Gents has anyone done a deployment that involves some sort of failover mechanism. We are running a bunch of clients now and are worried about a hardware failure situation. We do not really run a ltsp installation but run under the same type of architecture. We can buy really cheap machines 100 buck a piece from a local used computer supplier complete with drive. We do a kick start install on the machine which installs linux and x windows with telnet and ftp and finally a mosix kernel. The kick start program modifies the inittab to point the x to query gdm on our main terminal application server (dual 1GHZ).
These terminals totally run our manufacturing operations so we need a fail-over solution. I am proposing running three 1u rack machines with dns round robin in the front and rsync the application directories from a 1u designated master. In a failure condition the clients connected would fail and x would respawn trying to connect until it got a operating node. Has anyone tried such a setup? Does it sound like a reasonable solution? Clifford Baeseman Greenheck Fan Corporation Linux Operations _______________________________________________________________ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _____________________________________________________________________ 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
