IBM DS6000 Series manual Avoiding single points of failure, Figure B-22 Single points of failure

Prior to multipath being available, some customers used OS/400 mirroring to two sets of disks, either in the same or different external disk subsystems. This provided implicit dual-path as long as the mirrored copy was connected to a different IOP/IOA, BUS, or I/O tower. However, this also required two copies of data. Since disk level protection is already provided by RAID-5 or RAID-10 in the external disk subsystem, this was sometimes seen as unnecessary.

With the combination of multipath and RAID-5 or RAID-10 protection in the DS6000, we can provide full protection of the data paths and the data itself without the requirement for additional disks.

Avoiding single points of failure

In Figure B-22,there are fifteen single points of failure, excluding the iSeries itself and the DS6000 storage facility. Failure points 9-12 will not be present if you do not use an Inter Switch Link (ISL) to extend your SAN. An outage to any one of these components (either

planned or unplanned) would cause the system to fail if IASPs are not used (or the applications within an IASP if they are).

1.IO Frame

2.BUS

3. IOP		6.	Port
4. IOA				14. Host Adapter
		7.	Switch
5. Cable		8.	Port
		9. ISL
10.	Port			13. Cable
11.	Switch
12.	Port

Figure B-22 Single points of failure

When implementing multipath, you should provide as much redundancy as possible. As a minimum, multipath requires two IOAs connecting the same logical volumes. Ideally, these should be on different buses and in different I/O racks in the iSeries. If a SAN is included, separate switches should also be used for each path. You should also use Host Adapters in different I/O drawer pairs in the DS6000. Figure B-23 on page 344 shows this.

Appendix B. Using the DS6000 with iSeries 343