More info on lockup problem with arla 0.21

Neulinger, Nathan R. nneul at umr.edu
Thu Feb 4 21:02:24 CET 1999


It locked up on me again. I was however able to turn on debugging. It looks
like the machine is OK, but all AFS access is hanging. When I enabled arlad
debugging with almost-all, I am getting a continuous stream (flooding
syslogs) of:

Feb  4 13:56:33 sysmon arla[357]: multi-sending wakeup: seq = 765152, error
= 0
Feb  4 13:56:33 sysmon arla[357]: worker 0: done
Feb  4 13:56:33 sysmon arla[357]: worker 0 waiting
Feb  4 13:56:33 sysmon arla[357]: worker 0: processing
Feb  4 13:56:33 sysmon arla[357]: Rec message: opcode = 4 (getnode), size =
292
Feb  4 13:56:33 sysmon arla[357]: Multi-send: opcode = 5 (installnode), size
= 384
Feb  4 13:56:33 sysmon arla[357]: multi-sending wakeup: seq = 765153, error
= 0
Feb  4 13:56:33 sysmon arla[357]: worker 0: done
Feb  4 13:56:33 sysmon arla[357]: worker 0 waiting
Feb  4 13:56:33 sysmon arla[357]: worker 0: processing
Feb  4 13:56:33 sysmon arla[357]: Rec message: opcode = 4 (getnode), size =
292
Feb  4 13:56:33 sysmon arla[357]: Multi-send: opcode = 5 (installnode), size
= 384
Feb  4 13:56:33 sysmon arla[357]: multi-sending wakeup: seq = 765154, error
= 0

There doesn't appear to be any XFS activity going on. The only thing that
got written to the log was:

Feb  4 13:58:18 sysmon kernel: xfs_syscall returns error: 0
Feb  4 13:58:18 sysmon kernel: xfs_pioctl
Feb  4 13:58:18 sysmon kernel: xfs_fh_to_dentry: dev: 0 inode: 25
Feb  4 13:58:18 sysmon kernel: xfs_syscall returns error: 7
Feb  4 13:58:18 sysmon kernel: xfs_message_receive opcode = 5
Feb  4 13:58:18 sysmon kernel: xfs_message_installnode
Feb  4 13:58:18 sysmon kernel: xfs_node_find
Feb  4 13:58:18 sysmon kernel: xfs_message_installnode: dp: ceabc7c4
Feb  4 13:58:18 sysmon kernel: xfs_message_installnode: fetching new node
Feb  4 13:58:18 sysmon kernel: new_xfs_node 0.536910899.799.2681
Feb  4 13:58:18 sysmon kernel: xfs_node_find

Killing arlad leads to an extremely unhappy kernel.

-------

As a side note, other than the cache parameters (which I have at 150M and
140M), is there much reason to go with anything other than the defaults in
arla.conf?

-- Nathan

------------------------------------------------------------
Nathan Neulinger                       EMail:  nneul at umr.edu
University of Missouri - Rolla         Phone: (573) 341-4841
Computing Services                       Fax: (573) 341-4216 





More information about the Arla-drinkers mailing list