Shard re-balance is happning with only incude_ip

Elastic Stack Elasticsearch
1

We have some issue one node and we remove the node by using Exclude_ip .Rebuild that nodes and successfully added that nodes with same ip amd hostname but when i am adding new nodes , re-balanced is not happening without incude_ip.

Any suggestion , how to proceed.

Thanks
Ashok

2

I am not sure I understand the question as you prefer to excluding a single node and then adding multiple nodes. Could you please explain the full sequence and what you see in greater detail?

The IPs to exclude are stored in the cluster state so when the node comes back to the cluster with the same IP it is still excluded. For that node, as you state, this need to be changed explicitly which I would say is expected. Does it also affect other nodes being added that have not previously been excluded?

3

Yes , it is affecting other nodes

4

Can you then please outline the exact sequence of actions so it can be reproduced?

5

When i have added again successfully without include command .But after restarting master node , it give me issue for whole cluster and also for new nodes.Now cluster is not balancing but it is in GREEN state.

"transient" : {
"cluster" : {
"routing" : {
"allocation" : {
"include" : {
"_ip" : "10.183.111.178"
},

6

Please provide some additional detail.

7

Please find the sequence of event

1: We have some issue with some nodes and have decommissioned the nodes with exclude command
2:Rebuild the decommission nodes with same ip and host-name
3:Add these nodes again by just as new nodes and that means without include command
4:Nodes are properly added and re balanced done
5: After some time , while adding additional new nodes , we need to restart the master nodes .New nodes are added but balanced not happened
6:To fix this i run include command
cluster.routing.allocation.include._ip":"New node 1 ip,new node 2 ip"
7:But this command added those nodes successfully and balanced starts with "node_concurrent_recoveries" but not as "cluster_concurrent_rebalance".This is issue .Cluster never balancing properly .It reached to water mark lavel and stopped the node_concurrent_recoveries.This created shards balanced improper and that means some may reached 200 shards but others are 100

Can i fix the issue by running ,
cluster.routing.allocation.include._ip":"*"

Thanks
Ashok

8

I think , below will slove the issue but i am not sure

"cluster.routing.allocation.exclude._ip" : ""

9

Any help please

10

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