The identifying features of a type 1 configuration (as shown in Figure 1) are:
Figure 1. Type 1 configuration
Dual Controller
Single Controller
The type 2 configuration can occur with or without switches, as shown in Figure 2 and Figure 3.
Figure 2. Type 2 configuration - With switches.
The identifying features of a type 2 configuration are:
Figure 3. Type 2 configuration - Without switches
In a type 1 configuration there are no externally managed switches to aid in debugging. The diagnostic tools available are the HBA Loopback test, FAStT MSJ (from the host adapter end) and the Controller Loopback test (from the RAID controller end). If you intend to diagnose a failed path while using the alternate path for production, be sure that you are familiar with the tools and the loop connections so that the correct portion is being exercised and you do not unplug anything in the active path.
For a type 2 configuration, use the features of the switches and the capability of MSCS to isolate resources from the bad or marginal path before beginning debug activities. Switches allow a view of log information that shows what problems have been occurring, as well as diagnostics that can be initiated from these managed elements. Also, a type 2 configuration has the capability to have more than one RAID controller unit behind a switch. In the diagnostic maps, the switches are referred to generically as concentrators. Figure 4 shows a type 2 configuration with multiple controller units.
Figure 4. Type 2 configuration
with multiple controller units
An example sequence for debugging a type 2 MSCS configuration is shown in the following sequence of figures.
You can attach multiple server pairs to the switches by using zoning or partitioning for pair isolation or combinations of type 1 and type 2 configurations. Break the larger configuration into its smaller sub elements and work with each piece separately. In this way you can remove the good path and leave only the bad path, as shown in the following sequence.
Figure 5. Offline controller B
Figure 6. All I/O flowing through controller A
Controller A is processing I/Os and is unavailable for diagnostics
The elements of the paths shown in Figure 7 are as follows:
Note: Some HBAs do not have integrated transceivers and therefore require an SFP.