Subject: Re: ypserv
To: Todd Vierling <tv@pobox.com>
From: Jasper Wallace <jasper@staff.argonet.co.uk>
List: current-users
Date: 09/21/1998 14:59:32
On Thu, 17 Sep 1998, Todd Vierling wrote:

> On Thu, 17 Sep 1998, C Kane wrote:
> 
> : I notice that, even in -current, NetBSD seems to be somewhat
> : behind in keeping up with changes to the ypserv source code.
> : 
> : NetBSD has ypserv.c version 1.7, but there is a ypserv version 1.12
> : available.
> 
> Those version numbers don't necessarily correlate.  Where is the 1.12?  The
> version numbers increment when _we_ make a change to the file, not an
> outside source.
> 
> : And, yes, ypserv still isn't working well for me.
> 
> What's wrong?  I'm using it regularly and can address your concerns.

did you reach a solution? I'm having problems with yp servers on NetBSD with
some Solaris 2.6 machines as clients - it works sporadicly at best and very
slowly then...

we've got 3 ypservers: a master and two slaves, all NetBSD i386 with the
980723 snapshot, 2 client netbsd machines again i386 with 980723
snapshot, and 3 sun machines as clients, a sparc 5, and Ultra 5, and an
Ultra II all running Solaris 2.6 HW 3/98 with the latest patch cluster.

The netbsd clients know of all 3 ypservers, the sun machines know of 2.
After a short while both the ypservers that the sun machines know about stop
working, the third that they don't touch keeps working fine.

I've got the latest ypserv source (as of yesterday) and compiled it up with
-DDEBUG.

watching the output with snoop shows that the servers get requests, sit on
them for a while (~ 1 min), and then reply with a domain_nonack_2 - this is
backed up by the ypserv.log.

I can provide snoop, tcpdump and ypserv.log output etc. if anyone wants it.

Hmmm wierd. on a hunch i edited nsswitch.conf on all the suns so they only
used nis for passwd and groups and put a resolv.conf in place. This seems to
have improved things, but not by much...



on a perhaps related(?) note, i'm running rpc.statd and rpc.lockd on one of
the NetBSD machines (the fileserver), and getting a lot of these in
/var/log/messages:

rpc.statd: Unsolicited notification from host <hostname>

where <hostname> is one of the 3 Solaris machines (they all do it).

and on the Solaris machines:

statd[186]: statd: cannot talk to statd at fileserver, RPC: Success(0)



--
Jasper Wallace                                Argo Interactive Group PLC
+44 (0)1243 815815                                 Unix Systems support,
+44 (0)411  264753                       Integration and Administration.