autofs errors AS3 server Sun client

[Jeffrey A. St. Pierre]

On Wed, 12 May 2004, Jeffrey A. St. Pierre wrote:

> I run a fairly large server for a group of researchers at the
> University of Colorado.  I have served most of my NFS Shares from
> a Sun E450 running Solaris 8.  Recently we setup a
> Penguincomputing Relion240 running RH AS3.  Attached to it is an
> adaptec 29160 SCSI card attached to a Promise RM8000 RAID (1.5
> TB).  This RAID is going to store all of our data to be nfs
> shared out, but I have to move stuff piecemeal and I will have a
> few nfs servers until I can centralize all data on the new
> system.
> 
> I just moved the first major project over to the RAID and shared
> it out.  The user of the project has to run Sun-compiled code, so
> the data automounts on the Sun and the job runs, slowly, probably
> due to NFS settings that I can tweak, but it runs.  The thing
> that worries me is the errors that are filling up my
> /var/log/messages on the Redhat Server:
> 
> May 12 09:04:03 <servername> rpc.rquotad: Can't find filesystem 
>                 mountpoint for directory /export/rd01/stack
> May 12 09:04:03 <servername> rpc.rquotad: No correct mountpoint
>                 specified.
> May 12 09:04:25 <servername> rpc.rquotad: Can't find filesystem
>                 mountpoint for directory /export/rd01/stack
> May 12 09:04:25 <servername> rpc.rquotad: No correct mountpoint
>                 specified.
> May 12 09:04:47 <servername> rpc.rquotad: Can't find filesystem
>                 mountpoint for directory /export/rd01/stack
> May 12 09:04:47 <servername> rpc.rquotad: No correct mountpoint
>                 specified.
> May 12 09:23:46 <servername> rpc.rquotad: Can't find filesystem
>                 mountpoint for directory /export/rd01/stack
> May 12 09:23:46 <servername> rpc.rquotad: No correct mountpoint
>                 specified.
> 
> where <servername> is my Redhat AS Relion server
> 
> As you may have guessed "stack" is the name of the project.  It
> automounts under /project/stack as per /etc/auto_master and
> /etc/auto_proj.  These errors crop up continuously.  When he
> actually runs code on the Sun client these messages spool every
> 15 seconds or so on my linux server.  Needless to say, by the
> time I transfer 100 or so projects over to the RAID, my messages
> file is going to overflow /var/log.
> 
> These are my settings on the linux server:
> 
>    rpms:
>    kernel-smp-2.4.21-9.0.3.EL
>    kernel-utils-2.4-8.37.1
>    nfs-utils-1.0.6-7.EL
>    autofs-3.1.7-41
> 
>    /etc/exports:
>    /export/rd01 aaa.bbb.ccc.0/255.255.255.0(rw,async)
>    ## Note: sync tried, results the same.
> 
>    where aaa.bbb.ccc.0 is my class C subnet 
> 
> and my settings on the Sun client:
> 
>    kernel patch level:
>    5.8 Generic_108528-27
> 
>    /etc/auto_master:
>    
>    # Master map for automounter
>    #
>    +auto_master
>    #/net          -hosts          -nosuid 
>    /autohome      /etc/auto_home  -rw,bg,intr,soft,actimeo=0
>    /homes         /etc/auto_users -rw,bg,intr,soft,actimeo=0
>    /projects      /etc/auto_proj  -rw,bg,intr,soft,actimeo=0
>    /misc          /etc/auto_misc  -rw,bg,intr,soft,actimeo=0
>    /-             /etc/auto_direct
> 
>    
>    /etc/auto_proj (Abbreviated entry):
> 
>    # Home directory map for automounter
>    # RCS: {$Id: auto_proj,v 1.161 2004/05/11 20:10:46 root Exp $}
>    #
>    +auto_proj
>    stack    -rw,rsize=8192,wsize=8192 \
>             <servername>.Colorado.EDU:/export/rd01/&
> 
> 
> So... does anyone see possible errors in my setup that would
> cause these bizzare error messages?  This is very critical,
> because I have to get things moved to the new server, but I can't
> have a 50GB messages file in 1 day, and also it forbodes worse
> things to come.  I'm assuming I just have something setup wrong.
> 
> Thanks,
> 
> Jeff
> 
> 

This is a self-reply.  I have gotten a little closer to solving
this issue, but I am still lacking some info.

OK... what I have discovered is that this is not a Sun/Linux
issue.  I can generate the errors from a linux box as well.
Essentially If I execute '/usr/bin/quota -v' on a client system
then the redhat nfs server spawns this error for each remotely
mounted directory:

 rpc.rquotad: Can't find filesystem mountpoint for directory
 /export/rd01/stack
 rpc.rquotad: No correct mountpoint specified.

Is this being caused because the client does not have an
/export/rd01/stack directory ( it automounts this directory
under /projects/stack)?  If so why does the error appear in the
server logs an not on the client?

As I have about 100 projects that I need to move to this server,
that's going to be a lot of error messages.  I guess I could just
turn rquotad off, but it is started up in the nfs init script,
which would mean commenting it out of that script rather than
just using chkconfig to turn it off.  I think the other option
may be just to deny rquota connections in /etc/hosts.deny.  Does
anyone agree with this or have another idea to figure out why
this is giving this error message?

Thanks,

Jeff