IMHO this isn't a valid issue for Nova and the delays seen with devices becoming writeable is more than likely with the hosts multipath configuration, storage fabric etc. Are we still seeing failures even with the additional retry logic added in os-brick?
IMHO this isn't a valid issue for Nova and the delays seen with devices becoming writeable is more than likely with the hosts multipath configuration, storage fabric etc. Are we still seeing failures even with the additional retry logic added in os-brick?