I run 3510FC and 2540 units in pairs. I build 2 5-disk RAID5 LUNs in each array, with 2 disks as global spares. Each array has dual controllers and I'm doing multipath.
Then from the server I have access to 2 LUNs from 2 arrays, and I build a ZFS RAID-10 set from these 4 LUNs being sure each mirror pair is constructed with LUNs from both arrays. Thus I can survive a complete failure of one array and multiple other failures and keep on trucking. Performance is quite good since using this in /etc/system: set zfs:zfs_nocacheflush = 1 And since recent ZFS patches for 10u4 which fixed FSYNC performance issues my arrays and servers are hardly breaking a sweat. I very much like that the arrays can handle lower-level problems for me like sparing and ZFS ensures correctness on top of that. This is for Cyrus mail-stores so availability and correctness are paramount in case you are wondering if all this belt & suspenders paranoia is worthwhile. If/when ZFS acquires a method to ensure that spare#1 in chassis#1 only gets used to replace failed disks in chassis#1 then I'll reconsider my position. Currently though there is no mechanism to ensure this so I could easily see a spare being pulled from the other chassis and leaving me with an undesirable dependency if I were doing ZFS with JBOD. This message posted from opensolaris.org _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
