I've finally got boot-from-SAN hosts running properly off my IBM DS4400 as of the RHEL5.1 beta (I can post my howto if anyone needs it), and I'm preparing to put a few of these hosts into production. The DS4400 is only capable of failover between paths, not true load balancing, but what I've successfully managed to do is: - Create two RAID-5 devices on my physical SAN - Assign each of those devices to a separate HBA - Use LVM to stripe (RAID 0) those individual SAN devices.
The result is an array with the reliability of RAID 5 and failover HBA's, but with two HBA's and SAN controllers doing the work. Overall, this seems to work well enough, but I'm wondering what any of the experts out there think. Is this safe? Is there a likely bottleneck I'm overlooking that negates the advantages? I'm wanting to put this on my primary fileserver which I only get to update once every so many years, so I'd like to get it right this from the start. Any suggestions would be fabulous. Thanks DC
_______________________________________________ rhelv5-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/rhelv5-list
