Posted | Started | Updated | Runtime |
Duration |
In Waiting |
Machine | Teuthology Branch | OS Type | OS Version | Nodes | Status |
---|---|---|---|---|---|---|---|---|---|---|---|
2023-11-03 14:14:54 | 2023-11-03 14:18:00 | 2023-11-03 14:54:41 | 0:36:41 | 0:22:44 | 0:13:57 | smithi | main | ubuntu | 20.04 | 3 | fail |
Description: orch:cephadm/workunits/{0-distro/ubuntu_20.04 agent/on mon_election/connectivity task/test_host_drain}
Sentry event: https://sentry.ceph.com/organizations/ceph/?query=e148ebcd92174bd1afb6ad1b89678b24
Command failed on smithi072 with status 22: 'sudo /home/ubuntu/cephtest/cephadm --image quay-quay-quay.apps.os.sepia.ceph.com/ceph-ci/ceph:054a55d156d3a2f0db54aab582507ae7417905e5 shell -c /etc/ceph/ceph.conf -k /etc/ceph/ceph.client.admin.keyring --fsid bd0b3be0-7a56-11ee-8db9-212e2dc638e7 -- bash -c \'set -ex\nHOSTNAMES=$(ceph orch host ls --format json | jq -r \'"\'"\'.[] | .hostname\'"\'"\')\nfor host in $HOSTNAMES; do\n # find the hostname for "host.c" which will have no mgr\n HAS_MGRS=$(ceph orch ps --hostname ${host} --format json | jq \'"\'"\'any(.daemon_type == "mgr")\'"\'"\')\n if [ "$HAS_MGRS" == "false" ]; then\n HOST_C="${host}"\n fi\ndone\n# now drain that host\nceph orch host drain $HOST_C --zap-osd-devices\n# wait for drain to complete\nHOST_C_DAEMONS=$(ceph orch ps --hostname $HOST_C)\nwhile [ "$HOST_C_DAEMONS" != "No daemons reported" ]; do\n sleep 15\n HOST_C_DAEMONS=$(ceph orch ps --hostname $HOST_C)\ndone\n# we want to check the ability to remove the host from\n# the CRUSH map, so we should first verify the host is in\n# the CRUSH map.\nceph osd getcrushmap -o compiled-crushmap\ncrushtool -d compiled-crushmap -o crushmap.txt\nCRUSH_MAP=$(cat crushmap.txt)\nif ! grep -q "$HOST_C" <<< "$CRUSH_MAP"; then\n printf "Expected to see $HOST_C in CRUSH map. Saw:\\n\\n$CRUSH_MAP"\n exit 1\nfi\n# If the drain was successful, we should be able to remove the\n# host without force with no issues. If there are still daemons\n# we will get a response telling us to drain the host and a\n# non-zero return code\nceph orch host rm $HOST_C --rm-crush-entry\n# verify we\'"\'"\'ve successfully removed the host from the CRUSH map\nsleep 30\nceph osd getcrushmap -o compiled-crushmap\ncrushtool -d compiled-crushmap -o crushmap.txt\nCRUSH_MAP=$(cat crushmap.txt)\nif grep -q "$HOST_C" <<< "$CRUSH_MAP"; then\n printf "Saw $HOST_C in CRUSH map after it should have been removed.\\n\\n$CRUSH_MAP"\n exit 1\nfi\n\''