Kestra keeps using old revision of a flow about kestra HOT 4 CLOSED

Thanks, if needed you can send me additional information by DB on slack.

Hopefully, I just reproduced the issue, I'm not sure how I reproduced it but my local environment is now using an old revision so I can try to investigate more easily.

from kestra.

Noticed the same issue again - this time with Realtime Kafka trigger. Kestra keeps using old revision unless the whole flow is disabled/re-enabled

from kestra.

Hi @aku I struggle to reproduce the issue and looking at the code I didn't saw any race that could explain this.

Can you give me more context to help me understand the issue:

• How do you deploy Kestra: standalone or separate process?
• On which infrastructure do you deploy Kestra: Docker compose, Kubernetes, other?
• How many replicas of each components?
• Which backend do you use (Postgres, MySQL, H2, Kafka/Elasticsearch)?

from kestra.

Hi @loicmathieu

Since I have multiple critical issues that are not reproducible using a fresh install, I'm thinking of exporting my local data (DB, S3, configs) and share it with your team. Also, I will describe exactly how I set everything up (unit files, configs, etc.)

As for your questions:

• I deploy separate components - webserver, scheduler, executor, worker using Ansible playbooks
• Everything is running on 4 Debian12 VMs without a docker.
• 1 webapp + 1 scheduler in a dedicated VM, 1 executor + 1 worker in a separate VM (x3)
• PG + S3, I'm running community edition

from kestra.