Here is the Job definition I'm using and the debug log I have pasted at https://pastebin.com/2s2eTkxj
Please let me know if there are any difficulties in checking the log.

Job Definition:
device_type: ADT-HYBRID
job_name: Flash and Run FCT Test
timeouts:
  job:
    minutes: 50
  action:
    minutes: 20
  connection:
    minutes: 10
visibility: public
actions:
- command:
     name: relay_pwr_on
     timeout:
        minutes: 1
- deploy:
    failure_retry: 3
    to: flasher
    images:
      package:
              url: https://
- boot:
    method: u-boot
    commands:
       - setenv factorymode 1
       - boot
    failure_retry: 3
    auto_login:
      login_prompt: 'login: '
      username: root
      password_prompt: 'Password:'
      password: root
      login_commands:
      - ifconfig eth0
      - sysinfo
    prompts:
    - 'root@hon-grip'
    - 'root@GRIP'
    transfer_overlay:
        download_command: wget -S
        unpack_command: tar -C / -xzf
- test:
    failure_retry: 3
    timeout:
      minutes: 20
    definitions:
    - repository: ssh://git@github.com/gop/lava-tests.git
      from: git
      path: lava-test/hybrid-test.yaml
      name: test

Best Regards
Pavan Kumar


On Thu, Feb 8, 2024 at 3:12 PM Milosz Wasilewski <milosz.wasilewski@foundries.io> wrote:
On Thu, Feb 8, 2024 at 3:15 AM pawanKumar <pawan4749@gmail.com> wrote:
>
> Hi Milosz
>
> Yes, I understand that.
>
> But in my case I'm exporting results to bamboo using rest api. My issue is about failure retry case, if the action was failure for the first time it shows result as fail. Even though it is pass at second time, Rest API takes the first failure and displays as failure test result in bamboo.
>
> In this case implementation of failure retry would be not useful. So, is there a way to override/ignore the first failure case.

Sounds like a bug, do you have an example job with this behaviour?

Best Regards,
Milosz

>
>
> Thanks.
>
> On Wed, 7 Feb, 2024, 2:34 pm Milosz Wasilewski, <milosz.wasilewski@foundries.io> wrote:
>>
>> Pawan,
>>
>> I think you're confusing 2 different things. Job status only says if
>> all actions in the job completed successfully. For example if your
>> test script runs for too long, lava will call a timeout and end the
>> job as 'incomplete'. Similarly if your infrastructure doesn't work
>> properly (power_on command doesn't power the device on, etc). On the
>> other hand test results can have 4 values: pass, fail, skip, unknown.
>> The last one is only telling that something went wrong with data
>> collection.
>>
>> To sum up, failed test doesn't mean that the test job is incomplete.
>> These are statuses of 2 different things.
>>
>> Best Regards,
>> Milosz
>>
>> On Wed, Feb 7, 2024 at 4:08 AM pawanKumar <pawan4749@gmail.com> wrote:
>> >
>> > Hi Team,
>> >
>> > I was running the lava jobs and I found a difference in the status after job completion with failures. There were failures in
>> > i) The interactive test and lava overlay test [cases] but still the status of the job shows as "COMPLETE".
>> > ii) I tried an action failure retry, when action failure for the first time, it shows the result as failure but job status as complete. The retry- failure status should be ignored if it fails for the first time. I convert the test results into JUNIT using rest api. In the junit result it displays failure for the test case.
>> > My question is, How can we customize the status of the job or ignore the test result in case of failure-retry?
>> > Any Advice please.
>> >
>> >
>> >
>> > Best Regards
>> > Pavan Kumar
>> > _______________________________________________
>> > lava-devel mailing list -- lava-devel@lists.lavasoftware.org
>> > To unsubscribe send an email to lava-devel-leave@lists.lavasoftware.org