Switch to using proxy_cache_lock_age

This allows us to cache the response in the event a file hasn't been downloaded yet
Set to 2m as a really long time to download a single 1m slice from upstream

Based on thoughts from https://github.com/lancachenet/generic/pull/113
This commit is contained in:
VibroAxe 2021-10-09 11:07:12 +00:00 committed by GitHub
parent 2fa012f667
commit 51753881f9

View file

@ -10,6 +10,10 @@
# this really happens, otherwise we end up wasting bandwith # this really happens, otherwise we end up wasting bandwith
# getting the file multiple times. # getting the file multiple times.
proxy_cache_lock on; proxy_cache_lock on;
# If it's taken over a minute to download a 1m file, we are probably stuck!
# Allow the next request to cache
proxy_cache_lock_age 2m;
# If it's totally broken after an hour, stick it in bypass (this shouldn't ever trigger)
proxy_cache_lock_timeout 1h; proxy_cache_lock_timeout 1h;
# Allow the use of state entries # Allow the use of state entries
@ -28,3 +32,4 @@
# 40G max file # 40G max file
proxy_max_temp_file_size 40960m; proxy_max_temp_file_size 40960m;