Ceph

nuke-on-error: true
overrides:
  ceph:
    branch: next
    conf:
      client:
        debug ms: 1
        log max new: 1
        rbd cache: true
      global:
        ms inject socket failures: 5000
      osd:
        debug ms: 1
        debug osd: 20
    fs: ext4
    log-whitelist:
    - slow request
roles:
- - mon.a
  - osd.0
  - osd.1
  - osd.2
- - mds.a
  - osd.3
  - osd.4
  - osd.5
- - client.0
tasks:
- chef: null
- clock: null
- ceph:
    log-whitelist:
    - wrongly marked me down
    - objects unfound and apparently lost
- thrashosds:
    timeout: 1200
- rbd_fsx:
    clients:
    - client.0
    ops: 500

looks like osd is sending stat messages, at some point the mon stops receiving and acking them. eventually marks the osd down (after the 900s timeout)