0

We are running a Oracle 11.2 DB on Redhat 6.8, and disk array is configured as RAID5, which is using a Lenovo 730i RAID adapter. I checked RAID policy and found that "Read Ahead Always" and "Always WriteBack" are set.

Now we are getting high read waiting time from Zabbix disk monitoring, but write waiting time is fine. r_await w_await

We are using a deadline I/O scheduler for the OS, and even set the read_expire parameter to 20, but didn't fix the problem. Is this a common behavior for RAID5 disk array?

Jason Yang
  • 21
  • 2
  • RAID5 should not be used in any production environment nowadays. With today's hard disks, it is a huge data-loss risk. You should start planning to move to RAID6 or some other more reliable disk system. Do not spend time on tweaking RAID5 performance. – Tero Kilkanen Mar 17 '21 at 08:13
  • @TeroKilkanen Thanks for your reply！So it is a common behavior, right? I know RAID5 won't bring good performance but didn't expect that bad... – Jason Yang Mar 18 '21 at 03:29
  • I don't know if that is common behavior. I am just highlighting that you have a huge risk of data loss with RAID5 and because of that you should migrate to some other data storage ASAP. – Tero Kilkanen Mar 18 '21 at 14:39

0 Answers0