Managing Disk Subsystems using IBM TotalStorage Productivity Center

434 Managing Disk Subsystems using IBM TotalStorage Productivity Center

Changing the number of entries viewed in the Event Log

The number of event entries that are displayed can be controlled by specifying:

 Total number of entries displayed

 Time interval for entries displayed

By default, the Event Log displays the last 100 events over the last 24 hours.

To change the number of entries displayed, select Options → Set Log View Count from the

menu bar. The maximum number of event entries that can be viewed is equal to the

maximum size of the event log.

To change the time interval for entries displayed, select Options → Set Time Range from the

menu bar.

9.2 Replication Manager problem determination

In this section, we present some methods and log sources for troubleshooting the Replication

Manager component.

Debugging for almost all Replication Manager problems requires the WebSphere Application

Server trace, and most problems also require an ICAT trace at a minimum. In general, RM

provides two major functions:

 Setting up a copy session

 Controlling a copy session

The majority of Replication Manager problems are symptoms of a problem with the

underlying interface to the hardware. Replication Manager communicates with ICAT, which

for ESS communicates with ESSNI, which talks to Copy Services, which talks to the actual

ESS microcode. Any breakdown in communications along this path causes Replication

Manager to behave incorrectly or not function at all. These are the major categories of

interface problems:

1. Lack of state changes (indications) coming from the ESS to Replication Manager. Each

time a copy relationship on a volume changes state, Replication Manager is notified using

an indication. An indication is a CIM event which is delivered asynchronously from the

actual underlying event. Replication Manager uses indications to update its knowledge of

the physical copy relationships as they change dynamically.

Loss of indications causes Replication Manager to appear to be “stuck” or not to have

worked at all. If indications do not arrive after the Start operation is performed on a

session, then the session stays in “defined” state and does not appear to be operating.

Another symptom of loss of indications is that the volume relationship states as reported

by the Copy Services application on the ESS do not match the states reported by

Replication Manager.

2. Unexpected freeze (suspend) operations can be seen when Replication Manager loses

connection to the ICAT or ICAT loses connection to the ESS for longer than 90 seconds.

Replication Manager periodically checks for all ESSs being alive, and initiates a freeze

when there is an active session on an ESS and that ESS does not respond to a presence

check. Upon a timeout when waiting for ESS status, Replication Manager performs a

freeze operation since the timeout could be the first symptom of a disaster.

3. Extremely long durations can be seen to display the session properties panel or the path

status panel. Replication Manager issues hardware queries to display these panels.

Under some circumstances when the ICAT or ESSNI does not respond to the query, the

user will see an extremely long response time, or the user interface may hang completely.