You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Shawn Heisey (Issue Comment Edited) (JIRA)" <ji...@apache.org> on 2011/11/22 18:32:40 UTC

[jira] [Issue Comment Edited] (SOLR-2906) Implement LFU Cache

    [ https://issues.apache.org/jira/browse/SOLR-2906?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13155280#comment-13155280 ] 

Shawn Heisey edited comment on SOLR-2906 at 11/22/11 5:32 PM:
--------------------------------------------------------------

The only static warming I have is an all docs search with a sort parameter (and no filter query), which precaches my sort.  I do have autowarming configured, but this behavior happens from initial solr startup.  Things seem to behave correctly with warming - size is autowarmCount, inserts are zero.

FastLRUCache behavior:
queryResultCache: size is one higher than inserts
documentCache: size is one higher than inserts
filterCache: size and inserts are identical

LFUCache behavior:
queryResultCache: size is one higher than inserts
documentCache: size is one higher than inserts
filterCache: inserts is higher than size by a variable amount

I've seen 10 (on two different runs) and 2 (on the most recent run) as the difference between inserts and size.

The FastLRUCache behavior is seen on my production servers with production queries, the LFUCache behavior is on my test server with a benchmark script providing the queries.  I suppose there might be something weird about my canned queries that makes the filterCache behave differently, but the original source of the queries was a production Solr log at level INFO.

                
      was (Author: elyograg):
    The only static warming I have is a *:* search with a sort parameter (and no filter query), which precaches my sort.  I do have autowarming configured, but this behavior happens from initial solr startup.  Things seem to behave correctly with warming - size is autowarmCount, inserts are zero.

FastLRUCache behavior:
queryResultCache: size is one higher than inserts
documentCache: size is one higher than inserts
filterCache: size and inserts are identical

LFUCache behavior:
queryResultCache: size is one higher than inserts
documentCache: size is one higher than inserts
filterCache: inserts is higher than size by a variable amount

I've seen 10 (on two different runs) and 2 (on the most recent run) as the difference between inserts and size.

The FastLRUCache behavior is seen on my production servers with production queries, the LFUCache behavior is on my test server with a benchmark script providing the queries.  I suppose there might be something weird about my canned queries that makes the filterCache behave differently, but the original source of the queries was a production Solr log at level INFO.

                  
> Implement LFU Cache
> -------------------
>
>                 Key: SOLR-2906
>                 URL: https://issues.apache.org/jira/browse/SOLR-2906
>             Project: Solr
>          Issue Type: Sub-task
>          Components: search
>    Affects Versions: 3.4
>            Reporter: Shawn Heisey
>            Priority: Minor
>         Attachments: ConcurrentLFUCache.java, LFUCache.java, SOLR-2906.patch, TestLFUCache.java
>
>
> Implement an LFU (Least Frequently Used) cache as the first step towards a full ARC cache

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org