My `iblinkinfo` dont work on host

As the title says, I execute sudo iblinkinfo on the host-side of server A and it reports the error:

ibwarn: [9475] _do_madrpc: send failed; Invalid argument

ibwarn: [9475] mad_rpc: _do_madrpc failed; dport (DR path slid 0; dlid 0; 0)

/var/tmp/rdma-core/rdma-core-54mlnx1/libibnetdisc/ibnetdisc.c:811; Failed to resolve self

discover failed

However, if I execute the same command on the dpu(Bluefield 2) side of server A, it succeeds.

In fact, when I send RDMA requests from another server B to the host side and dpu side of server A, they fail on the host side of A and succeed on the dpu side of A.

Could someone please provide me some ideas to solve the problem, thanks a lot.

P.S.

  • The link state are normal as ibstatus works fine
  • The ofed version of host and dpu side are both MLNX_OFED_LINUX-5.4-1.0.3.0
  • Both host and dpu side are equipped with Ubuntu OS