Project

General

Profile

Actions
Copy link

Bug #1504

closed

kclient workunit failed on kclient (can't set layout on new pool?)

Added by Josh Durgin over 12 years ago. Updated over 12 years ago.

Status:
Won't Fix
Priority:
Normal
Assignee:
Greg Farnum
Category:
-
Target version:
v0.35
% Done:

0%

Source:
Tags:
Backport:
Regression:
Severity:
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

Logs are in teuthology:~teuthworker/archive/nightly_coverage_2011-09-05/630

Partial output from the test:

2011-09-05T01:27:11.164 INFO:teuthology.task.workunit.client.0.err:+ mkdir layout_test
2011-09-05T01:27:11.165 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test show_layout
2011-09-05T01:27:11.165 INFO:teuthology.task.workunit.client.0.err:layout not specified
2011-09-05T01:27:11.166 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test set_layout -u 1048576 -c 1 -s 1048576
2011-09-05T01:27:11.167 INFO:teuthology.task.workunit.client.0.err:+ touch layout_test/file1
2011-09-05T01:27:11.189 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test/file1 show_layout
2011-09-05T01:27:11.191 INFO:teuthology.task.workunit.client.0.err:+ diff new_layout temp
2011-09-05T01:27:11.226 INFO:teuthology.task.workunit.client.0.err:+ echo hello, I'm a file
2011-09-05T01:27:11.305 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test/file1 show_layout
2011-09-05T01:27:11.362 INFO:teuthology.task.workunit.client.0.err:+ diff new_layout temp
2011-09-05T01:27:11.363 INFO:teuthology.task.workunit.client.0.err:+ touch layout_test/file2
2011-09-05T01:27:11.482 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test/file2 show_layout
2011-09-05T01:27:11.539 INFO:teuthology.task.workunit.client.0.err:+ diff new_layout temp
2011-09-05T01:27:11.541 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test/file2 set_layout -u 1048576 -c 2 -s 2097152
2011-09-05T01:27:11.770 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test/file2 show_layout
2011-09-05T01:27:11.821 INFO:teuthology.task.workunit.client.0.err:+ diff file2_layout temp
2011-09-05T01:27:11.822 INFO:teuthology.task.workunit.client.0.err:+ echo hello, I'm a file with a custom layout
2011-09-05T01:27:11.823 INFO:teuthology.task.workunit.client.0.err:+ touch layout_test/file3
2011-09-05T01:27:11.923 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test/file3 show_layout
2011-09-05T01:27:11.980 INFO:teuthology.task.workunit.client.0.err:+ diff new_layout temp
2011-09-05T01:27:11.981 INFO:teuthology.task.workunit.client.0.err:+ ceph osd pool create newpool
2011-09-05T01:27:12.042 INFO:teuthology.task.workunit.client.0.out:2011-09-05 01:26:52.713491 mon <- [osd,pool,create,newpool]
2011-09-05T01:27:12.641 INFO:teuthology.task.workunit.client.0.out:2011-09-05 01:26:53.313236 mon2 -> 'pool 'newpool' created' (0)
2011-09-05T01:27:12.647 INFO:teuthology.task.workunit.client.0.err:+ ceph mds add_data_pool 3
2011-09-05T01:27:12.912 INFO:teuthology.task.workunit.client.0.out:2011-09-05 01:26:53.584090 mon <- [mds,add_data_pool,3]
2011-09-05T01:27:13.339 INFO:teuthology.task.workunit.client.0.out:2011-09-05 01:26:54.010818 mon2 -> 'added data pool 3 to mdsmap' (0)
2011-09-05T01:27:13.344 INFO:teuthology.task.workunit.client.0.err:+ cephfs layout_test/file3 set_layout -p 3
2011-09-05T01:27:13.346 INFO:teuthology.task.workunit.client.0.err:Error setting layout: Invalid argument

Actions
Copy link
 #1

Updated by Greg Farnum over 12 years ago

Did you check that it was running on a new kernel? (And that, you know, teuthology worked?)

Actions
Copy link
 #2

Updated by Josh Durgin over 12 years ago

Yes, the newest kernel was used on all clients (so says the log and the machines themselves.)

Actions
Copy link
 #3

Updated by Greg Farnum over 12 years ago

  • Assignee set to Greg Farnum
Actions
Copy link
 #4

Updated by Greg Farnum over 12 years ago

  • Status changed from New to Won't Fix

I've run this a number of times and can't reproduce it. The only thing I can think of is that maybe the kernel's mdsmap is out of date when it tries to set the pool, so it fails.
Which I'm not sure is something we want to go to the trouble of finding a solution for, since we'd have to query the monitors on every bad pool ID input (ie, super-easy DOS). :/

Actions
Copy link
 #5

Updated by Sage Weil over 12 years ago

  • Target version changed from v0.36 to v0.35
  • Translation missing: en.field_position set to 1
  • Translation missing: en.field_position changed from 1 to 889
Actions
Copy link

Also available in: Atom PDF