falcon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "sandeep samudrala (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FALCON-348) Add shutdown hook for Falcon
Date Tue, 08 Sep 2015 07:15:46 GMT

    [ https://issues.apache.org/jira/browse/FALCON-348?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14734345#comment-14734345
] 

sandeep samudrala commented on FALCON-348:
------------------------------------------

[~peeyushb]: With this patch ServiceInitializer.destroy gets called via contex.destroy which
was not happening earlier as mentioned in FALCON-905.
This service destroys all the services one after the other as mentioned in the startup.properties
and metadata service is also part of the same. I have updated the review board entry with
how it looks in the logs with shutdown

{noformat}
2015-09-07 18:08:02,984 INFO  - [Thread-0:] ~ Service destroyed: org.apache.falcon.rerun.service.LateRunService
(ServiceInitializer:65)
2015-09-07 18:08:02,984 INFO  - [Thread-0:] ~ Destroying service: org.apache.falcon.metadata.MetadataMappingService
(ServiceInitializer:58)
2015-09-07 18:08:02,985 INFO  - [Thread-0:] ~ Shutting down graph db (MetadataMappingService:202)
2015-09-07 18:08:03,046 INFO  - [Thread-0:] ~ Service destroyed: org.apache.falcon.metadata.MetadataMappingService
(ServiceInitializer:65)
2015-09-07 18:08:03,046 INFO  - [Thread-0:] ~ Destroying service: org.apache.falcon.service.LogCleanupService
(ServiceInitializer:58)
{noformat}

> Add shutdown hook for Falcon
> ----------------------------
>
>                 Key: FALCON-348
>                 URL: https://issues.apache.org/jira/browse/FALCON-348
>             Project: Falcon
>          Issue Type: Improvement
>          Components: webapp
>    Affects Versions: 0.5
>            Reporter: Venkatesh Seetharam
>            Assignee: sandeep samudrala
>         Attachments: FALCON-348.patch
>
>
> A shutdown hook is missing and services are shutdown in a brute force mode. I'd like
to see activemq, titan and such are shutdown gracefully so we don't lose data and see ugly
exceptions in the logs.



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

Mime
View raw message