Failing to ensureIndex on sharded Mongo

I run a mongoDb, sharding it on 4 different shards (+ 1 router and + 1 config). Sometime ago I had to replace a shard, and everything went ok. But now, every time I run:

db.summoners.ensureIndex( { plus: 1 } )

I get this error:

{
    "raw" : {
            "10.132.156.211:27017" : {
                    "numIndexesBefore" : 3,
                    "note" : "all indexes already exist",
                    "ok" : 1
            },
            "10.132.156.218:27017" : {

            },
            "10.132.145.51:27017" : {
                    "numIndexesBefore" : 3,
                    "note" : "all indexes already exist",
                    "ok" : 1
            },
            "10.132.157.195:27017" : {
                    "numIndexesBefore" : 3,
                    "note" : "all indexes already exist",
                    "ok" : 1
            },
            "10.132.137.214:27017" : {
                    "numIndexesBefore" : 3,
                    "note" : "all indexes already exist",
                    "ok" : 1
            }
    },
    "ok" : 0,
    "errmsg" : "{ 10.132.156.218:27017: \"result without error message returned : {}\" }"
}

The problem is that the given shard, 10.132.156.218:27017, does not exists anymore.

It's not even listed on the shard list!

{ "_id" : "shard0000", "host" : "10.132.156.211:27017" }
{ "_id" : "shard0002", "host" : "10.132.145.51:27017" }
{ "_id" : "shard0003", "host" : "10.132.157.195:27017" }
{ "_id" : "shard0004", "host" : "10.132.137.214:27017" }

What should I do?

PS:: MongoDB shell version: 2.6.5

--

mongos> db.summoners.stats()
{
    "sharded" : true,
    "systemFlags" : 1,
    "userFlags" : 1,
    "code" : 11002,
    "ok" : 0,
    "errmsg" : "exception: socket exception [CONNECT_ERROR] for 10.132.156.218:27017"
}