Dan Hecht has posted comments on this change.

Change subject: IMPALA-2831: Bound the number of scanner threads per scan node.
......................................................................


Patch Set 2:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/4174/2/be/src/exec/hdfs-scan-node.cc
File be/src/exec/hdfs-scan-node.cc:

Line 741:   }
if we do it this way, let's factor the set_max_quota() out so it's clear that 
that we're now always doing that:

if (query option set) {
   num_scanner_threads = query-option
} else {
   num_scanner_threads = CpuInfo::num_cores();
}
set_max_quota(num_scanner_threads + 1);

but this makes me a little nervous still due to the side-effect on join thread. 
 This isn't all that different than changing FLAGS_num_threads_per_core to be 
1, right?

The alternative is to move this logic into ThreadTokenAvailableCb to not take a 
token if we've already spun up enough threads.  

Not saying we should change this approach necessarily, but wondering why you 
chose this approach over those two alternatives.


-- 
To view, visit http://gerrit.cloudera.org:8080/4174
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I191988ad18d6b4caf892fc967258823edcf9681f
Gerrit-PatchSet: 2
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Michael Ho <[email protected]>
Gerrit-Reviewer: Dan Hecht <[email protected]>
Gerrit-Reviewer: Michael Ho <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-HasComments: Yes

Reply via email to