From: Wim Godden <[EMAIL PROTECTED]> > I've got a problem getting some diskless clients to run. I want to > migrate some servers from standalone to having the same storage. I've > got a 160GByte RAID-system running Linux which would contain all data > for all the servers. To minimize the risk of servers going down due to > hard drive problems, I've decided to have them booting from the > RAID-system as well. > The problem is : how do I get this done ? I want to be able > to setup the > systems without too much hacking around, as I need to add servers > regularly. The servers all have different configurations (some of them > are webservers, some of them are mailservers, etc.) so I can't use the > same nfs root directory.
The lazy way I did it, when I wanted that for a while, was to use the same diskless boot infrastructure that runs the LTSP clients to boot the main servers. However, they get given a different NFS root that corresponds to a tree that you're maintaining over on the RAID server. In order to bring the whole thing back up, after a power failure, you need a DHCP and TFTP server that isn't diskless. The obvious thing is to make the RAID server provide those services ... only (not much traffic). > Is there a way to start from a standard distribution > (RedHat/Slackware/Suse) and build a diskless client (in fact a server) > based on it ? There was a package that allows a standard debian system to run diskless. I don't know much about it, but it apparently re-arranges bits of filesystem. ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _____________________________________________________________________ Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto: https://lists.sourceforge.net/lists/listinfo/ltsp-discuss For additional LTSP help, try #ltsp channel on irc.openprojects.net
