The hastop -all command on VCS cluster node with AlternateIO resource and StorageSG having service groups may leave the node in LEAVING state

On a VCS cluster node with AlternateIO resource configured and StorageSG attribute contain service groups with Zpool, VxVM or CVMVolDG resources, hastop -local or hastop -all commands may leave the node in "LEAVING" state.

This issue is caused by lack of dependency between service group containing LDom resource and service groups containing storage resources exported to logical domain in alternate I/O domain scenarios. In this scenario VCS may attempt to stop the storage service groups before stopping logical domain which is using the resources.

Workaround: Stop the LDom service group before issuing hastop -local or hastop -all commands.