You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@logging.apache.org by Dominik Psenner <dp...@apache.org> on 2018/02/03 10:47:36 UTC

[log4net] dotnet test host process crashes or hangs when run by jenkins docker container

Hi,

I am reaching out to you wearing the hat of the apache log4net community.
I'll give you some context first.

We are in the process of automating the builds and tests for our various
targeted frameworks. We still struggle with gettibg the netstandard-1.3
tests to work. Some time ago I already had a conversation with Rob Prouse
about this. I was able to find a good combination of arguments to get the
test results logged into a trx fike that can be consumed by jenkins. The
tests are run on a ubuntu machine that starts up a docker container based
on microsoft/dotnet:1.1-sdk.

What we observe is that the test suite is run and when a test fails, the
tedt process does not terminate and hangs there until jenkins kills the
pipeline after the 4 hour timeout. The last lifesign is this message:

    [exec] The active test run was aborted. Reason: Unable to
communicate with test host process.
Sending interrupt signal to process
Cancelling nested steps due to timeout
After 10s process did not stop


Unfortunately we are unable to reproduce the issue when running the test
locally and we have yet found no way to find out what actually goes wrong
or what we could do to avoid the issue. Would you please help us
troubleshoot and find a solution to this situation?

The source can be found here:
https://github.com/apache/logging-log4net/tree/feature/cd-pipeline

The pipeline configuration is in the Jenkinsfile. The test is run via nant.
So the build target can be found at the very end of this file:
https://github.com/apache/logging-log4net/blob/feature/cd-pipeline/tests/nant.build

And last but not least, the full jenkins build console output can be found
here:
https://builds.apache.org/job/logging-log4net/job/feature%252Fcd-pipeline/192/console

Best regards,
Dominik

Re: [log4net] dotnet test host process crashes or hangs when run by jenkins docker container

Posted by Dominik Psenner <dp...@apache.org>.
Hi!

I've got an update on this topic. In the past days I had the chance to work
on the build pipeline and today I have diagnostical test logs at hand.
Unfortunately I've no idea what causes the following output:

```

runtests-netstandard-1.3:

     [exec] Build started, please wait...
     [exec] Build completed.
     [exec]
     [exec] Test run for
/home/jenkins/jenkins-slave/workspace/log4net_feature_cd-pipeline-76KUCPODUF6LCE45226EBUR4GNVLYPMYVC23Z4ITBOMNJT3CA2WA/netstandard/1.3/log4net.tests/bin/Debug/netcoreapp2.0/log4net.tests.dll(.NETCoreApp,Version=v2.0)
     [exec] Microsoft (R) Test Execution Command Line Tool Version 15.7.0
     [exec] Copyright (c) Microsoft Corporation.  All rights reserved.
     [exec]
     [exec] Starting test execution, please wait...
     [exec] Logging Vstest Diagnostics in file:
/home/jenkins/jenkins-slave/workspace/log4net_feature_cd-pipeline-76KUCPODUF6LCE45226EBUR4GNVLYPMYVC23Z4ITBOMNJT3CA2WA/netstandard/1.3/log4net.tests/test.diagnostics.log
     [exec] The active test run was aborted. Reason:
     [exec]
     [exec] Test Run Aborted.
     [exec] /home/jenkins/jenkins-slave/workspace/log4net_feature_cd-pipeline-76KUCPODUF6LCE45226EBUR4GNVLYPMYVC23Z4ITBOMNJT3CA2WA/tests/nant.build(869,10):
     [exec] External Program Failed: dotnet (return code was 1)
     [move] 3 files moved.

BUILD SUCCEEDED - 1 non-fatal error(s), 2 warning(s)

```

The following location [1] contains the diagnostical logfiles for the build
that produced the output and I would greatly appreciate any help with
diagnosing what's going wrong here.

Cheers,
Dominik

[1]
https://builds.apache.org/job/logging-log4net/job/feature%252Fcd-pipeline/lastSuccessfulBuild/artifact/package/tests/bin/netstandard/1.3/log4net.tests/

2018-02-09 14:23 GMT+01:00 Dominik Psenner <dp...@gmail.com>:

> *bump*
>
> Has anyone out there an idea how we could troubleshoot the issue or any
> ideas what else we could do?
>
>
>
> On 2018-02-03 11:51, Dominik Psenner wrote:
>
>> The first message was rejected by the mailing list, please see the
>> forward.
>> Please cc the dev at logging.apache.org mailing list.
>>
>> ---------- Forwarded message ----------
>> From: "Dominik Psenner" <dp...@apache.org>
>> Date: 3 Feb 2018 11:47 a.m.
>> Subject: [log4net] dotnet test host process crashes or hangs when run by
>> jenkins docker container
>> To: <nu...@googlegroups.com>
>> Cc: <de...@logging.apache.org>
>>
>> Hi,
>>
>> I am reaching out to you wearing the hat of the apache log4net community.
>> I'll give you some context first.
>>
>> We are in the process of automating the builds and tests for our various
>> targeted frameworks. We still struggle with gettibg the netstandard-1.3
>> tests to work. Some time ago I already had a conversation with Rob Prouse
>> about this. I was able to find a good combination of arguments to get the
>> test results logged into a trx fike that can be consumed by jenkins. The
>> tests are run on a ubuntu machine that starts up a docker container based
>> on microsoft/dotnet:1.1-sdk.
>>
>> What we observe is that the test suite is run and when a test fails, the
>> tedt process does not terminate and hangs there until jenkins kills the
>> pipeline after the 4 hour timeout. The last lifesign is this message:
>>
>>      [exec] The active test run was aborted. Reason: Unable to
>> communicate with test host process.
>> Sending interrupt signal to process
>> Cancelling nested steps due to timeout
>> After 10s process did not stop
>>
>>
>> Unfortunately we are unable to reproduce the issue when running the test
>> locally and we have yet found no way to find out what actually goes wrong
>> or what we could do to avoid the issue. Would you please help us
>> troubleshoot and find a solution to this situation?
>>
>> The source can be found here: https://github.com/
>> apache/logging-log4net/tree/feature/cd-pipeline
>>
>> The pipeline configuration is in the Jenkinsfile. The test is run via
>> nant.
>> So the build target can be found at the very end of this file:
>> https://github.com/apache/logging-log4net/blob/feature/cd-pipeline/tests/
>> nant.build
>>
>> And last but not least, the full jenkins build console output can be found
>> here: https://builds.apache.org/job/logging-log4net/job/feature%
>> 252Fcd-pipeline/192/console
>>
>> Best regards,
>> Dominik
>>
>> --
>> Dominik Psenner
>>
>>