The l3 suite has failures with some vms getting a dhcp address. The vms are stuck sending dhcp offers and do not get a response. Somewhere in the pipeline the request/response fails. Other subnets in the same test suite work fine. This usually occurs in the l3 suite, but sometimes happens in other suites.

This worked until about three weeks ago. At that point many genius and netvirt patches were going in to fix transaction leaks and other issues. It seems that some timing was changed just enough.

https://logs.opendaylight.org/releng/jenkins092/netvirt-csit-1node-openstack-ocata-upstream-stateful-nitrogen/501/log_l3.html.gz#s1-t11-k2-k2-k1-k7-k1

Sending discover...
Sending discover...
Sending discover...