dockermesosmesos-chronos

Docker hangs and gets corrupted on reboot

We are running a scheduling engine with docker, chronos & mesos.

Running 2 mesos slaves on each node. Sometimes, too many Jobs gets executed on each node and docker becomes unresponsive and docker gets corrupted on rebooting the server. Is there anything wrong with the setup? Not sure, why docker hangs and gets corrupted on reboot?

Thanks

Solution

  • Running Docker containers won't work properly because restarting one agent will cause Docker containers managed by the other agent to be deleted.

    Check out --cgroups_root flag in https://github.com/apache/mesos/blob/master/docs/configuration/agent.md This flag only applies to MesosContainerizer (can be used to launch Docker containers).