Oracle in mirror'd or RAID'd storage arrays (SUN)

From: Nigel Corbin <nigelc_at_gatwick.geoquest.slb.com>
Date: 1996/12/23
Message-ID: <59mak0$rg4@snlsu1.london.sinet.slb.com>#1/1

Hi,

I have a client using 7.2.3 on Sun/Solaris. They use a Sparc Storage Array with a logical volume manager (Veritas, I think). The machine is a multi-cpu SparcServer, with lots of memory. It runs extremely slowly - and appears to be killed by any user running a simple insert/update. Query performance is fine. Any writing to the database saturates the i/o system completely, with the rollback tablespace file getting a hammering. I have no way of monitoring redo logs, but suspect that they might also be a problem.

Does anyone know what causes this sort of problem - I have seen notes saying not to put redo logs on a storage array (or mirrored/RAID'ed storage) - but thats all I've got.

Nigel Corbin

---
================================================================================
Nigel Corbin                            |       GeoQuest Data Management


                                        |       Schlumberger House

phone: 	+44(0)1293 556517               |       Buckingham Gate
fax  :	+44(0)1293 556599               |       Gatwick Airport
email: 	nigelc_at_gatwick.geoquest.slb.com |       RH6 0NZ


                                        |       England

================================================================================

Received on Mon Dec 23 1996 - 00:00:00 CST