GitHub user ottobackwards reopened a pull request:

    https://github.com/apache/incubator-metron/pull/344

    METRON-538  Ensure proper shutdown, exceptions of integration test 
components

    This is an effort to eliminate spam errors from our testing and hopefully 
improve our testing at the same time.
    
    This PR will track the commits for this effort, and allow for timely 
comment and review.
    
    So far the things we are looking for or seeing are
    
    - problems with shutdown dependencies with the testing components like yarn 
and zk
    - problems were we cannot completely mock out components.  Even if the 
tests pass, we are spammed with errors because of the attempts to connect to 
components that are part of the test because they are not needed.
    
    
    Because of Apache Storm Issue STORM-1730
    https://issues.apache.org/jira/browse/STORM-1730 it is not possible
    to shutdown the LocalServer we use for storm without leaking threads.
    These threads will output error messages when they 'wake up' and try to 
connect
    with curator - to a zk embedded server that has already been shutdown.
    This issue is fixed in Storm 1.0.3, but that is not released yet
    
    
    Because of this unresolved issue  ZOOKEEPER-1237: 
https://issues.apache.org/jira/browse/ZOOKEEPER-1237 we also may see errors like
    Running org.apache.metron.enrichment.integration.EnrichmentIntegrationTest
    
    2016-11-09 22:25:50,293 ERROR [SyncThread:0] server.NIOServerCnxn 
(NIOServerCnxn.java:sendBuffer(178)) - Unexpected Exception: 
    
    java.nio.channels.CancelledKeyException
    
        at sun.nio.ch.SelectionKeyImpl.ensureValid(SelectionKeyImpl.java:73)
    
        at sun.nio.ch.SelectionKeyImpl.interestOps(SelectionKeyImpl.java:77)
    
        at 
org.apache.storm.shade.org.apache.zookeeper.server.NIOServerCnxn.sendBuffer(NIOServerCnxn.java:151)
    
        at 
org.apache.storm.shade.org.apache.zookeeper.server.NIOServerCnxn.sendResponse(NIOServerCnxn.java:1081)
    
        at 
    
    
    
    My goal is to keep this PR mergable with every commit as I go.  And as a 
bonus, each travis build will hopefully show the progress.

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/ottobackwards/incubator-metron METRON-538

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/incubator-metron/pull/344.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #344
    
----
commit 6d230bada58bf297365b289affc375a48ce1da84
Author: Otto Fowler <[email protected]>
Date:   2016-11-03T17:44:32Z

    In Maas Integration, be sure to shutdown zookeeper AFTER yarn, to avoid 
extra errors when shutting down yarn

commit fb79f16b71f781ba02c7c7b9ec33d17444bfe320
Author: Otto Fowler <[email protected]>
Date:   2016-11-03T19:46:29Z

    Make HBaseBolt testable by allowing building with a provided table provider 
as opposed to building from string, this allows mocking.  This eliminates log 
spam of errors when the runtime class still attempts and fails to connect to 
zookeeper, which is not running or needed for this test.  I did not change the 
function for withTableProvider() because I do not know all the consequences of 
changing the flux configurations

commit acad2e879d425cbc9b744f0e771cae7c27de12cd
Author: Otto Fowler <[email protected]>
Date:   2016-11-09T20:27:09Z

    ensure shutdown order
    
    close zk clients
    
     use try with resources consistantly in ConfigurationUtils

commit b7146aca29b69c7a039e057db81eb1378ac01db5
Author: Otto Fowler <[email protected]>
Date:   2016-11-09T20:27:54Z

    merge tracking branch apache/master

commit b876aa2ecbed437ef24d7daaa5456696718d2144
Author: Otto Fowler <[email protected]>
Date:   2016-11-10T04:37:56Z

    remove some commented out code
    
    create enrichment topic instead of relying on auto.create to avoid possible 
error output/logging

commit a894fdffbe20d8a6bdc19fa09f7665ca093ad0fd
Author: Otto Fowler <[email protected]>
Date:   2016-11-11T14:43:14Z

    Add the ability to get the current logging level, so that you can get, 
change, and reset to previous in tests
    
    set logging level to SEVERE when waiting on metadata propogation to cut 
down on possible errors

commit e98b7b21b906bcceb1f00616d65ac9d68f283de2
Author: Otto Fowler <[email protected]>
Date:   2016-11-11T15:22:27Z

    Move logging changes to wrap all of create topic

commit a22fbddc574c87f9fe40201ca8bfdb9efac3e20f
Author: Otto Fowler <[email protected]>
Date:   2016-11-11T16:43:05Z

    Turn logging off when creating topics.

commit 247d26a1a5c6d2d667c67864597e75eee02b09d4
Author: Otto Fowler <[email protected]>
Date:   2016-11-11T18:44:20Z

    added support for waiting until the broker state for the kafka server was 
running

commit 1cb15f8fc286da67b6a5f4aa4b76bd1db168847f
Author: Otto Fowler <[email protected]>
Date:   2016-11-11T19:49:16Z

    Attempt to change logging level for scala code via log4j

commit 1a9a4246c5e5c558e24055cd91a2c580ad906aea
Author: Otto Fowler <[email protected]>
Date:   2016-11-11T20:38:26Z

    turn off verbose logging, set shutdown order

commit 1f3821e7e235435f426af109b5d8e57d7dbe3ffa
Author: Otto Fowler <[email protected]>
Date:   2016-11-11T21:09:36Z

    remove verboselogging

----


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---

Reply via email to