All containers dying on a CoreOS node during high cpu/network usage

Fleet on CoreOS is known to stop all containers on a node, if that node is considered unresponsive due to high CPU/network load. This may cause failures in the mesos cluster.

Make sure to check your fleet logs if you see your containers stopping during high network/cpu conditions.

To resolve this problem you may have to tune your etcd and fleet parameters.

References:
https://www.mail-archive.com/user@mesos.apache.org/msg03107.html

Have more questions? Submit a request

Comments

Powered by Zendesk