Boinc-client has crashed, cmsearch is still running

[Lem]

Houston, I have a problem!

About 30 minutes ago, while I was uploading some results, boinc-client has crashed and died. I still have four cmsearch running processes, now depending from init. No, wait... now they're three, one has finished... and it seems its output files have been written correctly.

Since one of the three remaining tasks has already run for more than 16 days, which time I wouldn't like to waste, my question is: if I leave everything as is... I mean: if I leave the three processes running, without restarting boinc-client, and I wait till they finish, and after that I restart boinc-client (it will take some more days for the long one) will the results be uploadable?

Please, say "yes".

And: I won't suffer from the heartbeat issue, since boinc-client is dead, will I?

Bye.

[Ananas]

Hmmm, that's exactly the rare situation when the heartbeat thing actually should kick in - the only reason I can imagine that it did not do that already would be, that the communication through shared memory didn't work and the clients are running in standalone mode.

I cannot answer the question about how the BOINC client treats the results if you finish them in this mode, sorry :-/

edit : If just the BOINC manager (the visual interface) died but the BOINC core client is still there, restarting the manager should not do any harm to the results, it should connect to the still running core client. You can check that, the core client should be visible in the task list (Linux something like : ps -ef|grep -i boinc / windows : check the task manager for "boinc.exe"). But as you wrote that the project clients are connected to init now, I guess that isn't the case :´(

[Lem]

I confirm that boinc-client isn't running anymore.

Two of the four RNA XLL tasks have come to an end in the meantime.
I can say that each of them has written correctly its two output files (which are the only files that have been updated, since the crash, in the RNA /var/lib/boinc-client/slots/*/ directories).

No heartbeat problems till now with RNA. I did have a heartbeat issue with the einstein@home workunit running on my GPU (I can see it from its stderr.txt file), but stderr.txt files of rna workunits haven't been updated after the crash and everything seems smooth.

So I try to be confident.

I've started some folding@home work on the two free cores, so I won't waste any cpu spare time.

Thanks, and

bye.

[Ananas]

@Yoyo : When you tried to comment out the wrapper source lines that handle the heartbeat, did you do that on Linux only? We know that there was no effect on Windows - could the reason be that your change didn't find its way into the Windows binary?