Hello, Does anyone know if netfilter/xtables is able to modify the vlan priority bit of an ethernet frame? We have Cisco UCS deployed in our environment and within the fabric interconnect we can define different traffic classes (CoS). One of the capabilities of the CoS scheme in the UCS is we can honor a host administratively setting the CoS bit to something other than its default value. Eg. a CoS 0 (best-effort) adapter sending CoS 5 (Platinum) traffic. Our use case for this kind of mangling is to opportunistically priority different types of traffic based on their purpose. In our example we have file servers running redhat cluster for storage services, these servers push out a lot of bulk data to clients however there are certain types of lighter more latency sensitive traffic (cluster heartbeats / quorum pokes) which should be treated differently than the bulk data. Our CoS 5 traffic class is configured to be lossless and multicast optimized (special resources for enhanced processing of multicast messages). If we were able to mangle the vlan prio bit of certain types of traffic we could do some fairly powerful shaping of traffic within the FI of the UCS. Eg: iptables -t mangle -p udp -d a.b.c.d --dport 12345 -j L2COSMARK --priority 5 Any suggestions or direction on if this is possible would be greatly appreciated. -- To unsubscribe from this list: send the line "unsubscribe netfilter" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
