Crash after "stop" while mining

[thesnat21]

I had 16 cores mining, sent a stop command and got this:

biblepayd: /usr/include/boost/thread/pthread/recursive_mutex.hpp:113: void boost::recursive_mutex::lock(): Assertion `!pthread_mutex_lock(&m)' failed.

[OrbisPictus]

same here. with "genproclimit": 8, Ubuntu 16.04., Licht's package 1.1.4.3

[biblepay]

Btw, does this happen on every stop or just randomly?

[thesnat21]

i'd say 90% of the time?

It seems more frequent on higher thread counts. It almost seems like something clears before the mining threads are interrupted, cannot say for sure. I tried running in valgrind and it slowed it down too much to be able to reproduce.

[biblepay]

Ok, thanks, it almost seems like this started the day I added the 'setpodcunlockpassword' to biblepay-cli. I seem to recall getting complaints after that when the biblepayd shuts down... maybe. Not sure yet.

[fogoat]

How does one go about troubleshooting this issue? Is valgrind useful?

[biblepay]

Yes that would be helpful if someone could help out as we seem to have gotten busy all of a sudden.

Yes, you can run biblepayd in valgrind, and then issue the stop and you should see the assertion in the valgrind output with a line number; please give us the line number and we will find out why it failed.

[biblepay]

It would be cool if someone can valgrind the daemon for us (but it must be built from source) and give us the line # of the crash and the module filename.

[thesnat21]

Agreed, but in my previous attempts the client runs so slow in valgrind the issue doesn't occur..

[biblepay]

Lets try again someone, to make sure its not a speed issue. Also please ensure its running from a compiled copy.

[biblepay]

Does not happen on Evolution.