I'd like to report some trouble I had with the node configuration
(qlustar v. 10.1.1-2): I tried to add some nfs-mount as described here:
However, the script was not executed. It took me quite a while to figure
out that these scripts are run with 'run-parts' (see /lib/qlustar/nfs-rc
for those who are interested) and this does not accepct stripts with
'.sh' (see e.g. the first answer here:
for some explanation).
Probably it would be great to either use the '--regex'-option to
circumvent this behaviour or at least to tell users about it?
Moreover, I wondered if there is way to adjust the scripts in
'/lib/qlustar/'? Unfortunately, I didn't find their place on the headnode.
Is there a preferred way to change the name of a GPU gres type in
Qluman? For example, we have to use
to request that specific model of GPU. I would like to simplify that
to something like:
When I attempt to change the names from what the GPU Wizard had
originally set up, it makes things go badly, in that the the gres
resource is no longer recognized. Is there a clean way to do this?
I have a node where I've installed QluMan 10.1.1 so I can configure
the Cluster using the GUI. Now this node is NOT added to the cluster
(i.e. not currently booting any OS image) so it's somewhat "wasted".
As far as I understood the OS images are executed in the nodes in RAM
memory. So theoretically I could boot this node with an OS image and
later mount the fs to try qluman-10.1.1.19-singularity.sqsh however I
don't see how that could work because of lib compatibility. Is that
And I wonder if there is a container image already built specifically
to run the qluman.sqsh application. I couldn't find none so far and I
wanted to ask before starting to explore this path.
What do you recommend?