No locale meta information provided

On-Premise
1 
jed.js:253 Uncaught Error: No locale meta information provided.
    at a.dcnpgettext (jed.js:253)
    at a.gettext (jed.js:172)
    at N (locale.jsx:213)
    at Object.<anonymous> (index.jsx:134)
    at o (bootstrap:83)
    at Module.<anonymous> (apiMixin.jsx:13)
    at o (bootstrap:83)
    at Module.<anonymous> (indicators.jsx:22)
    at o (bootstrap:83)
    at Module.<anonymous> (installWizard.jsx:27)

Getting this error on UI console for sentry-onpremise setup. Are thoughts why this error would come. The the UI just get stuck after this

2

What is your browser language?

3

en-GB

4

And how are you installing Sentry?

5

So I will explain when exactly i am getting the issue. I have 1 machine where the sentry on-premise dockers are running. This machine is connected to 1 another machine which hosts the postgres and redis docker containers.
Now recently due to high CPU usage on my service machine, I added 1 more machine on which on-premise services are running and the system now works as a load balancer behind HAP. This new machine connects to the same database machine.
Now this issue comes when sentry services are running on both the machines. Its gone if on any the machine I stop the sentry docker containers, be it the new one or the old one.

Here is my assumption that may or may not be correct. May be not even close:
There is some unique identification through which services are connected to the database server. Keeping both the machines on is resulting into sharing the unique code due to with authentication fails.
Hope this help!

6

This sounds like a shared session issue between the web servers and Redis. If the frontend web servers are not the exact same version, all sorts of weird things can happen so I strongly urge you to make sure they are using the same version.

If this is not that, then I’m not sure if I understand your setup fully.

7

So I am using the same sentry version. There is common docker file getting used with this build
FROM sentry:9.1-onbuild
I also had the same thought that might be some session is getting shared but no idea how to fix it. If you can help me with some places where I can check

8

So this was a version issue. I found this message popping at the top

Celery workers are referencing a different version of Sentry (9.1.2 vs 9.1.1)

And this command fixed the issue

docker-compose run --rm web upgrade
2 Likes
9

That’s a bit suspicious as it only does DB migrations. I guess, you have also restarted/recreated your workers which then started to use the newer image?

10

So after I saw the error of Celery workers version on sentry dashboard, I googled a bit and found this command. I have 2 servers running sentry service and after running the command the 1st one started pointing to the 9.1.1 version but this command didn’t worked on the second server.
For second one then I had to remove all docker images, change the sentry version in Dockerfile from sentry:9.1-onbuild to sentry:9.1.1 and then run docker-compose again. After that the second server started pointing to the 9.1.1 version

11

@sunjha - any reason why you are not upgrading to 9.1.2 at least?

12

Yeah. I did that only. Accidentally mentioned 9.1.1 here

1 Like