[Beowulf] How to debug error with Open MPI 3 / Mellanox / Red Hat?

Faraz Hussain info at feacluster.com
Wed May 1 08:05:10 PDT 2019

Quoting John Hearns <hearnsj at googlemail.com>:

> Errrr..   you are not running a subnet manager?
> DO you have an Infiniband switch or are you connecting two servers
> back-to-back?

Unfortunately, I am not familiar with a subnet manager. These are  
sixteen machines in an HP enclosure. Fourteen of them are running RHEL  
6.9. Two we put centos 7.5 to test out. I assume the Mellanox stuff is  
all built-in to this enclosure? Or are there some configuration steps  
I need to do?

> Also - have you considered using OpenHPC rather tyhan installing CentOS on
> two servers?
> When you expand this manual installation is going to be painful.

Good idea about Open HPC. I have read about it and will check it out again.

More information about the Beowulf mailing list