Oracle Tablespace I/O High Avg Read(ms) in AWR

We're in the process of migrating to ExtremIO for our Oracle 12.1.0.2 upgrade (AIX 7.1 lpar); I am seeing extremely high Avg Read(ms) in the AWR reports for several of the tablespace/datafiles, for example:

Tablespace	Reads	Av Rds/s	Av Rd(ms)	Av Blks/Rd	1-bk Rds/s	Av 1-bk Rd(ms)	% Opt Reads	Writes	Writes avg/s	Buffer Waits	Av Buf Wt(ms)
PSINDEX	8,220	2	522.89	1.15	2	535.43	0	1,791	0	1	0
PSUNDOTS	3	0	13.33	1	0	13.33	0	2,679	1	0	0
GLLARGE	2,491	1	26,816.70	1	1	26,816.70	0	2	0	0	0
PSIMAGE2	1,859	1	2,310.48	2.88	0	8,073.29	0	187	0	0	0
APLARGE	1,318	0	0.3	15.91	0	10	0	2	0	0	0
PTTBL	478	0	56,259.14	1	0	56,259.14	0	801	0	1	0
SYSAUX	54	0	1.48	1	0	1.48	0	619	0	0	0

Not seeing this on our current VMAX system it's migrating from - any good advice appreciated, thanks.

Responses(1)

slucido

109 Posts

0

May 4th, 2017 06:00

Thanks for posting the question. A few observations:

load appears light but need the timeframe for more analysis
- Average of 2 reads a second on PSINDEX
- Average of 1 read per second on GLLARGE & PSIMAGE2

Ideas for more analysis:

We can review the AWR reports to get more of an accurate picture
- Please contact sam.lucido@emc.com if we can help. I'll work with you on having an Oracle expert provide more analysis.
Work with the XtremIO storage administrator for storage analysis including other hosts on the same ports
If virtual and using VMware vSphere then check with the VMware administrator

View All

No Events found!