Delay in /tmp/pid File Creation Causes unhealthy Status on Rocky Linux 8.6

[sinbumu]

I'm experiencing an issue with running Besu in a Docker/Podman container on Rocky Linux 8.6. The container stays in an unhealthy state for an extended period (approximately 9-10 minutes) before eventually switching to healthy. The delay appears to be related to the /tmp/pid file, which is not created immediately upon container startup, causing the health check to fail repeatedly.

Environment

• OS: Rocky Linux 8.6
• Container Engine: Podman (or Docker, also tested)
• Docker version: 27.2.0
• Besu version: hyperledger/besu:latest
• System Specs: 4 vCPUs, 16 GB RAM

Steps to Reproduce

1. Launch a Rocky Linux 8.6 instance.
2. Install Podman/Docker.
3. Pull the latest Besu Docker image.
4. Run the following command to start Besu:

sudo podman run -d --name besu_node \
  -e BESU_LOGGING=TRACE \
  -p 8545:8545 -p 8546:8546 -p 30303:30303 \
  hyperledger/besu:latest \
  --rpc-http-enabled --rpc-http-host=0.0.0.0 --host-allowlist="*"

5. Check the container health status (podman ps or docker ps) and observe that it remains in the unhealthy state.

Observed Behavior

The container stays in an unhealthy state for about 9-10 minutes, then transitions to healthy. During this time, the health check repeatedly fails with exitCode=1, which seems to be related to the /tmp/pid file not being available immediately.

Expected Behavior

The Besu container should create the /tmp/pid file promptly upon startup to allow the health check to succeed or provide an alternative health check method that accurately reflects the container's readiness state.

Logs

Here are relevant sections of the Docker logs showing repeated health check failures:

time="2024-11-12T06:58:15.854925216Z" level=debug msg="Health check for container done (exitCode=1)"
...
time="2024-11-12T06:58:20.905606739Z" level=debug msg="Health check for container done (exitCode=1)"

Additional Information

• Disk Performance: Disk write speed is around 140 MB/s, so storage performance does not seem to be the issue.
• Attempts with Alternative Configurations: Tried increasing health-interval and StartPeriod, but the delay in healthy state persists.
• Other OS: The same configuration on Ubuntu does not exhibit this delay, so this issue might be specific to Rocky Linux or RHEL-based systems.

Questions

• Could there be an OS compatibility issue with Rocky Linux, or specific configurations we should adjust?