DPR ERR 2077 on Dashboards during peak times

[omuelle1]

Hi,

We have a problem that during peak hours users are accessing a dashboard that consists of several reports. Some reports are having the DPR ERR 2077, and the user then has either to click refresh several times or restart the browser to see the full report. We are running Cognos 10.1.1



Has anyone seen this error before and has found a solution?

Thank you so much for your help,

Oliver

[bdbits]

Well, cannot see the full error, but googling the portion I can see it appears to be losing a link to the report process. This could be a timeout issue. My first step would be to check the Cognos logs around the time you experience the error. It will likely have more detailed information pointing to the actual problem.

[omuelle1]

Thank you. I went ahead and found the error in the logs:

http-9300-8     DISP    6235    2       Audit.Other.dispatcher.DISP.com.cognos.pogo.handlers.engine.ServiceLookupHandler        Request                 Failure <messages><me                                                                                           ssage><messageString>The server did something wrong</messageString></message><message><messageString>DPR-ERR-2077 Failed to forward the request because the associated extern                                                                                           al process with PID 18116 is unavailable, either because the process is shutting down, or this was an absolute affinity request and the requested process does not exist anym                                                                                           ore. For more information, enable the dispatcher and the external process detailed logs, and reproduce the conditions that caused the error.</messageString></message></messa                                                                                           ges>    http://developer.cognos.com/schemas/reportService/1.absoluteHandler trace back: [the_dispatcher] com.cognos.pogo.handlers.performance.PerformanceIndicationHandler [t                                                                                           he_dispatcher] com.cognos.pogo.handlers.logic.ChainHandler [service_lookup] com.cognos.pogo.handlers.engine.ServiceLookupHandler [load_balancer] com.cognos.pogo.handlers.log                                                                                           ic.ChainHandler [lb_forwarder] com.cognos.p2plb.clerver.LoadBalanceHandler [reportservicechain] com.cognos.pogo.handlers.logic.ChainHandler [warp_with_authenticate] com.cogn                                                                                           os.pogo.handlers.logic.ChainHandler [do_warp] com.cognos.pogo.handlers.logic.ChainHandler [warpmta] com.cognos.pogo.reportservice.ReportServerHandler


I don't really see what settings I could change to get rid of this.

[bdbits]

"The server did something wrong" is one of my all-time favorite log messages. Accurate but meaningless.

It definitely sounds like a process was terminated, perhaps due to a timeout or somebody killed a report server to try and resolve a performance issue. Are there any hints in entries just before or after the error? Is it possible someone is occasionally intentionally killing report servers?

[omuelle1]

Haha I like the server did something wrong as well!

I don't think someone kills any reports, it seems this happens mostly during peak hours. Like if I go to the dashboard right now, I won't see the issues but I will later tonight. I do not see anything surroing that error that would help.

[sdf]

http://www-01.ibm.com/support/docview.wss?uid=swg21615939