IB Fabric Errors causing failure In Exachk

Asset ID:	1-71-1954879.1
Update Date:	2017-07-11
Keywords:

Solution Type Technical Instruction Sure

Solution 1954879.1 : IB Fabric Errors causing failure In Exachk

Applies to:

Exadata X4-2 Hardware - Version All Versions and later
Exadata X3-2 Full Rack - Version All Versions to All Versions [Release All Releases]
Information in this document applies to any platform.

Goal

Avoid Exachk checking historical IB fabric errors

Solution

The problem with IB fabric errors is that they are cumulative. This means that over a long period of time, some errors will show up especially if any nodes or switches have been rebooted or cables moved etc.

Exachk has no way of knowing when these errors occurred and therefore it has to report them. Therefore, in order to ensure that exachk is examining the current state of the IB fabric and not an historical state, it is essential to clear the current IB fabric errors one or two hours before running exachk. The process for this is as follows:-

Log on to any IB leaf switch and run:-

1. ibclearerrors
2. ibclearcounters

Wait one or two hours for sufficient IB traffic to have flowed to "test" the IB fabric

3. Run exachk

This process ensures that exachk reports on the current performance of your IB fabric.

See also MOS document 1473622.1 for error counters specific to Virtual Lanes (VL) and MOS document 1538237.1 for general troubleshooting information.

Attachments

This solution has no attachment