Redis What if not fsync node.conf when update slots config ?

I migrate some slots in a huge redis cluster(500M*500S) for balance redis node load. As a pure memory cache & to save more money, we use cloud disk which is a shared, cheap hdd disk group. So, all the redis nodes in this redis cluster may share the same cloud disk.

When one slot config updated (run SETSLOT 10 Node <node ID>), all the redis nodes would write node.conf in one second, but our cloud disk is not able to process all of writes in a low latency(<100ms). Then, the fsync write node.conf(clusterBeforeSleep) in event loop may block the user requests.

Is fsync write node.conf to update slots config really necessary? If we remove the fsync when update slot config, what the worst scenario we may receive?

PS. We can't improve cloud disk as it supported by Cloud Service Vendor.