Username: coppinp
Cluster: Baobab
/dpnc/beegfs is not mounted on the Baobab login node
(I didn’t check the compute nodes)
(baobab)-[coppinp@login1 ~]$ ls /dpnc
(baobab)-[coppinp@login1 ~]$
On Yggdrasil or Bamboo:
(yggdrasil)-[coppinp@login1 ~]$ ls /dpnc
beegfs
(yggdrasil)-[coppinp@login1 ~]$
Dear Paul,
Yes it’s possible that you had some problems to connect to this folder we had a scheduled a maintenance this afternoon :
baobab scheduled maintenance 20th of march 2025
Sorry for inconvenance,
Best regards,
Please note that this issue is still ongoing and already started yesterday evening.
(baobab)-[coppinp@login1 ~]$ ls /dpnc/beegfs
ls: cannot access '/dpnc/beegfs': No such file or directory
We knew about the maintenance and therefore did not contact you immediately, as we were thinking that the issue could perhaps be automatically resolved when Baobab came back online after the maintenance.
To give more info: /dpnc/beegfs is a BeeGFS filesystem with the BeeGFS client running on grid06. It is mounted on Baobab, Yggdrasil, and Bamboo as a NFS filesystem from grid06:/mnt/beegfs
On Yggdrasil and Bamboo we see the filesystem fine:
(yggdrasil)-[coppinp@login1 ~]$ ls /dpnc/beegfs
ams atlas cta dampe etc fast herd neutrino pan share sys usage users
but on the Baobab login node it seems that the mount is missing/broken. Again, I did not check the computing node. Can you have a look?
Cheers,
Paul
Dear Paul,
As explained in the previous post there are no issue now. Maybe you have wrong path :
(baobab)-[root@login1 ~]$ df -kh /srv/beegfs/dpnc/
Filesystem Size Used Avail Use% Mounted on
beegfs_dpnc 568T 500T 69T 88% /srv/beegfs/dpnc
(baobab)-[root@login1 ~]$ ls /srv/beegfs/dpnc/
groups
Best regards,
Yes, /srv/beegfs/dpnc/ is working fine.
BUT I am talking about another filesytem: /dpnc/beegfs
/dpnc/beegfs is the correct path. It is a 900 TB server set up and maintained by Yann Meunier and myself from the DPNC. You can check on Yggdrasil and Bamboo that the mount is there and working fine. On Baobab the mount is not there. More details on this filesystem are provided in my previous message.
Thank you though for the late night reply!
All the best,
Paul
Dear Paul,
Sorry for confusion, I check now and server refuse connection :
(baobab)-[root@login1 dpnc]$ mount grid06.unige.ch:/mnt/beegfs /mnt
mount.nfs: access denied by server while mounting grid06.unige.ch:/mnt/beegfs
It’s working on nodes :
(baobab)-[root@gpu034 beegfs]$ df -kh .
Filesystem Size Used Avail Use% Mounted on
grid06.unige.ch:/mnt/beegfs 916T 688T 229T 76% /dpnc/beegfs
Could you please check server side?
Best regards,
Hi Gael,
I added
129.194.9.190(rw,async,mp,no_subtree_check,fsid=200,crossmnt)
to
grid06:/etc/exports
and ran
exportfs -ra
Can you try again? As to the cause, is it possible that the IP address of the baobab login node changed?
All the best,
Paul
Dear Paul,
It’s now working fine 
grid06.unige.ch:/mnt/beegfs on /mnt type nfs4 (rw,relatime,vers=4.2,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,clientaddr=129.194.9.190,local_lock=none,addr=192.33.218.206)
We did not change the ip of the login node since a long time.
Best regards,
Dear Gael,
Ok, odd. I will check with Yann Meunier to try and understand if anything changed on our side that I was unaware of, and to prevent any issues in the future.
Thanks a lot for the help!
All the best,
Paul
Dear Paul,
I checked with Yann, and it seems we change route during previous maintenance because public interface has now 10g but we did not expect this result.
Sorry for this issue.
Best regards,
Okay, now I understand better! Thanks