Ok, well I have been pouring over the archives for a little while now, and just haven't found what I was looking for so be prepared for some newbie ness. (Also went through the FAQ) Currently I have a goal to setup a nice little active-active fail-over cluster. We are going to be using this cluster for samba file service to a semi large group of users. I don't intend to use a GFS file system to actually access the same data at the same time (GFS is $$ correct?) so I was going to create separate partitions on a shared external scsi box (I am using a couple of Dell 2650's (with PERC 3/DC's Raid cards) and a PV 220 with number 73 gig drives) I was going to setup two separate raid sets, and just fail over a raid set in case of problems. But during my quest of finding out info, it seems like these cards can actually can share the same raid set at the same time (anybody else ever do this? am I off my rocker?). This leads me to think that LVM would be perfect for this scenario. That is if LVM can handle be accessed by the cluster. I was thinking this might be possible since I would be making separate LV's. Am right in thinking that this will save me a disk (one) from having do parity, and give me the option to move space around. Of course this could just ranting all around. =) Is this possible? Is it a horrible Idea? I realize there is going to be the possibility for the "split brain" problem, is there ways to guard against this? I was thinking of using some well timed external powerswitch boxes to help avoid this issue. Is it just not worth the risk? I am hoping some you people have had more experience with this might be able to send some wisdom! --Randy Schapel Educational Systems Mott Community College _______________________________________________ linux-lvm mailing list linux-lvm@sistina.com http://lists.sistina.com/mailman/listinfo/linux-lvm read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
