What is the reason for your question ? Maintenance ?
Do you mean a normal Failover or a Site-Failover ? In a normal takeover the aggregate is still available, but in a Site-Failover the aggregate can't be used.
But if this occurs (e.g. controller A fails) and controller B have no access to a certain aggregat of controller A at this time, a kernel panic is possible. right ?
It should not panic. Some aggregates will be missing after takeover, that's all. Of course if root aggregate itself is missing, no takeover is possible.
