You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@uima.apache.org by "Thilo Goetz (JIRA)" <ui...@incubator.apache.org> on 2008/08/13 10:48:44 UTC

[jira] Commented: (UIMA-1087) Embeddability: recover better from out-of-memory errors

    [ https://issues.apache.org/jira/browse/UIMA-1087?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12622138#action_12622138 ] 

Thilo Goetz commented on UIMA-1087:
-----------------------------------

I think it's generally a bad idea to try and recover from OOM Errors (or any other error, for that matter).  They affect the whole JVM, not just the thread they occur in.  All sorts of follow-on errors may ensue if you try to recover from an OOM situation.  It would be better to concentrate on avoiding OOM errors in the first place (maybe by allocating a certain amount of space to a CAS and throwing a regular runtime exception when that amount has been exceeded).


> Embeddability: recover better from out-of-memory errors
> -------------------------------------------------------
>
>                 Key: UIMA-1087
>                 URL: https://issues.apache.org/jira/browse/UIMA-1087
>             Project: UIMA
>          Issue Type: Improvement
>          Components: Core Java Framework
>    Affects Versions: 2.2.2
>            Reporter: Marshall Schor
>            Priority: Minor
>
> I have heard reports that some embedders of UIMA run it in a separate process, because sometimes a large document or misbehaving annotator can use up all the java memory, resulting in an unrecoverable out-of-memory error that "crashes" the whole JVM.  Investigate / implement ways that the framework might be able to more gracefully recover from this kind of error, without crashing the JVM.    Please post ideas about how to do this as comments to this Jira.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.