[jira] [Resolved] (ACCUMULO-3084) Delays between calls to logGC causes tserver to fail

["Eric Newton (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/ACCUMULO-3084?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Eric Newton resolved ACCUMULO-3084.
-----------------------------------

       Resolution: Fixed
    Fix Version/s: 1.7.0

> Delays between calls to logGC causes tserver to fail
> ----------------------------------------------------
>
>                 Key: ACCUMULO-3084
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-3084
>             Project: Accumulo
>          Issue Type: Bug
>          Components: tserver
>    Affects Versions: 1.5.0, 1.5.1, 1.6.0
>            Reporter: Eric Newton
>            Assignee: Eric Newton
>             Fix For: 1.5.2, 1.6.1, 1.7.0
>
>          Time Spent: 1h 10m
>  Remaining Estimate: 0h
>
> A complex iterator initialization took a long time (hours).
> The SimpleTimer cleanup was blocked on the tablet lock.
> After the scan was stopped and the SimpleTimer ran the logGCInfo saw a large jump in GC time.  In this case it was 60s over the course of 2 hours between checks.
> The tabletserver determined that GC was causing unacceptable delays and killed itself.
> The check should take into account the time difference between checks, and probably log a warning if the time difference is much larger than expected.



--
This message was sent by Atlassian JIRA
(v6.2#6252)