* Chris Hofstaedtler <z...@debian.org> [210124 22:24]: > I don't see why this would "just fail randomly". I don't think its a > timing thing. Maybe this worker is weird in a different way.
Having now looked at the (not very granular) timestamps, it looks like the 30sec "service did not finish starting" timeout from systemd. I could see that happening for pdns, but not for pdns-recursor. There has to be something extremely weird going on, or if it's really a timing / I/O thing, then -that worker- is overloaded beyond being useful. Lets see what the pdns-recursor 4.4.2-4 test run brings. Paul, can you expedite that test run? Possibly in a way it runs on the problematic worker? Chris
