RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2025-01-16 07:52:29 +00:00
Code Issues Packages Projects Releases Wiki Activity
2,438 Commits 216 Branches 394 Tags 1.6 GiB
e1597a036f
Commit Graph

2113 Commits

Author SHA1 Message Date
Zack Cerza
3e65d18261 Add Remote.user attribute 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
3352b58d30 Use Remote.shortname in logs 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
30d1d518d5 Make Remote.shortname actually short 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
36fe6a5830 Remote.hostname doesn't have to be a property 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
b2648b21af Fix PEP-8 issues 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
8567352360 Pass hostname to execute() 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
429553053f Use 'true' instead of 'echo online' 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
d0f7a47bf9 Add Remote.ensure_online() 
If the connection is alive, do nothing. If not, reconnect. Allow any
exceptions to bubble up to the caller. This is intended to fix unhelpful
'Bad File Descriptor' errors we were seeing when VMs go down.

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
3adb7d4629 Use Remote.hostname in logs 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:22 -05:00
Zack Cerza
29d3299497 Consolidate log file setup into shared function 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:21 -05:00
Zack Cerza
5dbce8b68f Use Unicode format strings 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-10 09:10:21 -05:00
Zack Cerza
085c508fd4 Revert "Revert "Show hostname instead of IP in errors"" 
This reverts commit 10fee0e368.

Conflicts:
	teuthology/orchestra/run.py
 2014-05-10 09:10:21 -05:00
Warren Usui
b7a7383d75 Allow .teuthology.yaml to set downburst path 
If .teuthology.yaml defines downburst, _get_downburst_exec()
now returns that value as the path to the downburst executable.

Also cleaned up code in create_if_vm.  ctx.downburst_conf was
never defined, so the code that handled the AttributeError
was always being executed.

Fixes: 6921

Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-05-09 17:20:26 -07:00
Zack Cerza
e2adb1fce1 Merge pull request #254 from ceph/wip-7707-wusui 
Use master as default for debian upgrade.
 2014-05-09 13:17:31 -05:00
Zack Cerza
8460c7a8d9 Force log lines to be interpreted as UTF-8 
Any invalid UTF-8 byte will be replaced with a Unicode replacement
character: U+FFFD or '�'

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-09 10:45:01 -05:00
Zack Cerza
cd7f268d9d Use binary flag for paramiko ChannelFiles 
This works around http://tracker.ceph.com/issues/8313

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-09 09:24:22 -05:00
Warren Usui
c0ba105453 Use master as default for debian upgrade. 
Make sure that uri is defined for debian upgrades.
Use master as default.

Added _get_uri_() which consolidates checking and assignment of
sha1, tags, and branch values.

Fixes: 7707
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-05-08 19:12:16 -07:00
Warren Usui
6fbf98bb0c Further clarify 'too many values to unpack' error. 
Many errors in yaml configurations cause ValueError to get thrown
with the message 'too many values to unpack.'  A previously reverted
change tried to handle all these situations and print an appropriate
message.

The current behavior of throwing the ValueError exception and exiting
is probably what we still want in these cases.  So instead of handling
the error, the code now checks for the exception at the top-most call
and displays appropriate messages in log.error and in the ctx.summary
data itself.

Fixes: 7510
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-05-08 10:57:06 -07:00
Josh Durgin
bca32ef54b Merge pull request #252 from ceph/wip-fsx-krbd 
rbd_fsx: expose krbd and related fsx options

Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
 2014-05-07 18:14:54 -07:00
Zack Cerza
bb170c1bf7 Merge pull request #249 from ceph/wip-8284 
rados.py: Add pool_snaps option for ceph_test_rados test command
 2014-05-07 13:20:40 -05:00
Ilya Dryomov
b105a07ac6 rbd_fsx: expose krbd and related fsx options 
Expose

-K (enable krbd mode) through 'krbd',
-Z (use direct IO) through 'direct_io',
-U (disable randomized striping) through 'randomized_striping',
-H (disable discard ops) through 'punch_holes',

-r readbdy (read alignment) through 'readbdy',
-w writebdy (write alignment) through 'writebdy',
-h holebdy (discard alignment) through 'holebdy'.

