This injects backfill full as opposed to lowering the backfill full ratio
Signed-off-by: David Zafman <dzafman@redhat.com>
--- /dev/null
+roles:
+- - mon.a
+ - mgr.x
+ - osd.0
+ - osd.1
+ - osd.2
+- - osd.3
+ - osd.4
+ - osd.5
+ - client.0
+openstack:
+ - volumes: # attached to each instance
+ count: 3
+ size: 10 # GB
+override:
+ ceph:
+ conf:
+ mon:
+ osd default pool size: 3
+ osd min pg log entries: 5
+ osd max pg log entries: 10
+tasks:
+- install:
+- ceph:
+ log-whitelist:
+ - but it is still running
+ - missing primary copy of
+ - objects unfound and apparently lost
+ - overall HEALTH_
+ - \(OSDMAP_FLAGS\)
+ - \(SLOW_OPS\)
+ - \(PG_
+ - \(OBJECT_MISPLACED\)
+ - \(OSD_
+ - \(OBJECT_
+ - \(TOO_FEW_PGS\)
+ - \(POOL_BACKFILLFULL\)
+- thrashosds:
+ op_delay: 30
+ clean_interval: 120
+ chance_down: .75
+ min_live: 5
+ min_in: 5
+ chance_test_backfill_full: .5
+- radosbench:
+ clients: [client.0]
+ time: 1800
+ type: rand
+ objectsize: 1048576