2

I'm just set up a small load balanced architecture in AWS using two Large EC2 instances in the same availability zone with a non-sticky ELB in front. Amazon reports that both instances are health and I see the healthchecks returning 200 on both servers, yet all traffic is being routed to one of the servers. If I pull that server out of rotation traffic goes to the other server, but when I put it back in it all goes back to that one server. This is live traffic - not a load test - so its not related to single-client-IP or sticky sessions.

Any suggestions for what I can look at or how to begin to debug the elastic load balancer setup? Thanks

1
  • I'm going to assume your backends are (were as this is old) in different AZs and you had a limited set of clients which were caching DNS or you were using an A record for the ELB. – Nathan V Sep 8 '15 at 22:15
1

Is there a stickiness policy for the ELB? Check in the AWS Console, EC2 > Load Balancers > [your ELB] > Description. Under Port Configuration it should say "Stickiness: Disabled".

2

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.