Netboot Mailing List (by thread)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Sharing NFS file system (making it read only)

Dvir Oren wrote:
I want to share the same file system for all of my etherboot clients.
 I realize that some parts have to be different, and I intend to create
ramdisks for them, and mount them.
Tada! :) hehe you think like me.  That is exactly what i did.  / and /usr we mounted nfs RO, and /home was RW.  During rc.sysinit i copied the ro version of /var into a newly e2fsck'd  /dev/ram0 then mounted it over /var.  was very tiny.
However, if I make the NFS file system read only, I come across a
lot of problems:

1.  I need /dev to be rw.

I never ran into that problem, i created links out of dev into the rw /var, for progs that need fifos n stuff (like syslogd). and used a remote syslog to be the receptor for all system logs.
2.  I can't have /etc/passwd (and the related files) read only, if I
want to add users, etc.  I can't have them link to somewhere else,
passwd claims that it can't lock them.
Why? #1 all my passwd files were the same bare minimum and i used NIS for "real" users, but even if you needed to have real users in the passwd file they can all be created from one  central location (wherever "/etc" is available rw, like the fileserver)..
Has anybody succeeded in making the file system read only?  I'm
finding it extremely difficult.
Yep, me. i've used it for workstations and for MPI/PVM clusters.
Hopefully this isn't entirely off topic.

Dvir Oren

This Mail was sent to netboot mailing list by:
"Dvir Oren" <>
To get help about this list, send a mail with 'help' as the only string in
it's body to If you have problems with this list,
send a mail to

It is by Linux alone I set my machine in motion.
      It is by the open source alone that kernel acquires speed,
       the modules acquire patches.
      The patches become a new release.
It is by Linux alone I set my machine in motion.

For requests or suggestions regarding this mailing list archive please write to