Hi Tom,
On 15 Feb 2016, at 23:50, Tom Ritter tom@ritter.vg wrote:
I do. My checkout is dd3d01fae752ff0c400ec55a700dfa11c31546d6 - which I believe is HEAD of master. I haven't done any tricks consciously - I know Mike has configured things so it runs in a virtualenv now but I had it running before that, so I did not. The README worked fairly well.
I'm running it with SQLAlchemy-0.7.10 It will take a few days to accumulate enough data to generate a usable output file, and up to a week to get really up and moving. If you restart the whole thing you may lose data and start from scratch - so I avoid stopping my scanners.
Sometimes the scanners get stuck, so I had set up a job that would display when they loop around, which I output here: https://bwauth.ritter.vg/bwauth/AA_scanner_loop_times.txt It appears my scanner.9 is broken heh. I'm not sure what to do about that, I guess I'll kill that individual job and start it back up in screen....
A community member has looked closely at my data to see if they could diagnose why things would sometimes get stuck, or relays would fall through the cracks, which is why I have my logs symlinked there. In general there's a lot of things broken with the code right now, but it functions 'well enough' to keep the network running. Just not well enough to prevent it from getting stuck sometimes or to measure all relays consistently.
I wrote a thing that would notify me if things got too out of date (here: https://github.com/tomrittervg/checker/blob/master/samplejobs/BWAuthChecker.... ) but it's obviously not very good. I need to correct it because sometimes the file is 3 hours out of date legitimately (I think?) and obviously I'm not checking the loop-around time.
Let me know if there's anything else I can answer for you!
no, thanks indeed, Unless you know anything about the efforts to produce an improved bwauth codebase? In the meantime, I'll try to get the tip of master set up on jessie.
Thanks!
Sebastian