Today i experienced a weird situation and wanted to get your opinions what kind of reason could create such situation.
Today one of our 9 Cisco switches a "WS-C2950S" (we also got 2 other WS-C2950S on same network) stop responding icmp ping packages. When i tried to telnet the switch its network was unreachable but i was able to see its existance from other switches by "sh cdp neig". So i decided to fix the situation on a suitable night time work, checking by console cable or even rebooting the device.
But this morning i saw the error resolved itself right after exact 23 hours of time:
Tue 03/06 8:57 AM Ping Up
Mon 03/05 9:57 AM Ping Down
Then i started to wonder... what this could possibly be about??
We have like 40 clients behind that switch and there was no communication problem during the problem.
I would need to know a bit more about the switch, like is its management on the same VLAN as everything else, where were you pinging from, etc. Was it unreachable from everywhere, or just from your machine? My guess would be something to do with the ARP cache. Next time it happens, clear down the ARP cache on your machine, or on the router if it is on a different VLAN.
Do you have any routers ion your topology that could be doing proxy ARP?
Don't forget that a switch is switch, and will still work even if its management is unavailable at layer-3.
Indeed we have like 15 vlans and if switch management port is on vlan A, monitoring system is on vlan B and my client is on vlan C. Those vlans got communication normally and all other switches mng ports are also on Vlan A.
I also have neighbour router but those switches and router got 1 year+ uptime on monitoring system logs which makes me think that could be a new problem. On the other hand theres no new config change on any device.
As alternative a rogue device taking same ip could cause such problem if its not pingable... but that vlan A is not switchport accessed to no interface at all.
I ll check about "proxy arp", need some reading.