Recently one of my elasticsearch cluster (firstly a client, then a data node) died because of no disk space. There is no warning, no auto evaculation. I understand it could be limitation for docker. My question is, how can I easily enable log rotate in rancher way? And how can I get notified if my node doesn’t work?
My solution is to add “–log-opt max-size=50m” to DOCKER_OPTS in /etc/init.d/docker
Will keep monitoring if issue is resolved.\
BTW, I’m using 1.1.0-dev5 build.