- log_href:
http://qa-proxy.ceph.com/teuthology/yuriw-2024-08-29_14:15:40-rados-wip-yuri8-testing-2024-08-28-1632-squid-distro-default-smithi/7880215/teuthology.log
- archive_path:
/home/teuthworker/archive/yuriw-2024-08-29_14:15:40-rados-wip-yuri8-testing-2024-08-28-1632-squid-distro-default-smithi/7880215
- description:
rados/cephadm/workunits/{0-distro/centos_9.stream agent/on mon_election/connectivity task/test_host_drain}
- duration:
0:11:52
- email:
yweinste@redhat.com
- failure_reason:
- flavor:
default
- job_id:
7880215
- kernel:
- last_in_suite:
False
- machine_type:
smithi
- name:
yuriw-2024-08-29_14:15:40-rados-wip-yuri8-testing-2024-08-28-1632-squid-distro-default-smithi
- nuke_on_error:
- os_type:
centos
- os_version:
9.stream
- overrides:
- admin_socket:
- branch:
wip-yuri8-testing-2024-08-28-1632-squid
- ceph:
- conf:
- global:
- mon election default strategy:
3
- mgr:
- debug mgr:
20
- debug ms:
1
- mgr/cephadm/use_agent:
True
- mon:
- debug mon:
20
- debug ms:
1
- debug paxos:
20
- osd:
- debug ms:
1
- debug osd:
20
- flavor:
default
- log-ignorelist:
-
\(MDS_ALL_DOWN\)
-
\(MDS_UP_LESS_THAN_MAX\)
-
MON_DOWN
-
mons down
-
mon down
-
out of quorum
-
CEPHADM_STRAY_HOST
-
CEPHADM_STRAY_DAEMON
-
CEPHADM_FAILED_DAEMON
- log-only-match:
- sha1:
1ef864504b8875c83ee6c2c5fedc13315bebf7f5
- ceph-deploy:
- conf:
- client:
- log file:
/var/log/ceph/ceph-$name.$pid.log
- mon:
- install:
- ceph:
- flavor:
default
- sha1:
1ef864504b8875c83ee6c2c5fedc13315bebf7f5
- selinux:
- allowlist:
-
scontext=system_u:system_r:logrotate_t:s0
-
scontext=system_u:system_r:getty_t:s0
- workunit:
- branch:
wip-yuri8-testing-2024-08-28-1632-squid
- sha1:
1ef864504b8875c83ee6c2c5fedc13315bebf7f5
- owner:
scheduled_yuriw@teuthology
- pid:
- roles:
-
['host.a', 'mon.a', 'mgr.a', 'osd.0', 'osd.1']
-
['host.b', 'mon.b', 'mgr.b', 'osd.2', 'osd.3']
-
['host.c', 'mon.c', 'osd.4', 'osd.5']
- sentry_event:
- status:
pass
- success:
True
- branch:
wip-yuri8-testing-2024-08-28-1632-squid
- seed:
2672
- sha1:
1ef864504b8875c83ee6c2c5fedc13315bebf7f5
- subset:
111/120000
- suite:
rados
- suite_branch:
wip-yuri8-testing-2024-08-28-1632-squid
- suite_path:
/home/teuthworker/src/github.com_ceph_ceph-c_1ef864504b8875c83ee6c2c5fedc13315bebf7f5/qa
- suite_relpath:
qa
- suite_repo:
https://github.com/ceph/ceph-ci.git
- suite_sha1:
1ef864504b8875c83ee6c2c5fedc13315bebf7f5
- targets:
- smithi043.front.sepia.ceph.com:
ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBAnKQsVHasl99lWJQPvf8eyFudRoFHdv1NT9d4FbNpf4O6p6kgPpq/enOobwmyX7Ag0W7Y+SWpnvVOOOm9vEQZQ=
- smithi062.front.sepia.ceph.com:
ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBEHDYij7qDpKav/IIE3xMVqCdwgdQzbkh5TKL+IpIgiSj7En2MYLjV8KaNJktyUXEK6AE7Q/bWohwTTUAzGvQqY=
- smithi099.front.sepia.ceph.com:
ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBDS3Cgv6sUB9S7Mvd69h39x9TAW/1yWDOjYHnhkmn+8ZEZ2IS0267Rct1TY6JzcwckFD9RH3CbDUgjEbInqow5E=
- tasks:
-
- internal.buildpackages_prep:
-
- 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:
-
- internal.base:
-
- internal.archive_upload:
-
- internal.archive:
-
- internal.coredump:
-
- internal.sudo:
-
- internal.syslog:
-
- internal.timer:
-
- pcp:
-
- selinux:
-
- ansible.cephlab:
-
- clock:
-
- pexec:
- all:
-
sudo dnf remove nvme-cli -y
-
sudo dnf install nvmetcli nvme-cli -y
-
- install:
-
- cephadm:
-
- cephadm.shell:
- host.a:
-
set -ex
HOSTNAMES=$(ceph orch host ls --format json | jq -r '.[] | .hostname')
for host in $HOSTNAMES; do
# find the hostname for "host.c" which will have no mgr
HAS_MGRS=$(ceph orch ps --hostname ${host} --format json | jq 'any(.daemon_type == "mgr")')
if [ "$HAS_MGRS" == "false" ]; then
HOST_C="${host}"
fi
done
# One last thing to worry about before draining the host
# is that the teuthology test tends to put the explicit
# hostnames in the placement for the mon service.
# We want to make sure we can drain without providing
# --force and there is a check for the host being removed
# being listed explicitly in the placements. Therefore,
# we should remove it from the mon placement.
ceph orch ls mon --export > mon.yaml
sed /"$HOST_C"/d mon.yaml > mon_adjusted.yaml
ceph orch apply -i mon_adjusted.yaml
# now drain that host
ceph orch host drain $HOST_C --zap-osd-devices
# wait for drain to complete
HOST_C_DAEMONS=$(ceph orch ps --hostname $HOST_C)
while [ "$HOST_C_DAEMONS" != "No daemons reported" ]; do
sleep 15
HOST_C_DAEMONS=$(ceph orch ps --hostname $HOST_C)
done
# we want to check the ability to remove the host from
# the CRUSH map, so we should first verify the host is in
# the CRUSH map.
ceph osd getcrushmap -o compiled-crushmap
crushtool -d compiled-crushmap -o crushmap.txt
CRUSH_MAP=$(cat crushmap.txt)
if ! grep -q "$HOST_C" <<< "$CRUSH_MAP"; then
printf "Expected to see $HOST_C in CRUSH map. Saw:\n\n$CRUSH_MAP"
exit 1
fi
# If the drain was successful, we should be able to remove the
# host without force with no issues. If there are still daemons
# we will get a response telling us to drain the host and a
# non-zero return code
ceph orch host rm $HOST_C --rm-crush-entry
# verify we've successfully removed the host from the CRUSH map
sleep 30
ceph osd getcrushmap -o compiled-crushmap
crushtool -d compiled-crushmap -o crushmap.txt
CRUSH_MAP=$(cat crushmap.txt)
if grep -q "$HOST_C" <<< "$CRUSH_MAP"; then
printf "Saw $HOST_C in CRUSH map after it should have been removed.\n\n$CRUSH_MAP"
exit 1
fi
teuthology_branch:
main
verbose:
True
pcp_grafana_url:
priority:
99
user:
yuriw
queue:
posted:
2024-08-29 14:20:10
started:
2024-08-29 18:42:13
updated:
2024-08-29 19:01:12
status_class:
success
runtime:
0:18:59
wait_time:
0:07:07