[SOLR-9199] ZkController#publishAndWaitForDownStates logic is inefficient

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 6.2, 7.0
• Assignee: Hrishikesh Gadre
• Reporter: Hrishikesh Gadre
• Created: 08/Jun/16 21:38
• Updated: 26/Aug/16 13:58
• Resolved: 09/Jun/16 20:27

Description

The following logic introduced as part of SOLR-8720 is inefficient.

https://github.com/apache/lucene-solr/blob/6c0331b8309603eaaf14b6677afba5ffe99f16a3/solr/core/src/java/org/apache/solr/cloud/ZkController.java#L687-L712

Specifically,

• foundStates flag is set to TRUE before the for loop.
• In the for loop we check if any replica on this node is not in the DOWN state. If yes, then foundStates = FALSE
• If foundStates == TRUE then we break out of the loop and return.

The problem here is that once foundStates is set to FALSE, it is never reset to TRUE. Hence we end up spending the whole 60 secs iterating over the collections even though all the replicas are marked as DOWN in later iterations.