IBM System Storage DS4000, Hardware Manual

Page: 439 / 566

Chapter 7. Advanced maintenance, troubleshooting, and diagnostics
421
Missing path(s) to host
When multiple logical drives associated with different hosts (or host groups) suddenly failover
onto their non-preferred controllers in the same direction, then the fault is likely to be closer to
the DS4000 or DS5000 storage subsystem. In this case, the first place to check is the
Recovery Guru summary of problems to see whether there are any other outstanding failures,
such as failed SFP. If not, then take a look in the Major Event Log (MEL) for a possible
controller reset or other host-side incidents at the time of failover. It is also be worth checking
the status and error count on the Fibre Channel switch ports that connect to the DS4000 or
DS5000 storage subsystem. This could expose some marginal links to the host port(s) on the
DS4000 or DS5000 controllers.
If logical drive(s) associated with a single host or host group fail over onto their non-preferred
controllers while other logical drives remain unaffected, then the fault is likely to closer to the
host itself. In this instance, we need to focus just on the affected host(s) and paths to it.
If storage partitions are used, then the Storage Subsystem Profile and the host properties in
the mapping section shows the WWPN of each HBA device in a host that the controllers are
configured to see, as shown in Figure 7-63.
Figure 7-63 Host properties
Unfortunately, this does not provide a real-time view of the actual devices that are currently
accessible. One way to check this is through the SAN switches. Here, we need to look at what
devices are logged into the fabric and also what devices are zoned to see each other. This
can sometimes be difficult with larger SAN configurations.
There is an alternate way to obtain a real-time view of devices accessible from each controller
by analyzing the Capture State Information output. This is a text file containing low-level shell
data from both controllers. The Capture State Information file can be generated from the
Storage Manager Subsystem Management window by selecting Advanced

Troubleshooting

Capture State Information.