Signed-off-by: Ilya Dryomov <ilya.dryomov@inktank.com>
 2014-05-07 17:55:09 +04:00
Zack Cerza
a1838b2a0f Rewrite most file-retrieval functions 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-06 14:59:52 -05:00
Zack Cerza
01cf3671b4 Fix linter errors 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-05-06 14:59:52 -05:00
Warren Usui
8bed6ab625 FIx mktemp dir and redundant Paramiko connecting. 
Use previously initialized connection for sftp_get calls.
Use local directory for tarball temp file location.
 2014-05-06 14:59:52 -05:00
Warren Usui
36b07b8aee Use SFTPClienti get for long reads/writes 
Modified remote.py to use the paramiko SFTPClient get
method to extract long files (mostly tar files) from
the remote host.  The code formerly saved the data
in a long local string which was very inefficient.

Fixes: 8261
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-05-06 14:59:52 -05:00
Zack Cerza
091d1fe48d Revert "Revert "Clean up remote.py and misc.py changes."" 
This reverts commit 02504c3fd2.
 2014-05-06 14:59:52 -05:00
Zack Cerza
0268487057 Revert "Revert "Handle raw data I/O."" 
This reverts commit 0c8a3e2334.
 2014-05-06 14:59:51 -05:00
Sandon Van Ness
7b1eec94f1 Use longer varchar for locked_by in DB. 
Signed-off-by: Sandon Van Ness <sandon@inktank.com>
 2014-05-06 12:06:56 -07:00
David Zafman
ea3bef1e1d rados.py: Add pool_snaps option for ceph_test_rados test command 
Fixes: #8284

Signed-off-by: David Zafman <david.zafman@inktank.com>
 2014-05-06 10:30:36 -07:00
Sage Weil
4ad23dc5f8 rgw: fix indentation for cache_pools 
Fixes: #8295
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-05-06 09:22:49 -07:00
Sage Weil
33b4bfc764 ceph_manager: reset osd weights to 1 when waiting for clean 
If we leave the weights adjusted, we can get PGs stuck in a remapped state
because we are probabilistically rejecting placement.  Avoid this by
setting on in osds to 1.0 before waiting for clean.

Signed-off-by: Sage Weil <sage@inktank.com>
 2014-05-05 17:05:10 -07:00
Zack Cerza
0c8a3e2334 Revert "Handle raw data I/O." 
This reverts commit 257e1459fa.
 2014-05-05 10:03:29 -05:00
Zack Cerza
02504c3fd2 Revert "Clean up remote.py and misc.py changes." 
This reverts commit 74eff43ee1.
 2014-05-05 10:03:18 -05:00
Sage Weil
794c9465da ceph_manager: fix float stringification 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-05-03 07:24:32 -07:00
Sage Weil
a723ddf5db ceph_manager: fix typo 
From ce7fa1839f. Tested this time.

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
 2014-05-02 21:18:38 -07:00
Sage Weil
ce7fa1839f ceph_manager: randomly reweight in osds 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-05-02 16:32:53 -07:00
Sage Weil
c5da7b21f7 rgw: option to create a cache pool 
64mb for now!

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
 2014-05-01 14:36:26 -07:00
Alfredo Deza
f74eea7b9e Merge pull request #247 from ceph/requests-sessions 
Use a requests.Session object for retries instead of safe_while
 2014-05-01 13:17:00 -04:00
Zack Cerza
3fa6271f1a Calculate a timeout to use based on queue size 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-30 14:30:43 -05:00
Zack Cerza
041666b07a Add --runs, to print only run names 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-30 14:30:43 -05:00
Zack Cerza
a9d7aa3514 Refactor teuthology.beanstalk 
This architecture will make it easier to add new functionality.

