Brock Noland has posted comments on this change.

Change subject: [tests] fix tests - two cpus and fresh install
......................................................................


Patch Set 2:

Yep sounds good to me. I haven't fully investigated why the change fixes the 
test for me.

The cache is sharded by the numbers of cpus. Based on the following comment I 
was thinking I am thinking I edge case where I didn't see any hits but would 
with 4-procs.

      // Otherwise, we expect to have gotten some hits.
      // If our cache were a perfect LRU implementation, then we would actually
      // expect 0 hits here as well, since we are accessing the entries in
      // exactly the same order as we inserted them, and thus would evict
      // an entry before we look for it again. But, our LRU cache is sharded
      // so we expect to get some hits on the second time.

based on that plus this comment:

      // Basic test for the CompilationManager code cache.
      // This runs a bunch of compilation tasks and ensures that the cache
      // sometimes hits on the second attempt for the same projection.

I am thinking it's a safe change, but obviously important to verify that is 
actually a safe change.

-- 
To view, visit http://gerrit.cloudera.org:8080/4446
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I81b70f63923078d449f6541a61b292517e49877d
Gerrit-PatchSet: 2
Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-Owner: Brock Noland <br...@phdata.io>
Gerrit-Reviewer: Brock Noland <br...@phdata.io>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Todd Lipcon <t...@apache.org>
Gerrit-Reviewer: Will Berkeley <wdberke...@gmail.com>
Gerrit-HasComments: No

Reply via email to