utah fails after reboot, but test continues to run on SUT

Bug #1179531 reported by Chris Gagnon
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
UTAH
Fix Released
High
Andy Doan

Bug Description

Description:
utah returns failure after testcase with reboot step, but test continues to run on SUT after utah has failed.

the console log can be found here:
http://10.97.2.10:8080/job/mir-glmark-daily-new/54/console

I've also attached the /var/lib/utah folder

Related branches

Revision history for this message
Chris Gagnon (chris.gagnon) wrote :
Changed in utah:
importance: Undecided → High
Andy Doan (doanac)
Changed in utah:
status: New → Confirmed
Revision history for this message
Andy Doan (doanac) wrote :

I think this is because our ssh.py code now throws a UTAHProvisioningException, but our is_utah_done logic is still just trying to catch socket errors. This causes the retry logic to fail. Still digging into this.

Revision history for this message
Andy Doan (doanac) wrote :

changing that exception seems to get things going. the log gets cluttered with messages of utah waiting for it to be done, but that's expected.

Andy Doan (doanac)
Changed in utah:
assignee: nobody → Andy Doan (doanac)
Revision history for this message
Chris Gagnon (chris.gagnon) wrote :

I've worked around this issue with this hack in my jenkins job.

get_results () {
rsync -avz --remove-source-files -e 'ssh -i /var/lib/jenkins/.ssh/utah -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no' jenkins@$machine:/tmp/results.json . || RETCODE=1
}

until get_results && cat -n results.json | grep "fragment-steps=0" | grep 'vertex-steps=0'; do
    sleep 10
done

Andy Doan (doanac)
Changed in utah:
status: Confirmed → Fix Committed
Changed in utah:
milestone: none → 0.12.4
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

Bug watches keep track of this bug in other bug trackers.