azslurm scale fails with ModuleNotFoundError: No module named 'slurmcc' and NFS not mounted on scheduler VM

I’m trying to configure/use autoscaling on an Azure HPC Slurm cluster, but on the scheduler VM azslurm scale fails with a missing Python module, and it also looks like the expected NFS share(s) are not mounted.

Environment:

• Node: compular-scheduler (Slurm scheduler VM on Azure)

Running as root via sudo -i

Disk and mounts look like this:

root@compular-scheduler:~# df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/root        62G   36G   27G  58% /
tmpfs           7.9G     0  7.9G   0% /dev/shm
tmpfs           3.2G  1.1M  3.2G   1% /run
tmpfs           5.0M     0  5.0M   0% /run/lock
/dev/sda15      105M  6.1M   99M   6% /boot/efi
tmpfs           1.6G  4.0K  1.6G   1% /run/user/1000

I was expecting to see one or more NFS-mounted filesystems for the shared storage (e.g. /shared, /apps, or similar), but they do not appear in df -h on the scheduler.

When I run azslurm scale I get: <private data>

The azslurm script points to a virtual environment under /opt/azurehpc/slurm/venv and imports slurmcc:

$ head /opt/azurehpc/slurm/venv/bin/azslurm
#!/opt/azurehpc/slurm/venv/bin/python

import os

if "SCALELIB_LOG_USER" not in os.environ:
    os.environ["SCALELIB_LOG_USER"] = "slurm"
if "SCALELIB_LOG_GROUP" not in os.environ:
    os.environ["SCALELIB_LOG_GROUP"] = "slurm"

from slurmcc.cli import main

Python in that venv is:

$ which python
/opt/azurehpc/slurm/venv/bin/python

But slurmcc does not seem to be installed there:

$ python -m pip list | grep -i slurm
# (no output)

$ python -m pip show slurmcc
WARNING: Package(s) not found: slurmcc

So it looks like the azslurm CLI is present, but the underlying slurmcc package is missing from /opt/azurehpc/slurm/venv. At the same time, the scheduler VM does not show any NFS-mounted shared storage in df -h, which might indicate the Slurm/Azure integration or provisioning did not complete correctly.

My questions:

1. What is the correct way to (re)install or repair the slurmcc package and azslurm environment on an Azure HPC Slurm scheduler VM?
2. Is there an official script/extension or documented procedure to re-run the Azure Slurm connector / autoscaling installation on an existing scheduler without breaking the cluster?
3. Should the scheduler normally have NFS-mounted shared storage visible in df -h (e.g. /shared, /apps, or similar)? If yes, what is the recommended way to verify and/or re-mount the expected NFS shares on the scheduler node (As I have important data on the nfs disk)?

Any guidance on restoring a working azslurm scale command and ensuring the scheduler’s NFS mounts are correctly configured would be appreciated.I’m trying to configure/use autoscaling on an Azure HPC Slurm cluster, but on the scheduler VM azslurm scale fails with a missing Python module, and it also looks like the expected NFS share(s) are not mounted.