[
https://issues.apache.org/jira/browse/HIVE-29532?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Denys Kuzmenko updated HIVE-29532:
----------------------------------
Description:
*Problem Statement* - Hive Compaction Initiator not initiating compaction for
few tables.
Every Initiator cycle hits exactly 1000 candidates (COMPACTOR_FETCH_SIZE limit)
and ignores the rest on the next cycle
In findPotentialCompactions, the condition TC.CTC_TIMESTAMP >=
CURRENT_TIMESTAMP - INTERVAL 300 SECOND, returns rows where CTC_TIMESTAMP is
within the last 300 seconds, i.e., the last 5 minutes.
The next cycle doesn't pick this table due to the above condition hence it
remains in the table itself.
was:
*Problem Statement* - Hive Compaction Initiator not initiating compaction for
few tables.
Every Initiator cycle hits exactly 1000 candidates and ignores the rest on the
next cycle
> FindPotentialCompactions method is Truncating out some records, even though
> corresponding entries still exist in COMPLETED_TXN_COMPONENTS
> -----------------------------------------------------------------------------------------------------------------------------------------
>
> Key: HIVE-29532
> URL: https://issues.apache.org/jira/browse/HIVE-29532
> Project: Hive
> Issue Type: Bug
> Affects Versions: 4.0.0, 4.1.0, 4.0.1, 4.2.0
> Reporter: Denys Kuzmenko
> Priority: Major
>
> *Problem Statement* - Hive Compaction Initiator not initiating compaction
> for few tables.
> Every Initiator cycle hits exactly 1000 candidates (COMPACTOR_FETCH_SIZE
> limit) and ignores the rest on the next cycle
> In findPotentialCompactions, the condition TC.CTC_TIMESTAMP >=
> CURRENT_TIMESTAMP - INTERVAL 300 SECOND, returns rows where CTC_TIMESTAMP is
> within the last 300 seconds, i.e., the last 5 minutes.
> The next cycle doesn't pick this table due to the above condition hence it
> remains in the table itself.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