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-30 14:30:43 -05:00
Zack Cerza
adc51e1c69 Drop usage of safe_while 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-29 16:41:45 -05:00
Zack Cerza
ea9c034f3c Use the new ResultsReporter.session object 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-29 16:17:08 -05:00
Zack Cerza
ec72137e1b Add a requests.Session object to ResultsReporter 
By default it is set up to retry requests 10 times

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-29 16:17:08 -05:00
Zack Cerza
61e469b632 Remove unused timeout arg to ResultsReporter init 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-29 16:17:07 -05:00
Warren Usui
74eff43ee1 Clean up remote.py and misc.py changes. 
Fixed method names to be non-redundant (remote_mktemp in remote is
now just mktemp, for example), and made some parameters be more
self descriptive.  Added some docstrings.  Fixed sudo setting in
get_file method.

Made chmod independent of the actual sftp file getting.
Do not do extra file copy if non-sudo read is needed.
Fixed some parameter names. Made sure temp files are removed.
 2014-04-29 16:14:11 -05:00
Warren Usui
257e1459fa Handle raw data I/O. 
Paramiko 1.13.0 checks data in the Channel and fails if
invalid UTF-8 characters are sent.  The teuthology/misc.py
functions that piped cat output (get_file) and piped tar
output (pull_directory and pull_directory_tarball) formerly
did not work for Paramiko 1.13.0.

Code was changed to use SFTPClient to copy raw data.  The
remote_mktemp and function was changed to be a method of the
remote object.  Remote object methods to copy files and write
tar files were also added.

In misc.py, get_file() was changed to be a wrapper around the
remote object call.  Pull_directory() and pull_directory_tarball
were changed to use the remote object methods as well.

The remote methods used to send the data makes use of the
SFTPClient interface in Paramiko.

The remote_mktemp function was also modified slightly to insure
that new-lines do not appear in temp file names.

Fixes: 8116
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-04-29 16:14:11 -05:00
Zack Cerza
3faeb08db2 When deleting all of a run's jobs, delete the run 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-29 15:18:42 -05:00
Gregory Meno
5339c1f2ee Changes so these are not installed and still removed 2014-04-29 15:18:42 -05:00
Gregory Meno
ce778848a5 Changes invocation of serialize_remote_roles to internal task to avoid being run during nuke 2014-04-29 05:09:07 -07:00
Gregory Meno
1532af440e Moves node: remote mapping to the internal task. 2014-04-29 05:09:07 -07:00
Gregory Meno
d71a87452f These will likely go somewhere better before merge 2014-04-29 05:09:07 -07:00
Gregory Meno
063b6a27f2 Fixes #8050 Adds a cluster.yaml that is written by interactive task 2014-04-29 05:09:07 -07:00
Sage Weil
f261687f29 valgrind: fix tcmalloc suppression for trusty 
Fixes: #8225
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-04-28 15:57:36 -07:00
Sage Weil
bab84d45ab Revert "valgrind.supp: be less picky about library versions" 
This reverts commit f895d16c9e.
 2014-04-28 15:57:36 -07:00
Warren Usui
694827bc0c Allow scrubbing while thrashing 
Added ability to implement scrubbing while thrashing
(scrub_interval in config can be set to an interval
similar to how clean_interval is set).  Defaults to 0,
which indicates that no scrubbing will take place.
Add scrub_interval description to thrashosds docstring.

Fixes: 7199
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-04-28 11:00:38 -07:00
Zack Cerza
2cbe1dc0f8 Only attempt to use sudo if necessary 2014-04-28 11:12:29 -05:00
Zack Cerza
f102e4943d Post last_in_suite jobs, but delete when run 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-25 15:22:30 -05:00
Zack Cerza
5de353e73f Update unit test for Cluster.__repr__() 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-25 15:01:41 -05:00
Zack Cerza
e6e287446a Fix Cluster.__repr__() 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-25 11:36:43 -05:00
Zack Cerza
af209851e6 Don't push last_in_suite jobs to paddles 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-24 10:46:11 -05:00
Samuel Just
e64d8314ac task/: add tests for ec and rep mark_unfound_lost delete 
Signed-off-by: Samuel Just <sam.just@inktank.com>
 2014-04-22 17:39:29 -07:00
Sandon Van Ness
9078513c0b Fix for #8115 
Increase boot disk size per #8115 where monitors shut down due to
/ being full on vm machines.

