We saw a problem with installing ece 3.5.1 on a new ec2 instance in aws. All pre-reqs were followed but despite this we kept seeing "unknown host" exeptions in the logs in /mnt/data/elastic/services/%hostname%/logs/runner.
We also noticed a behavior where name resolution inside a container was not working, despite /etc/hosts and /etc/resolv.conf being correct.
The fix for us was to 'disable' the nscd daemon or,
- sudo systemctl stop nscd
- sudo systemctl disable nscd
Once this was done and the ece install was re-started, the containers started to have name resolution again.
(Example test was, docker exec %podname% ping %podname% which had preivously failed. )