You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "wangjingying (Jira)" <ji...@apache.org> on 2021/11/30 12:19:00 UTC
[jira] [Commented] (HBASE-26495) Limit the length of initial resultSets for HBase thrift service
[ https://issues.apache.org/jira/browse/HBASE-26495?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17451095#comment-17451095 ]
wangjingying commented on HBASE-26495:
--------------------------------------
[~zhangduo] Please take a look at this PR when you have time. Thank you very much !
> Limit the length of initial resultSets for HBase thrift service
> ---------------------------------------------------------------
>
> Key: HBASE-26495
> URL: https://issues.apache.org/jira/browse/HBASE-26495
> Project: HBase
> Issue Type: Improvement
> Components: Thrift
> Affects Versions: 1.4.14
> Reporter: wangjingying
> Priority: Major
> Fix For: 1.4.14
>
>
> When the client requests getScannerResults or getScannerRows, set up a large numRows will cause ArrayIndexOutOfBoundsException ，and may cause OOM.
> For example
> {code:java}
> conn := cli.(*hbase.THBaseServiceClient)
> tscan := &hbase.TScan{
> StartRow: []byte("prefix_1"),
> StopRow: []byte("prefix_2"),
> Columns: []*hbase.TColumn{
> {
> Family: []byte("N"),
> Qualifier: []byte("title"),
> },
> },
> }
> scanresult, err := conn.GetScannerResults([]byte("rowkey1"), tscan, 10000000){code}
> Here initialize the length of resultSets using the minimum value of "hbase.thrift.max.scanner.results.init.length" and numRows。
--
This message was sent by Atlassian Jira
(v8.20.1#820001)