[
https://issues.apache.org/jira/browse/ACCUMULO-3858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14562154#comment-14562154
]
Josh Elser commented on ACCUMULO-3858:
--------------------------------------
bq. Hmm... seems like we should be able to figure out a nearly 30% difference
Yeah, that was my hunch too, but to see it oscillate back down to mid-500s
without making any other changes was... odd. I'm not entirely convinced we
weren't just fooled as to what "normal" was (perhaps you are more confident
than I am) -- but I definitely agree it's worth investigating.
I plan to commit some changes to stabilize this test for the present, but I'll
open up another to make sure we look into this more.
> WatchTheWatchCountIT failed with too few watchers
> -------------------------------------------------
>
> Key: ACCUMULO-3858
> URL: https://issues.apache.org/jira/browse/ACCUMULO-3858
> Project: Accumulo
> Issue Type: Test
> Components: t, test
> Reporter: Josh Elser
> Assignee: Josh Elser
> Fix For: 1.8.0, 1.7.1
>
>
> Saw a test failure in automated testing
> {noformat}
> Total watches was not greater than 500, but was 495
> {noformat}
> If we had less watchers than we expected, I'm guessing that means that each
> table wasn't fully online yet and thus not all of the watches were set for
> each tablet.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
