On 8/5/22 16:29, Phelan, Michael wrote:
>> -----Original Message-----
>> From: Ilya Maximets <[email protected]>
>> Sent: Friday 5 August 2022 15:08
>> To: Phelan, Michael <[email protected]>
>> Cc: [email protected]; Stokes, Ian <[email protected]>; Aaron Conole
>> <[email protected]>; ovs-dev <[email protected]>
>> Subject: Re: [ovs-build] |fail| pw1664033 [ovs-dev] system-traffic: Don't run
>> IPv6 header modification test on kernels < 5.19.
>>
>> On 8/5/22 15:59, Phelan, Michael wrote:
>>> Hey Ilya,
>>> I've tried recreating the test failure but haven't had any success so I'm 
>>> not sure
>> why exactly this test failed on the CI run. I did notice that it also failed 
>> on
>> another CI run with the output here:
>> https://mail.openvswitch.org/pipermail/ovs-build/2022-August/024240.html.
>>
>> Do you have a full log for this test?  This tail of a log file doesn't show 
>> the actual
>> failure.
> 
> Unfortunately no, the logs of each individual unit test aren't retained.

Can we make a full log being posted with the report or at least the
actual diff that failed?

> 
>>
>>>
>>> It looks like this test was only merged to master yesterday through this
>> commit:
>> https://github.com/openvswitch/ovs/commit/2ff43c78c685dd7ce8bbef61ef52a
>> 4f9ce98bfc7. It could be that the test is unreliable but I will keep an eye 
>> out for
>> anymore failures over the next few days.
>>
>> The test should be fairly reliable in the userpsace testsuite, I can't 
>> reproduce the
>> failure on my 2 systems (rhel8 and ubuntu 22.04).
>> And there is no much difference between this test and other similar
>> ipv6 test around it.  SO, I'm not sure what is going on.

I think, I found the issue with the test.  It failed on one of my
systems after a few hundreds of runs.  The issue appears to be
incorrect setting of the neighbor cache for fc00::1.  It is set
up in the wrong namespace, so if the ND request didn't have a chance
to be sent before we clear OVS rules, it can not be resolved.

Will test for a bit more with the fix and post another patch.

>>
>> The same test in check-kernel will fail without the fix applied, but you're 
>> not
>> running check-kernel IIUC.
> 
> That is correct, we are not running check-kernel.
> 
>>
>>>
>>> Let me know if there is anything more I can do.
>>>
>>> Kind regards,
>>> Michael.
>>>
>>>
>>>> -----Original Message-----
>>>> From: Ilya Maximets <[email protected]>
>>>> Sent: Friday 5 August 2022 13:52
>>>> To: Phelan, Michael <[email protected]>
>>>> Cc: [email protected]; Stokes, Ian <[email protected]>; Aaron
>>>> Conole <[email protected]>; ovs-dev <[email protected]>
>>>> Subject: Re: [ovs-build] |fail| pw1664033 [ovs-dev] system-traffic:
>>>> Don't run
>>>> IPv6 header modification test on kernels < 5.19.
>>>>
>>>> On 8/5/22 14:51, Ilya Maximets wrote:
>>>>> On 8/5/22 14:44, [email protected] wrote:
>>>>>> Test-Label: intel-ovs-compilation
>>>>>> Test-Status: fail
>>>>>> http://patchwork.ozlabs.org/api/patches/1664033/
>>>>>>
>>>>>> AVX-512_compilation: failed
>>>>>> DPLCS Test: fail
>>>>>> DPIF Test: success
>>>>>> MFEX Test: success
>>>>>> Errors in DPCLS test:
>>>>>
>>>>>> 9. system-traffic.at:229: 9. datapath - ping6 between two ports
>>>>>> with header modify (system-traffic.at:229): FAILED
>>>>>> (system-traffic.at:252)
>>>>>
>>>>
>>>> CC: s/ovs-build/ovs-dev/
>>>>
>>>>> Hi, Michael.  Could you, please, check what is wrong with this test?
>>>>>
>>>>> The tail of the testsuite log doesn't contain any useful
>>>>> information, it doesn't even show what exactly failed.
>>>>>
>>>>> Best regards, Ilya Maximets.
>>>
> 

_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to