Signed-off-by: Sandon Van Ness <sandon@inktank.com>
 2014-04-22 12:35:11 -07:00
Zack Cerza
1448cdf5fc Work around #8166 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-21 10:11:15 -05:00
Zack Cerza
0e90c69fd4 watch_tube() belongs to the beanstalk module 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 16:06:32 -05:00
Zack Cerza
e9a1c778f1 Update requests version 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 15:33:26 -05:00
Zack Cerza
5dbc64264e s/wait-for-package/wait_for_package/ 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 15:03:37 -05:00
Zack Cerza
03b8cdac38 Refactor try_delete_jobs() 
Also tweak its error message
 2014-04-17 12:27:42 -05:00
Zack Cerza
d12e6f4e00 Be slightly less verbose about logging 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
741c773be6 Look for archive_base in config 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
ee33192f70 When deleting jobs, also delete them from paddles 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
8fdea4d132 Submit queued jobs to paddles 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
8a4de411a6 Rename teuthology.queue to teuthology.worker 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
1449e753ab Use shared methods to connect to beanstalkd 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
165f5d5345 When killing a run, delete paddles jobs 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
66a2742211 Add methods for querying and deleting jobs 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 12:27:41 -05:00
Zack Cerza
fc94879498 safe_while: Don't sleep() on the first attempt 
This was causing unnecessary delays in several places

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 11:08:36 -05:00
Zack Cerza
e323392778 Pass -D flag to teuthology report 
Fixes an issue where tests run on old teuthology branches that died for
uncommon reasons were not being marked as dead

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-17 09:43:07 -05:00
Zack Cerza
4b9202bc2f Update to use psutil 2.x API 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-16 15:21:11 -05:00
Zack Cerza
8b93c03fc9 Generate subtasks instead of copy/pasting them 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-16 09:45:07 -05:00
Zack Cerza
761d76937d Don't run apache functions if not using apache 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-16 09:45:07 -05:00
Zack Cerza
089dda1599 Optionally use civetweb instead of apache 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-16 09:45:07 -05:00
Zack Cerza
823219bb27 Don't pass apache's config to radosgw 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-15 14:28:04 -05:00
Zack Cerza
12af2abb4c Rename some functions and variables 
This is to make the refactoring a little smoother and easier to read.

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-15 13:12:00 -05:00
Zack Cerza
f82f663782 Fix all but one of the PEP-8 issues 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-04-15 10:58:36 -05:00
Warren Usui
4c0151367e Improve unlock error messages. 
Added messages if the hostname is invalid, and if
the user is not the owner of the lock.

Fixes: 6295
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-04-14 17:21:37 -07:00
Warren Usui
eef2bf672b Fix kvm issues for Trusty 
Change kvm reference to qemu-system-x86_64 and use raw format
in kvm/qemu command.  Tested on both Trusty and Precise.

Fixes: 7825
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-04-14 16:07:27 -07:00
Dan Mick
6dc29904c6 Change status to _status everywhere (fix regression) 
0550dd1191 didn't add '_' to 'status'
everywhere it was needed.

