[CASE-OPENED] Softkill issued by the agent while running a report
We have an application, which hits the oracle db to generate the report. For certain criteria the reports generates successfully, but for certain we are getting the below trace in edaprint.log
03/31/2009 10:12:12 request by cmrpip000002 for explicit connect to agent (DEFAULT) 03/31/2009 10:12:12 connecting cmrpip000002 u=null,tscomid=1,sesid=1 03/31/2009 10:16:50 request by edachkup to notify connection timeout for cmrpip000002 03/31/2009 10:16:50 Agent (tscomid=1) statistics: 03/31/2009 10:16:50 tscomid connect state active query time last cmd last 03/31/2009 10:16:50 > executed cl pid foc/ext 03/31/2009 10:16:50 user sesid 03/31/2009 10:16:50 > io 03/31/2009 10:16:50 1 -- stopping 1 Mar 31 10:16:50 TABLE -- 03/31/2009 10:16:50 > 6185 0/0 03/31/2009 10:16:50 number of sessions: 1 03/31/2009 10:16:50 last command: TABLE 03/31/2009 10:16:50 last response time: 0:00:00.000 03/31/2009 10:16:50 last errnum: 0 03/31/2009 10:16:50 last executed class: -- 03/31/2009 10:16:50 last masterfile name: SQLOUT 03/31/2009 10:16:50 focus i/o: 0 03/31/2009 10:16:50 external database input: 0 03/31/2009 10:16:50 number of transactions or hli commands: 0 03/31/2009 10:16:50 total execution time: 0:00:02.720 03/31/2009 10:16:50 size of process image in memory (kb): 55288 03/31/2009 10:16:50 priority: 2 03/31/2009 10:16:50 agent deactivated (tscomid=1, pid=6185) 03/31/2009 10:17:07 request by edachkup to notify soft kill required for agent (tscomid=1) 03/31/2009 10:17:07 request by edachkup to notify connection timeout for cmrpip000002 03/31/2009 10:17:07 TERM signal received, pid=6185 03/31/2009 10:17:07 tscomid=1, sesid=1: Terminating tscom3 agent, pid=00006185 03/31/2009 10:17:07 request by t3rp6185 to notify disconnect of sesid=1 03/31/2009 10:17:07 disconnect cmrpip000002 u=null,tscomid=1,sesid=1 03/31/2009 10:17:07 request by edapmon to notify agent is stopped (tscomid=1)
PS: We are getting the output of the query within seconds while running in Oracle, but when we run it via the webfocus console, it times out with the above trace.
Even changing the agent_refresh =1 or cpu_limit=-1 doesnt help.
A case has been raised to IBI with the server trace too.
I would appreciate any faster responseThis message has been edited. Last edited by: Kerry,
March 31, 2009, 04:02 AM
GamP
If the query finishes within seconds from the Oracle side, I'd try to find out what the differences are in the SQL generated and the fast Oracle SQL. Once you know these differences, you could decide to do an sql passthru request in stead of something else. Which might speed things up from the oracle side, and still provide the benefits of webfocus.