Upgrade from 1.1.1 to 1.2.0

[aauer]

Describe the problem/error/question

Deployed a k8s adaptation of n8n with GitHub - 8gears/n8n-helm-chart: A Kubernetes Helm chart for n8n a Workflow Automation Tool. Easily automate tasks across different services..
Could deploy versions until 1.1.1.

The migration to 1.2.0 is struggling. The Redis connection can’t initialize to an AWS Elasticache with this version (or a greater).
The 1.2.0 integrate the Redis cluster management, maybe it has other integrations which could crash the connections as my redis isn’t deployed as a cluster.

What is the error message (if any)?

The logs are blocked to those.

Loading config overwrites [ '/n8n-config/config.json' ]
UserSettings were generated and saved to: /home/node/.n8n/config
2023-09-05T08:01:02.763Z | debug    | Lazy Loading credentials and nodes from n8n-nodes-base "{\n  credentials: 345,\n  nodes: 433,\n  file: 'DirectoryLoader.js',\n  function: 'loadAll'\n}"
2023-09-05T08:01:02.869Z | debug    | Initialising Redis client of type subscriber connection with host: my-redis-master.default.svc.cluster.local and port: 6379 "{ file: 'RedisServiceHelper.js', function: 'getRedisStandardClient' }"

The landing page is only prompting : n8n is starting up. Please wait
Deployment in local isn’t blocking, only productions on AWS are.
Also, if removing scaling (redis) n8n can start.

Information on your n8n setup

• n8n version: 1.1.1 to 1.2.0 and greater
• Database : PostgreSQL
• **Running n8n via (Docker, npm, n8n cloud, desktop app): Helm **

[Jon]

Hey @aauer,

Welcome to the community

As far as I know we didn’t make any changes with Redis between 1.1.1 and 1.2.0 so it sounds like this could be something environmental.

Which version of Redis are you using and what happens if you just leave n8n for a bit to start up does it eventually start?

[aauer]

Hi @Jon and thanks for your quick feedback.

I’m currently using redis in v6.2. I was waiting few hours without any success …
My env var isn’t changing between those 2 versions, i’m not sure where to look at …

[Jon]

Hey @aauer,

I would imagine 6.2 is ok, Have you tried upgrading to 1.4 to see if that is any different?

[aauer]

It’s still the same with more than 2hours passed

Loading config overwrites [ '/n8n-config/config.json' ]
UserSettings were generated and saved to: /home/node/.n8n/config
2023-09-05T09:58:00.667Z | debug    | Lazy Loading credentials and nodes from n8n-nodes-base "{\n  credentials: 344,\n  nodes: 431,\n  file: 'DirectoryLoader.js',\n  function: 'loadAll'\n}"
2023-09-05T09:58:00.923Z | debug    | Initialising Redis client of type subscriber connection with host: master.xxxx and port: 6379 "{ file: 'RedisServiceHelper.js', function: 'getRedisStandardClient' }"

As i don’t have any more output i can’t know where to search … I tried to change any kind of env variables to make it work without any success …

[netroy]

I think this might be caused by TLS on the redis connection.

Can you please try this custom docker image n8nio/n8n:redis-tls, and set the env variable QUEUE_BULL_REDIS_TLS to true ?

[aauer]

Works like a charm ! You’re amazing
As indication what are the considerations to make it works ?

[netroy]

This was fixed in fix(core): Add support for in-transit encryption (TLS) on Redis connections by netroy · Pull Request #7047 · n8n-io/n8n · GitHub and will be included in the 1.6.0 release tomorrow.

[jan]

New version [email protected] got released which includes the GitHub PR 7047.