Re: [slurm-users] job_desc_msg_t and job_record_t

Hi Wu, As the name itselff suggests, job_desc_msg_t structure is used to communicate the data about job between different components of slurm. job_record_t is the one that is used by slurm to store the data about a job locally. I hope this explanation helps. It depends, if updating one effects

Re: [slurm-users] X11 forwarding issues

Hi Russell Jones, did you try to stop firewall on the client cluster-cn02 ? Patrick Le 16/11/2020 à 19:20, Russell Jones a écrit : > Here's some debug logs from the compute node after launching an > interactive shell with the x11 flag. I see it show X11 forwarding > established, then it ends wit

Re: [slurm-users] X11 forwarding issues

Hello, check sshd settings (here are ours): X11Forwarding yes X11DisplayOffset 10 *X11UseLocalhost no* Add PrologFlags in slurm.conf: PrologFlags=x11 Cheers, Barbara On 11/16/20 7:20 PM, Russell Jones wrote: Here's some debug logs from the compute node after launching an interactive shel

[slurm-users] job_desc_msg_t and job_record_t

Hello, We see there are two job data structures job_desc_msg_t and job_record_t, and are wondering the relation between them. What is each of them used for or how are them used? If we update the value of one of the entries in job_desc_msg_t, will the corresponding value change in job_record_t?

Re: [slurm-users] X11 forwarding issues

Here's some debug logs from the compute node after launching an interactive shell with the x11 flag. I see it show X11 forwarding established, then it ends with a connection timeout. [2020-11-16T12:12:34.097] debug: Checking credential with 492 bytes of sig data [2020-11-16T12:12:34.098] _run_pr

[slurm-users] Update of this parameter is not supported: Account

Hi all. I can’t seem to change the account or QOS on a pending job: $ scontrol update job where JobId=133628 set Account=def-eoliver QOS=normal Update of this parameter is not supported: Account=def-eoliver Request aborted $ squeue -j 133628 JOBID USER ACCOUNT NAME

[slurm-users] accounting database question

Hi List, apologies if this has been asked before (or is obvious) - I did do some reading & searching but can't quite figure the best way to achieve this. Background - we have two productions clusters, both running SLURM. They are not currently a multi-cluster setup; they are not running the s

Re: [slurm-users] X11 forwarding issues

Hello, Thanks for the reply! We are using Slurm 20.02.0. On Mon, Nov 16, 2020 at 10:59 AM sathish wrote: > Hi Russell Jones, > > I believe you are using a slurm version older than 19.05. X11 forwarding > code has been revamped and it works as expected starting from the 19.05.0 > version. > >

Re: [slurm-users] X11 forwarding issues

Hi Russell Jones, I believe you are using a slurm version older than 19.05. X11 forwarding code has been revamped and it works as expected starting from the 19.05.0 version. On Mon, Nov 16, 2020 at 10:02 PM Russell Jones wrote: > Hi all, > > Hoping I can get pointed in the right direction her

[slurm-users] X11 forwarding issues

Hi all, Hoping I can get pointed in the right direction here. I have X11 forwarding enabled in Slurm, however I cannot seem to get it working properly. It works when I test with "ssh -Y" to the compute node from the login node, however when I try through Slurm the Display variable looks very diff

[slurm-users] [SOLVED] failed to send msg type 6002: No route to host

Hi, just to give a feedback about this problem: it was a firewall problem. When using "salloc/srun" from the login node, the login node must accept connections from the compute nodes even if nor slurmd nor slucrmctld are running on this login node. I had to widely open the firewall on the interna

Re: [slurm-users] Heterogeneous GPU Node MPS

I read that also, however the RTX cards are not really pre-Volta and when I run the mps-server, the nvidia-smi tool gives me e.g. +--- --+ | Processes: GPU Memory | | GPU