Switches, Hubs, and Modems
cancel
Showing results for 
Search instead for 
Did you mean: 

Ping problem in certain VLANs

Bjorn Tore Paulen
Frequent Advisor

Ping problem in certain VLANs

I have a large fibre network based on 5304s (~100). Within the admin VLAN I can ping without packet loss at all. But on certain VLANs I get ~4-6% packet loss on ~10000 packets. This VLAN is one with two hosts (/30) only. The admin-VLAN is /20. IP addresses is set on the switches.

Any clues anyone? Tips on troubleshooting?
5 REPLIES
Gonzo Granello
Valued Contributor

Re: Ping problem in certain VLANs

Some more details would be nice.... Assuming that you run GIG fiber, there should be no loss unless that one VLAN is a; very busy and b; going over a link that is heavily used by other VLans. That alone is also not really a reason for loss, however some other factors (like backplane usage / utilization) or simply oversubscription by using more than 4 Gig into one module could be the reason. For trouble shooting, i recommend as a first step to just look at the counters and find out what the bandwith usage of teh links in question is, might be wise to "group" some connections on the blades involved rather than have those going to the backplane. Also a good think to check, are you using trunked links ? If so, make sure they are not oversubed since they do not really aggregate bandwith as most people think they do. It is very possible to oversub them, since once a connection is established it will stay on that physical link, hence running out of bandwith will drop packets. Of course as always - check FW versions and make sure they are the latest and the same if possible.

Hope that helps - or at least point you in the right direction....

///GG
most time the day i have to mask my contempt for the a-holes in charge......
Bjorn Tore Paulen
Frequent Advisor

Re: Ping problem in certain VLANs

Thanks.

I use 1000BaseLX all over. Traffic <20Mbps. No error counters. No link-loss. No trunks. Only L2, ~30Vlans many in which only 2 hosts (/30).

Could seem like a ARP update /MAC issue. It's just so random.

Have involved HP Tech - will keep you posted.
OLARU Dan
Trusted Contributor

Re: Ping problem in certain VLANs

1. About Layer 1:

Check the fibre patch panels. If you have 62.5 micron core fibre you should have 62.5 micron pigtails and 62.5 micron patchcords. If you have 50 micron core fibre you should have 50 micron pigtails and 50 micron patchcords.

Don't laugh: I inherited a few 62.5 micron fibre cables having 50 micron pigtails; when using 62.5 micron patchcords I could not get anywhere, and when usin 50 micron patchcords I got randomly disconnected - just like you say.

2. About 1000LX transceivers, on multimode fibre (both 50 and 62.5 micron):

- don't use them for distances of a few tens of meters (I use them over 100 meters)
- you can go to maximum 300 meters if you don't use Mode Conditioning Patchcords (MCP).
- you can go to maximum 550 meters if you use MCP.
- if you use MCP: make sure you use one at each end, and make sure the core diameter of the MCP matches the one for the fibre cable and pigtails.
Bjorn Tore Paulen
Frequent Advisor

Re: Ping problem in certain VLANs

In this network we are only using singlemode fiber (9/125). I can`t really see that the fiber itself should be an factor in this problem. The link stays up all the time, the problem affects really just one VLAN.
Tyson_1
Occasional Contributor

Re: Ping problem in certain VLANs

We are experiencing a very similiar problem. In our case, the devices are on the same VLAN, but a different subnet. We notice the problem usually because our monitoring software will indicate that the device is unreachable. When checking, in every case, the 5300 switch will report an incorrect MAC address for the unreachable device. The MAC address that is always shown is the MAC address of our ISP router. After a few minutes, the 5300 will drop the incorrect ARP entry and everything will be fine again. If you receive a resolution from HP Support, please post it to the forum!