From owner-freebsd-current  Thu Apr 10 04:03:52 1997
Return-Path: <owner-current>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id EAA29050
          for current-outgoing; Thu, 10 Apr 1997 04:03:52 -0700 (PDT)
Received: from nexis.net (customer-1.ican.net [198.133.36.101])
          by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id EAA29034
          for <freebsd-current@freebsd.org>; Thu, 10 Apr 1997 04:03:33 -0700 (PDT)
Received: from localhost (james@localhost)
	by nexis.net (8.8.5/8.8.5) with SMTP id HAA06126;
	Thu, 10 Apr 1997 07:01:33 -0400 (EDT)
Date: Thu, 10 Apr 1997 07:01:33 -0400 (EDT)
From: James FitzGibbon <james@nexis.net>
To: Erik Manders <erik@il.ft.hse.nl>
cc: freebsd-current@freebsd.org
Subject: Re: NFS + Quotas in current
In-Reply-To: <199704100935.LAA06963@charm.il.ft.hse.nl>
Message-ID: <Pine.BSF.3.95q.970410065322.6080A-100000@nexis.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

On Thu, 10 Apr 1997, Erik Manders wrote:

> > This is with a 2.2.x client ?  When I try to mount with quotas :
> > 
> > [nebula:james] ~ (66) # mount -t nfs -o rw,noatime,userquota,-b 
> > nexis.net:/usr/home /mnt
> > 
> > I get this:
> > 
> > nfs: -o userquota: option not supported
> > 
> This is true, sort of. Mount_nfs doesn't support the userquota option.
> 
> I'm running a slightly patched 2.2.1-RELEASE and my setup is as follows:
> 
> Server /etc/exports:
> /var/mail         -maproot=root:wheel nfscharm nfstop
> 
> Server /etc/fstab:
> /dev/sd1a         /var/mail    ufs  rw,userquota=/usr/quotas/mail.user 1 2 
> 
> Client /etc/fstab:
> nfs:/var/mail     /var/mail    nfs  rw,intr                            0 0
> 
>   Both machines have quota code enabled in the kernel, but only the
> server has check_quotas=YES in /etc/sysconfig. In addition, the server
> runs rpc.rquotad. I added it to /etc/rc.local, but you can have inetd
> run it.
> 
>   You basically let the fileserver enforce quotas and tell the clients
> when limits have been reached. rpc.rquotad runs to let quota(1) work.

I've tried doing this, but the server never tells the clients when the
quotas has been reached.  I can run quota(1) no problems, but if I have a
5mb quota on a directory and I try to write more than 5mb from the client,
the client *thinks* everything went just fine.   Meanwhile, the server
just stopped writing data at 5mb.

Can you run a really quick test on your setup to verify if I've done
something wrong ?  Set up a small (<10mb quota) on the server.  On the
client, run a little perl program like this :

--START--
#!/usr/bin/perl
#

for( $i=0; ; $i++ ) {

	print "This is test line $i\n";

}
---END---

>From the client, run that redirected to a file on the quota'd filesystem. 
At the same time, run a 'tail -f' on the same file from the server. When I
did this, I could watch the quota go up from either machine.  The tail
process showed new lines being added (the increasing number obviously
being the indicator.  However, when the quota limit was reached, the
client never knew.  It just kept on writing happy as can be. The tail
process just stopped.  No new lines were being committed to disk.

If your results differ in this respect, I'd really appreciate knowing.  At
least then I know it's possible, and that I can with effort get it to
work.  If not, I have to re-think what I'm trying to do.

> Something on this subject should be put into the FAQ or the handbook.

Agreed.

Thanks again for your help.

--
j.