Hi Taher, All,
I put a small BuildBot documentation at
https://svn.apache.org/repos/asf/ofbiz/tools/documentation/BuildBot I'll now
create a wiki page to import it
Everybody is welcome to try and improve it
Jacques
Le 07/01/2018 à 23:54, Jacques Le Roux a écrit :
Le 07/01/2018 à 21:08, Taher Alkhateeb a écrit :
Sorry, I'm not sure what you mean by a transient error?
A test error which happens only once. The same code (no commits between) is
used few minutes later and it works.
If you want to know more about possible such issues please ask infra.
I was told in the pass that such things happen. Why do you think I use IRC? I
did not discover this by myself.
How did you determine that it is a transient error,
Because it works locally and, when it's launched from IRC just after, the same
code passes the tests on Buildbot also.
and did you identify what is causing it?
That I have not ideas and you need to ask infra. There is one sentence I often heard when
asking: "shit happen" (you certainly know about the fan too)
I was also told that disk and especially network issues happen. The ASF is far from owning all the machines and IIRW the Buildbot machines:
https://ci.apache.org/
Why are we having more of these errors than in the past?
I think you mix repetitive false warning, which was fixed with INFRA-15394,
with these kinds of issues:
1. svn fails to update
2. tests pass locally but not on Buildbot
which happen from time to time and not specially more now. Then IRC, with the
restriction I explained below, is the solution.
Jacques
On Sun, Jan 7, 2018 at 8:41 PM, Jacques Le Roux
<jacques.le.r...@les7arts.com> wrote:
Le 07/01/2018 à 15:44, Taher Alkhateeb a écrit :
I am suspecting this is still a broken build system rather than a
genuine error. Can you confirm?
It's not a "broken build system" but was a transient error as you can see
there
https://ci.apache.org/builders/ofbizTrunkFrameworkPlugins
If tests are running okay locally
Yes they are, at least now...
then we still have a faulty build logic.
Nope, it's not a "faulty build logic." just a transient error on build
servers
When you see that tests pass locally and not on Buildbot it's most certainly
an issue with servers.
Those are hard workers and make errors from time to time (which shows that
not only human make errors)
When something like that happens, I get to IRC (using
https://webchat.freenode.net)
There using a recognisable username (I use jleroux) I get to the ofbiz
channel.
I wait for ofbiz-bot to appear, click on query and then I can make a request
in the chat line to restart a scheduler.
A request is of the form
"force build ofbizTrunkFrameworkPlugins"
you put the text you want to appear after this expression, something like
"forces manual build after Buildbot weird 85 errors on labels :-o"
I hope this clarify. I'll put a .md file in tools\documentation\BuildBot and
will import is as HTML in a new wiki page for BuildBot
I hope this will help us to get to a fruitful collaboration :)
BTW we are still waiting for something to happen, see
https://issues.apache.org/jira/browse/INFRA-15785
Jacques
On Sun, Jan 7, 2018 at 5:20 PM, <build...@apache.org> wrote:
The Buildbot has detected a new failure on builder
ofbizTrunkFrameworkPlugins while building . Full details are available at:
https://ci.apache.org/builders/ofbizTrunkFrameworkPlugins/builds/10
Buildbot URL: https://ci.apache.org/
Buildslave for this Build: lares_ubuntu
Build Reason: downstream
Build Source Stamp: [branch ofbiz/ofbiz-framework/trunk] 1820457
Blamelist: jleroux
BUILD FAILED: failed shell_4
Sincerely,
-The Buildbot
