Hi,
With delivery of May, we can now defined timer for each test of a job. Some of our tests are not correctly executed and then timeout expired. But we want to execute next test and not stop the job with an incomplete verdict How can we configure this behavior in yaml ?
See following example of yaml used
Thanks in advance for your answer Regards
[Description: Description: Description: Description: Description: Description: logo_big5]
Florence ROUGER-JUNG | TINA: 166 7356 | Tel: +33 2 44 02 73 56 | Mobile: +33 6 13 49 38 02 STMicroelectronics | MCD Auriga Building | 9-11, rue Pierre-Félix Delarue | 72100 Le Mans | FRANCE
On 22 May 2017 at 08:37, Florence ROUGER JUNG florence.rouger-jung@st.com wrote:
Hi,
With delivery of May, we can now defined timer for each test of a job.
Some of our tests are not correctly executed and then timeout expired.
Timeouts are fatal - if any part of a LAVA test job times out, that is an error, so the job will be incomplete.
Failure-retry exists to attempt the same operation again but if there are "allowable" timeouts, that must be handled by the test writer using a custom script that knows exactly how long to wait before giving up, clean up after the broken operation and then exit zero so that the next stage can occur. The test definition then calls that script instead of the original command.
Think about how this would be done if you were running the same test on a local system and create a script which can handle the failing test. In that script, when the failure is detected, if lava-test-case exists in $PATH, use that to report --result fail before exiting zero. If lava-test-case is not in $PATH, use echo or print etc to describe what has happened.
https://validation.linaro.org/static/docs/v2/writing-tests.html#write-portab...
On 22 May 2017 at 10:26, Neil Williams neil.williams@linaro.org wrote:
On 22 May 2017 at 08:37, Florence ROUGER JUNG florence.rouger-jung@st.com wrote:
Hi,
With delivery of May, we can now defined timer for each test of a job.
Some of our tests are not correctly executed and then timeout expired.
Timeouts are fatal - if any part of a LAVA test job times out, that is an error, so the job will be incomplete.
Failure-retry exists to attempt the same operation again but if there are "allowable" timeouts, that must be handled by the test writer using a custom script that knows exactly how long to wait before giving up, clean up after the broken operation and then exit zero so that the next stage can occur. The test definition then calls that script instead of the original command.
Think about how this would be done if you were running the same test on a local system and create a script which can handle the failing test. In that script, when the failure is detected, if lava-test-case exists in $PATH, use that to report --result fail before exiting zero. If lava-test-case is not in $PATH, use echo or print etc to describe what has happened.
https://validation.linaro.org/static/docs/v2/writing-tests.html#write-portab...
This approach doesn't work always. If you schedule 2 test and first misbehaves (stalls or freezes the target), the test shell eventually times out. In v1 the remaining test shells are executed after reboot. In v2, the test job is terminated. The only workaround for this issue I can think of at the moment is to run each test as a separate job.
milosz
--
Neil Williams
neil.williams@linaro.org http://www.linux.codehelp.co.uk/
Lava-users mailing list Lava-users@lists.linaro.org https://lists.linaro.org/mailman/listinfo/lava-users
lava-users@lists.lavasoftware.org
-
Florence ROUGER JUNG -
Milosz Wasilewski -
Neil Williams