Actions
Bug #19434
closedceph task's finally block failed with EACCES
Status:
Can't reproduce
Priority:
Urgent
Assignee:
-
Category:
qa
Target version:
-
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
I think this is because of the switch to ceph-mgr; somehow by the time it unwinds everything, ceph-mgr can no longer be accessed for the ceph pg dump from osd_scrub_pgs():
Traceback (most recent call last):
File "/home/dmick/src/ceph/teuthology/teuthology/contextutil.py", line 32, in nested
yield vars
File "/home/dmick/src/ceph/ceph/qa/tasks/ceph.py", line 1581, in task
osd_scrub_pgs(ctx, config)
File "/home/dmick/src/ceph/ceph/qa/tasks/ceph.py", line 1032, in osd_scrub_pgs
stats = manager.get_pg_stats()
File "/home/dmick/src/ceph/ceph/qa/tasks/ceph_manager.py", line 1685, in get_pg_stats
out = self.raw_cluster_cmd('pg', 'dump', '--format=json')
File "/home/dmick/src/ceph/ceph/qa/tasks/ceph_manager.py", line 1095, in raw_cluster_cmd
stdout=StringIO(),
File "/home/dmick/src/ceph/teuthology/teuthology/orchestra/remote.py", line 193, in run
r = self._runner(client=self.ssh, name=self.shortname, **kwargs)
File "/home/dmick/src/ceph/teuthology/teuthology/orchestra/run.py", line 414, in run
r.wait()
File "/home/dmick/src/ceph/teuthology/teuthology/orchestra/run.py", line 149, in wait
self._raise_for_status()
File "/home/dmick/src/ceph/teuthology/teuthology/orchestra/run.py", line 171, in _raise_for_status
node=self.hostname, label=self.label
CommandFailedError: Command failed on mira068 with status 13: 'sudo adjust-ulimits ceph-coverage /home/ubuntu/cephtest/archive/coverage timeout 120 ceph --cluster ceph pg dump --format=json'
Updated by Dan Mick about 7 years ago
2017-03-31 02:40:18.968409 7f6fbe3ed700 4 mgr.server handle_command prefix=pg dump
2017-03-31 02:40:18.968420 7f6fbe3ed700 20 is_capable service=pg command=pg dump read on cap
2017-03-31 02:40:18.968422 7f6fbe3ed700 10 mgr.server _allowed_command client.admin not capable
2017-03-31 02:40:18.968424 7f6fbe3ed700 1 mgr.server handle_command handle_command access denied
2017-03-31 02:40:18.968439 7f6fbe3ed700 0 log_channel(audit) log [INF] : from='client.4129 172.21.6.112:0/941668890' entity='client.admin' cmd=[{"prefix": "pg dump", "target": ["mgr", ""], "format": "json"}]: access denied
2017-03-31 02:40:18.968444 7f6fbe3ed700 1 mgr.server handle_command do_command r=-13 access denied
# ceph auth get client.admin
exported keyring for client.admin
[client.admin]
key = AQASv91Y3tRRFhAAShrqacDKMJ9X/MkvZ5DtpA==
auid = 0
caps mds = "allow *"
caps mon = "allow *"
caps osd = "allow *"
Excerpt from /etc/ceph/ceph.keyring:
[client.admin]
key = AQASv91Y3tRRFhAAShrqacDKMJ9X/MkvZ5DtpA==
auid = 0
caps mds = "allow *"
caps mon = "allow *"
caps osd = "allow *"
Updated by Dan Mick about 7 years ago
This might be because I had not yet included 7edca203d8b62fdbe44107ebfff45ab854200c4f, but was running ceph binaries that included the mgr change.
Actions