Bench new-site hangs on containerized deployment

ERPNext
1

I’m setting up a new environment using helm to deploy in AWS. All containers deploy successfully, the common_site_config.json gets written and has the rds_db option enabled. The RDS database itself has the following options set:

character-set-client-handshake = FALSE
character_set_server = utf8mb4
collation_server = utf8mb4_unicode_ci

I can exec into the new-site container and manually log into the database, so that much is working. I’ve confirmed that the kubernetes secret for the db is present and correct (I used that password to log into db as root). The new-site container (a site gets created by the chart during deployment) never finishes. The site db never gets created and a site directory doesn’t appear. Manually trying to run bench new-site just hangs indefinitely. Adding ‘–verbose’ to the bench command doesn’t give any insight. This feels like a db problem but I’m stumped on how to troubleshoot.

2

try local setup/development setup, you can add print statements or setup debugger and try things there.

3

Thanks for the tip! I did as suggested and was able to tease out a redis problem. I was using an external redis cache and it was demanding authentication. Disabling the auth requirement allowed the site installation to progress.