[Ocfs2-users] ocfs2 hangs during webserver usage

[Stephan Hendl]

After 9 days without an error we got a further webserver hang last night. Our loadbalancer distributed the requests to 3 servers wcl[2,3,4]. wcl3 got after 9 in the evening more traffic than normal but not very high. At all the time there were no Busy locks - that' the good message. But once wcl3 got higher traffic the webserver processes reported "-D" status with en empty "WIDE-WCHAN-COLUMN" on that server. At the same time the webserver processes at the other two node reported in the "WIDE-WCHAN-COLUMN": ocfs2_wait_for_status_completion.

wcl3 (that one with the traffic):
============================
Thu May  3 21:28:01 CEST 2007
  PID STAT COMMAND          WIDE-WCHAN-COLUMN
  417 D    kjournald        wait_on_buffer
26451 D    httpd            -
26471 D    httpd            -
30209 D    httpd            -
29227 D    httpd            -
24211 D    httpd            wait_on_buffer
14600 D    httpd            -

============================
Thu May  3 21:30:01 CEST 2007
  PID STAT COMMAND          WIDE-WCHAN-COLUMN
 2211 D    kjournald        wait_on_buffer
26451 D    httpd            -
26471 D    httpd            -
30209 D    httpd            -
  525 D    httpd            -
29227 D    httpd            -
14600 D    httpd            -
 2024 Ds   sadc             log_wait_commit

============================
Thu May  3 21:32:01 CEST 2007
  PID STAT COMMAND          WIDE-WCHAN-COLUMN
26451 D    httpd            -
26455 D    httpd            -
26471 D    httpd            -
13240 D    httpd            -
28115 D    httpd            -
30209 D    httpd            -
  525 D    httpd            -
29227 D    httpd            -
14600 D    httpd            -


wcl2 as well as wcl4:
============================
Thu May  3 21:28:01 CEST 2007
  PID STAT COMMAND          WIDE-WCHAN-COLUMN
 4935 D    httpd            o2net_send_message_vec

============================
Thu May  3 21:30:01 CEST 2007
  PID STAT COMMAND          WIDE-WCHAN-COLUMN
18590 D    httpd            ocfs2_wait_for_status_completion
 4935 D    httpd            ocfs2_wait_for_status_completion
 9507 D    httpd            ocfs2_wait_for_status_completion
10643 D    httpd            ocfs2_wait_for_status_completion
10310 D    httpd            ocfs2_wait_for_status_completion
10334 D    httpd            ocfs2_wait_for_status_completion
 2434 D    httpd            ocfs2_wait_for_status_completion

============================
Thu May  3 21:32:01 CEST 2007
  PID STAT COMMAND          WIDE-WCHAN-COLUMN
18590 D    httpd            ocfs2_wait_for_status_completion
 4935 D    httpd            ocfs2_wait_for_status_completion
 9507 D    httpd            ocfs2_wait_for_status_completion
10643 D    httpd            ocfs2_wait_for_status_completion
10285 D    httpd            ocfs2_wait_for_status_completion
10302 D    httpd            ocfs2_wait_for_status_completion
10310 D    httpd            ocfs2_wait_for_status_completion
10334 D    httpd            ocfs2_wait_for_status_completion
 2434 D    httpd            ocfs2_wait_for_status_completion

/proc/slabinfo as well as /proc/meminfo is also eavailable. Shell I open a bugzilla case?

Regards
Stephan