[NCLUG] help finding documentation for unique cluster setup

Mike Jensen jentfoo at gmail.com
Tue Jul 18 22:54:23 MDT 2006


Thanks a lot guys....that is all very great information....I have not had a
enough time to fully digest everything yet (worked a 14 hours day, and am
exhuasted).  First let me reply to a few things said on here.  As for hands
on experinced I do plan to setup a high availability cluster, just for the
pratice before hand.  But since this is so heavily documented I did not feel
I should waist your time on it.  I know what I am trying to do is not very
real world.  Also when I said load balancing I ment cpu load, not hard drive
load.  I figured since one hard drive array would be through a gigabit port
it would be too hard to distibute that.  But I will talk about that in my
next bit.

But based on what you guys said (and things you said that allowed me to do
some additional searching) here is what I am thinking:

First scrap the load balancing, too advanced and lets do baby steps.

Second I think for ease of use I am going to be using centos with redhat's
GFS.  Although, like I said I have not quite had time to look into unionFS
yet.  So after some research that may seem to be a better option.

Here is a question though, and I am sure this is a cluster noob question.  I
was looking into the GFS documentation and was looking at how to make and
mount GFS drives.
( http://www.redhat.com/docs/manuals/csgfs/browse/rh-gfs-en/ch-manage.html )
It looks like it needs a dev location for the drive.  How will I make a dev
location for the drive that is on another server?  Will this be made durring
the clustering process?  Right now GFS does not look _that_ hard.

Thank you guys....I will keep you posted on how it all goes, and I am sure I
will have more questions :-)



More information about the NCLUG mailing list