|
Problem IT05851
|
Status: Closed |
ASYNCHRONOUS WRITE TIME IS SIGNIFICANTLY UNDERSTATED IN DB2 MONITORING |
| product: |
DB2 FOR LUW / DB2FORLUW / A10 - DB2 |
| Problem description: |
pool_async_write_time is currently reporting an estimate of how
long pagecleaners have waited for any number of I/Os to
complete. However, documentation also suggests it is the
acccumulation of time spent for each I/O to complete. A formula
is provided where it is suggested that average async write time
can be determined by dividing asynchronous write time by the
number of asynchronous writes, which is incorrect given the
current information being returned. Pagecleaners use
asynchronous I/O, which involves submitting I/Os in batches, and
it becomes more difficult to apportion time to specific I/Os (we
may wait for the longest one in a batch to be complete before
processing any of the completed I/Os). Regardless, a more
accurate measure of I/O performance would be provided if the
end-to-end times of each I/O were accumulated. |
| Problem Summary: |
****************************************************************
* USERS AFFECTED: *
* All DB2 systems are affected *
****************************************************************
* PROBLEM DESCRIPTION: *
* See Error Description *
****************************************************************
* RECOMMENDATION: *
* Upgrade to DB2 Version 10.1 Fix Pack 5 *
**************************************************************** |
| Local Fix: |
use operating system tools to monitor write performance |
| Solution |
Problem first fixed in DB2 Version 10.1 Fix Pack 5 |
| Workaround |
see Local Fix |
| Timestamps |
Date - problem reported :
Date - problem closed :
Date - last modified :
| 01.12.2014
23.07.2015
23.07.2015 |
| Problem solved at the following versions (IBM BugInfos) |
|
| Problem solved according to the fixlist(s) of the following version(s) |
| 10.1.0.5
|
 |
