Hi,

We use Focus db's for our report caster catalogs. When we run a very large job (that fails) and I want to view the log of the job via the report caster console I get the following massage "Not enough memory for this action" . I can open a log for a differnet job. I think that there are to many entries in botlog2 for this job. Can that happen? And is there a way not to log everything into botlog2 for a job execution?

We are chasing a :
BTP1010 Task error:Pcb2.: java.net.SocketException: Connection reset
BTP1010 No report to distribute.

Error and I think they are related.

Can reproduce this on our test and production environments.

Thank you
Phillip du Plooy

Everything on the report caster console page gets executed using servlets.
That means that the data being returned also goes through the servlets.
These things use the appserver's memory settings (usually tomcat), so it may help to increase tomcat's min/max heap sizes to allow for a higher memory usage.
I'm not really an expert in this area, but it is worth a try.

This Pcb2.: java.net.SocketException error is that from the original job or is it being generated when retrieving the log?

We use New Atlanta, and the max heap size is already set to 1024m (start value is 512m).

The java error is from the original job ( chrased agent) and not from retrieving the log.

Thank you
Phillip du Plooy

An alternative is to generate the masters for the caster log tables,
you can then table these via a simple WebFOCUS request

Philip:
If the objective is to figure out what the problem is, change the Focezxec to turn ECHO on... Schedule the Focexec thru RC setting the notification to be 'On error' with Full notification to your email address. This is on the Settings tab of the schedule. This way, you will get whatever is logged in BOTLOG as an email (in addition to going to BOTLOG and BOTLOG2).