[ 
https://issues.apache.org/jira/browse/SOLR-12672?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16582832#comment-16582832
 ] 

Erick Erickson commented on SOLR-12672:
---------------------------------------

Interesting idea. Some random thoughts:
 * I can see "issues" if the process prevented indexing, might lead to 
recoveries that are worse than the original problem.
 * What about a "rolling disruption"? What I have in mind here is just 
preventing _queries_ from going to a node that's "disrupted". Something like 
"while the node is disrupted, don't use for serving queries". I'm thinking that 
you'd do some kind of rolling disruption and have all the _non disrupted_ nodes 
be the only ones that served queries. NOTE: I have no idea how 
easy/hard/practical this is....

 ** Not sure how to communicate this to other nodes, using ZooKeeper seems like 
it would lead to a zillion alerts being triggered.....
 ** Maybe what'd happen is the node that's disrupted somehow forwards the 
request somewhere else or replies immediately with "I'm busy, ask someone else".

I'm running from one place to another so I haven't thought about the 
implementation so feel free to ignore...

> Implement Synchronized Disruption into Solr
> -------------------------------------------
>
>                 Key: SOLR-12672
>                 URL: https://issues.apache.org/jira/browse/SOLR-12672
>             Project: Solr
>          Issue Type: New Feature
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Trey Cahill
>            Priority: Trivial
>         Attachments: Synchronized Disruption in Solr.pdf
>
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> On large Solr clusters, at any given time, there is probably an instance 
> running garbage collection.  By implementing a synchronized disruption across 
> the entire cluster, the response times of a large cluster should decrease as 
> it helps prevent random instances from running GC while the rest of the 
> cluster is responding to a request.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to