Signed-off-by: Dan Mick <dan.mick@inktank.com>
 2014-04-10 14:53:10 -07:00
Sage Weil
364d0b4091 ceph.conf: longer client_moutn_timeout (which also applies to librados) 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-04-10 13:24:41 -07:00
Alfredo Deza
0550dd1191 more statuses redefinition fixes 
Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
 2014-04-09 16:56:19 -04:00
Alfredo Deza
30f3b01a0d rename variable to avoid collision 
Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
 2014-04-09 16:36:18 -04:00
Alfredo Deza
b476ec297e remove unused import 
Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
 2014-04-09 16:27:25 -04:00
Alfredo Deza
15a3acb713 fix redefinition of region 
Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
 2014-04-09 16:26:47 -04:00
Yuri Weinstein
b3218eeb09 changed mon data avail warn to avoid monitor storage warnings 2014-04-08 18:21:56 -07:00
Sage Weil
a0c1952d58 ceph.conf: don't force lockdep on mds 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-04-07 16:02:33 -07:00
Zack Cerza
2e997aa7b1 Report job updated time 
When reading job information from archive files, use the mtime of
teuthology.log as the job updated time

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-31 23:12:34 -05:00
Zack Cerza
796af560a3 Handle jobs with no targets 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-31 21:07:24 -05:00
Zack Cerza
addfed2da8 When ignoring a raised exception, at least log it 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-31 17:46:27 -05:00
Zack Cerza
4efb57d3bd Provide real error messages for unfound (sub)tasks 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-31 09:57:19 -05:00
Sage Weil
f895d16c9e valgrind.supp: be less picky about library versions 
...so that this works on trusty

Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-29 09:39:46 -07:00
Josh Durgin
dfdeb5c3c0 radosgw-agent: coerce max-entries config to a string 
This is necessary to pass it as a command line argument.

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-27 17:22:00 -07:00
Yehuda Sadeh
b6ad5c6ce2 rgw: idle timeout config moves to the external server line 
Signed-off-by: Yehuda Sadeh <yehuda@inktank.com>
 2014-03-27 10:24:00 -07:00
Zack Cerza
158f9ba1ff Revert "Lines formerly of the form '(remote,) = ctx.cluster.only(role).remotes.keys()'" 
This reverts commit d693b3f895.
 2014-03-27 11:35:28 -05:00
Warren Usui
d693b3f895 Lines formerly of the form '(remote,) = ctx.cluster.only(role).remotes.keys()' 
and '(remote,) = ctx.cluster.only(role).remotes.iterkeys()' would fail with
ValueError and no message if there were less than 0 or more than 1 key.
Now a new function, get_single_remote_value() is called which prints out
more understandable messages.

Fixes: 7510
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
Signed-off-by: Warren Usui <warren.usui@inktank.com>
 2014-03-26 18:43:48 -07:00
Zack Cerza
5aa5566b3a Don't explode when finding an empty job 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-26 12:33:40 -05:00
Zack Cerza
859e66b38a Merge pull request #232 from ceph/wip-suite-priority 
suite: allow priority to be set
 2014-03-26 12:16:17 -05:00
Sage Weil
67844e0ef3 suite: allow priority to be specified when scheduling 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-26 10:09:12 -07:00
Josh Durgin
b2cf052a08 rgw: allow overriding options 
This is mainly useful for toggling ec/replicated pools right now.

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-25 18:04:35 -07:00
Josh Durgin
812e48a148 radosgw-admin: skip data sync tests when only metadata is configured 
Fixes http://tracker.ceph.com/issues/7818

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-25 14:57:50 -07:00
Josh Durgin
f89c0d59fb rgw: fix ec-data-pool config 
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-25 14:57:50 -07:00
Zack Cerza
f250da841f Add log statement to find_run_info() 
Sometimes it takes a while.

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-25 15:50:03 -05:00
Zack Cerza
b3d0c19f5f Exit gracefully on KeyboardInterrupt 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-25 15:40:56 -05:00
Zack Cerza
03974bf583 Add --description flag 
Default to not printing job descriptions. Add -d/--description flag to
control that behavior. Change --delete's short option to -D. Other small
output tweaks.

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-25 15:38:34 -05:00
Zack Cerza
381ecea8b6 Fix lots of linter errors 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-25 15:27:54 -05:00
Zack Cerza
70ab6e0315 Add --preserve-queue to teuthology-kill 
For any cases where we just want to kill the running jobs and not delete
any scheduled jobs. This only applies to killing entire runs and is not
the default.

