Avg read time exceeded on data volume

[Former Member]

Dear experts,

We are currently running MaxDB 7.7.04.29 on Red Hat Linux 5.1 for our BW instance.

The data and log volumes and SAP binaries are all utilizing NetApp storage, configured on a

single LUN within a single FlexVol running on a single filer. I realize this is not the desired

configuration as both SAP and NetApp recommend separating data, log, and binaries on separate

LUNs. We intend to reconfigure per said recommendation.

That said, do you think this single LUN configuration to be the source of our I/O performance problems?

If not, what other areas should I investigate?

We are getting this error from DBAN.prt, results of the db analyzer:

• W2 Avg read time 32.49 ms for 1378 reads on volume /sapdb/BWP/sapdata/DISKD0001

--Erick

[lbreddemann]

> I realize this is not the desired

> configuration as both SAP and NetApp recommend separating data, log, and binaries on separate

> LUNs. We intend to reconfigure per said recommendation.

>

> That said, do you think this single LUN configuration to be the source of our I/O performance problems?

> If not, what other areas should I investigate?

> We are getting this error from DBAN.prt, results of the db analyzer:

>

> * W2 Avg read time 32.49 ms for 1378 reads on volume /sapdb/BWP/sapdata/DISKD0001

Hi there!

From DB point of view, there is no way to tell why the IO performance is bad.

The best the database can do is to tell that it is bad (or good).

Anyhow, the setup you described had been one of the 'usual suspects' for IO issues - so, yes it's quite possible that your IO issues are caused by this.

I'd propose to implement the planned change and check the IO performance again afterwards.

regards,

Lars

[Former Member]

Lars,

Thank you for your feedback.

We'll give it a shot and see if it resolves our problem.

--Erick