RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2025-01-11 13:41:02 +00:00
Code Issues Packages Projects Releases Wiki Activity
5,615 Commits 214 Branches 394 Tags 1.3 GiB
47a41286a5
Commit Graph

284 Commits

Author SHA1 Message Date
Kefu Chai
3ad90bfbdb rados/singleton-nomsgr/all/11429: wait until clean before checking log 
to make sure that load_pgs() is finished before checking its output

Fixes: http://tracker.ceph.com/issues/16157
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2016-09-06 17:50:30 +08:00
Loic Dachary
52017e6707 openstack: add disks where necessary 
Ref: http://tracker.ceph.com/issues/17152

Signed-off-by: Loic Dachary <loic@dachary.org>
 2016-08-29 10:33:55 +02:00
Josh Durgin
b76c8903fb Merge pull request #1091 from athanatos/wip-16413 
11429: wait_for_clean between restarting the osds and starting the bench

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2016-07-22 14:28:59 -07:00
Sage Weil
ec439c01ea rados/thrash/workloads/cache-agent-big: set min_size=2 on m=1 k=2 ec pool 
Otherwise min_size will be 3 and we will wedge on a single down
osd.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-06-24 14:44:00 -04:00
Samuel Just
c4ad6b3b33 11429: wait_for_clean between restarting the osds and starting the bench 
ceph.restart now marks the osds down, so the objects are actually being
created while slowest of the osds boots. That causes a ton of 1 byte
objects to be created in a degraded state and causes the cleanup to take
a long time.  Also, reduce length of bench since it's only being used
to ensure the osds came up correctly.

Signed-off-by: Samuel Just <sjust@redhat.com>
 2016-06-22 09:25:25 -07:00
Sage Weil
11ee1a8cfd rados/objectstore/objectstore.yaml: skip bluestore tests 
Avoid breaking rados suite while we are developing rapidly.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-06-18 12:51:32 -04:00
Yuri Weinstein
3977b8ffdb Fixes http://tracker.ceph.com/issues/16334 
Fixes http://tracker.ceph.com/issues/16332
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
 2016-06-15 13:43:49 -07:00
Samuel Just
7e06dd3bdc suites/rados: add test for 16113 
Signed-off-by: Samuel Just <sjust@redhat.com>
 2016-06-06 10:59:01 -07:00
Sage Weil
ce927692b1 rados: vary mon kv backend 
Use just rocksdb for thrash to avoid blowing up the matrix.  Leave
leveldb for the ec thrash suites.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-05-15 19:20:32 -04:00
ZhenyuLeng
1eaa002691 Test osd_discard_disconnected_ops 
Signed-off-by: ZhenyuLeng <lengzhenyu@ict.ac.cn>
 2016-05-13 20:49:17 +08:00
Samuel Just
0cade258c8 Merge pull request #986 from ceph/wip-crush-drop-legacy 
rados: test jewel instead of legacy tunables

Reviewed-by: Samuel Just <sjust@redhat.com>
 2016-05-11 11:12:03 -07:00
Samuel Just
a4f8dc91bc Revert "Revert "Rados: Tracker-13531-resolve stuck peering"" 2016-05-10 10:27:32 -07:00
Sage Weil
1b687fcff6 rados: test jewel instead of legacy tunables 
Legacy tunables semi-frequently turn up crush mappings that only
return 1 instead of 2 results.  Switch to covering jewel tunables
explicitly instead.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-05-09 08:31:24 -04:00
Sage Weil
84aeb9e89e rados: debug rgw test 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-05-06 12:03:14 -04:00
Sage Weil
2a44c3d20d drop broken name length config args 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-05-05 09:07:36 -04:00
Samuel Just
18f2aa09ff Revert "Rados: Tracker-13531-resolve stuck peering" 2016-05-02 20:39:39 -07:00
shylesh kumar
e3bb551067 Rados: wip-13531:Resolve stuck peering 
Signed-off-by: shylesh kumar <shylesh.mohan@gmail.com>
 2016-05-02 12:41:14 -04:00
Sage Weil
f21651662e rados/monthrash: debug client on librados api tests 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-04-24 16:39:04 -04:00
Sage Weil
b095cc6c54 rados/objectstore: add ceph_test_keyvaluedb test 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-04-22 14:08:10 -04:00
Sage Weil
e265bee1c9 rados: specify short names for misc jobs 
These need to specify short names bc they don't have all osds on
a dedicated block device.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-04-15 08:59:01 -04:00
Sage Weil
32719a60f7 rados: specify xfs 
If no fs is specified, the OSD is created on /, and the scratch
devices aren't used at all.  That means ext4, and the long file
name lameness.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-04-14 14:31:44 -04:00
Sage Weil
2b7dd172c2 rados/basic: test over xfs (not just btrfs) 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-04-13 13:14:04 -04:00
Sage Weil
897e52ee82 Merge pull request #944 from athanatos/wip-lfn-upgrade 
LFN upgrade tests and a few other misc fixes
 2016-04-11 14:01:21 -04:00
Josh Durgin
6903f1cd31 admin malloc interface testsReviewed-by: Josh Durgin <jdurgin@redhat.com> 
admin malloc interface tests

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
 2016-04-07 15:38:22 -07:00
Samuel Just
7e53203e80 rados/singleton-nomsgr: add lfn upgrade tests 
Upgrade from hammer/infernalis to x and verify lfn objects are valid
across upgrade.