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-25 10:18:33 -05:00
Zack Cerza
c404be6162 Postpone creation of logger object 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-25 10:07:52 -05:00
Josh Durgin
47ba1711d7 rgw: add option to use erasure coding instead of replication 
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-24 18:19:41 -07:00
Josh Durgin
0cb00b1fb9 radosgw-agent: default to 1000 max entries 
Using a small number of max entries is only needed for one test, so
make it configurable. Setting it too low makes s3tests fail.

Fixes http://tracker.ceph.com/issues/7742

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-24 14:54:36 -07:00
Josh Durgin
3e1683064c internal: cleaner fix for binary gibberish in logs 
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-24 14:26:36 -07:00
Josh Durgin
b677bdd1bb internal: ignore binary junk in kernel logs 
NUL bytes can show up here after powercycle testing

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-24 14:14:40 -07:00
Josh Durgin
190d8183e9 ceph: ignore daemon types that aren't configured 
not every cluster needs an mds

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-24 14:14:40 -07:00
Sage Weil
752a76fb48 radosgw-agent: use our mirror instead of github 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-21 13:50:10 -07:00
Sage Weil
9b6f93a5b2 Merge remote-tracking branch 'gh/firefly' 2014-03-21 09:12:40 -07:00
Sage Weil
4e2f36fcd8 rgw: handle empty conf case 
2014-03-20T23:30:28.964 ERROR:teuthology.contextutil:Saw exception from nested tasks
Traceback (most recent call last):
  File "/home/teuthworker/teuthology-firefly/teuthology/contextutil.py", line 27, in nested
    vars.append(enter())
  File "/usr/lib/python2.7/contextlib.py", line 17, in __enter__
    return self.gen.next()
  File "/home/teuthworker/teuthology-firefly/teuthology/task/rgw.py", line 83, in ship_config
    idle_timeout = conf.get('idle_timeout', 30)
AttributeError: 'NoneType' object has no attribute 'get'

This happens with job yamls like:

- rgw:
  - client.0

(where it is a list and not a dict of clients)

Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-21 08:36:24 -07:00
Sage Weil
7dcfdd288f Merge remote-tracking branch 'gh/firefly' 2014-03-20 16:50:15 -07:00
Sage Weil
3a2b77c142 rgw: allow fastcgi idle timeout to be adjusted 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-20 16:09:52 -07:00
Sage Weil
daa5429278 Merge branch 'firefly' 2014-03-20 10:35:30 -07:00
Sage Weil
7088885ecd internal: ignore ext4 recovery msg 
plana43 kernel: [   21.179205] EXT4-fs (sda1): INFO: recovery required on readonly filesystem

is harmless.

Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-18 15:23:16 -07:00
Sage Weil
5679ae2a1f Merge branch 'firefly' 2014-03-18 15:01:04 -07:00
Sage Weil
8e748c0c2e ceph.conf: tolerate 500ms of clock drift (up from 250) 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-18 15:00:55 -07:00
Sage Weil
fb7b1f8768 Merge remote-tracking branch 'gh/firefly' 2014-03-17 20:56:51 -07:00
Loic Dachary
6acfa6d638 ceph_manager: update ec_pool creation parameters 
As of https://github.com/ceph/ceph/pull/1477 the erasure code parameters
are controled via the osd erasure-code-profile set command instead of
being inlined and duplicated in the crush ruleset creation and the pool
creation. There is no need to create the crush ruleset, it is done
implicitly.

Signed-off-by: Loic Dachary <loic@dachary.org>
 2014-03-18 00:48:53 +01:00
Samuel Just
2b90be2b2b ceph_manager: default to 16 initial pgs per pool 
1 is really too few for anything to happen at a
reasonable rate.

