incubator-blur-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aaron McCurry (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (BLUR-315) Fix max heap guard on row fetch
Date Thu, 05 Dec 2013 12:50:36 GMT

    [ https://issues.apache.org/jira/browse/BLUR-315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13840057#comment-13840057
] 

Aaron McCurry commented on BLUR-315:
------------------------------------

I think that we can relate them but the problem is the heap calculator is not exact and really
can't be, it's there mainly to prevent insanely large Rows from returning too much data. 
It was designed with the mindset that a single Record wasn't going to blow things up in the
first place, so perhaps we should simply throw a warning if the row fetch limit is larger
than the frame size?  It's not going to be exact however.

Any other ideas on how to do this better?

Aaron

> Fix max heap guard on row fetch
> -------------------------------
>
>                 Key: BLUR-315
>                 URL: https://issues.apache.org/jira/browse/BLUR-315
>             Project: Apache Blur
>          Issue Type: Bug
>          Components: Blur
>    Affects Versions: 0.2.0, 0.2.1, 0.2.2
>            Reporter: Tim Williams
>
> We have a guard on a row fetch that limits to 10mb.  If the user is only selecting a
single document from the row, it currently bypasses that heap guard (BlurUtil.fetchDocuments()).
 We should move the guard to check all documents - even the first.
> On a related note, maybe there's a way to relate the guard to our thrift frame size -
which seems to be the constraining factor?



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message