PMS Information Systems
Welcome to PMS Information Systems - IBM i (AS/400) Forum !!!

Get Answers for all your queries on IBM i (AS/400).

Watch Disk-Related performance problems using DISK WATCHER

View previous topic View next topic Go down

Watch Disk-Related performance problems using DISK WATCHER

Post  maran on Sat Dec 26, 2015 3:06 pm

Watch Disk-Related performance  problems using DISK WATCHER


IBM Disk Watcher helps identify disk-related performance problems on our IBM i


Why Disk Watcher?
Disk Watcher function provides disk data to help identify the source of disk-related performance problems on our  IBM i* platform. This tool gathers detailed information associated with I/O operations to disk units.


What does it provide?


Disk Watcher provides data beyond that available in tools such as Work with Disk Status (WRKDSKSTS), Work with System Status (WRKSYSSTS) and Work with System Activity (WKSYSACT).
Disk Watcher, like other tools, provides data on disk I/O, paging rates, CPU use and temporary storage use. But Disk Watcher goes beyond this by simultaneously collecting program, object and job/thread/task information associated with disk I/O operations.

How to invoke?

The command STRDW is used to invoke the tool Disk watcher.

What does it collect?

Disk Watcher collects data at two detail levels - statistics and trace. A Disk Watcher collection can include one or both of these collection types, specified on the type (TYPE) parameter.

Where does it write the collected data?
The collected data is written to a set of database files that begin with QAPYDW.

What parameters do affect the performance of Disk Watcher?

We will face the special performance issues when collecting program and object resolution data in files QAPYDWPGMR and QAPYDWOBJR.
We can specify *ALL or *RESIDENT on the Object Information (OBJINF) parameter of the command.


OBJINF(*RESIDENT)
When *RESIDENT is selected, Disk Watcher will collect all program and object information currently in memory (a.k.a. resident storage). The performance expense associated with this option is relatively small.

OBJINF(*ALL)
When *ALL is selected, the performance impact of the collection will be significant. When Disk Watcher attempts to collect program and object information that isn't in memory, it'll cause page faulting. This will slow the collection performance and can affect the system as a whole. It's important to remember that the performance impact depends heavily on the amount of data being collected as well as the system's activity level.

Collection Interval
A dynamic interval option (*DYN) is available that will adjust the interval during the collection. The interval will be lengthened if the buffers aren't filling up between samples and shortened if any I/O data is missed. When using the *DYN option, the interval will never exceed two minutes.

What factors limits the data collection?
Limiting your collection should improve the tool's performance and reduce the amount of data collected for analysis. STRDW allow us to limit our collection in several ways.

1.ASP number (ASP) or ASP device name (ASPDEV
2.Disk unit number (DSKUNIT)
3.Storage pool identifier (POOLID).
4.I/O type (IOTYPE) - Collect data for a specific type of I/O operation (i.e. read operations, write operations or all other types of operations). This option applies only to trace data.
5.Collection condition (COLCOND) - Collect only I/O operations whose duration meets a specific condition. This option applies only to trace data.
6.Collection range (COLRNG) - Collect only I/O operations whose duration falls into a specified time range. This option applies only to trace data.

Thanks to SystemiMagazine
avatar
maran
Admin

Posts : 442
Join date : 2009-07-24

View user profile http://pmsinformationsystem.forumotion.net

Back to top Go down

View previous topic View next topic Back to top

- Similar topics

 
Permissions in this forum:
You cannot reply to topics in this forum