Hi

I'm looking for a way of limiting the swap usage of tasks running as
docker containers. My use case is exactly the same as in MESOS-1662 but
the flag introduced to fix that does not apply to docker from what I can
tell.

Setting a custom docker flag to limit total memory to be the same as the
memory resource is a possible work around but doing that puts the
Responsibility to of keeping the values in sync on the one running tasks
through e.g marathon and not on the operator of the cluster.

Any ideas? How hard would it be to change so that flag also applies for
docker?

Attachment: signature.asc
Description: PGP signature

Reply via email to