- log_href:
http://qa-proxy.ceph.com/teuthology/gabrioux-2024-09-07_07:25:09-orch:cephadm-wip-guits-reef-2024-09-06-1902-distro-default-smithi/7893493/teuthology.log
- archive_path:
/home/teuthworker/archive/gabrioux-2024-09-07_07:25:09-orch:cephadm-wip-guits-reef-2024-09-06-1902-distro-default-smithi/7893493
- description:
orch:cephadm/workunits/{0-distro/centos_9.stream agent/off mon_election/classic task/test_host_drain}
- duration:
0:13:40
- email:
gabrioux@ibm.com
- failure_reason:
- flavor:
default
- job_id:
7893493
- kernel:
- last_in_suite:
False
- machine_type:
smithi
- name:
gabrioux-2024-09-07_07:25:09-orch:cephadm-wip-guits-reef-2024-09-06-1902-distro-default-smithi
- nuke_on_error:
- os_type:
centos
- os_version:
9.stream
- overrides:
- admin_socket:
- branch:
wip-guits-reef-2024-09-06-1902
- ceph:
- conf:
- global:
- mon election default strategy:
1
- mgr:
- debug mgr:
20
- debug ms:
1
- mgr/cephadm/use_agent:
False
- 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:
03cef208dc8a39e0b8a1806b4763767a2ff62439
- ceph-deploy:
- conf:
- client:
- log file:
/var/log/ceph/ceph-$name.$pid.log
- mon:
- install:
- ceph:
- flavor:
default
- sha1:
03cef208dc8a39e0b8a1806b4763767a2ff62439
- selinux:
- whitelist:
-
scontext=system_u:system_r:logrotate_t:s0
- workunit:
- branch:
wip-guits-reef-2024-09-06-1902
- sha1:
f368955f75d280ab61cfdc8a737fe7204b5a0c20
- owner:
scheduled_gabrioux@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-guits-reef-2024-09-06-1902
- seed:
7999
- sha1:
03cef208dc8a39e0b8a1806b4763767a2ff62439
- subset:
1/8
- suite:
orch:cephadm
- suite_branch:
wip-guits-reef-2024-09-06-1902
- suite_path:
/home/teuthworker/src/github.com_ceph_ceph-c_f368955f75d280ab61cfdc8a737fe7204b5a0c20/qa
- suite_relpath:
qa
- suite_repo:
https://github.com/ceph/ceph-ci
- suite_sha1:
f368955f75d280ab61cfdc8a737fe7204b5a0c20
- targets:
- smithi018.front.sepia.ceph.com:
ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBG9bp5jnTIoD1ShDgGtgSdmE3OkkbwabmX4Qv3//BoExL8FsIZ9H+T7M8GB2E3a7KmctGU3tkI3Xu+VzV9Aqc6I=
- smithi090.front.sepia.ceph.com:
ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBOIb/ru/F3Bhp4Z79BKHS7lHTmBcQTJAJeDxIZ4Q1tZeikFcE9Pt10AGkD0WYTdio9J/otUBd83YtOVK8QN4iBc=
- smithi191.front.sepia.ceph.com:
ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBPhKWCmzNdUs3JhxykW+PEcGj9K53Hupk7U6tt4DNlfTts1wl4o5TVMDIo7OI68fXx4uwPqgoO6cAeBWuGIFPGA=
- 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:
70
user:
gabrioux
queue:
posted:
2024-09-07 07:25:32
started:
2024-09-07 07:57:59
updated:
2024-09-07 08:21:18
status_class:
success
runtime:
0:23:19
wait_time:
0:09:39