On Tuesday, July 30, 2019 at 7:08:43 PM UTC-7, Dave S wrote: > > > > On Monday, July 29, 2019 at 3:06:54 PM UTC-7, Dave S wrote: >> >> >> >> On Tuesday, April 23, 2019 at 8:00:44 PM UTC-7, Dave S wrote: >>> >>> I run web2py under nginx and uwsgi. The latter runs in master mode >>> (rather than emperor mode), and I think it is what drives my question: >>> >>> *How do I control the size of the web2py log files?* >>> I have seen the files grow without bound (until I did a manual log >>> rotate), but currently the web2py/logs/web2py.log file is being rotated for >>> me after 1-3 requests. The uwsg conf file web2py.ini specifies a >>> log-maxsize, but that's for the uwsgi logs, and the web2py logs are being >>> kept under 300 bytes. >>> >>> (300 bytes x 6 rotations with a few requests per second means I don't >>> have to capture error clues) >>> >>> Versions: >>> web2py 2.1.5.4 for the moment, as I didn't jump on the .17 versions, and >>> the .18s kept jumping every time I began to be less inert >>> nginx 1.14.1 >>> uwsgi 2.0.15 (master=true, processes=4) >>> >>> /dps >>> >>> >> Having had to restart the server recently, I studied this again. >> >> It appears that when you start uwsgi, [web2py dir]/logs/web2py.log grows >> normally until it reaches a limit, and then each uwsgi worker starts >> autorotating it, and each rotation gives you 1-3 requests worth of logging, >> and this goes on until the next restart. >> >> My uwsgi is set to autorotate logs, but I'd expect that to only deal with >> uwsgi's logs. Is the behavior above controlled by appfactory() in >> gluon/main.py? >> > > In order to debug an issue, I've been using the "touch routes.py" trick, > and that is a more graceful way of restarting the web2py.log. It appears > that autorotates start at about 1,000,000 bytes (but it's difficult to > catch the actual turning point). The file used for uwsgi configuration has > a log size value of 2,000,000 which fits the size of the uwsgi.logs as > rotated by uwsgi. > > After spending more time looking at appfactory() and also the Python logging library, I can now report:
- the 1000000 comes from my own logging.conf (probably a default value) - setting the delay parameter (= 1) in the args list for rotating file handlers has ameliorated the issue. With that set, logs growing to the limit happens often enough that I can debug much more easily. /dps -- Resources: - http://web2py.com - http://web2py.com/book (Documentation) - http://github.com/web2py/web2py (Source code) - https://code.google.com/p/web2py/issues/list (Report Issues) --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/web2py/060861f6-539c-4adf-9d76-e64997c72ced%40googlegroups.com.
