Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 3 Apr 2000 14:07:54 +0200
From:      Brad Knowles <blk@skynet.be>
To:        Jaap Akkerhuis <jaap@nic.nl>
Cc:        Stephen Montgomery-Smith <stephen@math.missouri.edu>, Coleman Kane <cokane@one.net>, Jeff Fisher <jeff@jeffenstein.org>, freebsd-stable@FreeBSD.ORG
Subject:   Re: What is needed in /stand
Message-ID:  <v0422080cb50e38666c2b@[195.238.1.121]>
In-Reply-To: <200004031123.NAA71586@114046.kema.nl>
References:  <200004031123.NAA71586@114046.kema.nl>

next in thread | previous in thread | raw e-mail | index | archive | help
At 1:23 PM +0200 2000/4/3, Jaap Akkerhuis wrote:

>  The standard solution is to have a small /tmp on root, use that
>  during the booting phase, and at the end of that, nfs mount /usr/tmp
>  on top of that. Of course, one shouldn't leave any file open on
>  the original /tmp.

	This would be a problem for any filesystem mounted on top of 
/tmp, but since /usr/tmp is not supposed to be cleaned on reboot and 
/tmp is, and /usr/tmp is intended to be a place where relatively 
large temporary files can be stored (while /tmp should be reserved 
for only smaller temporary files), I believe that it is best to keep 
them completely separate and distinct from each other.

	Not to mention the fact that I might want to enable async, 
softupdates, or mfs mounts on /tmp, while I might not want to have to 
do the same for all of /usr.


	Besides, file locking becomes impossible in -STABLE once you've 
mounted it with NFS (we don't have a working lockd, although work in 
this area is progressing in -CURRENT), and NFS writes generally suck 
when compared to local writes.

	If you absolutely, positively, *MUST* unify your /usr/tmp and 
/tmp directories or else you will go stark raving mad and murder 
everyone on the planet, then IMO a symlink is a better idea than 
doing an NFS mount.

	But better still is to make no attempt to unify them, and instead 
have a separate and dedicated filesystem that gets mounted on /tmp, 
and you can use softupdates, async mounts, or mfs, or nothing extra 
at all -- all depending on your particular needs and desires.

>                      There are other ways to accomplish similar
>  effects, for instance, mount_union.

	Which still has sufficient problems to cause it to be 
disrecommended even in -CURRENT (check the comments in your LINT 
kernel).

	I believe work is going on in this area and may have progressed 
to the point where it won't cause kernel panics the moment you try to 
enable it and reboot, but I don't know how much further things may 
have gotten than that.

--
   These are my opinions -- not to be taken as official Skynet policy
======================================================================
Brad Knowles, <blk@skynet.be>                || Belgacom Skynet SA/NV
Systems Architect, Mail/News/FTP/Proxy Admin || Rue Colonel Bourg, 124
Phone/Fax: +32-2-706.13.11/12.49             || B-1140 Brussels
http://www.skynet.be                         || Belgium


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?v0422080cb50e38666c2b>