impala-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Tim Armstrong (Code Review)" <ger...@cloudera.org>
Subject [Impala-ASF-CR] IMPALA-4835: Part 1: simplify I/O mgr mem mgmt
Date Wed, 08 Nov 2017 21:42:06 GMT
Tim Armstrong has posted comments on this change. ( http://gerrit.cloudera.org:8080/8414 )

Change subject: IMPALA-4835: Part 1: simplify I/O mgr mem mgmt
......................................................................


Patch Set 4:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/8414/4/be/src/runtime/disk-io-mgr.cc
File be/src/runtime/disk-io-mgr.cc:

http://gerrit.cloudera.org:8080/#/c/8414/4/be/src/runtime/disk-io-mgr.cc@797
PS4, Line 797:       reader->disk_states_[disk_queue->disk_id].DecrementRequestThread();
> The scan range might be cancelled in DiskIoMgr::HandleReadFinished as well.
That's a good question. I don't understand whether it's a bug or not. It's definitely suspicious.
I think that in some case re-scheduling the scan range is enough to get the cleanup to happen
next time the scan range is picked up, but I'm not sure. I'm looking at simplifying error
propagation here (amongst other things) so this should change to something more obviously
correct. Currently it's really confusing because the error is propagated in HandleReadFinished()
by both ScanRange::Cancel() and enqueueing a BufferDescriptor with the error status.



-- 
To view, visit http://gerrit.cloudera.org:8080/8414
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: If5cb42437d11c13bc4a55c3ab426b66777332bd1
Gerrit-Change-Number: 8414
Gerrit-PatchSet: 4
Gerrit-Owner: Tim Armstrong <tarmstrong@cloudera.com>
Gerrit-Reviewer: Dan Hecht <dhecht@cloudera.com>
Gerrit-Reviewer: Tianyi Wang <twang@cloudera.com>
Gerrit-Reviewer: Tim Armstrong <tarmstrong@cloudera.com>
Gerrit-Comment-Date: Wed, 08 Nov 2017 21:42:06 +0000
Gerrit-HasComments: Yes

Mime
  • Unnamed multipart/alternative (inline, 8-Bit, 0 bytes)
View raw message