Signed-off-by: Samuel Just <sjust@redhat.com>
 2016-04-07 15:35:30 -07:00
Samuel Just
da9c65b1af rados_api_tests: reduce the osd_max_object_name_len value 
This test can be run against ext4, which maxes out under
1k.

Signed-off-by: Samuel Just <sjust@redhat.com>
 2016-04-07 15:35:30 -07:00
Sage Weil
5500a69443 Merge pull request #932 from athanatos/wip-journal-defaults 
morepggrow: enable filestore and journal soft backoffs
 2016-04-04 13:28:51 -04:00
Sage Weil
8e11f0d827 suites: radosbench 180 -> 150s 
I hit an ENOSPC after 150s on bluestore.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-31 12:27:11 -04:00
Loic Dachary
a7194e5685 openstack: thrash_cache_writeback_proxy_none.yaml need bigger disks 
http://tracker.ceph.com/issues/15324 Fixes: #15324

Signed-off-by: Loic Dachary <loic@dachary.org>
 2016-03-30 18:05:08 +02:00
Sage Weil
935a7eb229 rados/singleton-nomsgr/all/full-tiering: s/readforwad/readproxy/ 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-30 08:23:31 -04:00
Sage Weil
d823ff6864 rados/singleton: thrash proxy (not forward) cache mode 
See #12814

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-30 08:23:31 -04:00
Loic Dachary
d9a8f49fc5 openstack: define the resources required to run a rados suite 
Signed-off-by: Loic Dachary <loic@dachary.org>
 2016-03-30 01:45:14 +02:00
Gregory Farnum
2f5c76cf56 Merge pull request #898 from ceph/wip-15206 
test_msgr: improve debug ms level
 2016-03-28 09:24:42 -07:00
Sage Weil
2b3ec8de8d suites: run rados bench for max 180 seconds 
If we want a longer runtime than that, run it several times in
sequence.

This limits the amount of disk space we'll consume.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-26 09:18:41 -04:00
Haomai Wang
322b79fd26 test_msgr: improve debug ms level 2016-03-20 11:54:49 +08:00
Samuel Just
eff7adaf78 morepggrow: enable filestore and journal soft backoffs 
Signed-off-by: Samuel Just <sjust@redhat.com>
 2016-03-18 14:57:59 -07:00
Sage Weil
e0a1c38ed0 suites/rados/thrash: vary min_{read,write}_recency_for_promote 
Hopefully this will shake out #15171

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-17 09:03:23 -04:00
Sage Weil
c619feb7ad rados/singleton-nomsgr: run singleton upgrade tests on ubuntu 
Fixes: #15139
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-16 13:16:51 -04:00
Sage Weil
764946bb0f Merge pull request #880 from ceph/wip-rados-vs-mds 
suites: remove unused mdss from RADOS tests
 2016-03-16 08:57:06 -04:00
John Spray
a3d78d0516 suites: fix assumed filesystem pool name 
Tests shouldn't really assume the names of
cephfs pools, but because it's hardcoded in
the yaml here I'll just update it.

Signed-off-by: John Spray <john.spray@redhat.com>
 2016-03-15 16:55:10 +00:00
John Spray
77e35c01fd suites: remove unused mdss from RADOS tests 
These daemons were apparently unused in tests that
created an MDS daemon but never use a fuse task
to access it.

Signed-off-by: John Spray <john.spray@redhat.com>
 2016-03-15 13:14:05 +00:00
Sage Weil
3625b8c438 rados/singleton-nomsgr/all/11429: no mds 
The newer cephfs logic wants to use the 'status' command on the
daemon, but that doesn't exist on firefly.  Luckly we don't need
it for this test.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-10 12:06:21 -05:00
Samuel Just
88e86d77a2 rados: Enable odsync write for pggrow thrasher 
Signed-off-by: Samuel Just <sjust@redhat.com>
 2016-03-10 07:45:30 -08:00
Sage Weil
7784ffe61e rados/thrash/workloads/radosbench: shorten 300s->200s 
Signed-off-by: Sage Weil <sage@redhat.com>
 2016-03-08 11:14:55 -05:00
Samuel Just
e70a633b4f Merge pull request #832 from athanatos/wip-lost 
Wip lost

Reviewed-by: Sage Weil <sage@redhat.com>
 2016-02-29 15:53:12 -08:00
Zack Cerza
0441f100a0 Merge pull request #837 from ceph/wip-valgrind 
suites: debuginfo: true for valgrind
 2016-02-26 08:38:49 -07:00
Sage Weil
0436306cbe Merge pull request #833 from ceph/wip-async 
Revert "basic/msgr: remove async and random for now"
 2016-02-26 10:01:53 -05:00
Sage Weil
f274737dfa suites: debuginfo: true for valgrind 
This lets us get line numbers in the valgrind xml output.

Signed-off-by: Sage Weil <sage@redhat.com>
 2016-02-26 09:46:27 -05:00
Kefu Chai
9a3252c17e tasks/scrub_test: add test for get-inconsistent-* commands 
Signed-off-by: Kefu Chai <kchai@redhat.com>
 2016-02-22 18:35:50 +08:00
Samuel Just
ef75c8971e rados: add lost-unfound upgrade tests 
Signed-off-by: Samuel Just <sjust@redhat.com>
 2016-02-19 12:28:36 -08:00