You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@omid.apache.org by "Ohad Shacham (JIRA)" <ji...@apache.org> on 2018/02/25 12:27:00 UTC

[jira] [Resolved] (OMID-87) Fix BatchPool initialization

     [ https://issues.apache.org/jira/browse/OMID-87?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Ohad Shacham resolved OMID-87.
------------------------------
    Resolution: Fixed

> Fix BatchPool initialization
> ----------------------------
>
>                 Key: OMID-87
>                 URL: https://issues.apache.org/jira/browse/OMID-87
>             Project: Apache Omid
>          Issue Type: Bug
>    Affects Versions: 0.8.2.0
>            Reporter: Francisco Perez-Sorrosal
>            Assignee: Francisco Perez-Sorrosal
>            Priority: Critical
>              Labels: batch, batchpool, creation, fix
>             Fix For: 0.9.0.0
>
>
> When initializing the ObjectPool for Batch java objects in BatchPoolModule.java, an Apache's GenericObjectPool is used. The pool configuration lacks the call to config.setMaxIdle(int), which provokes that when the Batch objects are returned to the ObjectPool during the pre-creation phase, those above the default max idle number of objects in the pool (which is 8 according to the [Apache documentation|https://commons.apache.org/proper/commons-pool/api-1.6/org/apache/commons/pool/impl/GenericObjectPool.html]) are destroyed.
> If at some point the TSO needs to access those Batch objects between 8 and the number specified in the Omid configuration (numConcurrentCTWriters), they will be created on the fly, which it was we want to avoid when pre-creating the ObjectPool during the TSO initialization.



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