Connection (public external) problems on Baobab after maintenance 28 August 2026

Primary informations

Username: $singhgu3
Cluster: $baobab

Description

Thank you for performing the maintenance and making the cluster available since today 28th August. Howoever I am facing some public internet connection problems on baobab.

I am running some gpu jobs on the cluster that I launched some hours ago at ealy afernoon, which require a public internet connection to upload metrics of my experiments and some images of simulation on wandb. I am noticing that my logs are being saved locally in my directory on the cluster.

Howoever the connection is not stable or is being cut, I believe. I see that my jobs are still running, but on the wandb platform i see that logs are stopped. When launching my jobs they connect to my wandb user logs few logs, but after few minutes I see the no logs on wandb, and the wandb jobs are being automatically stopped. But on the other side, inspecting my jobs on the cluster, they are all still running.

I also notice the following error logged in the logging file :/home/users/s/singhgu3/grey-box-interpolants multirun/2026-08-28/16-04-47/.submitit/12096158_2/12096158_2_0_log.out

--- Logging error ---
--- Logging error ---

Can you please confirm me that there is stil some connection issue on baobab?

Best regards,

Gurjeet

Not sure if it is connected but in my job on Baobab I got:

IOError: [Errno 121] Remote I/O error
/var/spool/slurmd/job12097655/slurm_script: line 28: echo: write error: Remote I/O error

line 28 is simply: ‘echo $(date) - All done.’

Update 31-08: For reference, I talked with other members of our group and their jobs are also all failing. It seems to be a systematic issue since Baobab came back online.

Yes, I am also facing this error. I am trying to load a model from PyTorch using pytorch.load(...)

I am having the exact same error I/O: [Errno 121] Remote I/O error.
“”"
Evaluating outputs/swe/2b5f4757ab806d37135d863f38907365/st-inr-vf/black_box/b0abd8f3a588687e1279150830b19171/seed_4/best_chkpt.pth
Batch size for evaluation: 15, Number of batches: 7
skipped ([Errno 121] Remote I/O error)
“”"

I have a similar problem on Baobab also.

WARNING: skipping mount of /srv/beegfs/scratch/users/s/strassep/output: stat /srv/beegfs/scratch/users/s/strassep/output: communication error on send
INFO: Terminating squashfuse_ll after timeout
INFO: Timeouts can be caused by a running background process
FATAL: container creation failed: mount hook function failure: mount /srv/beegfs/scratch/users/s/strassep/output->/srv/beegfs/scratch/users/s/strassep/output error: while mounting /srv/beegfs/scratch/users/s/strassep/output: while getting stat for /srv/beegfs/scratch/users/s/str
assep/output: stat /srv/beegfs/scratch/users/s/strassep/output: communication error on send
srun: error: gpu018: task 0: Exited with exit code 255

Dear Users,

Since the Baobab migration, we have been encountering storage issues following an update to the latest major release. You can track the status here:

We are actively working with vendor support to find a solution.

In the meantime, please feel free to use other clusters (Bamboo or Yggdrasil) until the issue is resolved.

We apologize for the inconvenience.

Best regards

Dear Users,

We have updated a system parameter, and the storage issue should now be resolved.
Please let us know if you experience any further I/O performance or stability issues.

Best regards,