Problem with unassigned Shards


(Andreas Wachter) #1

Hello,

I came to work from my holidays today and have some troubles with my shards.
> {
> "cluster_name" : "elasticsearch",
> "status" : "yellow",
> "timed_out" : false,
> "number_of_nodes" : 5,
> "number_of_data_nodes" : 4,
> "active_primary_shards" : 5535,
> "active_shards" : 5545,
> "relocating_shards" : 0,
> "initializing_shards" : 0,
> "unassigned_shards" : 11060,
> "delayed_unassigned_shards" : 0,
> "number_of_pending_tasks" : 0,
> "number_of_in_flight_fetch" : 0,
> "task_max_waiting_in_queue_millis" : 0,
> "active_shards_percent_as_number" : 33.3935561577838
> }

Any ideas on how to fix that?


(Mark Walkom) #2

You really need to reduce that shard count, it's way too high!

However check your master logs for why.


(David Pilato) #3

Have a look at the pending tasks and your logs.

That being said you have a way too much shards per node IMO.


(Andreas Wachter) #4

I found the problem my diskspace-watermark was too low.

How can I reduce it? I can't change the number shards once they are configured, if I remember it correct.


(Mark Walkom) #5

If you are on 5+ you can use _shrink.


(system) #6

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.