Is consistent hash load balancing supported in multi-cluster service mesh?

I have two clusters that are configured as mutli-primary with separate networks. I put the exact same app in both clusters. The configuration included a destination rule that specifies consistent hash load balancing based on a cookie. It all appears to work until I try shutting down the pods in one of the clusters. When I do that I get some weird OPENSSL TLS error. Is this a bug potentially? Or is consistent hash load balancing not supported in multi cluster mode?