Signed-off-by: Samuel Just <sam.just@inktank.com>
 2014-03-17 16:11:39 -07:00
Samuel Just
847462b4b1 ceph_manager::wait_for_clean: reset timeout if we make progress 
Fixes: #7748
Signed-off-by: Samuel Just <sam.just@inktank.com>
 2014-03-17 16:11:36 -07:00
Josh Durgin
a4261af71b Merge pull request #227 from ceph/wip-7749 
Pass '--object-sync-timeout 30' to radosgw-agent

Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-17 12:10:27 -07:00
Zack Cerza
ef2edcd3a9 Pass '--object-sync-timeout 30' to radosgw-agent 
http://tracker.ceph.com/issues/7749

Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-17 14:00:08 -05:00
Zack Cerza
7f177b3a55 Merge pull request #226 from ceph/wip-7714 
prevent undefined values in finally block
 2014-03-17 13:42:27 -05:00
Zack Cerza
98e27d7b13 Fix max_job_time timeout 
Signed-off-by: Zack Cerza <zack.cerza@inktank.com>
 2014-03-17 11:19:02 -05:00
Alfredo Deza
57c8fd9687 prevent undefined values in finally block 
Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
 2014-03-17 10:49:31 -04:00
Sandon Van Ness
d3c294d206 Make distro kernels work on Trusty. 
Some grub changes and package names caused them to fail on trusty.

Signed-off-by: Sandon Van Ness <sandon@inktank.com>
 2014-03-17 09:42:11 -05:00
Sage Weil
d72335fe9a Merge remote-tracking branch 'gh/firefly' 2014-03-13 13:15:04 -07:00
Sage Weil
1ccabd802f lock: allow -a with --brief 
Signed-off-by: Sage Weil <sage@inktank.com>
 2014-03-12 08:57:23 -07:00
Sandon Van Ness
b455846d9f Handle newer btrfstools. 
Newer btfs userland tools needs a -f like xfs instead of a prompt
for yes. Trusty needs this change.

Signed-off-by: Sandon Van Ness <sandon@inktank.com>
 2014-03-11 18:54:02 -07:00
wusui
e5b65b4578 Merge pull request #222 from ceph/wip-trusty 
Handle newer btrfstools.
 2014-03-11 18:48:13 -07:00
Sandon Van Ness
a5a94ae310 Handle newer btrfstools. 
Newer btfs userland tools needs a -f like xfs instead of a prompt
for yes. Trusty needs this change.

Signed-off-by: Sandon Van Ness <sandon@inktank.com>
 2014-03-11 18:15:12 -07:00
Samuel Just
1c611330cd radosbench: cleanup will probably take longer than writing the objects 
With thrashing, cleanup will require recovering the degraded objects
first, while the initial write out doesn't since they they are new
objects.

Signed-off-by: Samuel Just <sam.just@inktank.com>
 2014-03-11 13:55:32 -07:00
wusui
a9a826adb8 Merge pull request #220 from ceph/wip-print 
Add print task.
 2014-03-11 09:41:39 -07:00
Dan Mick
69ed31cc34 lock.py: allow --brief to stand on its own (without --list) 
I use --brief all the time, and it's annoying to have to type
the useless --list at the same time.

Signed-off-by: Dan Mick <dan.mick@inktank.com>
 2014-03-10 19:38:46 -07:00
Yuri Weinstein
ed6a200104 Add print task. 
Made a task that simply prints the output that is given to it as
an argument. Can be used like any other task (under sequential,
etc...)

IE:

tasks:
- print: "String"
- chef: null
- print: "Another String"

Signed-off-by: Sandon Van Ness <sandon@inktank.com>
Signed-off-by: Yuri Weinstein <yuri.weinstein@inktank.com>
 2014-03-10 16:45:03 -07:00
Josh Durgin
7cb750c4cc radosgw-admin: test data sync with more than max-entries objects 
For now hard code the max-entries for the radosgw-agent, since
this is the only test that is affected by it.

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
 2014-03-10 00:29:50 -07:00