[
https://issues.apache.org/jira/browse/HBASE-21504?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
xuming updated HBASE-21504:
---------------------------
Summary: When Using FIFOCompactionPolicy, a compaction may commit a initial
hfile whose maxTimeStamp is Long.MAX_VALUE. These initial hfile will never be
archived by FIFOCompactionPolicy (was: When Using FIFOCompactionPolicy, in
some case, doing a compaction will commit a new hfile, but it's maxTimeStamp is
Long.MAX_VALUE.The hfile will never be archived)
> When Using FIFOCompactionPolicy, a compaction may commit a initial hfile
> whose maxTimeStamp is Long.MAX_VALUE. These initial hfile will never be
> archived by FIFOCompactionPolicy
> ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-21504
> URL: https://issues.apache.org/jira/browse/HBASE-21504
> Project: HBase
> Issue Type: Bug
> Components: Compaction
> Affects Versions: 2.1.0
> Reporter: xuming
> Priority: Critical
> Attachments: 1.patch
>
>
> When i use FIFOCompactionPolicy, and if all hfiles(>1) are TTL expired in a
> region, once we do a compaction on the region, the compaction policy will
> select the latest hfile to do compaction.But beacuse the hfile is already TTL
> expired, compactor will write a new hfile(whose entry counter is 0 and
> maxTimeStamp is Long.MAX_VALUE). The new hfile will never be TTL expired, so
> we can not archive it by FIFOCompactionPolicy forever.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
