Skip to content

cover failover use cases #12

Description

@norg

There are scenarios where two (or more) interfaces are in an active/passive mode. So interface0 receives traffic to inspect and interface1 does not. But sometimes interface1 has the LINK UP and receives a small portion of (management) traffic. In those cases it's best if all cores and queues in cluster_qm mode are assigned to the interface with the traffic and avoid the other completely. But in the failover case where interface0 looses the traffic and interface1 receives it, we need to swap that.

So one part is to make sure the swap will happen but it needs to be defined how it's detected, since a simple link or traffic check won't work. So adding some sort of threshold, maybe even customized via some value/template, could ensure how much traffic is seen as "inactive" for an interface.

The other part is to make sure it is such a scenario and not just a switch of amount of traffic. This might be solved by a flag. Because it could also be that just some parts of the traffic is gone or has swapped which is nearly impossible to maintain.

Metadata

Metadata

Assignees

Labels

enhancementNew feature or request

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions