[Solved] Rancher on AWS - huge EBS IOPS usage


We have been using rancher for around 15 days and its awesome!
We are running it on Amazon WS but now we have a problem with the cluster pricing. We have 2 machines - a master node - (t2.small) and node (t2.micro). There is agent running also on master node. The cluster is using a db.t2.small RDS MySQL instance. The problem is that for just 80 hours in which the cluster was almost idle (~100 visits), it created 350 million IOPS to the EBS volume. Do you have any idea if we can reduce the IOPS usage? I will update the question once I know if the IOPS are against the database volume or the node’s volume. All volumes are EBS.


Problem is solved!

Just in case anyone is having the same problem:
The load was on the master node. We were using rancher/server:v1.1.0-rc5. Once i upgraded to the stable 1.1.0 the usage dropped significantly! You can see on the monitoring details:

Rancher is awesome, thanks for the great software and keep it up!

1 Like