[Aureanus]

We have a remote site that is having some odd connection issues.

When users from this site connect to our exchange server (and a another reports server) they cannot pull down email but can send it.

Investigation showed that whilst the client on the remote site could browse the server using the FQDN or short name, the server couldn't do the same to the client - but could if the IP address was used.

Further investigation showed that on both the servers a /32 route to the client PC was been added to the routing table. Removing the route allowed the connection to be resorted (and the /32 route would then add itself back into the table) for 10 minutes.

We found that if we statically added a route into the table, that was identical to the one we removed then the client could always connect. As there are multipule clients at this site we tried adding a /24 route and removing the /32, but found that the client could not then pull emails down - they could only do it if a /32 is in place.

The entries from the routing table are below

Network Destination Netmask Gateway Interface Metric
0.0.0.0 0.0.0.0 172.22.3.247 172.22.0.13 10
10.55.2.0 255.255.255.0 172.22.3.247 172.22.0.13 1

Default Gateway: 172.22.3.247

It only works when we add the line below into the table:

10.55.2.8 255.255.255.255 172.22.3.247 172.22.0.13 1

I am baffled by this behaviour as they all point to the same gateway.