LoadBalancers fail after a day or so

I have 2 load balancers assigned to my web containers which run on ports 3080 and 3443. I am not exposing any ports explicitly for the web application and am using the Load Balancer application to map 80 -> 3080 and 443 -> 3443. After some arbitrary amount of time, the Load Balancer stops forwarding packets to the web container. Running v1.5.2 rancher server.

Has anyone:

  • Had this happen?
  • Come up with a solution?


The Load Balancers do not appear to be failing. Rather, the Rancher’s Managed Network appears to be at issue. As a test, I removed the Load Balancers and then exposed ports 80 and 443 directly from the web container(s). The containers still failed to present my application. Then, I toggled from managed network to bridge and it started to work immediately. Are there any logs I can submit to help the team troubleshoot this issue?


Specifying bridge network allows the containers to present the appropriate ports successfully. Two days later and all is working well still. This implies some intermittent problem with the Rancher Managed Networking.

I don’t use rancher’s network at all anymore…it’s too buggy…service discovery or ipsec network randomly fails if I’m running more then 20-30 containers…I’ve been waiting for over a year for rancher to become more stable and scalable…so far it hasn’t happened

@bacheson, thanks for the comment. I’m finding that not many rancher users seem to speak up about issues and there seems to be a culture of denial among devs. It’s odd.