Bad cluster slots output after migration + node crashed

[er...@intentiq.com]

Hi,

We see the following output in redis-cli cluster slots command : 

.....

12) 1) (integer) 8868

     2) (integer) 8872

     3) 1) "172.31.14.11"

        2) (integer) 6379

        3) "076aa8bd853703507df1a0303059bfc54a63cda1"

     4) 1) ""

        2) (integer) 0

        3) "a3b2aac20c56a8499b596ed425525eaa112d2a97"

 13) 1) (integer) 9961

     2) (integer) 9965

     3) 1) "172.31.14.11"

        2) (integer) 6379

        3) "076aa8bd853703507df1a0303059bfc54a63cda1"

     4) 1) ""

        2) (integer) 0

        3) "a3b2aac20c56a8499b596ed425525eaa112d2a97"

 14) 1) (integer) 11053

     2) (integer) 11057

     3) 1) "172.31.14.11"

        2) (integer) 6379

        3) "076aa8bd853703507df1a0303059bfc54a63cda1"

     4) 1) ""

        2) (integer) 0

        3) "a3b2aac20c56a8499b596ed425525eaa112d2a97"

....

It seems this cause Jedis to think there is a node "" with hostport ':0'

I think this cluster state happened after after a node crashed while transfering its slots to another.

Initially the slots of the node where busy, but we asaign them to another node (cluster setslot) and since then the cluster works fine except the above odd output.

Can this issue be fixed?

Thanks,

Erez