In case the container is not spawned with SYS_RESOURCE, setting unlimited size for coredumps will never work.
Since the defaults in the script is to abort on any uncaught errors - I propose we catch the ulimit error, and notify the user about the possible missing capability.
In case the container is not spawned with
SYS_RESOURCE
, setting unlimited size for coredumps will never work. Since the defaults in the script is to abort on any uncaught errors - I propose we catch theulimit
error, and notify the user about the possible missing capability.This fixes issues like https://github.com/zalando/postgres-operator/issues/2387