[ https://issues.apache.org/jira/browse/HBASE-22460?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Viraj Jasani updated HBASE-22460: --------------------------------- Release Note: Leaked store files can not be removed even after it is invalidated via compaction. A reasonable mitigation for a reader reference leak would be a fast reopen of the region on the same server. Configs: 1. hbase.master.regions.recovery.check.interval : Regions Recovery Chore interval in milliseconds. This chore keeps running at this interval to find all regions with configurable max store file ref count and reopens them. 2. hbase.regions.recovery.store.file.ref.count : This config represents Store files Ref Count threshold value considered for reopening regions. Any region with store files ref count > this value would be eligible for reopening by master. Default value -1 indicates this feature is turned off. Only positive integer value should be provided to enable this feature. > Reopen a region if store reader references may have leaked > ---------------------------------------------------------- > > Key: HBASE-22460 > URL: https://issues.apache.org/jira/browse/HBASE-22460 > Project: HBase > Issue Type: Sub-task > Affects Versions: 3.0.0, 1.5.0, 2.3.0 > Reporter: Andrew Kyle Purtell > Assignee: Viraj Jasani > Priority: Minor > Fix For: 3.0.0, 2.3.0 > > > We can leak store reader references if a coprocessor or core function somehow > opens a scanner, or wraps one, and then does not take care to call close on > the scanner or the wrapped instance. A reasonable mitigation for a reader > reference leak would be a fast reopen of the region on the same server > (initiated by the RS) This will release all resources, like the refcount, > leases, etc. The clients should gracefully ride over this like any other > region transition. This reopen would be like what is done during schema > change application and ideally would reuse the relevant code. If the refcount > is over some ridiculous threshold this mitigation could be triggered along > with a fat WARN in the logs. -- This message was sent by Atlassian Jira (v8.3.4#803005)