Occasional engine errors
JonathanBrouwer opened this issue · 6 comments
I am every so often getting the following error:
[=====|= ] 7 cores, 8 queued, latest: https://lichess.org/igg8JtJs#23
E: Engine error: unexpected end of file
W: Worker 4 waiting for engine to shut down after error. Context: https://lichess.org/igg8JtJs#24
W: Aborting batch FCRIeuUA.
Sometimes I can go a few hours without having this error, other times I have multiple in one minute.
Any idea what could cause this?
System information:
Operating System: Manjaro Linux
KDE Plasma Version: 5.20.4
KDE Frameworks Version: 5.77.0
Qt Version: 5.15.2
Kernel Version: 5.11.0-1-MANJARO
OS Type: 64-bit
Processors: 8 × Intel® Core™ i7-6700K CPU @ 4.00GHz
Memory: 15,6 GiB of RAM
Graphics Processor: GeForce GTX 970/PCIe/SSE2
Just to confirm, which fishnet --version is this?
fishnet 2.2.0
grepped the journalctl of my 2 instances of fishnet and could not find any 'Engine error'. System has older processors so fishnet is using stockfish-x86-64-sse41-popcnt , if that makes a difference. hth
Alright, thanks. I can't reproduce the issue locally on the same game, even though Stockfish analysis is supposed to be deterministic.
Let's exclude that it's due to the order of distributing positions to processes. Can you please run your instance with --cores 1 for a while, and paste any errors here, if they still happen?
Hm I just got another error which I haven't seen before
[=== | ] 7 cores, 5 queued, latest: https://lichess.org/vjaAlkWd#15*** stack smashing detected ***: terminated
I've had some issues with my cpu before, so I think this might actually be hardware problems. (Especially considering you can't reproduce this, and stack smashing seems like a very unlikely bug for me to run into)
I'll try running a few stress tests and seeing if I can get them to fail
These were indeed most likely hardware errors. Sorry for bothering you!