Use case of UTILIZENODE API

[ChienHuaWang <ch...@sap.com>]

While exploring the UTILIZENODE API to move replicas, it would depend on the
preferences & autoscaling policies defined.
But wondering what's the priority for its decision? Let's say I define
maximize freedisk & minimize heapUsage, and also set-cluster-policy as
example in doc. {"replica": "<2", "shard": "#EACH", "node": "#ANY"}
 
Would it check violation first? How it pick replica to delete and add?
Sometimes I observed "No node can satisfy the rules", even I tried to move
to a new node.

If I want to rebalance for new/existing node, what's the best approach? It
seems I have limited control of UTILIZENODE API, would this be suggested to
use in production? 

Appreciate if anyone can share the use case of UTILIZENODE API to move
replicas.




--
Sent from: https://lucene.472066.n3.nabble.com/Solr-User-f472068.html

Re: Use case of UTILIZENODE API

[ChienHuaWang <ch...@sap.com>]

For someone who is interested in this API

It seems a known issue that UTILIZENODE does not enforce policy rules as
found in Jira
https://issues.apache.org/jira/browse/SOLR-12050

Please feel free to comment.



Regards,
Chien



--
Sent from: https://lucene.472066.n3.nabble.com/Solr-User-f472068.html