Lost Communication with Valence

[ktoole]

We are getting an error when two of us are on that started yesterday saying that "Valence has lost communication with the server". It attempts to reconnect and sometimes comes back, but may take quite sometime to reconnect. We can't firgure out what is wrong.

It looks like when you run debug, it stops all the other users using the same QZSRCGI program.

[richard.milone]

Are you saying this only happens when you're debugging a job?

[ktoole]

Acctually we have discovered this is happening when we are just running an application also. We could not find an error on the iSeries and neither (only two of us are using it right now) was running debug. It had been just fine when only I was using the system, but now with two of us this issue has appeared. We also found that session variables saved in RPG were not there in some cases, while others saved at the same time were there. I don't know if this is related, but I thought I would mention it.

[richard.milone]

The CNX support team will contact you directly to review your application.

[ktoole]

Just so others know, when you run debug with a service job in RDi, if you have multiple stores loading through the same RPG program (calls to your iSeries) multiple thread jobs will show up in RDi. There will be one thread in the debug screen for each call. Make sure that each of the threads complete. If you do not, CGI jobs will still be running for them waiting for debug to continue on your iSeries machine. You can see the CGI jobs that you have locked up in EXTW status by using WRKACTJOB SBS(QHTTPSVR) and looking for the PGM-QZSRCGI jobs. If the current user in these jobs is not QTMHHTP1 and is your user id on the iSeries, you may need to end this job with *immed to clear the problem. It takes about 2 minutes on our system for the job to clear once you have told it to end *immed.

[richard.milone]

ktoole, thanks for contributing that info to the forum.