Index.routing.allocation.disable_allocation in Elastic Search 5.2

Elastic Stack Elasticsearch
#1

Hi,

I got a problem of many "unassigned shards". So through some reading I tryed to do :
index.routing.allocation.disable_allocation= false

This returned me:

index.routing.allocation.disable_allocation please check that any required plugins are installed

Looking around I didnt find such plugins. Also I didnt yet see a reference in ES 5.X for "index.routing.allocation.disable_allocation". Is there such an option in ES 5.2?

If no, what should I do?

Also if someone knows from what can this problem can happen, and what other solutions are there?

#2

Simply setting this won't help unless you have already disabled allocation.

So, why are these unallocated? What do your logs show? How many shards, are they primary or replica?

#3

I though I found a magic answer :slight_smile:

I dont know why, the machines that I am working on are not from our company but an outside one. From my understanding they copied the VM of the ES a few times, and copied the VMs of from where the filebeats sends the logs to ES

What logs?

I actualy don't understand enough in shards. So I can't answer the 3rd question.

When looking at the logs of logstash, I get:

error: "failed to derive xcontent" (but I am not sure this is connected)
(Thanks on the quick reply)

#4

What does _cat/allocation show?

#5

Since the machine I am working on is secure I cant copy paste from there or work with internet therefore I will not copy but type in the releavent parts:

I got a bunch of numbers then KppASWN
and then I got:
51 UNASSIGNED

#6

How many nodes in your cluster?

#7

When I did:

_nodes/process it shows me only one node:

"_nodes": {
"total" : 1,
... ),

cluster_name : elasticsearch,
nodes : {
"details of only one node"
}

#8

And what does _cat/health show?

#9

1490862395 "time" elasticsearch yellow 1 1 51 51 0 0 51 0 - 50.0%

#10

by the way:
when I typed in _cat /allocation:

one of the numbers that was in kppASWN... was 51 and also it was in the UNASSIGNED section

#11

Then everything is ok.

You cannot assign replica and primary shards to a single node, so to clear that up you need another node, or to remove the replicas.

#12

So if I open a new ElasticSearch instance (which opens a new node) this will fix the problem?

A few questions regarding this:

1- how to erase replicas?
2 - how did this happen how did replicas come upon?
3 - how opening a new node will solve it, does ES know how to allocate shards automatically?

#13

I probably will erase the replicas for a fast answer and build later a new node (another ES)

#14

Replica's are added by default and will be automatically assigned by ES if you add more nodes to the cluster.
You can dynamically change the number of replicas using something like;
curl -XPUT localhost:9200/*/_settings -d '{ "index" : { "number_of_replicas" : 0 } }'.

#15

ummm I got:

image.png1688×420 16.3 KB

Sorry that it is a picture and not the text (security reasons)

#16

Remove the -d.

1 Like
#17

(Also in curl I was unsuccesful

#18

Thanks! It is not 100% healthy!

#19

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