[lustre-discuss] Question about Install OFED with --with-nvmf

Xinliang Liu xinliang.liu at linaro.org
Wed Jun 28 02:56:41 PDT 2023


Hi Shuobin,

On Tue, 27 Jun 2023 at 15:01, 王烁斌 via lustre-discuss <
lustre-discuss at lists.lustre.org> wrote:

> Hi~
>
> I want to use RDMA between sever and storages.
>
> I try this and get error as follow:
> *root at 172-0-37-44 rdma]# nvme discover -t rdma -a 192.168.11.12 -5 4420*
> *Failed to write to /dev/nvme-fabrics: Invalid argument*
>
> log message in kernel :
> *26 20:44:22 172-0-37-44 kernel: nvme fabrics: no handler found for
> transport rdma.*
>
> After investigation, it was found that the issue was due to the "OFED"
> driver not being installed with the "-- with nvmf" option. So I tried the
> following installation method:
> *./mlnxofedinstall --add-kernel-support  --kmp --with-nvmf*
>
> *but I get this:*
> *WARNING: NVMEoF is not supported over kernel
> 4.18.0-425.3.1.el8_lustre.x86_64, will continue installation without it.*
>
>
> *so,How should I locate the problem?*
>

You can check the script where the WARNING prints, I think it should check
the kernel configuration something like this:
$ egrep "CONFIG_NVME_FABRICS|CONFIG_NVME_RDMA"  -rn
/boot/config-4.18.0-477.10.1.el8_8.aarch64
1810:CONFIG_NVME_FABRICS=m
1811:CONFIG_NVME_RDMA=m

Best,
Xinliang


>
> Thanks!
> _______________________________________________
> lustre-discuss mailing list
> lustre-discuss at lists.lustre.org
> http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20230628/e883062c/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 60689 bytes
Desc: not available
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20230628/e883062c/attachment-0001.png>


More information about the lustre-discuss mailing list