I’m trying to use AWS RDS Aurora (MySQL)
It seems that I need to run temporal-sql-tool manually. GitHub - temporalio/helm-charts: Temporal Helm charts
I wonder how to automate this properly?
I will use a fixed temporal docker image version, so how to make sure I get the right version of schema and tool? I need to git clone the repo with a specific tag? Or is this tool and schema it uses in the docker image already?
Lastly, I’m providing admin credential to temporal docker image, why not have this automatically setup when temporal service start for the first time?
How do I run this? I tried the following and none works
docker run -it --rm temporalio/admin-tools:1.7.0 bash
docker run -it --rm temporalio/admin-tools:1.7.0 sh
docker run -it --rm temporalio/admin-tools:1.7.0 /bin/bash
@Wenquan_Xing
No I did not figure out anything
I need to automate this
I tried
docker run --rm temporalio/admin-tools:1.7.0 /usr/local/bin/temporal-sql-tool
But the output is really strange
Something is wrong with this docker image
If I have to run it in the background then use exec to connect to it then there is no way to automate it
The entry point of the docker image is set to
“Entrypoint”: [“tail”,"-f","/dev/null"]
So it is not possible to run other commands
This docker image is meant to run forever
We need a different docker image in order to do just run temporal-sql-tool just once
Also my other question: why do we have to do this at all? Can’t this be part of the bootstrap?
HI, I got the same issue when trying to create the temporal db from the temporal-admintools container.
The SQL DB is based on AWS Aurora MySQL instance
This is the command I’m running within the container:
Update from me: I’m using the service docker image to setup the schema. And this is the exact image and version I’m using to run the service so no mis-match problem at all. I don’t trust setup schema from temporalio/admin-tools:x.y.z will work for temporalio/server:x.y.z. So here is what I’m running and hope it helps:
Using temporalio/server:x.y.z
And I’m not running this manually. I put this into a K8s job and run it once when deploy. So if I updated the image this happens automatically. This might be a bad idea.