@ross "Because each write of a raidz is striped across the disks the effective IOPS of the vdev is equal to that of a single disk. This can be improved by utilizing multiple (smaller) raidz vdevs which are striped, but not by mirroring them."
So with random reads, would it perform better on a raid5 layout since the FS blocks are written to each disk instead of a stripe? With zfs's implementation of raid10, would we still get data protection and checksumming? "How many luns are you working with now? 15? Is the storage direct attached or is it coming from a storage server that may have the physical disks in a raid configuration already? If direct attached, create a pool of mirrors. If it's coming from a storage server where the disks are in a raid already, just create a striped pool and set copies=2." We're not using a SAN but a Sun X4270 with sixteen SAS drives (two dedicated to OS, two for ssd, raid 11+1. There's a total of seven datasets from a single pool. -- This message posted from opensolaris.org _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
