On Fri, Aug 25, 2000 at 10:46:31PM -0400, Stephan Koledin wrote:
> Was there any successful resolution to this problem? I'm seeing what appears
> to be the same symptoms on my second server (Redhat6.2, Coda-5.3.8 built
> with krb5 support).
> 
> - kauth2/auth2 refuses to run, giving the "base PDB record does not exist."
> error
> - codasrv dies immediately after loading entire data space into memory
> - updateclnt appears to be running fine, but VLDB and VRDB aren't being
> replicated
> - I am using kerberos for auth - is there anything special I need to do for
> a replicated server?
> 
> TIA for any help resolving this, it's driving me crazy...

The following patch to /etc/rc.d/init.d/update.init should fix it.

The updatesrv on the SCM is serving files from the wrong directory. By
not specifying a path it uses the default path which should be fine.

Jan

--8<-----------------------------------------------------------------
RCS file:
/afs/cs/project/coda-src/cvs/coda/coda-src/scripts/update.init,v
retrieving revision 4.6
retrieving revision 4.8
diff -u -u -r4.6 -r4.8
--- update.init 2000/06/24 03:30:20     4.6
+++ update.init 2000/08/24 09:32:11     4.8
@@ -57,11 +57,11 @@
 
        if [ "`cat /vice/hostname`" = "`cat /vice/db/scm`" ]; then
                echon "updatesrv "
-               updatesrv -p $DBDIR &
+               updatesrv &
        fi
 
        echon "updateclnt "