Hello, I am setting up an app server and want to maximize SQL performance as much as possible without hurting transfer of files 20-100k too much. I am running mysql 3.22.20a on a raid5 array with three 4GB disks and a chunk size of 128k I am running sql-bench now, concentrating on select performace, and I am getting somewhat better performance than I did off of a single drive. My main question is: what chunk size should I use for this type of application? Also, I know the stride value used for mke2fs is important, I read the docs and used a stride of of 32... Here is my raid5.conf and my mke2fs cmdline raiddev /dev/md0 raid-level 5 nr-raid-disks 3 nr-spare-disks 0 chunk-size 128 device /dev/sdb1 raid-disk 0 device /dev/sdc1 raid-disk 1 device /dev/sdd1 raid-disk 2 mke2fs -b 4096 -m 1 -R stride=32 /dev/md0 obinfo: kernel 2.2.3, raidtools 0.90 19990309, patch raid0145-19990309-2.2.3 dual p2 450, symbios scsi card, 4GB UWSCSI drives. If anyone has any recommendations, I would appreciate hearing them. It is a pain to rebuild the raid array and rerun the sql-bench tests... Chris
