90 lines
3.9 KiB
HTML
90 lines
3.9 KiB
HTML
<!DOCTYPE HTML>
|
|
<html lang="en-US">
|
|
<head>
|
|
<title>README for Quickstart Logs</title>
|
|
</head>
|
|
<body>
|
|
<h1>How to recreate this job</h1>
|
|
<p> Please refer to the <a href="README-reproducer-quickstart.html">recreation
|
|
instructions</a>
|
|
</p>
|
|
|
|
<h1>How to figure out what went wrong?</h1>
|
|
<p>Check the console log and search for <b>PLAY RECAP</b>. There are sometimes
|
|
multiple ansible runs in a job, usually the last one is the relevant.
|
|
If no <b>PLAY RECAP</b> text is found that usually means an infra failure
|
|
before Quickstart could even start. Try rechecking or asking on <i>#tripleo</i>
|
|
if there's an ongoing infra issue.</p>
|
|
|
|
<p>Look for a line above the <b>PLAY RECAP</b> that starts with
|
|
"<b>fatal:</b>". If no such line is found, try searching for other PLAY RECAP
|
|
lines or other error outputs.</p>
|
|
|
|
<p>If this "fatal" line contains the execution of a shell script and redirects
|
|
to a log, check which machine that task ran on. Look under that node's
|
|
directory in the logs to find the file.</p>
|
|
|
|
<p>Example output:<br/>
|
|
<br/><code>
|
|
fatal: [<b>undercloud</b>]: FAILED! => {"changed": true, "cmd": "set -o pipefail &&
|
|
/home/zuul/<b>overcloud-prep-images.sh</b> 2>&1 | awk '{ print
|
|
strftime(\"%Y-%m-%d %H:%M:%S |\"), $0; fflush(); }' >
|
|
/home/stack/<b>overcloud_prep_images.log</b>", "failed": true, "rc": 1}<br/>
|
|
<br/>
|
|
PLAY RECAP *********************************************************************<br/>
|
|
</code></p>
|
|
|
|
<p>In this case the <code>overcloud-prep-images.sh</code> script failed, which
|
|
is redirected to <code>/home/zuul/overcloud_prep_images.log
|
|
</code> on the undercloud.</p>
|
|
|
|
<p>If this is a different Ansible error, that could mean either an infra
|
|
problem (often has <b>UNREACHABLE</b> in the line) or a bug in Quickstart. Ask
|
|
on <i>#tripleo</i> to get help or open a bug on
|
|
<a href='https://bugs.launchpad.net/tripleo/+filebug'>Launchpad</a>. Add the
|
|
"ci" tag if it's a CI issue and "quickstart" if you suspect that the bug is in
|
|
Quickstart itself.</p>
|
|
|
|
<h1>Links to common log files</h1>
|
|
<ul>
|
|
<li><a href='undercloud/home/zuul/'>undercloud/home/zuul/</a>
|
|
- the source and log output of all templated shell scripts</li>
|
|
<li><a href='undercloud/var/log/'>undercloud/var/log/</a> -
|
|
directories and files copied from /var/log on the undercloud.
|
|
If other overcloud/subnodes exist, similar $node/var/log
|
|
directories will also exist in these logs.</li>
|
|
<li><a href='undercloud/var/log/extra/'>undercloud/var/log/extra/</a> -
|
|
extra system details like package list, and cpu info gathered from the
|
|
undercloud</li>
|
|
<li><a href='undercloud/var/lib/mistral'>undercloud/var/lib/mistral</a> -
|
|
output of all ansible used by config-download to drive the overcloud deployment</li>
|
|
<li><a href='stackwiz'>stackwiz</a> - tempest test results</li>
|
|
<li><a href='docs/build/'>docs/build/</a> - autogenerated documentation</li>
|
|
</ul>
|
|
|
|
<h1>Variables used in the job run</h1>
|
|
<p>The logs contain files showing variables used in the job run.</p>
|
|
<ul>
|
|
<li><a href='undercloud/var/log/extra/dump_variables_vars.json.txt.gz'>undercloud/var/log/extra/dump_variables_vars.json.txt.gz</a>
|
|
- contains the variables used in the running the actual test</li>
|
|
<li><a href='releases.sh'>releases.sh</a>
|
|
- the output of the script setting release-related variables</li>
|
|
<li><a href='playbook_executions.log'>playbook_executions.log</a>
|
|
- prints out the complete commands, with all expanded arguments,
|
|
to run the playbooks</li>
|
|
</ul>
|
|
|
|
<h1>Dry Run option</h1>
|
|
<p>As a debugging step, a job can be run manually with '-dryrun'
|
|
appended to the job name. When the "playbook dry run" option is invoked,
|
|
the playbooks will not execute and collect logs will not run but
|
|
certain log files, including 'toci_env_args_output.log', which
|
|
contains the environment variables used in the job, and
|
|
playbook_executions.log will still be produced in the logs
|
|
directory for inspection. This option serves to assist with
|
|
debugging and to test the testing scripts themselves.</p>
|
|
</body>
|
|
</html>
|
|
|
|
|