-
Type: Task
-
Resolution: Done
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
Labels:
The page on troubleshooting Replica Sets needs additional information.
Things to include are:
1) Running 'rs.status()' on all nodes in the set to check the status of replication
2) Running 'db.printReplicationInfo()' on all nodes in the set to check the size of the oplog
3) Running 'db.printSlaveReplicationInfo()' on all nodes in the set to check the replication delay
4) Using 'mongo' to test the connection from every node to every other node (both directions)
Some notes about this last:
- There must be connectivity from every node to every other node in order for replication to work
- Problems with networking or firewall rules can prevent this connectivity, and hence prevent replication from working
- Given that there are three nodes in the set, one node running on each of hosts A, B, and C, the process would be:
- On host A:
- Run "$ mongo --host B --port 27017"
- Run "$ mongo --host C --port 27017"
- Repeat the process on hosts B & C (each trying to connect to the other two hosts)
- If any of the connections fails, there's a networking or firewall issue that needs to be diagnosed separately
- On host A: