That's a normal warning message, you can ignore it.

The error is in the next line: No heartbeat from core client. This means there is a problem with RPC communication between the core client and the science application. Since the client is no longer controlling it, the science application kills itself.

If this happens rarely, it isn't a problem. The core client will restart the science, and pick up from the last checkpoint. You will see a message:
Result (result name) exited with zero status but no 'finished' file
If this happens repeatedly you may need to reset the project.

If it happens frequently, then it is a major problem. Often it is caused by a software conflict, and we will need to do some troubleshooting.

Maybe I'm not being clear. That error appears for every Valid and Pending result on the website's Result Status page. My results look fine, it's just this one link that looks funky. And there is nothing that says "No heartbeat from core client. "

Here's a more complete piece of the log from my desktop client
That is just the initial set up stage and after there should be pages of process progress entries ending with the </stderr_txt>. If there is not with that 'exited', it's hard to believe such a work unit managed to get in the Result Status page without an 'Error'. If there is reams of records and you just snipped, then your client recovered. Revert to Didactylos initial post as cause finding is needed on frequent occurrence.

Can we take it that you don't have that problem often enough to notice, and you do, in fact, have absolutely nothing to worry about?

If so: that's great. Happy crunching.

And if you don't mind, please will you edit your original post and insert [COMPLETED] after the title? Thanks.

For information only:

The stderr section isn't really intended for members to use. For successful results, it is pretty much ignored altogether. It is no more and no less than what the science application logs to stderr while it is running. Some science apps log more, some log less. We had one beta application that logged insane amounts of trivial information!

So, what good is it? When something goes wrong, we (or the techs) can examine the trace of exactly what the science application did, where it crashed and how it crashed. It's exceedingly useful for giving a headstart in troubleshooting or debugging.

It is marked "Result Log" because it is the metainformation that gets returned with the results files. As you observe, it's nearly all stderr log.

And a final observation: if you are trying to troubleshoot a failed unit that hasn't uploaded, you can find the stderr.txt file in the slot directory.