On Wed, 27 Mar 2019 at 16:31, Fabiano Ferronato fabiferro@hotmail.com wrote:
Em ter, 26 de mar de 2019 às 17:18, Milosz Wasilewski milosz.wasilewski@linaro.org escreveu:
On Tue, 26 Mar 2019 at 16:12, Antonio Terceiro antonio.terceiro@linaro.org wrote:
On Fri, Mar 15, 2019 at 05:36:15PM +0100, Fabiano Ferronato wrote:
Hi all,
after submitting a watchjob request for a test already finished in lava side, the test status (OK, Fail) is visible in /admin but fetch is not happening automatically. It works if I start fetch manually.
curl 0.0.0.0:5000/api/watchjob/lava/63_test/ver0315/board1 --header "Auth-Token: xxxxx76d4a6a5646abfa9d53xxxxxxxxx" --form "backend=lava_swtest" --form testjob_id=38721
Could you explain how can I make it to fetch automatically?
Do you have all the required daemons running?
I think it's not only that. I also noticed that jobs added with 'watchjob' API are not fetched on LAVA events. They are (probably) only checked on schedule defined in the backend but I didn't verify if this is the case. So I think there is a bug. It was never high enough on my priority list to check closer.
milosz
This is documented in https://squad.readthedocs.io/en/latest/install.html
Please let us know if you have suggestions on how to make the documentation more clear _______________________________________________ Squad-dev mailing list Squad-dev@lists.linaro.org https://lists.linaro.org/mailman/listinfo/squad-dev
Squad-dev mailing list Squad-dev@lists.linaro.org https://lists.linaro.org/mailman/listinfo/squad-dev
Hi Antonio and Miloz! Thanks for answering.
I'm happy to say that Squad / Lava basic integration is working. I'm using 'watchjobs' for now because our setup is multinode. So, I can't start multinode tests but I can 'watchjobs' and arrange the information in a suitable way. I can say that SQUAD is a very nice and useful tool.
Previously, I said that fetch was not working, latter I've found that it was taking approximately 1h to happen and now it is taking less than 5 minutes . I still don't know why.
Hmm, this sounds weird. 1h looks reasonable as this is default timeout after which the job results are fetched. 5 min suggests that the jobs are fetched after ZMQ event. You can shorten the polling timeout from 1h to sth shorter. I can't explain te 5 min for the moment.
Now I'm trying to understand the Threshold Settings and how to send related notifications.
there are 2 timeouts. 1st timeout is a 'grace period' no notifications will be send before it expires. This means it's time to add your test jobs to the build. Second timeout is meant to trigger the report unconditionally (whether all results are collected or not).
About documentation, it's ok for what I've used so far (installation, submit, watch). Could be interesting to include links to git as examples. I've found really helpful examples there. I've found no documentation about the Docker image.
Yes, that's a known problem. We're not using docker images in production yet, so there is very little written about that. I created docker-compose that starts LAVA and SQUAD: https://github.com/mwasilew/lava-docker-compose
I've tried Ansible scripts but it didn't work from the very beginning (something related to encrypted files). I'm not used to Ansible, so this could be a noob thing.
it might be related to version of ansible. I had similar problem with encrypted files when using older version - they were used as if they were plain text.
milosz
Best Regards, Fabiano