<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
Hi all,</div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
<br>
</div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
I hit the same issue and pushed a fix for it. See:</div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
<a href="https://jira.whamcloud.com/browse/LU-19820">https://jira.whamcloud.com/browse/LU-19820</a></div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
<a href="https://review.whamcloud.com/c/fs/lustre-release/+/63536">https://review.whamcloud.com/c/fs/lustre-release/+/63536</a></div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
<br>
</div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
Aurélien</div>
<div id="appendonsend"></div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>De :</b> lustre-discuss <lustre-discuss-bounces@lists.lustre.org> de la part de Mark Dixon via lustre-discuss <lustre-discuss@lists.lustre.org><br>
<b>Envoyé :</b> jeudi 22 janvier 2026 10:33<br>
<b>À :</b> Christopher J Orr <cjorr@purdue.edu><br>
<b>Cc :</b> lustre-discuss@lists.lustre.org <lustre-discuss@lists.lustre.org><br>
<b>Objet :</b> Re: [lustre-discuss] DKMS build broken with NVIDIA doca packages</font>
<div> </div>
</div>
<div class="BodyFragment"><font size="2"><span style="font-size:11pt;">
<div class="PlainText">External email: Use caution opening links or attachments<br>
<br>
<br>
Hi Christopher,<br>
<br>
We previously used a similar approach but, with the (very welcome!!) move<br>
to using DKMS on EL, DOCA now supports multiple kernels at the same time<br>
and so maintains a per-kernel ofa source directory - a blanket default is<br>
no longer appropriate.<br>
<br>
In fact, on one of my test hosts /usr/src/ofa_kernel/default ended up<br>
becoming a dangling link. Not sure if that was a bug, or if DOCA has given<br>
up on it.<br>
<br>
Unless Jon gets there first, I'll get a ticket opened when I get to it.<br>
<br>
Best,<br>
<br>
Mark<br>
<br>
On Wed, 21 Jan 2026, Christopher J Orr wrote:<br>
<br>
> [You don't often get email from cjorr@purdue.edu. Learn why this is important at
<a href="https://aka.ms/LearnAboutSenderIdentification">https://aka.ms/LearnAboutSenderIdentification</a> ]<br>
><br>
> [EXTERNAL EMAIL]<br>
><br>
> This is how I ended up fixing it on Lustre 2.14.0_ddn191 on Rocky 9.7<br>
> with DOCA-OFED.<br>
><br>
> ------------------------------------------------------------------<br>
> --- lustre-dkms_pre-build.sh.orig 2026-01-06 16:55:25.428285300 -<br>
> 0500<br>
> +++ lustre-dkms_pre-build.sh 2026-01-06 18:00:28.357307490 -0500<br>
> @@ -9,8 +9,9 @@<br>
><br>
> case $1 in<br>
> lustre-client)<br>
> + [ -f /etc/sysconfig/lustre ] && . /etc/sysconfig/lustre<br>
> SERVER="--disable-server"<br>
> - KERNEL_STUFF=""<br>
> + KERNEL_STUFF="${KERNEL_STUFF:-}"<br>
> ;;<br>
><br>
> lustre-zfs|lustre-all)<br>
> ------------------------------------------------------------------<br>
><br>
> ...and then, add<br>
> KERNEL_STUFF="--with-o2ib=/usr/src/ofa_kernel/default/"<br>
> ...to /etc/sysconfig/lustre<br>
><br>
> I hope this helps!<br>
> Thanks,<br>
> Christopher Orr<br>
><br>
