- Newest
- Most votes
- Most comments
Based on the error logs showing "Server poll timeout. failed" messages, this appears to be a network connectivity issue despite the successful endpoint tests. This type of error typically indicates that the replication agent is unable to maintain a stable connection with the replication server.
Several conditions could cause these errors even when initial endpoint tests succeed:
-
Intermittent network connectivity issues: While the initial connection tests may succeed, the actual replication process requires a sustained connection. Temporary network disruptions, latency spikes, or packet loss could cause these timeout errors.
-
Network configuration problems: Security groups, network ACLs, route tables, or gateway configurations might allow initial connectivity tests to succeed but block sustained traffic patterns required for replication.
-
Firewall or proxy interference: There might be firewalls, proxy servers, or SSL offloading rules that are intercepting or blocking the sustained connection needed for replication, even though they allow the initial connectivity test.
-
SSL/TLS handshake issues: The initial connection might succeed, but there could be problems with maintaining the SSL/TLS connection over time.
-
Resource constraints: The replication server might be experiencing high load or resource constraints, causing it to be unresponsive to the agent's requests.
To troubleshoot, you could:
- Capture network traces during the replication attempt
- Check for any proxy servers or firewalls that might be interfering
- Review all network components between the source and target
- Perform extended SSL connectivity tests between your source environment and the replication server
Sources
Why am I receiving a "FAILED TO CONNECT AGENT TO REPLICATION SERVER" error in my AWS MGN or AWS DRS replication initiation steps? | AWS re:Post
answered a year ago
Relevant content
- AWS OFFICIALUpdated 3 months ago
There are 11 rows involved - there are no resource constraints.
Can anyone provide an example where this "Network configuration problems: Security groups, network ACLs, route tables, or gateway configurations might allow initial connectivity tests to succeed but block sustained traffic patterns required for replication." is a possibility. ACLs either work or they don't. SG's either work or they don't.
Either the target and source tests are completely bogus or this advice is. Both cannot be sound