missing ifup-ib in latest release?

Hi … I have some old cluster nodes that were working fine under previous versions of CentOS 7 (I think it was CentOS 7.3 before update) but after doing a recent update to CentOS 7.5 I can’t seem to get the interface to come up. I reinstalled the latest MLNX_OFED drivers (MLNX_OFED_LINUX-4.3- which installed properly. I see the card in lspci and the kernel modules seem to be loaded as well. However, I can’t seem to bring up the interface. Doing an ifup I get this:

ifup ib0

ERROR : [/etc/sysconfig/network-scripts/ifup-eth] Device ib0 does not seem to be present, delaying initialization.

Which seemed weird to me that it was trying to use the ifup-eth code instead of the ifup-ib code to bring up the interface. When I looked for this file I don’t see it on the system with the mlnx_ofed software installed. If I don’t install mlnx_ofed and just leave the CentOS drivers installed the card comes up fine. I also notice this comes from the rdma-core package from CentOS:

rpm -qf /etc/sysconfig/network-scripts/ifup-ib


When I look at the mlnx_ofed installed machine I don’t see an rdma-core package…

rpm -qa | grep rdma




So I’m wondering if I am missing something with this? Previous versions I didn’t seem to have any issues with getting it installed and using it. Anyone have some advice as to what I should look at further to figure this out? Thanks,

Hi ,

Could you please check if ib0 interface found under “ifconfig -a” ?

If not, i suggest the following:

  1. invoke mst start → mst status → ifconfig and check again

  2. Try to restart the interfaces:

  • /etc/init.d/openibd restart

  • opensm start or start the SM on the switch

  1. If the above still not working , create interface manually :

vi /etc/sysconfig/network-scripts/ifcfg-ib0









Hi… not sure why this came up and sent me a message today about it. Yea that solution didn’t work as the IB interfaces aren’t working properly. I suspect it’s related to this:




I discovered this problem with the kernel after a few weeks of posting this. I need to look back at this and see if this issue has been fixed in the latest kernel or if it’s still broken.