You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@asterixdb.apache.org by "Ian Maxon (Jira)" <ji...@apache.org> on 2024/02/23 18:08:00 UTC
[jira] [Updated] (ASTERIXDB-3360) Honor ByteBuffer position and limit when writing compressed B-Tree pages
[ https://issues.apache.org/jira/browse/ASTERIXDB-3360?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ian Maxon updated ASTERIXDB-3360:
---------------------------------
Labels: triaged (was: )
> Honor ByteBuffer position and limit when writing compressed B-Tree pages
> ------------------------------------------------------------------------
>
> Key: ASTERIXDB-3360
> URL: https://issues.apache.org/jira/browse/ASTERIXDB-3360
> Project: Apache AsterixDB
> Issue Type: Improvement
> Components: STO - Storage
> Affects Versions: 0.9.9
> Reporter: Wail Y. Alkowaileet
> Assignee: Wail Y. Alkowaileet
> Priority: Major
> Labels: triaged
> Fix For: 0.9.9
>
>
> Compressed page writer always clear the to-be-written ByteBuffer position and limit to be the entire buffer. This could be fine for writing pages in row format. However, in columnar format, this might incur compressing and writing unrelated data (as the ByteBuffer may contain data from a previous write).
> Instead of clearing the ByteBuffer position and limit at the compressed page writer, it should be done during the call to BufferCache#confiscatePage(), which is the case currently. The compressed page writer, however, should not re-clear whatever limit is provided.
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)