Poor performance Oracle 10g on vFiler

[ManojKaparwan]

Hi

We are facing currently massive performance issues on one of the Oracle DB running on Solaris 10 Sparc and NetApp vFiler.

Host has 6 databases

5 databases have - separate 5 IP for the unseryling storage volumes and connected to single vFiler.

last 6th one also has got separate IP but the disk ServiceTime is around 20  seconds ( yes seconds ! ) 

Host has got 02 network interfaces and IPMP is configured in ACTIVE-STANDBY mode

Database version is - 10.2.0.5

Questions

1) Is it recommended to use ACTIVE-ACTIVE configuration of IPMP for highly transactional databases which does massive IO? if configured are there any known side effects ?

2) Is there any know bug on Solaris 10 Sparc which might be causing issues here ?( however looks less likely as OS is doing well for other 5 databases on the same host/os)

regards

Manoj  

[asulliva]

Hello Manoj,

Can you provide any insight into your aggregate configuration and volume layout?  E.g. how many disks in the aggregate(s), what volumes are on which aggregates, are there other workloads sharing the same aggr/vol, protocol(s) used, etc.  Additionally, what type of workload is seeing the extremely high latency?  Is it log files, data files, backup, etc.?

Andrew