[
https://issues.apache.org/jira/browse/IMPALA-3200?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16576542#comment-16576542
]
ASF subversion and git services commented on IMPALA-3200:
---------------------------------------------------------
Commit 4af3a7853e9d450675a61bff5b031a12e7a05172 in impala's branch
refs/heads/master from [[email protected]]
[ https://git-wip-us.apache.org/repos/asf?p=impala.git;h=4af3a78 ]
IMPALA-7333: remove MarkNeedsDeepCopy() in agg and BTS
This takes advantage of work (e.g. IMPALA-3200, IMPALA-5844)
to remove a couple of uses of the API.
Testing:
Ran core, ASAN and exhaustive builds.
Added unit tests to directly test the attaching behaviour.
Change-Id: I5f5b8a418d4816f603a64da6287ec392dbd4603f
Reviewed-on: http://gerrit.cloudera.org:8080/11156
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> Replace BufferedBlockMgr with new buffer pool
> ---------------------------------------------
>
> Key: IMPALA-3200
> URL: https://issues.apache.org/jira/browse/IMPALA-3200
> Project: IMPALA
> Issue Type: Improvement
> Components: Backend
> Affects Versions: Impala 2.6.0
> Reporter: Tim Armstrong
> Assignee: Tim Armstrong
> Priority: Blocker
> Labels: resource-management
> Fix For: Impala 2.10.0
>
>
> We want to replace BufferedBlockMgr, a query-wide buffer pool, with a new
> BufferPool that is shared between all queries. The goals are:
> * Support for guaranteed reservations: i.e. if a reservation is granted, the
> buffer pool will fulfil it (unless the OS is unable to fulfill the buffer
> pool's memory requirements).
> * Simplified interaction between reservations and pins (if you have a
> reservation, you can pin, if you don't, you can't)
> * Support for increasing reservations (up to a planner-specified limit).
> * Support for smaller buffer sizes with similar performance (so we can reduce
> minimum memory requirement to execute spill-to-disk algorithm)
> * Support for larger buffers to support wide rows
> * Reduced reliance on TCMalloc, which isn't suited to management of large
> buffers (e.g. see IMPALA-2800)
> * Better transfer model for buffer pool pages, so we can implement
> transfer-of-ownership consistently for row batches (instead of mixing
> transfer with MarkNeedsToReturn()).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
