[lustre-discuss] Mounting troubles on Rocky 9.6 and 2.17.50 tag

[Aurelien Degremont]

Hi Stepan,

I'm suspecting a configuration issue. Could you describe more in details your setup (nodes, lnet config, for server and clients).

llmount.sh is starting everything locally, on the local node by default.
Then you have this "not available" and client keeps connecting that needs to be fixed.

Small I/O working and large I/O not working, and both related to errors makes me think the small I/O goes to the client cache, and the large I/O are flushed sooner, and the system is reporting an error right away. The cached I/O will eventually fail when the system will try to flush them.

What is your setup: client, server, "lctl list_nids" on each. Also check everything is "UP" in "lctl dl" output on each node.
Also, what was your format options for your MDT and OSTs?


Aurélien

________________________________
De : lustre-discuss <lustre-discuss-bounces at lists.lustre.org> de la part de Stepan Beskrovnyy via lustre-discuss <lustre-discuss at lists.lustre.org>
Envoyé : mercredi 14 janvier 2026 12:31
À : lustre-discuss at lists.lustre.org <lustre-discuss at lists.lustre.org>
Objet : [lustre-discuss] Mounting troubles on Rocky 9.6 and 2.17.50 tag

External email: Use caution opening links or attachments

Hi Everyone!

Building from source lustre tag v2.17.50 on custom Rocky Linux 9.6
Linux Kernel version 5.14.0-570.58.1 .
Building from source via tutorial https://wiki.whamcloud.com/pages/viewpage.action?pageId=427393157
Build goes well, kernel modules loading successfully.
llmount.sh goes well too, and fio benchmark works successfully on test filesystem
Then, I start mounting mdt and oss on nodes, I get error (See first picture) that lustre cant found mount for OST(with mdt exact same problem). [image.png]
But mounting processes seemed to be ok.
After mounting all filesystem and trying to cp something on it, I have problems like (See picture two) connect is not available and lustre cant find target. [image.png]
If I copy small files like 10Kb, copying ends successfully, with the same errors.
If I copy bigger directory, like ior, then copying freeze with errors like on second picture.
Any thoughts ? What do I do wrong ? How can I debug this?
Thanks,
Stepan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20260114/2e816b83/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 33731 bytes
Desc: image.png
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20260114/2e816b83/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image.png
Type: image/png
Size: 171891 bytes
Desc: image.png
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20260114/2e816b83/attachment-0003.png>