Closed Miwer closed 2 years ago
also error.log is not rotated. I can see in the logrotate rule it utilizes the recursive log symbolic link in the log directory (rule is "/data/logs/*/error.log"), but the symbolic link is wrong - it points to itself creating an infinite loop, when trying to do anything with it.
I believe the symbolic link should instead point to "." (the current directory)
bash-5.0# ls -l /data/log/log*
lrwxrwxrwx 1 root users 3 Nov 21 12:20 /data/log/log -> log
bash-5.0# cd /data/log/log
bash: cd: /data/log/log: Symbolic link loop
My /appdata/NginxProxyManager/log and logs folder contain logs from months ago. Are these rotated/purged after a certain amount of time?
There seem to be some discrepancy in the /etc/nginx/conf.d/default.conf file. the 8080 host is logging to /data/logs/fallback_access.log while the 4443 host is logging to /data/logs/fallback-access.log
This causes two fallback access logs to be created, and one of them is not rotated due to the rule setup in logrotate.d
Good catch, I will fix that.
Also error log on the 4443 host is set to /dev/null crit - I think it should point to fallback_error.log like the 8080 host, just as the access log is probably meant to point to the same log file on both the 8080 and 4443 host.
I think this is correct. Because there is no valid SSL certificate to present when reaching a non-configured host and because we forcefully terminate the connection (return 444
), errors are ignored since they are not relevant.
also error.log is not rotated. I can see in the logrotate rule it utilizes the recursive log symbolic link in the log directory (rule is "/data/logs/*/error.log"), but the symbolic link is wrong - it points to itself creating an infinite loop, when trying to do anything with it.
I will fix the bad symlink. However, this is not the cause of error.log
not being rotated. The problem seems to be with the logrotate file path definition: it should be /data/logs/error.log
instead of /data/logs/*/error.log
.
This should be fixed with the latest Docker image.
(running latest version)
There seem to be some discrepancy in the /etc/nginx/conf.d/default.conf file. the 8080 host is logging to /data/logs/fallback_access.log while the 4443 host is logging to /data/logs/fallback-access.log
This causes two fallback access logs to be created, and one of them is not rotated due to the rule setup in logrotate.d
(also I thought only "log" directory was supposed to exist, not "logs", so maybe there's an additional fix there?)
Also error log on the 4443 host is set to /dev/null crit - I think it should point to fallback_error.log like the 8080 host, just as the access log is probably meant to point to the same log file on both the 8080 and 4443 host.