Through DB2 we can get the average time in ms disk access is having DB2. These times are crucial for the detection of a IO problem with DB2 instance.

Usually we take into consideration that a value close to 2-3ms is good, more than 10ms can indicate problems.

Avg ms/write:

select trunc(decimal(sum(pool_write_time))/decimal(

(sum(pool_data_writes)+sum(pool_index_writes))),3)

from sysibmadm.snaptbsp

 

Avg ms/read:

select trunc(decimal(sum(pool_read_time))/decimal(

(sum(pool_data_p_reads)+sum(pool_index_p_reads))),3)

from sysibmadm.snaptbsp

To obtain correct values, you must have activated monitor switches, you can view the status with the sentence “get monitor switches”.