[
https://issues.apache.org/jira/browse/HIVE-10128?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14387951#comment-14387951
]
Gunther Hagleitner commented on HIVE-10128:
-------------------------------------------
maybe it's because it's late, but it seems you're not actually using the
threadLocalHashMap in the container. That seems to kind of defeat the purpose.
Also, I think threadLocalHashMap is a confusing name for a "getter".
Is there a reason to keep the non thread safe methods around? You could just
force everyone to use a getter object w/ position to access right? There
wouldn't be a perf overhead as far as I can tell.
> BytesBytesMultiHashMap does not allow concurrent read-only access
> -----------------------------------------------------------------
>
> Key: HIVE-10128
> URL: https://issues.apache.org/jira/browse/HIVE-10128
> Project: Hive
> Issue Type: Bug
> Reporter: Gopal V
> Assignee: Sergey Shelukhin
> Fix For: llap
>
> Attachments: HIVE-10128.01.patch, HIVE-10128.02.patch,
> HIVE-10128.patch, hashmap-after.png, hashmap-sync-source.png, hashmap-sync.png
>
>
> The multi-threaded performance takes a serious hit when LLAP shares
> hashtables between the probe threads running in parallel.
> !hashmap-sync.png!
> This is an explicit synchronized block inside ReusableRowContainer which
> triggers this particular pattern.
> !hashmap-sync-source.png!
> Looking deeper into the code, the synchronization seems to be caused due to
> the fact that WriteBuffers.setReadPoint modifies the otherwise read-only
> hashtable.
> To generate this sort of result, run LLAP at a WARN log-level, to avoid all
> the log synchronization that otherwise affects the thread sync.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
