Questions to Determine

• What are the configured BGP hold timer and keepalive interval on both sides?
• Are the hold timer expirations correlating with periods of packet loss on the link?
• What do the interface error counters show (CRC errors, input errors, drops)?
• Is the physical link experiencing intermittent errors since the media converter change?
• Are BGP keepalive packets being lost due to link-layer errors, causing the hold timer to expire?
• What is the link utilization -- could congestion be delaying keepalives?
• Would increasing the hold timer to a more standard value (e.g., 90/30) reduce flapping?
• Is BFD (Bidirectional Forwarding Detection) configured, and if so, is it conflicting with BGP timers?
• Could the media converter be introducing packet loss or latency?
• What does the BGP event log show for the timing of UP/DOWN transitions?