Hi. I’ve got one issue I can’t crack. Suddenly, no...
# gooddata-cn
m
Hi. I’ve got one issue I can’t crack. Suddenly, no insights can be loaded. Not even one simple attribute. Without any changes to the DS or LDM. I can query the DB without any problem. Can I access some logs to find out what can be the problem? I couldn’t find anything in the docs. Cheers.
z
is there any error in the dev tools?
m
Hi, Zuzka. Just one error for Execute (404). I’ve also tried testing the DS via API and that was ok.
j
take workspace id from the afm execute call from the web console and try to access /api/entities/workspaces/<id>
sorry, nonsense, I can see the catalog in your AD, workspace must exist
it seems like something is not running in your deployment. Is it all-in-one image or k8s deployment?
Btw have you changed anything? e.g. any upgrade?
m
It’s AIO (Community Edition). There was really no update. I’ve been just adding some insights on the dashboard, everything has been working fine. Suddenly, probably after some dash save, no insight can be loaded.
I’m waiting for Petr to check the docker log for some issues.
j
docker logs contain the root cause. Let's wait
m
First of all, restart helped. I’m sharing docker logs for investigation: https://docs.google.com/document/d/1003deF5vEMNzJozY40Ld0QnqU63dopHtXqga9axR_kk/edit?usp=sharing
j
@Robert Moucha @Ondrej Stumpf can you look into the logs ^^^? Seems like Calcique was not in a good shape...
o
I can't see root cause in the logs, the services are now complaining they cannot reach Calcique, but the reason why must have happened earlier. Any chance there is something interesting in earlier logs? Perhaps an OOM? Or (and the behavior does ring a bell for me) - a Pulsar error?
❤️ 1
m
@Petr Šimeček Are you able to search the earlier logs, please?
p
I don't know. "our" Gooddata.cn is running in the Linux screen on our temporary DigitalOcean VM. I captured the logs directly from stdout. If there is a default dir with log files, I can provide it easily. If not, I need a guidance how to deploy it properly then 'docker compose up'. Sorry for hassles, once we push it closer to production, our SRE will take care of it. It is just a POC at this stage.
r
Well, the 404 is confusing, I already sen this before. The correct error in this caase should be 503 or so. The 404 is actually a secondary error, when nginx reverse proxy could not find the file /usr/share/nginx/html/50x.html 😞
Filed internal issue NAS-3565 for fixing the missing error document