'cannot lint path-mtu' false positive when global-physnet-mtu is used instead

Bug #1958901 reported by Andrea Ieri
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Juju Lint
Incomplete
Undecided
Unassigned

Bug Description

From the charm config documentation:

    path-mtu:
      Default: null
      Description: |
        Maximum size of an IP packet (MTU) that can traverse the
        underlying physical network infrastructure without fragmentation
        when using an overlay/tunnel protocol. This option allows
        specifying a physical network MTU value that differs from the
        default global-physnet-mtu value.
      Type: int

In a simple cloud where the overlay network uses mtu 9000 the simplest (and valid) config is to set global-physnet-mtu to 9000 and leave path-mtu unset.

In such a scenario juju-lint will produce the following false positive:

[WARNING] [/tmp/openstack-bundle.yaml] [manual/manual] Application neutron-api has no config for 'path-mtu', cannot determine if == 9000.

Changed in juju-lint:
status: New → Fix Committed
status: Fix Committed → New
Revision history for this message
Gabriel Cocenza (gabrielcocenza) wrote :

I think in that case it's a question of creating a new rules file for the simple cloud.
bug 1988851 is setting global-physnet-mtu to 9000 and path-mtu >= 1558 for ovn and path-mtu >= 1550 for ovs.

Changed in juju-lint:
status: New → Incomplete
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.