Netflow on Nexus 3048

[mramj499]

I'm trying to enable Netflow on our 3048s.  I set up the sflow using these commands:

conf t
feature sflow
sflow sampling-rate 4096
sflow counter-poll-interval 60
sflow data-source interface e1/15-16
sflow collector-port 6343
sflow collector-ip 10.10.128.125 vrf management mgmt0
sflow agent-ip 10.11.128.12

When I run "sh sflow", here's the output, but it looks like the collector IP is changing to 0.0.0.0 as well as agent-ip is changing to that too.  

HQN-C3048-R1-2# sh sflow
sflow sampling-rate : 4096
sflow max-sampled-size : 128
sflow counter-poll-interval : 60
sflow max-datagram-size : 1400
sflow collector-ip : 0.0.0.0 , vrf : management
sflow collector-port : 6343
sflow agent-ip : 0.0.0.0
sflow data-source interface Ethernet1/15
sflow data-source interface Ethernet1/16

And when I run "sh sflow stat", this is what I get:

HQN-C3048-R1-2# sh sflow stat
Total Packets : 62078969896
Total Samples : 0
Processed Samples : 0
Dropped Samples : 0
Sent Datagrams : 0
Dropped Datagrams : 93

HQN-C3048-R1-2# sh run | sec sflow
feature sflow
sflow counter-poll-interval 60
sflow collector-ip 10.10.128.125 vrf management
sflow data-source interface Ethernet1/15
sflow data-source interface Ethernet1/16

Here's "show inventory"

HQN-C3048-R1-2#

NXOS compile time: 7/5/2018 11:00:00 [07/05/2018 15:57:12]

Hardware
cisco Nexus 3048 Chassis
Intel(R) Celeron(R) CPU P4505 @ 1.87GHz with 3903336 kB of memory.
Processor Board ID FOC20223XL0

Device name: HQN-C3048-R1-2
bootflash: 1880320 kB
usb1: 0 kB (expansion flash)

Kernel uptime is 481 day(s), 8 hour(s), 40 minute(s), 14 second(s)

Last reset at 951983 usecs after Wed Feb 13 22:58:10 2019

Reason: Disruptive upgrade
System version: 6.0(2)U6(8)
Service:

plugin

Nothing is actually reaching my Solarwinds NPM server at 10.10.128.125.  I have proper route in place to that NPM server and can ping it:

HQN-C3048-R1-2# ping 10.10.128.125 source-interface mgmt 0
PING 10.10.128.125 (10.5.128.125): 56 data bytes
64 bytes from 10.10.128.125: icmp_seq=0 ttl=124 time=1.407 ms

Don't know what I'm missing with my config on 3048.  Any ideas anyone?  Any help/suggestions are appreciated.

[Reza Sharifi · Accepted solution]

sflow collector-ip 10.10.128.125 vrf management mgmt0

Are you using the mgmt0 interface to send data to NPM?

Can you try a regular port (not in vrf management) and test with?

HTH 

[mramj499]

Instead of mgmt0, I had to put 10.11.128.128.  Now, Solarwinds NTA is saying Netflow info detected and the interfaces added, but it's not detecting any Netflow traffic.

Do you see anything wrong with my config?

HQN-C3048-R1-2# sh run | i sflow
feature sflow
sflow counter-poll-interval 60
sflow collector-ip 10.10.128.125 vrf management source 10.11.128.128
sflow agent-ip 10.11.128.128
sflow data-source interface Ethernet1/15
sflow data-source interface Ethernet1/16
sflow data-source interface port-channel52

HQN-C3048-R1-2# sh sflow
sflow sampling-rate : 4096
sflow max-sampled-size : 128
sflow counter-poll-interval : 60
sflow max-datagram-size : 1400
sflow collector-ip : 10.10.128.125 , vrf : management source 10.11.128.128
sflow collector-port : 6343
sflow agent-ip : 10.11.128.128
sflow data-source interface Ethernet1/15
sflow data-source interface Ethernet1/16
sflow data-source interface port-channel52

[Reza Sharifi]

The config looks correct. 

What if you try one interface as the source like this example?

feature sflow
sflow sampling-rate 5000
sflow max-sampled-size 200
sflow counter-poll-interval 100
sflow max-datagram-size 2000
sflow collector-ip 192.0.2.5 vrf management
sflow collector-port 7000
sflow agent-ip 192.0.2.3
sflow data-source interface ethernet 1/5

[mramj499]

Never mind, figured out the Solarwinds issue.  Your suggestion was 99% there.  I could not put in mgmt0.  Had to put in the source IP for mgmt0.

[Reza Sharifi]

Glad to help and know it is all good!