[OpenAFS] bos restart problem

Heiko Schulz hrs@mathematik.uni-stuttgart.de
Sat, 2 Mar 2002 14:36:49 +0100 (CET)


I have a problem if I want to restart the bosserver. 
If I shutdown the bosserver then some fileserver processes are still
running. If I ignore this, kill the bosserver and start him again,
I am getting lost: After some time of salvaging (a too long time I
mean) bos status says "file server is running" and every thing looks
fine - but I cannot access the partition table of the server. I can
access the volumes by using "vos exa", but "vos parti" or "vos listvol"
does not work, this gives "communication failure" and so on.

The workaround for me at the moment is to kill all fileserver processes
after killing the bosserver. Then, starting the bosserver, the things are
working well (after a short time of salvaging).

This problem occurs in both cases: In the case of the automatically
restart and also if I shutdown and kill the bosserver by my self.

I am using kernel 2.4.17 and openafs 1.2.3.
I have this problem since I switch the afs database from afs 3.4a/3.6
to openafs. It is possible that there are some bad volumes (But
at this moment, everything looks well) or errors in the database?

That can I do? It is a good way to disable the afs (general) restart?
I cannot believe that.

The problem looks similar to Michael Aldrichs problem (moving afs server 
processes) - is this solved? In one of the last contributions
to that problem I found

> Ok, I think I'm a bit more awake, now...  This looks like something
> already has port 7007 bound.

Perhaps (at least in my case) these are the old fileserver processes?

Who can help me?

Heiko Schulz

Dr. Heiko Schulz
Mathematisches Institut A
Pfaffenwaldring 57
Universitšt Stuttgart

Tel.   : (0711) 685 5344
Fax    : (0711) 685 5599
E-Mail : schulz@mathematik.uni-stuttgart.de