The Qlustar team is pleased to announce the immediate availability of
OpenHPC 2.0 [1] on top of CentOS 8.3 as an update to Qlustar 11. You can
migrate your existing images and chroots to this new release in the
usual way [1]. New ones can be generated via QluMan.
Note that the disappearance of CentOS [3] as we knew it in the past,
will not have a negative impact on Qlustar users running CentOS based
nodes. There will be a simple migration path to whatever new RHEL
compatible distribution will turn out to be the best choice at the end
of 2021. Migration will involve nothing more than a usual Qlustar
security update.
[1] https://openhpc.community/openhpc-2-0-released/
[2] https://docs.qlustar.com/en-US/Qlustar_Cluster_OS/11.0/html-single/Administ…
[3] https://blog.centos.org/2020/12/future-is-centos-stream/
Hi, I would like to install cuda11 in a cluster. In order to do so, I would need to first install the latest nvidia driver 450.
In bionic the default is the 440 version; is there any way to update to the latest nvidia driver?
with best regards
A.
Hello List,
I'm wondering if anyone is mounting pre-existing (local) storage on a
Qlustar node, and if so, how you go about it? My initial guess was to
re-define the setup (xfs on LVM on mdraid) in the Disk Config assigned
to the node, and mark it with preserve_always, but then the node
refuses to boot correctly because setup-storage complains it cannot
preserve my LV as it doesn't exist.
I guess I could create per-node fstab stubs containing just the necessary
lines, push them to the nodes using common/image-files with
softgroups, and then add a common/rc.boot script to look for such
stubs and mount them, but maybe there's a more integrated way?
Thanks,
A.
--
Ansgar Esztermann
Sysadmin Dep. Theoretical and Computational Biophysics
http://www.mpibpc.mpg.de/grubmueller/esztermann