 |
"Initializing the cluster" |
852 |
- allow for the PVM hypervisor per-instance kernel and initrd paths |
853 |
- add a new command ``gnt-cluster verify-disks`` which uses a new |
854 |
algorithm to improve the reconnection of the DRBD pairs if the device |
855 |
on the secondary node has gone away |
856 |
- make logical volume code auto-activate LVs at disk activation time |
857 |
- slightly improve the speed of activating disks |
858 |
- allow specification of the MAC address at instance creation time, and |
859 |
changing it later via ``gnt-instance modify`` |
860 |
- fix handling of external commands that generate lots of output on |
861 |
stderr |
862 |
- update documentation with regard to minimum version of DRBD8 supported |
863 |
|
864 |
|
865 |
Version 1.2.0 |
866 |
------------- |
867 |
|
868 |
*(Released Tue, 4 Dec 2007)* |
869 |
|
870 |
- Log the ``xm create`` output to the node daemon log on failure (to |
871 |
help diagnosing the error) |
872 |
- In debug mode, log all external commands output if failed to the logs |
873 |
- Change parsing of lvm commands to ignore stderr |
874 |
|
875 |
|
876 |
Version 1.2b3 |
877 |
------------- |
878 |
|
879 |
*(Released Wed, 28 Nov 2007)* |
880 |
|
881 |
- Another round of updates to the DRBD 8 code to deal with more failures |
882 |
in the replace secondary node operation |
883 |
- Some more logging of failures in disk operations (lvm, drbd) |
884 |
- A few documentation updates |
885 |
- QA updates |
886 |
|
887 |
|
888 |
Version 1.2b2 |
889 |
------------- |
890 |
|
891 |
*(Released Tue, 13 Nov 2007)* |
892 |
|
893 |
- Change configuration file format from Python's Pickle to JSON. |
894 |
Upgrading is possible using the cfgupgrade utility. |
895 |
- Add support for DRBD 8.0 (new disk template ``drbd``) which allows for |
896 |
faster replace disks and is more stable (DRBD 8 has many improvements |
897 |
compared to DRBD 0.7) |
898 |
- Added command line tags support (see man pages for ``gnt-instance``, |
899 |
``gnt-node``, ``gnt-cluster``) |
900 |
- Added instance rename support |
901 |
- Added multi-instance startup/shutdown |
902 |
- Added cluster rename support |
903 |
- Added ``gnt-node evacuate`` to simplify some node operations |
904 |
- Added instance reboot operation that can speedup reboot as compared to |
905 |
stop and start |
906 |
- Soften the requirement that hostnames are in FQDN format |
907 |
- The ``ganeti-watcher`` now activates drbd pairs after secondary node |
908 |
reboots |
909 |
- Removed dependency on debian's patched fping that uses the |
910 |
non-standard ``-S`` option |
911 |
- Now the OS definitions are searched for in multiple, configurable |
912 |
paths (easier for distros to package) |
913 |
- Some changes to the hooks infrastructure (especially the new |
914 |
post-configuration update hook) |
915 |
- Other small bugfixes |
916 |
|
917 |
.. vim: set textwidth=72 : |
918 |
.. Local Variables: |
919 |
.. mode: rst |
920 |
.. fill-column: 72 |
921 |
.. End: |