1

We have 2 Alcatel 9600 core switches in a redundant configuration and 2 distribution switches (Alcatel 6850's) on another floor connected by fibre to the core switches. See the diagram below.

Pinging 10.46.1.6 from 10.46.1.5 I get packet loss. Pinging 10.46.1.8 from 10.46.1.5 I get packet loss. Pinging 10.46.1.9 from 10.46.1.5 I get packet loss.

Pinging 10.46.1.5 from 10.46.1.6 I get packet loss. Pinging 10.46.1.8 from 10.46.1.6 I get NO packet loss. Pinging 10.46.1.9 from 10.46.1.6 I get NO packet loss.

So the packet loss is only on the 2 fibre links from 10.46.1.5 and the LAG connection.

The green connection between the 9600's is LAG. The blue link as mentioned to the 6850's is MM fibre. The possibility that there is physical damage on both the LAG and MM fibre is remote in my opinion. Nothing has changed in the configurations and the "show health" is within range for 10.46.1.5 (See second image)

What could be the reason for the packet loss and how do I resolve it? My first move will be to restart the switch 10.46.1.5. Is there a way to diagnose/check the interfaces health via the CLI?

The switch is in working mode. Before rebooting, how do I confirm that the working configuration is saved and certified, because I have read that if the working and certified configuration files are different, the switch will boot in certified mode.

I am new to working on Alcatel switches and was not original involved with the installation and configuration of the switches. Any help would be appreciated.

Core and distribution switches Show Health for 10.46.1.5

1 Answer 1

0

Personally - I find it's rare that rebooting a switch fixes much of anything. Especially when it's something like intermittent packet loss - in my experience this is usually something with a physical root cause.

I would suggest checking your fibers - I've had something very similar that caused no end of grief, before we physically traced the fibers and found that someone had pulled on tight around a kink. Not enough to break it, but enough that it's attenuation - and thus error rate - increased significantly.

Cable tracing can be a chore, so the alternative and proof-of-diagnosis is to run a second cable along the floor, and plug it in to you two end points. My second port of call would be checking the SFPs (Small formfactor pluggable modules, sometimes referred to as GBIC) on your switch - they're small and replaceable, and are the other common-ish source of intermittent transmission problems.

3
  • After some further investigation I do not have a clue what is going on. Even just plugging directly into the switch and pinging it give me packet loss.
    – Rynardt
    Apr 16, 2014 at 13:40
  • Do you have spare SFPs you can swap out?
    – Sobrique
    Apr 16, 2014 at 13:55
  • I plugged directly into copper ports on the switch and had 30-40% packet loss. Could this be due to network congestion?
    – Rynardt
    Apr 17, 2014 at 9:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .