You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Zoltán Borók-Nagy (Jira)" <ji...@apache.org> on 2020/03/30 10:26:00 UTC
[jira] [Updated] (HIVE-23102) waitForCompactionToFinish can
potentially wait for too long
[ https://issues.apache.org/jira/browse/HIVE-23102?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Zoltán Borók-Nagy updated HIVE-23102:
-------------------------------------
Description:
AlterTableCompactOperation.waitForCompactionToFinish() has the following code fragment:
{noformat}
//double wait time until 5min
waitTimeMs = waitTimeMs*2;
waitTimeMs = Math.max(waitTimeMs, waitTimeOut);
{noformat}
Based on the comment ("double wait time until 5min") I think it should use Math.min() instead of Math.max().
It also affects the runtime of Impala tests that use Hive compaction, because they hang for at least 5 mins each time we compact a table. Workaround to this is to configure the wait time to a lower value.
But still, from the comment I think the original intention of the code was to put an upper bound to the wait time, and not a lower bound.
was:
AlterTableCompactOperation.waitForCompactionToFinish() has the following code fragment:
{noformat}
//double wait time until 5min
waitTimeMs = waitTimeMs*2;
waitTimeMs = Math.max(waitTimeMs, waitTimeOut);
{noformat}
Based on the comment ("double wait time until 5min") I think it should use Math.min() instead of Math.max().
It also affects the runtime of Impala tests that use Hive compaction, because they hang for at least 5 mins each time we compact a table.
> waitForCompactionToFinish can potentially wait for too long
> -----------------------------------------------------------
>
> Key: HIVE-23102
> URL: https://issues.apache.org/jira/browse/HIVE-23102
> Project: Hive
> Issue Type: Bug
> Reporter: Zoltán Borók-Nagy
> Priority: Major
>
> AlterTableCompactOperation.waitForCompactionToFinish() has the following code fragment:
>
> {noformat}
> //double wait time until 5min
> waitTimeMs = waitTimeMs*2;
> waitTimeMs = Math.max(waitTimeMs, waitTimeOut);
> {noformat}
> Based on the comment ("double wait time until 5min") I think it should use Math.min() instead of Math.max().
> It also affects the runtime of Impala tests that use Hive compaction, because they hang for at least 5 mins each time we compact a table. Workaround to this is to configure the wait time to a lower value.
> But still, from the comment I think the original intention of the code was to put an upper bound to the wait time, and not a lower bound.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)