Currently we suspect the GRE tunnel to be a 1-core-bottleneck.
It could be possible that the reason for the GRE tunnel to be seemingly pinned to one core could be related to excessive one-core overhead of a non-work-conserving QDISC. We should evaluate and try some fitting work-conserving QDISC.
Currently we suspect the GRE tunnel to be a 1-core-bottleneck. It could be possible that the reason for the GRE tunnel to be seemingly pinned to one core could be related to excessive one-core overhead of a non-work-conserving QDISC. We should evaluate and try some fitting work-conserving QDISC.