I was installing Fedora Core 1 in a lab with about 10 computers. The
first machine we did a simple CD install, then created an NFS share
(/var/ftp/pub/fedora-core-1) and copied all the CD data to it. We then
made a kickstart boot disk and proceeded to do NFS installs on all the
other stations, pretty straight forward stuff. About ten minutes into
it, the NFS installs froze. We waited a while and noticed some *very*
slow activity on some of the machines (like one package in 15 minutes).
Figuring it could be a simple disk access issue on the server, we killed
all but two of the installs and waited... no improvement was seen. We
then attempted to restart the NFS server with (service nfs restart). It
stopped everything okay, but then hung when attempting to restart nfs.
Out of frustration we then rebooted that machine and when it came up the
remaining installs started to fly through packages like they had
initially... for about ten minutes. Same deal again, installs hung and
we couldn't restart NFS. A reboot cleared it. I checked out
/var/log/messages but didn't see anything. This time it finished, but
in the meantime we installed the rest of the lab with the CD. Sorry I
don't have more detailed info, we were in a rush to get the lab back up
for a class.
I'll try to repeat this tonight. My hunch is its an NFS issue and I
just happened to run accross it with these installs.
Also I should mention a few months ago we installed RH9 in this lab
using NFS and kickstart just like this and had no issues.
Has anyone else noticed issues like this? Anyone done a bunch of
concurent NFS installs and *NOT* noticed this issue?
----- begin /etc/exports -----
/var/ftp/pub *(ro)
----- end /etc/exports -----
- Johnathan
