Disable nfs version4 in proxmox 8.xx

[waltar]

nfsd "hung" so looks not to network related, would more check to controller/cable/connections related to access your disks. Still not clear what happens here.
When nfsd cannot surf data it could be by reason of cannot read, cannot write or cannot resolve permission access for user/group/acl ...
Do you have any kind of win-ad (sssd/winbind) connection configured which is not answering in time ?

 

[phanos]

nfsd "hung" so looks not to network related, would more check to controller/cable/connections related to access your disks. Still not clear what happens here.
When nfsd cannot surf data it could be by reason of cannot read, cannot write or cannot resolve permission access for user/group/acl ...
Do you have any kind of win-ad (ssd/winbind) connection configured which is not answering in time ?

No nothing like that and the disk that this sharing the nfs directories is accessible on the server after the crash. I can read/write/execute etc on the same directories that are no longer mounted on the nfs clients.

Of course this I check after the crash. There is no way for me to check the time of the crash if the disk is accessible. Are you thinking that a disk might be becoming inaccessible at some point? If for some reason the hard disk becomes inaccessible during some read/writes could this cause the crash and the nfs server can not longer recover when the disk becomes accessible again?

Maybe an issue with the sata controller?

I have this in the motherboard

PCH Built-in Storage    Intel® C2550 : 2 SATA3 6.0 Gb/s, 4 SATA2 3.0 Gb/s
Additional SATA Controller    Marvell SE9172: 2 SATA3 6.0 Gb/s, support RAID 0, 1
Marvell SE9230: 4 SATA3 6.0 Gb/s, support RAID 0, 1, 10

 

[waltar]

Are you thinking that a disk might be becoming inaccessible at some point?

Yes, of course.

 

[phanos]

Yes, of course.

I see. If indeed nfs server becomes unresponsive and can not recover for this after then this could be it. I will run the nfs server 3 for a few days and see if this help somehow. In the meantime I will check to see any compatibility issues of my motherboard sata controller with linux. I remember in the past that I read something regarding marvel controller and linux but this was a long time ago and the previous proxmox was running fine so I never gave it enough thought.

If all this fails it could very well be a hardware error.

 

[phanos]

Hi after running nfs version 3 for about two weeks I am here to report that I experience no crash and everything is running better.

I also look into the Marvel Sata 9230 controller which I read online that other users reported in the past that it had some issues in some older linux distributions. I upgraded the firmware of this controller 4 days ago and everything is still running fine.

I might in the future switch to nfs4 to check if the firmware upgrade had any effect on my setup but right now I am reluctant to do so since everything is running smoothly. If I eventually test nfs4 it I will report back.

Thanks everyone for the help and comments.

 

[bessome]

we have same problem on proxmox 8.2 with nfs 4.2 version. With nfs version 3 is better but NFS freezes for 2 minutes and then recovers.
in logs many:

[Mon Oct 28 12:53:41 2024] rpc_check_timeout: 547 callbacks suppressed
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 not responding, still trying
[Mon Oct 28 12:53:41 2024] call_decode: 547 callbacks suppressed
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK
[Mon Oct 28 12:53:41 2024] nfs: server 10.18.98.123 OK

 

[waltar]

Maybe firewall packet filter problems between ... ??

 

[bessome]

The firewall is disabled, and this happens on a standalone test server when NFS is used locally. We migrated from Proxmox 6 to Proxmox 8, and this issue appeared in the new version; it didn’t occur before.

 

[bessome]

this issue does not occur with the new kernel 6.8.12-2-pve

 

[bessome]

Got excited too soon, after a few hours, timeout errors started appearing again, even though there’s no significant load.

 

[platini]

Hi after running nfs version 3 for about two weeks I am here to report that I experience no crash and everything is running better.

I also look into the Marvel Sata 9230 controller which I read online that other users reported in the past that it had some issues in some older linux distributions. I upgraded the firmware of this controller 4 days ago and everything is still running fine.

I might in the future switch to nfs4 to check if the firmware upgrade had any effect on my setup but right now I am reluctant to do so since everything is running smoothly. If I eventually test nfs4 it I will report back.

Thanks everyone for the help and comments.

How is that going? I have the exact same problem with proxmox 8.xx. Few months ago i had that problem and changed to kernel 6.5 and seemed to mitigate bit. In the last week i had two nfs hangups with similar panic msgs. Did disabling nfs4 worked for you?