Hi,
I have a application using the Cognos 10.2.2 SDK for generating reports into different report outputs which are saved to file system.
The distributed Cognos environment (dispatchers running in total 36 report service processes) are only using a few number of report processes 1/6 or maybe 2/6 report service processes on all dispatchers.
Sometimes reports are failing due to soap exceptions when trying to get report output:
RSV-BBP-0039 The request 'reportGetOutput_Request' failed because the conversation was removed from the system to make room for other conversations.
RSV-SRV-0042 Trace back:
RSReportService.cpp(763): RSException: CCL_CAUGHT: RSReportService::processImpl()
RSReportServiceMethod.cpp(259): RSException: CCL_RETHROW: RSReportServiceMethod::process(): reportGetOutput_Request
RSReportServiceHelper.cpp(857): RSException: CCL_THROW: RSReportServiceHelper::absoluteAffinityError()RSV-BBP-0038 The request 'reportGetOutput_Request' failed because the conversation was already canceled.
RSV-SRV-0042 Trace back:
RSReportService.cpp(763): RSException: CCL_CAUGHT: RSReportService::processImpl()
RSReportServiceMethod.cpp(259): RSException: CCL_RETHROW: RSReportServiceMethod::process(): reportGetOutput_Request
RSReportServiceHelper.cpp(857): RSException: CCL_THROW: RSReportServiceHelper::absoluteAffinityError()We also have problem with the following operations which are increasing in time after running our application for a while, not always for all report jobs but in total we have a time increase.
- logon (contentManagerService1.logonAsync)
- query for report (contentManagerService1.queryAsync)
- getting report parameters (reportService1.getParametersAsync)
- logoff (contentManagerService1.logoffAsync)
The servers are very powerful so I guess it's a matter of tuning our environment or maybe decrease the number of running reports in parallel.
Is 25 running reports in parallel to many if we in total have the ability to run 36 report service processes?
Maybe there is some bottleneck in our Cognos environment?
Theses are my setup, the dispatcher servers are running two installations of Cognos simultaneously:
Content Manager (Primary)
32 Core
98 GB Memory
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
Content Manager (Secondary) (Virtual Machine)
2 Core
8 GB Memory
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
Dispatcher Server 1 (Dispatcher1,Dispatcher2)
32 Core
98 GB Memory
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
Dispatcher Server 2 (Dispatcher3,Dispatcher4)
32 Core
98 GB Memory
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
Dispatcher Server 3 (Dispatcher5,Dispatcher6)
32 Core
98 GB Memory
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
Cognos Configuration\Environment\IBM Cognos Services\IBM Cognos\Maximum Memory for Websphere Liberty in MB: 1536
General report service configuration settings:
Maximum number of processes for the report service during non-peak period: 6
Maximum number of processes for the report service during peak period: 6
Number of high affinity connections for the report service during non-peak period: 2
Number of low affinity connections for the report service during non-peak period: 8
Number of high affinity connections for the batch report service during peak period: 2
Number of low affinity connections for the batch report service during peak period: 8
Queue time limit of the report service (seconds): 360