I don't seem to be recieving any jobs
WannesMalfait opened this issue · 4 comments
Hi
I just wanted to contribute but I think I did something wrong. I cloned the repo and then ran cargo run --release --.
After starting fishnet with target\release\fishnet.exe --auto-update --cores auto, it seems like I have not recieved any jobs even after running it for a while. This is the console output:
><> Checking for updates (--auto-update) ...
Checking target-arch... x86_64-pc-windows-msvc
Checking current version... v2.1.2-dev
Checking latest released version... v2.1.1
><> Fishnet 2.1.2-dev is up to date
Checking configuration ...
Endpoint: https://lichess.org/fishnet
Backlog: Join queue if user backlog >= 30s or system backlog >= 3600s
CPU features: POPCNT | SSE | SSE2 | SSSE3 | SSE41 | AVX2 | BMI2 | INTEL | SF_SSE2 | SF_SSSE3 | SF_SSE41_POPCNT | SF_AVX2 | SF_BMI2
Engine: stockfish-x86-64-bmi2.exe (for GPLv3, run: target\release\fishnet.exe license)
Cores: 7
Running (CTRL-C to stop) ...
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
><> fishnet/2.1.2-dev: 3500 knps??? (nnue), 0 batches, 0 positions, 0 total nodes
Thanks for running fishnet. It looks like your instance is configured to be idle, unless a backlog is building up:
Backlog: Join queue if user backlog >= 30s or system backlog >= 3600s
You can rerun the configuration dialog with cargo run --release -- configure and select to join unconditionally. If that doesn't resolve the issue, running with -v might show what's going on.
Thank you for the (very) fast response, this was indeed the problem. Just out of curiosity, what is the difference between user and system backlog?
The user queue is for humans requesting analysis. We try to keep it very low, because they might be waiting for it.
The system queue is automatically requested analysis, for example to run spot checks and select games for a deeper look with cheat detection. It doesn't matter if it builds up a bit at peak time, as long as it's cleared during the day.
Which queue is selected for the next job, depends on the backlog situation, configuration, and your clients performance.
Ok thank you