A few days ago I rewired two racks at my local datacenter. They are side by side and have about 40 servers.
I was using an old office type switch and two Dell PowerConnects 5324 and I wanted to remove the old switch (which had about 80% of the machines hooked to it) and rewire everything to use the two PowerConnects instead (which had about 20% of the machines hooked up to them).
Basically what's happening is that I can't SSH/FTP/Ping/etc from the servers (locally) to each other, but ALL servers are accessible remotely. Some will connect fine, but others won't (error: no route to host). It's extremely odd because the servers are all connected to the two Dell PowerConnect 5324 Switches, and the two switches are connected to each other. Most servers can access each other but some, seems very randomly picked, can't connect to other random servers even if they are on the same switch.
Ex:
Same Switch, Rack #1:
Server 1 can connect to Server 2 and Server 3 Server 2 can connect to Server 2 but not Server 3 Server 3 can connect to Server 2 but nto Server 3
My guess is that the PowerConnects have some type of a route cache, which is telling them that the old (office style) switch is handling the traffic for the request.
I've rebooted the switches, and router and it made no difference.
I'm stuck scratching my head here and I would really appreciate some feedback. Is there some type of cache on here that I can clear? Could it be the router doing this?
No settings were changed on the servers, router or switch. Firewalls are not causing this.
Thanks for your help, Luc