-
Type: Bug
-
Resolution: Unresolved
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
Replication
-
ALL
-
200
When we restart a node we normally wait for the first connection to that node to be healthy. We do that wait with the unbridged connection. However, the bridge can sometimes become confused during a restart and return a "stale" disconnection to the next command which comes up. We should be able to work around this by, if using a bridged test, making sure the bridged connection works after waiting for the unbridged connection.
It should be possible to turn this off with an option.
Edit: It seems like it's not a trivial task to make sure that the bridged connection works after waiting for the unbridged connection, as it would require more work in mongobridge itself. One other solution would be changing the test to use failCommand rather than mongobrige, but this would require more work in the rollback_test.js library to do so.
- is related to
-
SERVER-77987 Add comments on mongobridge's limitation with restarting nodes
- Closed
- related to
-
SERVER-96260 Fix mongobridge's limitations with restarted nodes
- Backlog