these tests don't need to be split across the job matrix the same way
that we split features like frontend, ssl, objectstore, etc. by
combining them, we can still test the whole matrix of features, but with
only 1/3 of the total jobs
Signed-off-by: Casey Bodley <cbodley@redhat.com>
+++ /dev/null
-os_type: centos # xenial valgrind buggy, see http://tracker.ceph.com/issues/18126
-tasks:
-- install:
- flavor: notcmalloc
- debuginfo: true
-- ceph:
-- rgw:
- client.0:
- valgrind: [--tool=memcheck]
-- ragweed:
- client.0:
- force-branch: ceph-master
- rgw_server: client.0
- stages: prepare
-- ragweed:
- client.0:
- force-branch: ceph-master
- rgw_server: client.0
- stages: check
-overrides:
- ceph:
- conf:
- global:
- osd_min_pg_log_entries: 10
- osd_max_pg_log_entries: 10
+++ /dev/null
-# see http://tracker.ceph.com/issues/20360 and http://tracker.ceph.com/issues/18126
-os_type: centos
-
-tasks:
-- install:
- flavor: notcmalloc
-- ceph:
-- rgw:
- client.0:
- valgrind: [--tool=memcheck]
-- s3tests:
- client.0:
- force-branch: ceph-master
- rgw_server: client.0
-overrides:
- ceph:
- conf:
- global:
- osd_min_pg_log_entries: 10
- osd_max_pg_log_entries: 10
- client:
- rgw lc debug interval: 10
+++ /dev/null
-# see http://tracker.ceph.com/issues/20360 and http://tracker.ceph.com/issues/18126
-os_type: centos
-
-tasks:
-- install:
- flavor: notcmalloc
-- ceph:
-- rgw:
- client.0:
- valgrind: [--tool=memcheck]
-- swift:
- client.0:
- rgw_server: client.0
--- /dev/null
+# see http://tracker.ceph.com/issues/20360 and http://tracker.ceph.com/issues/18126
+os_type: centos
+
+tasks:
+- install:
+ flavor: notcmalloc
+- ceph:
+- rgw:
+ client.0:
+ valgrind: [--tool=memcheck]
+- s3tests:
+ client.0:
+ force-branch: ceph-master
+ rgw_server: client.0
+- swift:
+ client.0:
+ rgw_server: client.0
+- ragweed:
+ client.0:
+ force-branch: ceph-master
+ rgw_server: client.0
+ stages: prepare,check
+overrides:
+ ceph:
+ conf:
+ global:
+ osd_min_pg_log_entries: 10
+ osd_max_pg_log_entries: 10
+ client:
+ rgw lc debug interval: 10