Can someone explain what TCP Listen Queue Size actually does for my listener? My issue is we are having a few (maybe 10 out of 10,000) TLS failures from Exchange to the ESA's which are relayed outside the ORG.
The TLS error in the logs from Exchange are the following:
TLS negotiation failed with error BufferNotEnough |
TLS negotiation failed with error MessageAltered |
So my question is if its possible the queue size is to small causing the handshake to break during TLS negotiation. The actual NDR bounceback states STARTLS must be initialized before accepting mail which tells me TLS was not accepted during this handshake.