-
Type:
Bug
-
Resolution: Cannot Reproduce
-
Priority:
Major - P3
-
None
-
Affects Version/s: 3.2.0
-
Component/s: mongodump
-
None
-
Environment:Windows 7
"mongodump.exe" -u XXX -p XXX --authenticationDatabase admin 2015-12-08T18:19:35.281-0700 Failed: error counting admin.system.indexes: all servers down/unreachable when querying: configsvr1:27001,configsvr2:27002,configsvr3:27003
The database in question was upgraded from 3.0.7, all server binaries are now running 3.2.0.
mongodump connects to a MongoS, which records the following errors on that connection:
2015-12-09T21:08:48.804-0700 I NETWORK [conn2] query on admin.$cmd: { query: "1", help: 1 } failed to: configsvr1:27001 (192.168.0.1) exception: "query" had the wrong type. Expected Object, found String
2015-12-09T21:08:48.804-0700 I NETWORK [conn2] query on admin.$cmd: { query: "1", help: 1 } failed to: configsvr2:27002 (192.168.0.1) exception: "query" had the wrong type. Expected Object, found String
2015-12-09T21:08:48.804-0700 I NETWORK [conn2] query on admin.$cmd: { query: "1", help: 1 } failed to: configsvr3:27003 (192.168.0.1) exception: "query" had the wrong type. Expected Object, found String
EDIT: note the first error message is misleading. None of the config servers are down, and the cluster is fully operational in all other regards.