Log: http://qa-proxy.ceph.com/teuthology/smithfarm-2018-04-08_15:24:44-upgrade:hammer-x-wip-jewel-backports-distro-basic-smithi/2371505/teuthology.log

Failure Reason:

"2018-04-08 15:41:33.585719 osd.1 172.21.15.178:6800/11970 1 : cluster [WRN] map e11 wrongly marked me down" in cluster log

  • kernel:
    • sha1: distro
    • kdb: True
  • tasks:
    • internal.check_packages:
    • internal.buildpackages_prep:
    • internal.lock_machines:
      • 2
      • smithi
    • internal.save_config:
    • internal.check_lock:
    • internal.add_remotes:
    • console_log:
    • internal.connect:
    • internal.push_inventory:
    • internal.serialize_remote_roles:
    • internal.check_conflict:
    • internal.check_ceph_data:
    • internal.vm_setup:
    • kernel:
      • sha1: distro
      • kdb: True
    • internal.base:
    • internal.archive_upload:
    • internal.archive:
    • internal.coredump:
    • internal.sudo:
    • internal.syslog:
    • internal.timer:
    • pcp:
    • selinux:
    • ansible.cephlab:
    • clock:
    • print: **** Install version lower than v0.94.4
    • install:
      • tag: v0.94.3
      • sha1: 64695235ff1ba4b33491329af2dc51863c1b3707
    • ceph:
      • cluster: ceph
      • log-whitelist:
        • slow request
        • scrub mismatch
        • ScrubResult
      • fs: xfs
      • conf:
        • mon:
          • mon warn on legacy crush tunables: False
          • debug mon: 20
          • debug paxos: 20
          • debug ms: 1
        • osd:
          • debug osd: 25
          • debug filestore: 20
          • debug journal: 20
          • debug ms: 1
      • sha1: 64695235ff1ba4b33491329af2dc51863c1b3707
    • print: *** Upgrade the target that runs osd.0 and osd.1 to -x while the target that runs osd.2 stays v0.94.3
    • install.upgrade:
      • osd.0:
    • print: *** Restart the mon.a so that it is post-hammer v0.94.4 and implements the CEPH_FEATURE_HAMMER_0_94_4 feature
    • ceph.restart:
      • daemons:
        • mon.a
    • print: *** Verify that osd.0 cannot restart because osd.1 and osd.2 are still < v0.94.4
    • ceph.restart:
      • daemons:
        • osd.0
      • wait-for-healthy: False
    • exec:
      • osd.0:
        • set -x success=false for delay in 1 2 4 8 16 32 64 128 256 512 1024 ; do if ceph daemon osd.0 log flush ; then if grep "one or more pre-v0.94.4 hammer" /var/log/ceph/ceph-osd.0.log ; then success=true break fi fi sleep $delay done $success || exit 1
    • print: *** Stop all OSDs and restart osd.0 and osd.1 which are >= v0.94.4
    • ceph.stop:
      • daemons:
        • osd.0
        • osd.1
        • osd.2
    • exec:
      • mon.a:
        • set -x ceph osd down osd.0 ceph osd down osd.1 ceph osd down osd.2
    • ceph.restart:
      • daemons:
        • osd.0
        • osd.1
      • wait-for-healthy: False
    • exec:
      • mon.a:
        • set -x success=false for delay in 1 2 4 8 16 32 64 128 256 512 1024 ; do if ceph osd dump | grep 'osd.1 up' && ceph osd dump | grep 'osd.0 up' ; then success=true break fi ceph osd dump sleep $delay done $success || exit 1 ceph osd dump | grep 'osd.2 down' || exit 1
    • print: *** Verify that osd.2 cannot restart because it is < v0.94.4 and all other OSDs are >= v0.94.4
    • ceph.restart:
      • daemons:
        • osd.2
      • wait-for-healthy: False
    • exec:
      • mon.a:
        • set -x success=false for delay in 1 2 4 8 16 32 64 128 256 512 1024 ; do ceph daemon mon.a log flush if grep "disallowing boot of pre-hammer v0.94.4 OSD" /var/log/ceph/*.log ; then success=true break fi sleep $delay ceph osd dump done $success || exit 1
