It's all about the answers!

Ask a question

'Connection reset by peer' by jbe

Clifton Pang (112) | asked Jun 21 '13, 4:52 p.m.
edited Jun 21 '13, 4:54 p.m.
I got the 'Connection reset by peer' from the build engine. This happens if the build engine has been idling for a long time (e.g. no build for several days). 

2013-06-21 01:02:48 [Jazz build engine] running on host: ********

2013-06-21 01:02:48 [Jazz build engine] Should build occur?

2013-06-21 01:02:48 [Jazz build engine] Yes: Pre-build participant "" would like to build.

2013-06-21 01:02:48 [Jazz build engine] Invoking pre-build participant ""

2013-06-21 01:02:48 [Jazz build engine] Accepting changes into workspace "Workspace, Build, *****, ****" ...

2013-06-21 01:02:54 [Jazz build engine] Deleting fetch destination "/export/Public/****" before fetching ...

2013-06-21 01:02:54 [Jazz build engine] Fetching files to fetch destination "/export/Public/****" ...

CRRTC3524W: Repository connection failed: CRJAZ0099I When accessing the URL "http://clm.********.com:80/ccm/service/" the following HTTP error occurred: "Connection reset by peer"

Is there any place to either change the timeout (like what we can do from RTC eclipse login property), or keep it activated?

One answer

permanent link
Spencer Murata (2.3k115970) | answered Jun 28 '13, 8:48 a.m.
 What version of the JBE is this?  The JBE should attempt to reauthenticate itself on the next run through the loop.  Previously though we did not attempt to logout and reconnect.  So if you have a 3.x engine it is expected, otherwise it should fix itself.


Clifton Pang commented Jun 28 '13, 12:16 p.m.

Spencer, the jbe engine is running TeamConcertBuild 4.0.1, on a Redhat Linux Enterprise Architect.  The jazz server our company is also on 4.0.1.  

Typically, when this happens, we just restart jbe process and then it starts working again, however this is buggy and we don't know how to quite resolve this.

Heather Fraser-Dube commented Jul 04 '13, 10:22 a.m.

It does look like the JBE was able to re-connect, the failure takes place after the build is started, an accept has taken place and it is likely now into the stage where the build workspace is loaded.  Was there further details in the build log related to the failure? a stack trace?

Spencer Murata commented Jul 08 '13, 8:19 a.m.

In any case, this should be a covered use case, so it would probably be best to open a support ticket so we can get the logs and see if we have a defect somewhere.


Your answer

Register or to post your answer.