accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Christopher Tubbs (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (ACCUMULO-3974) Modify Randomwalk Bulk module to catch ACCUMULO-3967
Date Tue, 13 Oct 2015 16:08:05 GMT

     [ https://issues.apache.org/jira/browse/ACCUMULO-3974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Christopher Tubbs updated ACCUMULO-3974:
----------------------------------------
         Priority: Major  (was: Critical)
    Fix Version/s:     (was: 1.6.5)
                       (was: 1.7.1)
       Issue Type: Test  (was: Bug)

I'm dropping this from Critical to Major, and dropping 1.6.x and 1.7.x from the fixVersion.
The reasoning is that it seems this is unlikely to get done for those older versions (especially
since it is still unassigned), and it doesn't strike me as critical to modify randomwalk to
catch a bug we're reasonably confident we've fixed and not likely to regress in those older
versions.

Of course, it's still important we add a test to prevent regressions in future versions, and
if any work is done on older versions, feel free to re-add the appropriate fixVersion(s).
I just don't want to keep bumping it because nobody's working on it.

> Modify Randomwalk Bulk module to catch ACCUMULO-3967
> ----------------------------------------------------
>
>                 Key: ACCUMULO-3974
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-3974
>             Project: Accumulo
>          Issue Type: Test
>          Components: test
>            Reporter: Josh Elser
>             Fix For: 1.8.0
>
>
> [~ecn] asked me after I committed a fix for ACCUMULO-3967 "why didn't randomwalk catch
this bug?"
> I think we could potentially have caught this if in Setup we pre-split the table to be
a large collection of sequentially increasing tablets. It's not a guarantee catch (since the
bug itself was only shown in the case of import failures and the tablet distribution hasn't
changed).
> Alternatively, we could copy the existing bulk module into a new module. In this module,
we remove the splitting and merging, instead keeping a static split distribution. This would
be almost guaranteed to eventually recreate the scenario. Adding in a chaotic balancer, even
more likely to reproduce.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message