[OpenAFS] getting there... any ideas on how to fix

Jerry Normandin Jerry.Normandin@dafca.com
Sun, 9 Dec 2007 14:18:24 -0800


root@eng02 logs]# tail FileLog
Sun Dec  9 16:39:50 2007 File server starting
Sun Dec  9 16:39:50 2007 afs_krb_get_lrealm failed, using dafca.com.
Sun Dec  9 16:39:50 2007 VL_RegisterAddrs rpc failed; The IP address
exists on a different server; repair it

I whish the logs were a little more verbose.  So  far it appears that my
AFS lack performance (I inherited the system.. I am attempting to fix
it) has been due to configuration.  I removed bogus host addresses from
the vldb, I tracked down the correct one and replaced it with that.
Also CellServDB had bogus entries, only the
vldb server ip addresses should be there.. all the addresses of the
servers were there (again I inherited this)
I removed the vlserver instance on eng02 since I created a new one on
eng03 and it works fine.  I deleted=20
the vldb database file on eng02.  I restarted and still get the same
error.
Is there docs somewhere that mention how to fis the?  the  fileserver
will not come up (this is a RO node anyway) and goes into a salvager
loop.  slavager runs to completion, the fileserver starts and dies due
to the above error, bos restarts slavager and this goes on and on. =20

Any ideas? THis would be appreciated.

VLDB access is now lighting fast... no more timouts.
I can hardley wait to run Bonnie again!