4.8.3 Using the verify-topology Utility

The verify-topology utility can be used to identify various network connection problems.

The problems you can diagnose using verify-topology include:

• Missing InfiniBand Network Fabric cable
• Missing InfiniBand Network Fabric connection
• Incorrectly-seated cable
• Cable connected to the wrong endpoint

The utility is available in the ibdiagtools directory on all servers. To view the options for the verify-topology utility, use the following command:

./verify-topology -h

[ DB Machine Infiniband Cabling Topology Verification Tool ]
Usage: ./verify-topology 
    [-v|--verbose]
    [-r|--reuse (cached maps)]
    [-m|--mapfile]
    [-ibn|--ibnetdiscover (specify location of ibnetdiscover output)]
    [-ibh|--ibhosts (specify location of ibhosts output)]
    [-ibs|--ibswitches (specify location of ibswitches output)]
    [-t|--topology [torus | fattree | halfrack] default is fattree]

Example 4-1 Using verify-topology to Identify Cables Seated Incorrectly

The following is an example shows the output when using the verify-topology utility. In the example, the error shows the cables are connected incorrectly. Both cables from the server are going to same InfiniBand Network Fabric switch. If the switch fails, then the server loses connectivity to InfiniBand Network Fabric network.

[ DB Machine Infiniband Cabling Topology Verification Tool ]

Bad link:Switch 0x21283a8371a0a0 Port 11A - Sun Port 11B
        Reason : 2.5 Gbps Speed found. Could be 10 Gbps
        Possible cause : Cable isn't fully seated in

Bad link:Switch 0x21283a89eba0a0 Port 11B - Sun Port 11A
        Reason : 2.5 Gbps Speed found. Could be 10 Gbps
        Possible cause : Cable isn't fully seated in

Is every external switch connected to every internal switch..........[SUCCESS]
Are any external switches connected to each other....................[SUCCESS]
Are any hosts connected to spine switch..............................[SUCCESS]
Check if all hosts have 2 CAs to different switches..................[ERROR]
Node trnA-db01 has 1 endpoints. (Should be 2)
Port 2 of this node is not connected to any switch

--------fattree End Point Cabling verification failed-----

Leaf switch check: cardinality and even distribution.................[ERROR]

Internal QDR Switch 0x21283a8371a0a0 has fewer than 4 compute nodes
It has only 3 links belonging to compute nodes
Check if each rack has a valid internal ring.........................[SUCCESS]