I have a strange problem here, maybe someone had the same or a similar problem and can give me some hints how to solve it. I have a host on my network, which won't reply to ping requests sporadically. According to the responsible admins, this problem is getting worse since we enabled routing on our 5412zl, which was done by an old cisco device before. As they blamed me being responsible for this, I configured a mirror for the affected switchport to make sure that there is no packet loss on the network at all. As I expected, all the ping-request packets came out at the mirrorport. The host simply didn't answer all of them. The other way round, if I ping something else from the host, the console says that there are timeouts - although I can see a ping reply for every request on the mirrorport.
There are absolutely no such problems at the entire rest of the network, but this host is making me mad! Additionally the problem also occurs if I am in the same subnet as the host, so I dont think it's a routing issue at all.
Obviously, if I connect my laptop to the host via a 8-port noname switch there are no such problems while pinging the host. Unfortunately, I can't do a packet trace on the host itself as it's an ethernet-adapter of an ibm zseries.
Maybe anyone has an idea what this problem could be based on, or how i should further investigate?
We seem to have similar problems with the 5406zl, only for multicast packets. Using an MPEG probe I see packet loss behind some 5406. Can't see anything on the interfaces. We will replace these units (20-30 units) with other vendors if we can't find any solution.
Paulen - I'd suggest you start a new thread to enquire about your problem, you're likely to get more help that way.
zmichl - Do you have the problem when you are pinging any host to/from the IBM server? The first thing I would check is if there are any errors reported on the port ("show interface x") and that the duplex settings match from the switch to the server ("show interface brief x" and "show interface conf")
I already checked port counters, there are only a small amount of drops rx - no collisions or something. both ports are set up to do auto negotiation. The cable itself has to be ok, as we renewed it last christmas. Atm, I am up to get someone which changes the interface of the host to fixed port-setting - maybe theres something wrong with the auto negotiation of the host...
