D&C GLug - Home Page

[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]

Re: [LUG] irc

 

Henry Bremridge wrote:
>
> Using Debian Squeeze
> 
> I don't use automount, but have been thinking about installing it for
> some time. Question:
> 
> - At the moment if I mount a USB then I have to be root. Is there a way (what
>   do I need to do?) to be able to mount as myself? ie no automount but I
>   would need to enter the mount command in a terminal?
> 
> - How does automount work? (can anyone recommend a dummies-guide link?)

There is more than one automount system around, which leads to half the
confusion.

If you have a GNOME desktop install gnome-volume-manager, then check
"System" "preferences" "removal drives and media". Basically you plug in
a device and it mounts it, and creates an icon on your desktop which
opens the GNOME file manager (or other actions as appropriate - run
f-spot photo importing for a camera, or a CD playing application etc).

Gordon covered mounting - the "user" option in fstab does the trick you
wanted for USB - I'd jusy use the GNOME volume manager - much easier.


Old style automounting is covered briefly here:

 http://www.kryogenix.org/writings/tech/autofs

But the old style automounting was mostly aimed at providing things like
application servers to Unix clusters and workstations, as well as
mounting distributed home directories.

i.e.

You have application "proE" installed on servers albert, fred, and james
in "/opt/proE". When someone accesses /opt/proE on their workstation it
automatically mounts "/opt/proE" from the fastest responding server, and
when it has been inactive unmounts it. Thus you only see application
unavailable if the server you are using crashes whilst you are using it
(even then things may recover depending on timeouts and application
behaviour), but you don't need to install software on every machine, and
you can take albert out of service and upgrade proE and if you are
careful no one notices or only people using it at the time.

Alternatively you can use similar in big ISP email configs to mount
maildirs to front end mail servers. Have 2 front end mail servers,
deliver email to "simon@xxxxxxxxxxx" we have 10 backend servers
providing IMAP service to users, in 5 pairs sharing common data storage.
Thus you can lose any two servers and still deliver email. Of course if
you lose any of the 5 data stores.....

500 users with home directories on 50 machines, use automounter to keep
a list, and rather than 50 mounts from each machine to each other it can
mount /home/henry from whichever machine has it one when I type "cp
~henry/example.txt ~/".

You have to be in a complex but well organised network of Unix or Linux
systems to appreciate this. You also have to sit around pondering the
various NFS failure modes. You can mount directories with hard and soft
fail. Possibly NFS has moved on since I used these sorts of tricks, and
we have a lot more options for redundancy of storage and so on.

I'm guessing most home users will find the gnome-volume-manager
sufficent. Now ask about fusermount, and sshfs ;)

 Simon

-- 
The Mailing List for the Devon & Cornwall LUG
http://mailman.dclug.org.uk/listinfo/list
FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html