+++ This bug was initially created as a clone of Bug #1740144 +++ When multiple PMD threads in OVS want to push packets through a vhost-user port, but the virtual machine did not initialise enough queues, PMD threads have to take a lock on the queues. This can result in packets losses because of the extra cost on handling the packets. This behavior is hard to figure out without having a deep understanding on the platform, flows and PMD configurations. Making this visible would help to debug these kinds of issues.
This enhancement has been verified: https://bugzilla.redhat.com/show_bug.cgi?id=1740144#c3
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:4206