><br>
> On Wed, 2026-01-21 at 16:16 +0000, Patrick Farrell via lustre-discuss<br>
> wrote:<br>
>><br>
>> ---- External Email: Use caution with attachments, links, or sharing<br>
>> data ----<br>
>><br>
>><br>
>><br>
>><br>
>> Folks, if you want to create a JIRA ticket, you can ask for an<br>
>> account. We're very happy to get contributions.<br>
>><br>
>><br>
>> Regards,<br>
>> Patrick<br>
>><br>
>><br>
>> From: lustre-discuss <lustre-discuss-bounces@lists.lustre.org> on<br>
>> behalf of Jon Marshall via lustre-discuss<br>
>> <lustre-discuss@lists.lustre.org><br>
>> Sent: Wednesday, January 21, 2026 9:36 AM<br>
>> To: Mark Dixon <mark.c.dixon@durham.ac.uk><br>
>> Cc: lustre-discuss@lists.lustre.org <lustre-discuss@lists.lustre.org><br>
>> Subject: Re: [lustre-discuss] DKMS build broken with NVIDIA doca<br>
>> packages<br>
>><br>
>><br>
>><br>
>><br>
>><br>
>><br>
>> Hi Mark,<br>
>><br>
>><br>
>> Thanks for confirming I'm not on my own - I've not got any further,<br>
>> other than starting to look at creating a dummy RPM package that fits<br>
>> the criteria Lustre is looking for! That or using a very clunky<br>
>> wrapper script around rpm itself to lie to the configure script. I<br>
>> actually have got this second approach working so there is nothing<br>
>> wrong with building against the doca packages, but its a bit annoying<br>
>> to automate the build process for our servers like this.<br>
>><br>
>><br>
>> I've not got access to create a Jira ticket myself either.<br>
>><br>
>><br>
>> Cheers<br>
>> Jon<br>
>><br>
>><br>
>> From: Mark Dixon <mark.c.dixon@durham.ac.uk><br>
>> Sent: Wednesday, January 21, 2026 12:23<br>
>> To: Jon Marshall <Jon.Marshall@cruk.cam.ac.uk><br>
>> Cc: lustre-discuss@lists.lustre.org <lustre-discuss@lists.lustre.org><br>
>> Subject: Re: [lustre-discuss] DKMS build broken with NVIDIA doca<br>
>> packages<br>
>><br>
>><br>
>><br>
>><br>
>> Hi Jon,<br>
>><br>
>> As it happens, I've been looking at the same thing. I hadn't spotted<br>
>> LU-18002 (thanks), but unfortunately it isn't enough to accommodate<br>
>> the<br>
>> move to dkms on rhel.<br>
>><br>
>> I don't know how far you've got since Monday, but there now seems a<br>
>> need<br>
>> for an explicit check of /usr/src/ofa_kernel (as it's no longer owned<br>
>> by a<br>
>> package) and the "find" for rdma_cm.h needs the -L flag to make sense<br>
>> of<br>
>> the new maze of twisty passages.<br>
>><br>
>> I think that a new jira ticket needs to be opened...<br>
>><br>
>> Cheers,<br>
>><br>
>> Mark<br>
>><br>
>><br>
>> On Mon, 19 Jan 2026, Jon Marshall via lustre-discuss wrote:<br>
>><br>
>>> [EXTERNAL EMAIL]<br>
>>> Hi,<br>
>>><br>
>>> I'm in the process of rebuilding lustre on Rocky 8.10 and have<br>
>>> noticed that NVIDIA have been messing around with their packages<br>
>>> again, now rebranding everything under the doca label. For LTS<br>
>>> purposes we're sticking with 2.15.8 for lustre, and I'm trying to<br>
>>> get this to build with NVIDIA DOCA 3.2.1 LTS.<br>
>>><br>
>>> The trouble is, it seems they have rename the package mlnx-<br>
>>> ofa_kernel-devel to mlnx-ofa_kernel-dkms. Looking at the DKMS<br>
