We're having serious issues with the EBS plugin. It looks as if the plugin doesn't perform the necessary checks the cluster-wide status of a volume. We've seen issues both where rancher tries to attach an already attached volume to a host, and where it assumes that the volume is already attached when it isn't.
Because of this we're having big trouble both when removing failed docker hosts and adding new ones.
We're having serious issues with the EBS plugin. It looks as if the plugin doesn't perform the necessary checks the cluster-wide status of a volume. We've seen issues both where rancher tries to attach an already attached volume to a host, and where it assumes that the volume is already attached when it isn't.
Because of this we're having big trouble both when removing failed docker hosts and adding new ones.