Problems after cloudformation stack install

Product Support
1

I recently subscribed to the ttne AWS Marketplace service and launched the cloudformation script. A few config options later, the script successfully finished and I got the ‘CREATE_COMPLETE’ message.

After a while of not being able to log into the web console, I checked the logs and found this message repeating in a neverending loop:

Apr 06 20:49:52 ip-10-0-0-4.ec2.internal systemd[1]: Started The Things Stack.
Apr 06 20:49:52 ip-10-0-0-4.ec2.internal systemd[1]: Starting The Things Stack…
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal tti-lw-stack[13602]: INFO Valid license issuer=tti license_id=aws-marketplace-0001 metering=AWS valid_from=2019-10-29 00:00:00 +0000 UTC
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal tti-lw-stack[13602]: INFO Setting up core component
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal tti-lw-stack[13602]: INFO Setting up Identity Server
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal tti-lw-stack[13602]: error:cmd/internal/shared:initialize_identity_server (could not initialize Identity Server)
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal tti-lw-stack[13602]: — error:pkg/errors:request (request to postgres://postgres:gRk5fp>CFD3&4(g@<DB_ID>.us-east-1.rds.amazonaws.com:5432/ttn_lorawan failed)
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal tti-lw-stack[13602]: url=postgres://postgres:gRk5fp>CFD3&4(g@<DB_ID>.us-east-1.rds.amazonaws.com:5432/ttn_lorawan
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal tti-lw-stack[13602]: — net/url: invalid userinfo
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal systemd[1]: lorawan-stack.service: main process exited, code=exited, status=255/n/a
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal systemd[1]: Unit lorawan-stack.service entered failed state.
Apr 06 20:49:53 ip-10-0-0-4.ec2.internal systemd[1]: lorawan-stack.service failed.
Apr 06 20:49:56 ip-10-0-0-4.ec2.internal systemd[1]: lorawan-stack.service holdoff time over, scheduling restart.

Any pointers would be very much appreciated!

2

It seems like the database isn’t accessible which is weird. Can you check if the is.database-uri field matches the username and password of the RDS database?

3

Actually, the password was the reason. While the data in my files for your suggested field was correct, I relaunched the script with a new DB password (without special characters in it), and voila! It worked.
Something to pass on to the QA team I guess.

Thank you for your help!

4

Yes Indeed. I think we can add a validation to the field.

5

This topic was automatically closed 24 hours after the last reply. New replies are no longer allowed.