Our scripts take care of creating the comm database. However, AWS RDS initializes empty, causing
ecs tasks running the keyserver image to fail. This uses a null_resource to setup the comm database before launching keyserver nodes
Depends on D12688
Details
Details
On terraform apply, I no longer need to mysql remotely into aws rds to create the comm database. The comm database is first created and then the ecs tasks are launched
Diff Detail
Diff Detail
- Repository
- rCOMM Comm
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
| services/terraform/self-host/aws_db.tf | ||
|---|---|---|
| 104 ↗ | (On Diff #42213) | what does this line do? |
| services/terraform/self-host/aws_db.tf | ||
|---|---|---|
| 104 ↗ | (On Diff #42213) | So this ensures that the command is constantly run on every terraform apply as the changing timestamp always triggers the null resource. This might be overkill, but the case I'm accounting for is if AWS RDS needs to be relaunched or the database was dropped (which I had to do for testing) and the null_resource isn't destroyed with it. I think it's safe to put in for now, and maybe we can rethink how useful it is later |
| services/terraform/self-host/aws_db.tf | ||
|---|---|---|
| 104 ↗ | (On Diff #42213) | Ended up removing this. When one need to switch ip addresses, this prevents you from running terraform apply |