Is this normal

[aali]

Are these port statistics normal?  Customer has  a Procruve 5412 connected to HP 2810G switch over a 1Gb copper link.  HP 2810G is connected to servers and Storage, and 5412 has most of the users connections and a few servers.  Network is pretty slow.  I wanted to check the connection between 5412 and 2810 and below are the statistics

P-E5412zl# show interfaces ethernet a7
hc Show summary of network traffic handled by the ports.
<cr>
HP-E5412zl# show interfaces ethernet a7 hc

-- MORE --, next page: Space, next line: Enter, quit: Control-C

Status and Counters - Port Counters for port A7

Name : HP 2810G
MAC Address : 001c2e-ef70f9
Link Status : Up
Totals (Since boot or last clear) :
Bytes Rx : 41778325286874 Bytes Tx : 19926808863407
Unicast Rx : 35406647268 Unicast Tx : 26028956748
Bcast/Mcast Rx : 58,477,940 Bcast/Mcast Tx : 677,011,749
Errors (Since boot or last clear) :
FCS Rx : 0 Drops Tx : 2,130,797
Alignment Rx : 0 Collisions Tx : 0
Runts Rx : 0 Late Colln Tx : 0
Giants Rx : 0 Excessive Colln : 0
Total Rx Errors : 0 Deferred Tx : 0
Others (Since boot or last clear) :
Discard Rx : 3356 Out Queue Len : 0
Unknown Protos : 0
Rates (5 minute weighted average) :
Total Rx (bps) : 19,853,120 Total Tx (bps) : 6,362,608
Unicast Rx (Pkts/sec) : 1,926 Unicast Tx (Pkts/sec) : 1,141
B/Mcast Rx (Pkts/sec) : 2 B/Mcast Tx (Pkts/sec) : 39
Utilization Rx : 01.98 % Utilization Tx : 00.63 %

[YenLin]

Hi aali, I would suggest you to show ip int brief to check this link's speed and duplex.

Basically, I believe one peer is Giga bit, then the link can negotiate to full and 1G,

( you can config the port to full and static speed, if they can't negotiate to full duplex and 100 M or 1G)

But, I have no idea why HP switch always have this situation.

If you check this like is 1G speed with full duplex, can you try to change this link to another port?

And try to use ipef to test this link.

It's abnormal to see too much Drop rate and Discard rate for a link as the information you pasted.

Good luck!

[Vince_Whirlwind]

There 3 things you can see from those port stats:

1./ Current utilisation is very low

2/ Stats have been collecting over a loing period of time

3/ Some very small proportion of your total overall traffic is showing up as TX drops that have occurred at some stage during that time.

You can't really draw any conclusions from that.

I'd suggest you clear your stats, and have another look at them during period of high utilisation.

This is where a network monitoring tool is necessary - something that will use SNMP to collect stats on utilisation and errors over time, using short (3s maybe) time intervals.

Failing that, put a continuous ping over the link: you should get 1ms replies and no packet loss.

[aali]

Thanks for the feedback.  I will clear the stats and go from there.