PRS-PRO-0511 A partial reinitialization of the XTS processor is required

Cognos BI 10.1.1

Why do we get so many fails in the Presentation service ?
You can see number of fails in Cognos connection web portal if you go to;
Launch Cognos Administration
Click on status tab
Click on system
On the top right side you can expand the metrics
and find the Request – Presentation service
Note the Number of failed requests

Presentation Service metrics deal with logons, display of reports and display of content (admin page/main portal page/portlets) within the portal.

If you have a user that for some reason are entering invalid userid/password information when trying to logon to the IBM Cognos portal, then this will be recorded as a failed presentation request and will show up as a failed request on the Presentation service. For this example, there may be 4 failed presentation requests generated. Users that cancel a report just as it is about to
be displayed will also register a failed presentation service request.

A high failure rate on Presentation service can probably be caused by the reinitialization of the XTS validation. (“PRS-PRO-0511 A partial reinitialization of the XTS processor is required.”) While the dispatcher is at the idle state, the XTS validator sends request out in a certain period for “health check” on dispatcher communication. If no other active you should find 2 failure
requests in Presentation service and then 3 XTS reinitialization from the log file.

“PRS-PRO-0511 A partial reinitialization of the XTS processor is required.” is the text to look for in cogserver.log file.

The key thing is to do some baseline analysis of the metrics that are captured over a typical day (good day) or a typical week (good week) to see the # of request failures on average. Then knowing this and also understanding that the system was able to handle the running of
reports/jobs without any issues (errors or slow run times of reports) during the day or the week, you can adjust your thresholds to look for abnormally high failure rates above the typical failure rates in a good day or week.

So, it is important that you do not set thresholds that are two high/low depending on the metric being measured especially if the metric is measuring things such as failed logons or cancelled report runs.

More Information:
The system performance metrics are gathered from the individual services (batch, presentation service etc.) that are running and processing requests and summarized in the Administration page of IBM Cognos. The information is not stored in the audit database or the cogserver.log but within the content store database itself.

The intended workflow for investigating failed requests is to check the cogserver.log file and use the information there to identify the cause. If one opens the log using the /bin/logviewV2.exe tool then you can filter the log file to only show the entries for the service of interest. Failed requests are highlighted in red and from there you can get the details on what request(s) failed, and the associated messages and error codes can direct you to potential root cause for the failures. Additionally, the errors can then be searched ( on in IBM knowledge base for a TechNote that will help resolve the issue.

If it is still not clear why the request failed or what action to take to resolve the issue after that, you should then seek further assistance from Customer Support.

This can happen if you change values in the file /templates/ps/portal/system.xml – have you done that?

If you found a cause to this error please post it at forum.