[OPNFLWPLUG-500] He plugin: Null Pointer exception while running link scale tests Created: 12/Jun/15  Updated: 27/Sep/21  Resolved: 17/Aug/16

Status: Resolved
Project: OpenFlowPlugin
Component/s: General
Affects Version/s: None
Fix Version/s: None

Type: Bug
Reporter: SANDEEP GANGADHARAN Assignee: Hariharan Sethuraman
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

External issue ID: 3706

 Description   

1) Start karaf
2) Do feature install odl-openflowplugin-flow-services-ui-li
3) Run mininet with 10 switches in fully mesh topology.
4) Check if controller goes out of memory.
5) Check if controller has null pointer exceptions in karaf.log
6) Check if controller has learned all the expected links (10*10-1)=90
7) If 4,5,6 works repeat step 3-6 with switch count 10 + step(5)
8) If any of the steps from 4 to 6 fail stop the test.
9) For Lithium Redesign I see Null Pointer exceptions @ switch count = 10
10) For Master branch I see Null Pointer exceptions @ switch count = 25
11) There robot results link for each branch is as below.

Lithium Redesign
https://jenkins.opendaylight.org/releng/job/openflowplugin-csit-periodic-1node-cds-link-scalability-daily-lithium-redesign-only-master/3/robot/report/log.html

Master Branch
https://jenkins.opendaylight.org/releng/job/openflowplugin-csit-periodic-1node-cds-link-scalability-daily-only-master/3/robot/report/log.html

 Comments   
Comment by Luis Gomez [ 12/Jun/15 ]

I already added this to the RC0 observed issues.

Comment by Michal Rehak [ 15/Jun/15 ]

Hi, could you retest?
Those NPEs were fixed by:
https://git.opendaylight.org/gerrit/#/c/22169/

In build#4 the expection is still there. But #5 and #6 failed.

Comment by SANDEEP GANGADHARAN [ 15/Jun/15 ]

(In reply to michal rehak from comment #2)
> Hi, could you retest?
> Those NPEs were fixed by:
> https://git.opendaylight.org/gerrit/#/c/22169/
>
> In build#4 the expection is still there. But #5 and #6 failed.

The test is running in CI on stable/lithium. Last test ran on June 15. I could see the same exception.

Comment by Michal Rehak [ 22/Jun/15 ]

Regarding Li-redesigned codebase on master branch I found no NPE in last log:

https://jenkins.opendaylight.org/releng/job/openflowplugin-csit-periodic-1node-cds-link-scalability-daily-lithium-redesign-only-master/13/artifact/karaf.log

But for He codebase on master branch there are many NPEs:
https://jenkins.opendaylight.org/releng/job/openflowplugin-csit-periodic-1node-cds-link-scalability-daily-only-master/13/artifact/karaf.log

Here the NPE is logged with WARN severity. There is good chance that this is harmless. According to logs I guess that this happened during reconnection. Is there any tested functionality failing with this?

Comment by Luis Gomez [ 10/Nov/15 ]

So the bug here is fix the code so it does not throw NPEs.

Comment by Abhijit Kumbhare [ 10/Nov/15 ]

Hari,

Can you take a look at this when you get a chance? May be after M5 if you do not have time right now.

Abhijit

Comment by Andrej Leitner [ 17/Aug/16 ]

I checked logs of last successful csit jobs for openflowplugin-csit-1node-periodic-link-scalability-daily*:

-helium-redesign-only-boron
-helium-redesign-only-carbon
-lithium-redesign-only-beryllium
-lithium-redesign-only-stable-lithium
-only-beryllium
-only-boron
-only-carbon
-only-stable-lithium

and found NullPointerException (in StatisticsManagerImpl) only in the last job which is for he-design of plugin in stable/lithium codebase. Since we are after Lithium SRs and the issue is not present in actual Beryllium/Boron/Carbon codebase I am closing this bug.

Generated at Wed Feb 07 20:32:39 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.