https://jira.opendaylight.org This file is an XML representation of an issue en-us 8.20.10 820010 22-06-2022 https://jira.opendaylight.org/browse/INTTEST-118 integration-test <p>It seems that some change came in mid-June that has reduced the<br/> max device count we can get with <a href="https://jenkins.opendaylight.org/releng/view/CSIT/job/netconf-csit-1node-scale-max-devices-only-magnesium/plot/Netconf%20Devices/" class="external-link" target="_blank" rel="nofollow noopener">this job</a> by aprox 50%</p> <p>The claim is strictly coming from looking at the graphs we get in jenkins:</p> <p><a href="https://jenkins.opendaylight.org/releng/view/CSIT/job/netconf-csit-1node-scale-max-devices-only-magnesium/plot/Netconf%20Devices/" class="external-link" target="_blank" rel="nofollow noopener">magnesium</a></p> <p><a href="https://jenkins.opendaylight.org/releng/view/CSIT/job/netconf-csit-1node-scale-max-devices-only-aluminium/plot/Netconf%20Devices/" title="https://jenkins.opendaylight.org/releng/view/CSIT/job/netconf-csit-1node-scale-max-devices-only-magnesium/plot/Netconf%20Devices/" class="external-link" target="_blank" rel="nofollow noopener">aluminium</a></p> <p> </p> <p>The problem could be anywhere, including some change in the basic infra<br/> environment. One clue is that both jobs seemed to have dropped in the same<br/> time frame.</p> INTTEST-118

max device count lost 50%

Bug Medium Resolved Done Jamo Luhrsen Jamo Luhrsen Mon, 13 Jul 2020 22:00:43 +0000 Thu, 16 Jul 2020 23:44:37 +0000 Thu, 16 Jul 2020 23:44:37 +0000 0 1 <p><a href="https://jira.opendaylight.org/secure/ViewProfile.jspa?name=jluhrsen" class="user-hover" rel="jluhrsen">jluhrsen</a> will run these jobs with an older distribution to see if we can eliminate the infra being the issue.</p> <p>re-running the jobs in the sandbox with a distro built back when scale results were<br/> better still resulted in low numbers, which ruled out some bug on the controller as<br/> a whole.</p> <p>I did find something that looks like a race condition in the test logic. The fix<br/> in in test code and the sandbox results are now coming back up in the 850 range.</p> <p>Also, I think maybe this scale drop maybe started happening right when our infra<br/> and test code was moving from python2 to python3. Maybe some performance gain in<br/> our test logic caused the issue to be more prevalent.</p> <p>closing this now.</p> Development Rank 0|i03tnb: