You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@aurora.apache.org by "Mehrdad Nurolahzade (JIRA)" <ji...@apache.org> on 2016/12/21 17:10:58 UTC

[jira] [Created] (AURORA-1865) Experiment with Increasing H2 QUERY_CACHE_SIZE

Mehrdad Nurolahzade created AURORA-1865:
-------------------------------------------

             Summary: Experiment with Increasing H2 QUERY_CACHE_SIZE
                 Key: AURORA-1865
                 URL: https://issues.apache.org/jira/browse/AURORA-1865
             Project: Aurora
          Issue Type: Story
          Components: Scheduler
            Reporter: Mehrdad Nurolahzade
            Priority: Minor


Aurora scheduler is not setting a value for {{QUERY_CACHE_SIZE}}, therefore it is relying on H2's default cache size of 8 statements. There might be potential performance gain from increasing H2 query cache size ([example discussion thread|https://groups.google.com/forum/#!topic/h2-database/pGG9HA5ckhw]).

Looking at execution count of queries (in {{INFORMATION_SCHEMA.QUERY_STATISTICS}} meta table), selecting host attributes query is repeated 2-3 orders of magnitude more than other queries, giving it a substantial cumulative execution cost. Scheduler needs to ensure all such high frequency statements are always cached.

Verify if increasing {{QUERY_CACHE_SIZE}} positively impacts scheduler performance.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)