[OpenAFS] volrestore problems

[Brian Davidson]

On Mar 14, 2006, at 3:15 PM, Jim Rees wrote:

> Maybe you're hitting the limit on volume name length?  If so that's a 
> poor
> error message you're getting.

That gets me closer -- my count was off by 1 character.  Now I get a 
segfault...

Tape mount callout routine is /usr/local/var/openafs/backup/file.sh
Tape unmount callout routine is /usr/local/var/openafs/backup/fileu.sh
Will dump to a file
Operator queries are disabled
Dump tape name check is disabled
Starting Tape Coordinator: Port offset 0   Debug level 0
Token expires: Tue Mar 14 20:42:31 2006


Restore
Thanks, now proceeding with tape reading operation.
Restoring volume dept.asd.facultystaff.res Id 3266573891 on server 
129.174.0.247 partition /vicepb .. done
Segmentation fault


Here's a backtrace (let me know if there's something more useful I 
should provide):

(gdb) bt
#0  0x0806cbf0 in strcpy ()
#1  0x0814dfd0 in ?? ()
#2  0x080641ea in strcpy ()
#3  0x0804d9e7 in strcpy ()
#4  0x0804dcb8 in strcpy ()
#5  0x4002bc6f in pthread_start_thread (arg=0x411bbbe0) at manager.c:284
#6  0x4002bd5f in pthread_start_thread_event (arg=0x411bbbe0) at 
manager.c:308


It looks like it's making it far enough along in the process that the 
volume is getting restored, but not far enough for it to get linked 
into VLDB, or to be brought online.

$ vos listvol afsfs1.ssg.gmu.edu vicepb | grep asd
dept.asd.facultystaff.res        3266573891 RW       6633 K Off-line


The server in question is running RedHat Advanced Server 2.1.  I'm in 
the process of building up its replacement, which will be RedHat AS 4.  
I guess I'll expedite that server upgrade, in hopes that this is a 
threading issue.


I'm still open to additional suggestions,

Brian