Hello,
The documentation to setup a agentic workflow on self hosted runner is a little bit unclear. I take this doc: https://github.github.com/gh-aw/reference/self-hosted-runners/
There is multiple elements on multiple part, there is 2 part for ARC ...
Anyway, I run runner on kubernetes with a container runner + a container dind. The docker socket is shared via an empty dir.
- I use to not have privileged mode on the runner container, is this really mandatory ?
- I need to share the /tmp directory between the runner container/docker container via an empty directory
- I notice I need iptables command else there is a failure
And with all that, I keep getting an error
[SUCCESS] Host-level iptables rules configured successfully
[INFO] Generating configuration files...
[INFO] API proxy sidecar enabled - API keys will be held securely in sidecar container
[INFO] API proxy will route through Squid to respect domain whitelisting
[INFO] CLI proxy sidecar enabled - connecting to external DIFC proxy at host.docker.internal:18443
[INFO] Starting containers...
Container awf-squid Creating
Container awf-api-proxy Creating
Container awf-api-proxy Created
Container awf-squid Created
Container awf-cli-proxy Creating
Container awf-cli-proxy Created
Container awf-agent Creating
Container awf-agent Created
Container awf-iptables-init Creating
Container awf-iptables-init Created
Container awf-squid Starting
Container awf-api-proxy Starting
Container awf-api-proxy Started
Container awf-squid Started
Container awf-squid Waiting
Container awf-squid Healthy
Container awf-cli-proxy Starting
Container awf-cli-proxy Started
Container awf-squid Waiting
Container awf-api-proxy Waiting
Container awf-cli-proxy Waiting
Container awf-api-proxy Healthy
Container awf-squid Healthy
Container awf-cli-proxy Error dependency cli-proxy failed to start
dependency failed to start: container awf-cli-proxy is unhealthy
Error: awf-cli-proxy container logs (last 50 lines):
[cli-proxy] Starting CLI proxy sidecar...
[cli-proxy] External DIFC proxy at host.docker.internal:18443
[cli-proxy] Starting TCP tunnel: localhost:18443 → host.docker.internal:18443
[cli-proxy] TLS certificate available
[cli-proxy] Combined CA bundle created at /tmp/proxy-tls/combined-ca.crt
[cli-proxy] gh CLI configured to route through DIFC proxy at localhost:18443
[tcp-tunnel] Forwarding localhost:18443 → host.docker.internal:18443
[cli-proxy] DIFC proxy probe failed (attempt 1/10, diagnosis=unknown), retrying in 1s...
[cli-proxy] DIFC proxy probe failed (attempt 2/10, diagnosis=unknown), retrying in 2s...
[cli-proxy] DIFC proxy probe failed (attempt 3/10, diagnosis=unknown), retrying in 4s...
[cli-proxy] DIFC proxy probe failed (attempt 4/10, diagnosis=unknown), retrying in 8s...
[cli-proxy] DIFC proxy probe failed (attempt 5/10, diagnosis=unknown), retrying in 16s...
[cli-proxy] DIFC proxy probe failed (attempt 6/10, diagnosis=unknown), retrying in 30s...
[tcp-tunnel] Connection from ::1:47780
[tcp-tunnel] Connection closed: ::1:47780
[tcp-tunnel] Connection from ::1:47792
[tcp-tunnel] Connection closed: ::1:47792
[tcp-tunnel] Connection from ::1:59336
[tcp-tunnel] Connection closed: ::1:59336
[tcp-tunnel] Connection from ::1:59352
[tcp-tunnel] Connection closed: ::1:59352
[tcp-tunnel] Connection from ::1:54366
[tcp-tunnel] Connection closed: ::1:54366
[tcp-tunnel] Connection from ::1:47170
[tcp-tunnel] Connection closed: ::1:47170
Error: [ERROR] Fatal error: Error: AWF firewall failed to start: awf-cli-proxy could not connect to the external DIFC proxy (or exited before establishing a connection). Failing fast to avoid repeated in-agent retries. The agent was never invoked. See awf-cli-proxy container logs above for details.
at Object.QD [as startContainers] (/usr/local/lib/awf/awf-bundle.js:781:4730)
at process.processTicksAndRejections (node:internal/process/task_queues:103:5)
at async ZN (/usr/local/lib/awf/awf-bundle.js:786:10770)
at async e.<anonymous> (/usr/local/lib/awf/awf-bundle.js:796:1139)
[INFO] Stopping containers...
My issue is the "check logs" ... I was checking the /tmp/gh-aw/sandbox/firewall/logs/cli-proxy-logs but no log file where there, so .... i'm stuck trying to understand if there is something missing
Hello,
The documentation to setup a agentic workflow on self hosted runner is a little bit unclear. I take this doc: https://github.github.com/gh-aw/reference/self-hosted-runners/
There is multiple elements on multiple part, there is 2 part for ARC ...
Anyway, I run runner on kubernetes with a container runner + a container dind. The docker socket is shared via an empty dir.
And with all that, I keep getting an error
My issue is the "check logs" ... I was checking the /tmp/gh-aw/sandbox/firewall/logs/cli-proxy-logs but no log file where there, so .... i'm stuck trying to understand if there is something missing