
legacy rtp reservation
I'm having some trouble seeing results after configuring legacy RTP reserve queueing. The simulated rtp traffic from R6 to R5
has an IPP of 1 while the web traffic from R1 to SW4 has a cos of 7. I get the same weights after configuring ip rtp reserve as show below
interface Serial0/1/0
bandwidth 128
ip address 155.1.45.4 255.255.255.0
ip mtu 156
ip rip advertise 10
tx-ring-limit 1
tx-queue-limit 1
fair-queue 64 32 1
max-reserved-bandwidth 100
ip rtp reserve 16384 16383 128
end
Rack1R4#sh queueing int s0/1/0
Interface Serial0/1/0 queueing strategy: fair
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: weighted fair
Output queue: 30/1000/64/0 (size/max total/threshold/drops)
Conversations 2/3/32 (active/max active/max total)
Reserved Conversations 1/1 (allocated/max allocated)
Available Bandwidth 0 kilobits/sec
(depth/weight/total drops/no-buffer drops/interleaves) 26/4048/0/0/0
Conversation 7, linktype: ip, length: 36
source: 155.1.146.1, destination: 155.1.108.10, id: 0x1D13, ttl: 254, prot: 6
(depth/weight/total drops/no-buffer drops/interleaves) 4/16192/0/0/0
Conversation 19, linktype: ip, length: 160
source: 155.1.146.6, destination: 155.1.45.5, id: 0x0229, ttl: 254,
TOS: 32 prot: 17, source port 56694, destination port 16385
I'm assuming it isn't hitting the reserve queue based on the conversation number of 19. It should be 41 based on the dynamic+link queues.
Rack1R4#show queueing fair
Current fair queue configuration:
Interface Discard Dynamic Reserved Link Priority
threshold queues queues queues queues
Serial0/1/0 64 32 1 8 1
Comments
Hi, eventhough I am writing almost three years too late, but this problem description was exactly the same like I had miself few minutes ago and here was no solution provided. The thing is that legacy RTP reserve option assigns flow to reserved queue only if port number is even - same like myself you had UDP traffic to 16385 port. I f you fix it to 16384 it would work properly.
Hope it helps to anyone in the future!