Redundancy configurations help protect your network from physical adapter failures as well as Virtual I/O Server failures.
With N_Port ID Virtualization (NPIV), you can configure the managed system so that multiple logical partitions can access independent physical storage through the same physical fibre channel adapter. Each virtual fibre channel adapter is identified by a unique worldwide port name (WWPN), which means that you can connect each virtual fibre channel adapter to independent physical storage on a SAN.
Similar to virtual SCSI redundancy, virtual fibre channel redundancy can be achieved using Multi-path I/O (MPIO) and mirroring at the client partition. The difference between traditional redundancy with SCSI adapters and the NPIV technology using virtual fibre channel adapters, is that the redundancy occurs on the client, because only the client recognizes the disk. The Virtual I/O Server is essentially just a pipe. The second example below uses multiple Virtual I/O Server logical partitions to add redundancy at the Virtual I/O Server level as well.
Example: Host bus adapter failover
This example uses Host bus adapter (HBA) failover to provide a basic level of redundancy for the client logical partition. The figure shows the following connections:
The storage area network (SAN) connects physical storage to two physical fibre channel adapters located on the managed system.
The physical fibre channel adapters are assigned to the Virtual I/O Server and support NPIV.
The physical fibre channel ports are each connected to a virtual fibre channel adapter on the Virtual I/O Server. The two virtual fibre channel adapters on the Virtual I/O Server are connected to ports on two different physical fibre channel adapters in order to provide redundancy for the physical adapters.
Each virtual fibre channel adapter on the Virtual I/O Server is connected to one virtual fibre channel adapter on a client logical partition. Each virtual fibre channel adapter on each client logical partition receives a pair of unique WWPNs. The client logical partition uses one WWPN to log into the SAN at any given time. The other WWPN is used when you move the client logical partition to another managed system.
The virtual fibre channel adapters always has a one-to-one relationship between the client logical partitions and the virtual fibre channel adapters on the Virtual I/O Server logical partition. That is, each virtual fibre channel adapter that is assigned to a client logical partition must connect to only one virtual fibre channel adapter on the Virtual I/O Server, and each virtual fibre channel on the Virtual I/O Server must connect to only one virtual fibre channel adapter on a client logical partition.
The client can write to the physical storage through client virtual fibre channel adapter 1 or 2. If a physical fibre channel adapter fails, the client uses the alternative path. This example does not show redundancy in the physical storage, but rather assumes it would be built into the SAN.
Note: It is recommended that you configure virtual fibre channel adapters from multiple logical partitions to the same HBA, or you configure virtual fibre channel adapters from the same logical partition to different HBAs.
Example: HBA and Virtual I/O Server failover
This example uses HBA and Virtual I/O Server failover to provide a more advanced level of redundancy for the client logical partition. The figure shows the following connections:
The storage area network (SAN) connects physical storage to two physical fibre channel adapters located on the managed system.
There are two Virtual I/O Server logical partitions to provide redundancy at the Virtual I/O Server level.
The physical fibre channel adapters are assigned to their respective Virtual I/O Server and support NPIV.
The physical fibre channel ports are each connected to a virtual fibre channel adapter on the Virtual I/O Server. The two virtual fibre channel adapters on the Virtual I/O Server are connected to ports on two different physical fibre channel adapters in order to provide redundancy for the physical adapters. A single adapter could have multiple ports.
Each virtual fibre channel adapter on the Virtual I/O Server is connected to one virtual fibre channel adapter on a client logical partition. Each virtual fibre channel adapter on each client logical partition receives a pair of unique WWPNs. The client logical partition uses one WWPN to log into the SAN at any given time. The other WWPN is used when you move the client logical partition to another managed system.
The client can write to the physical storage through virtual fibre channel adapter 1 or 2 on the client logical partition through VIOS 2. The client can also write to physical storage through virtual fibre channel adapter 3 or 4 on the client logical partition through VIOS 1. If a physical fibre channel adapter fails on VIOS 1, the client uses the other physical adapter connected to VIOS 1 or uses the paths connected through VIOS 2. If VIOS 1 fails, then the client uses the path through VIOS 2. This example does not show redundancy in the physical storage, but rather assumes it would be built into the SAN.
Considerations
These examples can become more complex as you add physical storage redundancy and multiple clients, but the concepts remain the same. Consider the following points:
To avoid configuring the physical fibre channel adapter to be a single point of failure for the connection between the client logical partition and its physical storage on the SAN, do not connect two virtual fibre channel adapters from the same client logical partition to the same physical fibre channel adapter. Instead, connect each virtual fibre channel adapter to a different physical fibre channel adapter.
Consider load balancing when mapping a virtual fibre channel adapter on the Virtual I/O Server to a physical port on the physical fiber channel adapter.
Consider what level of redundancy already exists in the SAN to determine whether to configure multiple physical storage units.
Consider using two Virtual I/O Server logical partitions. Since the Virtual I/O Server is central to communication between logical partitions and the external network, it is important to provide a level of redundancy for the Virtual I/O Server. Multiple Virtual I/O Server logical partitions require more resources as well, so you should plan accordingly.
NPIV technology is useful when you want to move logical partitions between servers. For example, in active Partition Mobility, if you use the redundancy configurations above, in combination with physical adapters, you can stop all the I/O activity through the dedicated, physical adapter and direct all traffic through a virtual fibre channel adapter until the logical partition is successfully moved. The dedicated physical adapter would need to be connected to the same storage as the virtual path. Since you cannot migrate a physical adapter, all I/O activity is routed through the virtual path while you move the partition. After the logical partition is moved successfully, you need to set up the dedicated path (on the destination logical partition) if you want to use the same redundancy configuration you had configured on the original logical partition. Then the I/O activity can resume through the dedicated adapter, using the virtual fibre channel adapter as a secondary path.
No comments:
Post a Comment