Mailing-List: contact dev-help@reef.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@reef.apache.org
Date: Fri, 9 Dec 2016 06:17:58 +0000 (UTC)
From: "Julia (JIRA)" <jira@apache.org>
To: dev@reef.apache.org
Message-ID: <JIRA.13023058.1479949791000.480793.1481264278378@Atlassian.JIRA>
In-Reply-To: <JIRA.13023058.1479949791000@Atlassian.JIRA>
References: <JIRA.13023058.1479949791000@Atlassian.JIRA> <JIRA.13023058.1479949791909@arcas>
Subject: [jira] [Updated] (REEF-1677) IMRU: Evaluators failed during
 WaitingForEvaluator phase don't count towards
 MaximumNumberOfEvaluatorFailures limit
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Fri, 09 Dec 2016 06:18:00 -0000


     [ https://issues.apache.org/jira/browse/REEF-1677?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Julia updated REEF-1677:
------------------------
    Labels: FT  (was: )

> IMRU: Evaluators failed during WaitingForEvaluator phase don't count towards MaximumNumberOfEvaluatorFailures limit
> -------------------------------------------------------------------------------------------------------------------
>
>                 Key: REEF-1677
>                 URL: https://issues.apache.org/jira/browse/REEF-1677
>             Project: REEF
>          Issue Type: Bug
>          Components: IMRU
>            Reporter: Mariia Mykhailova
>            Assignee: Mariia Mykhailova
>              Labels: FT
>             Fix For: 0.16
>
>
> In {{IMRUDriver}}, we keep track of the list of failed evaluator ids, and use its length when we need to find out how many evaluators failed. However, if evaluator failed in {{WaitingForEvaluator}} state we immediately remove it from this list and request another evaluator, effectively forgetting about the failure. Thus, even with lots of evaluators failing at this stage we'll never hit {{MaximumNumberOfEvaluatorFailures}} limit, and will keep requesting new evaluators indefinitely.
> I think we should just remove this list (with the additional benefit of reduced memory consumption) and replace it with a single counter which is never decremented. We're only using the values in the list for sanity checks.


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