Connection message in Enterprise Guide for long running jobs

[DJWanna]

Since moving to SAS9.4M6/Enterprise Guide 8.2 our users state that there is a connection statement that appears like the libnames have all been dropped. When the session comes back to SAS from Oracle, for instance, the libname will no longer be active and the job crashes. Has anyone else experienced this type of behavior? The only workaround we have tried is copying the libname statement after every step. We also have the keepalive option on the workspace server set to 600, which is beyond the default. Would it help to increase that option and how big can we make it?

Thank you for your advice.

[gwootton]

The keepalive setting for the workspace server described in the doc link below is for the IOM connection between the workspace server and Enterprise Guide, and would not impact the connection between the database and the workspace server session. If the connection between oracle and SAS is being closed this is likely from a setting in the Oracle configuration.

OBJECTSERVERPARMS System Option - Keepalive
https://go.documentation.sas.com/doc/en/bicdc/9.4/biasag/n16001intelplatform00srvradm.htm#n16030inte...

It looks like this previous post also describes the need to increase the oracle setting:

https://communities.sas.com/t5/General-SAS-Programming/SAS-Access-Oracle-timeout/td-p/386943

[DJWanna]

Thanks, Greg. Was the oracle settings needed for M6? Users are not seeing this in M3.

[gwootton]

I don't think this would be related to a SAS release version. Typically a "keepalive" is needed because some external device is actively monitoring for and closing idle connections. For example the workspace server keepalive is usually used to prevent a firewall/vpn from closing the TCP connection between Enterprise Guide and the Workspace Server due to lack of activity on that connection.

So, either Oracle is closing the session because it's not seeing any activity on it passed some timeout, or an external device like a firewall is doing this.