Steps to reproduce:
1. Setup a sharded cluster (to make this easier to reproduce, set chunk size to 1).
2. Keep on doing inserts and call get last error with a w parameter for each insert.
Eventually (specifically, right when an auto-split occurred), the shards field for the getLastError result would contain the config server and will have a corresponding object that looks like:
{
"n" : 0,
"connectionId" : 8,
"wnote" : "no replication has been enabled, so w=2+ won't work",
"err" : "norepl",
"ok" : 1
}
Some of drivers treat a non-null err field as an error and would throw an exception, which is bad.
- duplicates
-
SERVER-4874 unrecognized command: setShardVersion
-
- Closed
-
- is duplicated by
-
-
- Closed
-
-
-
- Closed
-
- is related to
-
-
- Closed
-