FIX the GITEA_RUNNER_GID envar, 2nd attempt
Another attempt to fix the GITEA_RUNNER_GID envar issue. See https://github.com/vegardit/docker-gitea-act-runner/issues/5 Tested this GITEA_RUNNER_GID envar fix without building the upstream image, as that was problematic to achieve locally. The fixed image requires either compose: 'user:root' OR restart: always (or unless-stopped). Why? The UID change in /opt/run.sh fails the first time, at its next sudo command. Basically: changing the running UID's /etc/passwd entry causes the next sudo command to fail. The container restart (without wiping the read/write layer) starts with the already changed /etc/passwd, and then continues without the sudo error. Signed-off-by: William Blew <william@kulian.org>
This commit is contained in:
parent
4c9806d71c
commit
c96a2301a8
|
|
@ -42,7 +42,7 @@ if [ -n "${GITEA_RUNNER_GID:-}" ]; then
|
|||
effective_gid=$(id -g act)
|
||||
if [ "$GITEA_RUNNER_GID" != "$effective_gid" ]; then
|
||||
log INFO "Changing GID of user [act] from $effective_gid to $GITEA_RUNNER_GID..."
|
||||
sudo usermod -o -g "$GITEA_RUNNER_GID" act
|
||||
sudo groupmod -o -g "$GITEA_RUNNER_GID" act
|
||||
fi
|
||||
fi
|
||||
sudo chown -R act:act /data
|
||||
|
|
|
|||
Loading…
Reference in New Issue