[
https://issues.apache.org/jira/browse/HIVE-22460?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16967481#comment-16967481
]
Hive QA commented on HIVE-22460:
--------------------------------
Here are the results of testing the latest attachment:
https://issues.apache.org/jira/secure/attachment/12984915/HIVE-22460.0.patch
{color:green}SUCCESS:{color} +1 due to 1 test(s) being added or modified.
{color:red}ERROR:{color} -1 due to 2 failed/errored test(s), 17566 tests
executed
*Failed tests:*
{noformat}
org.apache.hadoop.hive.llap.cache.TestCacheAllocationsEvictionsCycles.testRandomFragmentation
(batchId=364)
org.apache.hadoop.hive.ql.TestTxnCommands.testParallelInsertAnalyzeStats
(batchId=354)
{noformat}
Test results:
https://builds.apache.org/job/PreCommit-HIVE-Build/19294/testReport
Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/19294/console
Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-19294/
Messages:
{noformat}
Executing org.apache.hive.ptest.execution.TestCheckPhase
Executing org.apache.hive.ptest.execution.PrepPhase
Executing org.apache.hive.ptest.execution.YetusPhase
Executing org.apache.hive.ptest.execution.ExecutionPhase
Executing org.apache.hive.ptest.execution.ReportingPhase
Tests exited with: TestsFailedException: 2 tests failed
{noformat}
This message is automatically generated.
ATTACHMENT ID: 12984915 - PreCommit-HIVE-Build
> LRFU cache policy leaks locked buffers upon purge()
> ---------------------------------------------------
>
> Key: HIVE-22460
> URL: https://issues.apache.org/jira/browse/HIVE-22460
> Project: Hive
> Issue Type: Bug
> Components: llap
> Reporter: Ádám Szita
> Assignee: Ádám Szita
> Priority: Major
> Attachments: HIVE-22460.0.patch
>
>
> LRFU policy's purge() implementation is carefully not removing buffers that
> are currently locked (i.e. in use by some IO thread). So far that's good.
> However it won't keep track of such buffers after the purge() method has
> finished: it will always reset its heap and list, thereby forgetting
> information on these buffers. It will never be able to evict these in the
> future, even if they get unlocked and become eligible for eviction.
> This is problematic as:
> * Although eventually these buffers might be evicted by BufferAllocator, by
> the time that happens we have wasted space and time.
> * Meta information about the buffers will remain in CacheContentsTracker
> forever, wasting heap space too.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
