Comment 32 for bug 1596075

Is it possible to just do another run and capture the logs? It seems like there is some missing data (maybe a bug in oslo logging or in the logging collector?).

I see in the PCAP that an HTTP 500 was returned by neutron for request 'req-2530b6c6-7c5b-4b21-92e1-c5dcd034fa15' but there are no Neutron logs that show the traceback and the 500 being returned. Also the requests associated with the overlapping subnets are missing log statements from various components (e.g. one starts immediately at a retry).

(Sorry about all of these requests but I haven't been able to reproduce this behavior locally with 3 Neutron API servers behind nginx)