5.1.1 : All shards unassigned?

[dawiro]

Hi,
I have a new cluster up and running and when I started up logstash it started sending heartbeat data through. However, all shards (primary and replica) for the newly created index remain unassigned with no reason code shown.
Shard allocation is enabled so I'm confused by the turn of events. Can someone throw some light on this for me? Is there some other step required to get the cluster managing these shards?

Regards,
David

[warkolm (Mark Walkom)]

Do you have enough disk space?
Are all nodes of the same version?

[dawiro]

Hi,
Yes, there's 1TB per data node and all nodes are same version. I've just fired up kibana and see this:

# curl "$(hostname -I):9200/_cat/shards?pretty"
.kibana 0 p UNASSIGNED
.kibana 0 r UNASSIGNED

...on one of the data nodes. I'm not seeing anything logged yet to give me a clue as to what's going on.

Regards,
David

[dawiro]

On the master node I see:

[2017-01-05T14:36:08,242][INFO ][o.e.c.m.MetaDataCreateIndexService] [es_az1_master_001] [.kibana] creating index, cause [api], templates [], shards [1]/[1], mappings [server, config]
[2017-01-05T14:36:08,286][INFO ][o.e.c.r.a.AllocationService] [es_az1_master_001] Cluster health status changed from [YELLOW] to [RED] (reason: [index [.kibana] created]).

[dawiro]

Ah, it turns out that zone identification hadn't been set due to changes in 5.x. Looking better now...

[Archit_Saxena]

hey, what do you mean by zone identification? Can you please elaborate? I have the same problem