torjus/nixos-servers
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects 1 Releases Wiki Activity

unbound: tune timeouts for faster recovery after network outages

Browse Source 
Lower infra-host-ttl (900s → 120s) and tcp-reuse-timeout (60s → 15s)
so unbound recovers faster from upstream TLS forwarder failures
instead of staying stuck after ISP outages.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Torjus Håkestad
2026-03-12 01:53:11 +01:00
parent d1516ddd66
commit 5c111c8d78
1 changed files with 6 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
services/ns/resolver.nix
View File

@@ -38,6 +38,12 @@
do-udp = "yes"; do-udp = "yes";
do-tcp = "yes"; do-tcp = "yes";
extended-statistics = true; extended-statistics = true;
# Recover faster from upstream failures (e.g. ISP outage)
# Default 900s is too long - keeps marking servers as bad
infra-host-ttl = 120;
# Clean up stale TLS connections faster (default 60s)
tcp-reuse-timeout = 15;
}; };
remote-control = { remote-control = {
control-enable = true; control-enable = true;