for me this looks like instability in your network. is this LAN? which speed?
Did you put a trace for some time ?
some ping loops checking response times
what happens if you increase the onconfig parameter LISTEN_TIMEOUT ?
It is a LAN. The server is located on a different VLAN, clients are in another one. There is a layer 3 switch which routes both VLAN, there are not ACL configured.
I sniffed the network when the problem arrives, the only weird thing I see is lot of checksum errors packets from the server. We disabled the checksum offload in the NIC, but the problem doesn´t disappeared.
Well, during weekend we proceeded with the new format in the server HDD and reinstallation of informix....without satisfactory results....
The same symptom. This time, we also changed the server´s location to another subnet (it is located in a branch office), so users tried to connect to the server using a different IP but the problem still is there.
The LISTEN_TIMEOUT is configured to 60.
could be related with the replication system ?
We have two servers, the primary (server with problem) and a secondary which is working right now. So far, both server have the same configuration, the only difference would be that secondary server works in standalone (when primary is disconnected from network).