    • print: *** Upgrade the target that runs osd.2 to -x and verify the cluster is back to being healthy
    • install.upgrade:
      • osd.2:
    • ceph.restart:
      • daemons:
        • osd.2
      • wait-for-healthy: False
    • exec:
      • mon.a:
        • sleep 300
        • ceph osd set sortbitwise
        • ceph osd set require_jewel_osds
    • ceph.healthy:
  • verbose: True
  • pid:
  • duration: 0:13:16
  • owner: scheduled_smithfarm@teuthology
  • flavor: basic
  • status_class: danger
  • targets:
    • smithi178.front.sepia.ceph.com: ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDFglcQkZML/8rJWpe70OjFdOw6Nm711SzjcdBJAnFjxEF+T1rtnRAcyZ8TqutaNIy5CHzZyaESjl0ei0OD2xzU/xayZ3sVkS1TLzRpimjdqXOt/4qNND+s+pfaDIYIeaPQXFBQNXKPbr5GCWT5gBJN5DSBCzinBxOLEX87fKlpSpkvAVDt4LRWnLXFTm7YaJqtWo+rqVLIO3t8DZHUUuvuBcH2QJdZEK9B3nDYdMxmUhep83WNIu36bHSFB0/XS03mJ+MuHSvdexgfrb9IZye/fTPJ2ipCZsDh0TWsDLijTMseSgOmkuE2wHZtL1pvDt8/MI9noCGD1S9i3dYBQzd3
    • smithi179.front.sepia.ceph.com: ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDD3hzUVtc39JvNLD8R0RB1NMSgtYO+Tg8odoFasa7c15hyX9D6eAgeDPrD91MxMKs21tf7wEmOr3q9NS8+9TM7A21fV6Gov78UQHosoJbmmjiCQqiTlXzU0nsE10dUZfgkqtiG2diHlKZ1+PVMZXMETCihq5glM3A6Yy/NlGSz3PfZC33VUmgQUj+YFXOXz71tIwYjQljzdoih9k5QFSy/Y3QuV9ySYp9E+6GMHJTAHvELXPNrykczAM0B1VpqOKqdrSjr5IDPwQAcqWqRCKzr2cLiRHWazFVz+K063sEy1AMX8aWBw1RkOtygPf/8lQBu2GZkT/YrvFnxroupbKJx
  • job_id: 2371505
  • log_href: http://qa-proxy.ceph.com/teuthology/smithfarm-2018-04-08_15:24:44-upgrade:hammer-x-wip-jewel-backports-distro-basic-smithi/2371505/teuthology.log
  • suite_branch: wip-jewel-backports
  • wait_time: 0:10:43
  • os_version: 14.04
  • branch: wip-jewel-backports
  • pcp_grafana_url:
  • email: ncutler@suse.com
  • archive_path: /home/teuthworker/archive/smithfarm-2018-04-08_15:24:44-upgrade:hammer-x-wip-jewel-backports-distro-basic-smithi/2371505
  • updated: 2018-04-08 15:50:30
  • description: upgrade:hammer-x/v0-94-4-stop/{distros/centos_latest.yaml distros/ubuntu_14.04.yaml ignore.yaml v0-94-4-stop.yaml}
  • started: 2018-04-08 15:26:31
  • last_in_suite: False
  • machine_type: smithi
  • sentry_event:
  • posted: 2018-04-08 15:24:52
  • teuthology_branch: master
  • sha1: 64695235ff1ba4b33491329af2dc51863c1b3707
  • name: smithfarm-2018-04-08_15:24:44-upgrade:hammer-x-wip-jewel-backports-distro-basic-smithi
  • roles:
    • [u'mon.a', u'osd.0', u'osd.1']
    • [u'osd.2']
  • overrides:
    • ceph:
      • log-whitelist:
        • slow request
        • scrub mismatch
        • ScrubResult
      • conf:
        • mon:
          • mon warn on legacy crush tunables: False
          • debug mon: 20
          • debug paxos: 20
          • debug ms: 1
        • osd:
          • debug osd: 25
          • debug filestore: 20
          • debug journal: 20
          • debug ms: 1
      • sha1: 64695235ff1ba4b33491329af2dc51863c1b3707
    • ceph-deploy:
      • conf:
        • client:
          • log file: /var/log/ceph/ceph-$name.$pid.log
        • mon:
          • osd default pool size: 2
    • workunit:
      • sha1: 64695235ff1ba4b33491329af2dc51863c1b3707
      • branch: wip-jewel-backports
    • install:
      • ceph:
        • sha1: 64695235ff1ba4b33491329af2dc51863c1b3707
    • admin_socket:
      • branch: wip-jewel-backports
  • success: False
  • failure_reason: "2018-04-08 15:41:33.585719 osd.1 172.21.15.178:6800/11970 1 : cluster [WRN] map e11 wrongly marked me down" in cluster log
  • status: fail
  • nuke_on_error: True
  • os_type: ubuntu
  • runtime: 0:23:59
  • suite_sha1: 64695235ff1ba4b33491329af2dc51863c1b3707