We are considering using RAID for non-replicated volumes in AFS, home directories in particular, and are interested in hearing about any experience other sites might have in this area. The primary motivation for this is that recovery from backup in the event of a disk failure appears to be a rather expensive operation, both in terms of labor required to do it and the amount of downtime for the affected volumes. RAID promises the elimination of such costs (except in the event of multiple drive failures), albeit at roughly two or three times the cost per unit of storage plus significantly increased system complexity. Our current environment includes AFS servers on both RS/6000s (running AIX 3.2.5) and Suns (running 4.1.3). Right now we only have a few GB in AFS but ultimately we expect to have 100-200 GB, most of which will be non-replicated data. Tales based on real experience would be most appreciated, but any relavant comments are welcome. -- Karl Swartz |INet [EMAIL PROTECTED] SLAC Computing Services | or [EMAIL PROTECTED] 1-415/926-3630 |UUCP uunet!lll-winken!unixhub!kls -or- ditka!kls (SLAC and the US Dept. of Energy don't necessarily agree with my opinions.)
