stackhpc / ansible-slurm-appliance

A Slurm-based HPC workload management environment, driven by Ansible.
36 stars 15 forks source link

Should fix dependency on volume order #327

Closed sjpb closed 6 months ago

sjpb commented 7 months ago

Currently, the slurm app TF (and caas) relies (here, here) on the fact that block devices are attached in the order they are defined; so if labels are applied in that order they end up on the right devices/filesystems. This works for RL8.6 and RL8.8 genericcloud images, actually I don't think it did for RL8.7 and it doesn't for RL9.2. For that, labels are getting applied to the expected devices, it's just that the device ordering isn't consistent with the definition order.

There is a fix at https://github.com/stackhpc/ansible-slurm-appliance/commit/93c70c0175bbfb058e23f51a39e4db174058d46b, which would also mean we could get rid of caas's inspection of image properties to define the correct volume path.