Datasheet

15
2.5.7 Operational considerations
Remember that the IBM i host partition’s state may affect the IBM i client partition.
For instance, when the hosting partition is restarted, the client partition loses contact with its disk.
This results in a system reference code sequence of A6xx0255 shown for the client partition. The
client VSCSI adapter loses its virtual path to the server when the server is rebooted and any I/O
that was in flight is returned to the device drivers as "aborted". The client’s disk device driver will
retry the operation, and receive failure responses while the server is still unavailable. Eventually
the client’s device driver gives up and returns an unsuccessful response to the microcode’s
Storage Management code. That microcode puts up the 0255 SRC and retries the aborted
operations indefinitely until the disks are reconnected. There will be no data loss/data corruption
as long as the client partition remains active until the disk path is reconnected.
As for Ethernet traffic in this same scenario, the TCP subsystem in the hosting partition would
come down, so the virtual networking in the client partition would no longer have a job to
communicate with. So the client partition would lose communications during this scenario. The
management console supplied IBM i client console would need to be used to restart the
communications.
If the hosting partition is put into restricted state, the Ethernet communications for the client
partition would stop, but the disk activity, say for a running batch job, would continue to run. The
Network Server Description (NWSD) in the hosting partition is not automatically varied off and is
not part of an ended subsystem, so the storage spaces associated with the NWSD remain
available to the client partition.