N9K-C9348GC-FXP - Sub-Interface on 10G-LR slow transfers

[kkillby]

Hi,

So we have recently installed a N9K-C9348GC-FXP and utilised the sub-interface command with encapsulation dot1q which we know is supported.

However when testing this circuit we are finding that iperf is only giving us about 50.3Mbps on a 1G circuit over the 10G NNI Interface.  Our circuit provider as done intrusive testing up to our cabinet and shown that the port is capable of 10G and the dot1q vlan is capable of 1G

We are using a Single Mode Fibre on the 10G-LR SFP (Recently changed for another new cable)

But whilst our SMARTNET is being dealt with wondering if anyone has seen this behaviour or might know why this is so?

Thanks

[Georg Pauwen]

Hello,

can you post the output of 'show interfaces x' where 'x' is the subinterface ?

[kkillby]

Please see below:

Ethernet1/52.72 is up
admin state is up, Dedicated Interface, [parent interface is Ethernet1/52]
Hardware: 1000/10000/25000 Ethernet, address: 00b7.7150.30b7 (bia 00b7.7150.30eb)
Description: *** xxxxxxxx ***
Internet Address is x.x.x.x/30
MTU 1500 bytes, BW 1000000 Kbit, DLY 10 usec
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation 802.1Q Virtual LAN, Vlan ID 3072, medium is broadcast
Auto-mdix is turned off
EtherType is 0x8100

[Georg Pauwen]

Hello,

is that the full output ? I am looking for anything related to queueing and input. It looks like the output is truncated...

[kkillby]

Hey,

I know but that is the full output that this particular model of 9K gives you for sub-interfaces, below is the parent Interface:

Ethernet1/52 is up
admin state is up, Dedicated Interface
Hardware: 1000/10000/25000 Ethernet, address: 00b7.7150.30b7 (bia 00b7.7150.30eb)
Description: *** xxxxxxxxxx ***
MTU 1500 bytes, BW 10000000 Kbit, DLY 10 usec
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, medium is broadcast
full-duplex, 10 Gb/s, media type is 10G
Beacon is turned off
Auto-Negotiation is turned on FEC mode is Auto
Input flow-control is off, output flow-control is off
Auto-mdix is turned off
Rate mode is dedicated
Switchport monitor is off
EtherType is 0x8100
EEE (efficient-ethernet) : n/a
Last link flapped 1d01h
Last clearing of "show interface" counters 5w6d
4 interface resets
30 seconds input rate 43303456 bits/sec, 5507 packets/sec
30 seconds output rate 72631112 bits/sec, 7634 packets/sec
Load-Interval #2: 5 minute (300 seconds)
input rate 22.10 Mbps, 3.73 Kpps; output rate 83.95 Mbps, 7.89 Kpps
RX
5239218982 unicast packets 1762259 multicast packets 19528 broadcast packets
5241000769 input packets 5223973677757 bytes
2672406676 jumbo packets 0 storm suppression bytes
0 runts 1 giants 1 CRC 0 no buffer
2 input error 0 short frame 0 overrun 0 underrun 0 ignored
0 watchdog 0 bad etype drop 0 bad proto drop 0 if down drop
0 input with dribble 0 input discard
0 Rx pause
TX
4254980248 unicast packets 1232271 multicast packets 5155 broadcast packets
4256217674 output packets 3824729763029 bytes
1501924265 jumbo packets
0 output error 0 collision 0 deferred 0 late collision
0 lost carrier 0 no carrier 0 babble 0 output discard
0 Tx pause

[Reza Sharifi]

Hi,

Do you see a difference when you don't use a sub-interface?

Can you try the same test with a regular 1Gig interface?

HTH

[kkillby]

With this particular Interface no as it is a dot1q from the provider.

However we have other circuits plugged directly into their own ports and they seem to be working fine.

[Georg Pauwen]

Hello,

what is connected on the other side ? You could try and globally enable 'ip tcp path-mtu-discovery' on the switch...

[kkillby]

Hey,

It's our circuit providers 10G port to us.  Thanks something additional to add to our testing.

Cheers