vinzenz/hyperhive
vinzenz/hyperhive
Watch
0
0
Fork
You've already forked hyperhive
 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix(#2410): bump container@hive-ci TimeoutStartSec to survive slow runner registration

Browse source 
gitea-runner registration (hive-ci-prefetch, host-side) sits on the
boot-critical path -- nspawn readiness is gated on the runner
registering, itself a forge + core-token round trip that can wait up
to 60s for the core token. The default ~60s TimeoutStartSec can trip
mid-register (especially right after a .runner purge, since every
boot re-registers from scratch), killing the half-started container
and triggering a restart loop until the token/forge settle. Bump to
180s so one register attempt has room to finish.
This commit is contained in:
atlas 2026-07-13 17:44:44 +02:00 • committed by mara
commit 2bf6290aa0
1 changed files with 22 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

26
nix/host-modules/hive-ci.nix
View file

@ -351,10 +351,28 @@ in
# generated by the host `hive-tls-ca` service. Order the container after
# it so the bind source exists before nspawn sets the mount up (a
# condition-skipped/late CA would otherwise fail the container start).
systemd.services."container@hive-ci" = lib.mkIf useSelfSigned {
after = [ "hive-tls-ca.service" ];
requires = [ "hive-tls-ca.service" ];
};
systemd.services."container@hive-ci" = lib.mkMerge [
(lib.mkIf useSelfSigned {
after = [ "hive-tls-ca.service" ];
requires = [ "hive-tls-ca.service" ];
})
{
# gitea-runner registration (hive-ci-prefetch, host-side)
# sits on the boot-critical path — the container's nspawn readiness
# is gated on the runner registering, a forge + core-token round
# trip that itself waits up to 60s for the core token. The default
# nspawn `TimeoutStartSec` (~60s, systemd's `DefaultTimeoutStartSec`)
# can therefore trip mid-register, especially right after a
# `.runner`-purge (every boot re-registers from scratch — see
# hive-ci-prefetch above) or during a hive-c0re restart storm that
# delays the core token. systemd then kills the half-started
# container and reschedules, producing an observed ~60s restart
# loop until the forge/token settle. Bumping the timeout well past
# the prefetch's own 60s wait gives one register attempt room to
# actually finish instead of being killed mid-flight.
serviceConfig.TimeoutStartSec = "180s";
}
];
containers.hive-ci = {
autoStart = true;