>>> configure script, it is searching for:<br>
>>> O2IBPKG="mlnx-ofed-kernel-dkms"<br>
>>> O2IBPKG+="|mlnx-ofed-kernel-modules"<br>
>>> O2IBPKG+="|mlnx-ofa_kernel-devel"<br>
>>> O2IBPKG+="|compat-rdma-devel"<br>
>>> O2IBPKG+="|kernel-ib-devel"<br>
>>> O2IBPKG+="|ofa_kernel-devel"<br>
>>><br>
>>> And hence it can't find the package (underscore instead of hyphen),<br>
>>> which causes the build to fail.<br>
>>><br>
>>> Digging around the JIRA, I found<br>
>>> this<<a href=""></a>https://linkprotect.cudasvc.com/url?a=https%3a%2f%2fjira.whamc<br>
>>> loud.com%2fbrowse%2fLU-<br>
>>> 18002%3fjql%3dtext%2520~%2520dkms%2520ORDER%2520BY%2520created%2520<br>
>>> DESC&c=E,1,jSSRk0tXHMx8RQEMnGYEBCTdjBWE-<br>
>>> 7d4UZni7OYRCsspax3v09_1sRG4eF9iy77rKx5DppDWrhVsH9ZQ7lk_1OT3Wmb_XeUj<br>
>>> WfNuEPbhpR8,&typo=1> issue, but it looks to only have been fixed in<br>
>>> 2.16, which we've sort of ruled out at this stage. Looking at the<br>
>>> actual<br>
>>> patch<<a href=""></a>https://linkprotect.cudasvc.com/url?a=https%3a%2f%2freview.wh<br>
>>> amcloud.com%2fc%2ffs%2flustre-<br>
>>> release%2f%20%2f55625%2f4%2flnet%2fautoconf%2flustre-<br>
>>> lnet.m4&c=E,1,Wi5eGkf0dY16u2VrGeX06tAPDP6YCLAJhfgPURLolu4ssfvLF8Xiw<br>
>>> PpqpixQifO1NdxtNZ5tpz8FAqP5gd419t_Yvuu_c-<br>
>>> NzIAY1JvTjYeVLYQ,,&typo=1>, it seems pretty minor and I was<br>
>>> wondering if this could be back ported to 2.15 as well.<br>
>>><br>
>>> I can work around by building things myself, but I was hoping to be<br>
>>> able to yum install the packages direct from the whamcloud repos,<br>
>>> as this greatly simplifies my rollout.<br>
>>><br>
>>> Cheers<br>
>>> Jon<br>
>>><br>
>>><br>
>>> Jon Marshall<br>
>>><br>
>>> High Performance Computing Specialist<br>
>>><br>
>>><br>
>>><br>
>>> IT and Scientific Computing Team<br>
>>><br>
>>><br>
>>><br>
>>> Cancer Research UK Cambridge Institute<br>
>>><br>
>>> Li Ka Shing Centre | Robinson Way | Cambridge | CB2 0RE<br>
>>><br>
>>> Web<<a href="http://www.cruk.cam.ac.uk/">http://www.cruk.cam.ac.uk/</a>> |<br>
>>> Facebook<<a href="http://www.facebook.com/cancerresearchuk">http://www.facebook.com/cancerresearchuk</a>> |<br>
>>> Twitter<<a href=""></a>https://linkprotect.cudasvc.com/url?a=http%3a%2f%2ftwitter.<br>
>>> com%2fCR_UK&c=E,1,aCcWa5p892R3_9Lj1VLXiO9wgithO5AHQZh841zayJAVcOaCk<br>
>>> JC2gyGFMTpTADviZ3xtPn6klyCExiJqHjg1k5lzggxNNPrsaIis62wIBwOJ&typo=1><br>
>>><br>
>>><br>
>>><br>
>>> [Description: CRI Logo]<<a href="http://www.cruk.cam.ac.uk/">http://www.cruk.cam.ac.uk/</a>><br>
>>><br>
>>><br>
>> _______________________________________________<br>
>> lustre-discuss mailing list<br>
>> lustre-discuss@lists.lustre.org<br>
>> <a href="http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org">http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org</a><br>
><br>
_______________________________________________<br>
lustre-discuss mailing list<br>
lustre-discuss@lists.lustre.org<br>
<a href="http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org">http://lists.lustre.org/listinfo.cgi/lustre-discuss-lustre.org</a><br>
</div>
</span></font></div>
</body>
</html>