Hi, You're pretty far away with 1.2.1. Afaik, RHEL6 already supported more than 32 cores, so I don't think kernel 3.10.0-327 should have a problem dealing with 32C CPUs. My approach would be to start

We have a linux cluster that runs ohpc 1.2.1. This cluster was built in early 2017 and has been working great so far. The head node runs Centos 7.3 and the compute nodes run Centos 7.2 (when ohpc 1.2.

So I want to use my cpu as a consumable resource in slurm will this plugin work SelectType=select/cons_res SelectTypeParamenter=CR_Core in my slurm.conf file or is there something else I need to do?

Hi, I’m having some trouble with the networking. I have a 10GB ethernet interconnect and also an Omnipath 100g. At first I went with the open source yum repo omnipath / libpsm2 drivers and enabled IPo

Are there any plans to add support for Slurm 19.05.x version in the near future and if so in which release of OHPC ?

The next release version is 1.3.9? What’s the planned release date for 1.3.9, and what’s the OS version supported by 1.3.9?

Ive defined an alternative initrd to boot from using "tabedit boottarget": To use it, I need to specify os=boottarget inside "tabedit nodetype" : The problem is when I try to use nodeset $node osimage

User-submitted jobs get stuck in a "RUNNING" state indefinitely. If the user tries to cancel a job, the logs complain that the user isn't a superuser. Eventually I have to put the node down due to hun

Hello, I would like to know if someone is still using hostbased authentication for SSH instead of automatic generating SSH keypairs for users, or even not doing anything at all. If yes which approach

Hello all! I have generated a login node in our environment. The configuration is diskless in this moment, but I have seen that the space is nearly full. Could be a problem in the future? How could I

Hi, all I thought that scontrol reboot was delayed for the time specified in 'sbatch --delay-boot'. However, even without the --delay-boot option, scontrol reboot does not restart until the job is com

How would I go about using some of the osu mpitests available for testing my psm2 / omnipath fabric ? I have these ones I installed from the yum intel_opa repo: Whenever I try any of them I get this m

Hi, I think I'm missing something really obvious here but can't seem to work out what I've done - I have a nice test cluster setup with Slurm working for job allocation etc, but if I run a basic quick

Hello, someone had success using the Mellanox OFED instead of the inbox drivers with OpenHPC? YUM is completely broken after adding the MLNX OFED repository to a already deployed OpenHPC cluster: Just

Hi everyone, I am going to install a new ohpc cluster. as CentOS7.7 was released several days ago, I want to install ohpc 1.3.8.1 in CentOS7.7. will this be ok? Or will a new version (1.3.9?) be relea

Has anyone had any luck adding extra rpm's to the xcat kickstart install? I was able to add kernel-devel without an issue by adding it to /opt/xcat/share/xcat/install/centos/compute.centos7.pkglist bu

Hi OHPC users! We have a cluster OHPC 1.3.8 -CentOS 7.6 with one master node (hn01) and 3 computes nodes (cn[01-03]). In hn01 node Slurm works without problems for any user. But we had the problem whe

Hi group, do you have some info about CentOS 8 support by OpenHPC? Have you any experience of running it? -- Antanas

Hi Janne, thank you for pointing to the sources. I'll try patching it and will report / pull request depending on results. -- Antanas ________________________________________ Nuo: Janne Blomqvist <jan

Hi Patrick, there is no particular reason for XFS except CentOS/RHEL 'standard' . -- Antanas ________________________________________ Nuo: OpenHPC-users@groups.io <OpenHPC-users@groups.io> varduPatric