Consensus was that there is no "standard" for where to create mountpoints.
This response from Alan contained a bunch of good info that seems worth
passing on:
In the right circumstances, any file system will fill up.
Some are just more sensitive to it than others. One affect
to consider when picking a location is the affect it will
have on file name lookups when a server is down. Directly
off the root, the filename lookup will probably block waiting
for that particular server to respond. Out of the typical
directory search, and it won't get touched. For example,
with a bunch of NFS mounts under /mnt:
/mnt/this-host
/mnt/that-host
/mnt/some-other-host
/mnt will be searched just fine, even if one of more of
the hosts are down. But, move each of those up a level
and the search blocks at each one that is down. Now,
an ls(1) or pwd(1) that has to go through traverse /mnt
will be in just as much trouble, that's much less common
than having to traverse /.
My advice is, use a mount point that is most natural for
the file system. Archive or large data spaces, I typically
hang under /mnt. User file systems, I put on /usr/users.
It just depends on the data and the "name".
Thanks to the many of you who replied.
Bill Skulley
> -----Original Message-----
> From: Skulley, William
> Sent: Monday, December 17, 2001 3:39 PM
> To: Tru64 Mailing List (E-mail)
> Subject: Standard NFS mount point location?
>
> We have a set of old 4.0d servers we are migrating (slowly) to 5.1. A
> previous administrator created NFS mount points on the root directory. I
> have since seen issues that prove to me that this is not a good plan (root
> filling up spooks me). We have the standard 5.1 off-the-disk install with
> /usr and /var on the same partition and the remainder of the root disk
> taken up with root and swap. I was considering somewhere in /usr but
> wasn't sure if that was really any better a solution. Is there a standard
> place to create NFS mount points?
>
> Thanks
> Bill Skulley
>
>
Received on Mon Dec 31 2001 - 14:25:19 NZDT