1

I have to migrate my old Oracle 10g database system to a brand new 12c server. The vendor suggests to use two SSD disks in RAID1. Since I'm very happy with my current configuration (two SAS RAID10 arrays) I'm a little bit scared about this. Do you have suggestion/reccomendations/experiences to share?

closed as primarily opinion-based by Balazs Papp, JSapkota, mustaccio, Marco, RLF Jan 19 '17 at 20:37

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

1

There is an Oracle doc that you might want to look at, Doc ID 1681266.1. You want to make sure that you are using 4K block sizes for storage. If you are moving to a new server and you are running at least Redhat 6.5 or Windows 2012, than that should not be a problem. Using a 4K physical block size will give you better performance.

0

From my experience , for database with medium OLTP load storing datafiles on SSD is not paid out from cost/overall effectiveness perspective. Most of the reads are against data which is already in memory, number of disk reads is relatively small. Since it's possible to have multiple DB Writers , don't expect much improvement here as well .

On the other hand, I think putting redo logs on SSD will help systems with high write activity.

For sure, SSD will not make it worse, but you said you are happy with your current system, so why to fix if it's not broken ?

  • My system is more then 9 years old and it's time to migrate. – Cristian Veronesi Jan 20 '17 at 8:50
  • OLTP systems are sensitive to latency of the the redo log writes. The most important thing is to remove outliers in the writes to the redo logs. The most effective way to do this, is to have dedicated redo log devices. Most customers who have moved their redo logs to SSD have seen and improvement, not because SSD's are faster per se, but because they now have dedicated devices. Often they are coming from a shared SAN storage architecture. – BobC Feb 25 '17 at 4:25

Not the answer you're looking for? Browse other questions tagged or ask your own question.