[infinispan-issues] [JBoss JIRA] Updated: (ISPN-1290) Expiration thread wakeup interval default is too short
Manik Surtani (JIRA)
jira-events at lists.jboss.org
Tue Aug 2 12:43:23 EDT 2011
[ https://issues.jboss.org/browse/ISPN-1290?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Manik Surtani updated ISPN-1290:
--------------------------------
Summary: Expiration thread wakeup interval default is too short (was: FileCacheStore default purge interval too short)
Labels: expiry (was: )
Priority: Minor (was: Major)
Description:
The default wakeup interval is defined to be 5000 ms.
When adding a large amount of data to the cache, purging will block the whole application, because it starts - runs for a long time - stops and will be again started immediatly.
This leads to a "logical infinite loop" blocking the application.
This will get even worse when parallel purging (multiThreadedPugre / purgerThreads) is used)
This is legacy, when the thread was intended to handle eviction as well. Now that the thread only takes care of expiry, there is no need for this default to be this short. A more sensible default (60s) is used instead.
was:
The default purge interval is defined to be 5000 ms.
When adding a large amount of data to the cache, purging will block the whole application, because it starts - runs for a long time - stops and will be again started immediatly.
This leads to a "logical infinite loop" blocking the application.
This will get even worse when parallel purging (multiThreadedPugre / purgerThreads) is used)
Complexity: Low
Component/s: Eviction
> Expiration thread wakeup interval default is too short
> ------------------------------------------------------
>
> Key: ISPN-1290
> URL: https://issues.jboss.org/browse/ISPN-1290
> Project: Infinispan
> Issue Type: Quality Risk
> Components: Configuration, Eviction, Loaders and Stores
> Affects Versions: 4.2.1.FINAL, 5.0.0.CR8
> Reporter: Robert Stupp
> Assignee: Manik Surtani
> Priority: Minor
> Labels: expiry
> Fix For: 5.0.0.FINAL
>
>
> The default wakeup interval is defined to be 5000 ms.
> When adding a large amount of data to the cache, purging will block the whole application, because it starts - runs for a long time - stops and will be again started immediatly.
> This leads to a "logical infinite loop" blocking the application.
> This will get even worse when parallel purging (multiThreadedPugre / purgerThreads) is used)
> This is legacy, when the thread was intended to handle eviction as well. Now that the thread only takes care of expiry, there is no need for this default to be this short. A more sensible default (60s) is used instead.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
More information about the infinispan-issues
mailing list