FAQ

Use the Forgot Password? option. To set new password, enter your login or the e-mail address associated with your PLGrid account.

If the e-mail address linked to your PLGrid account is no longer valid or you do not have access to it and therefore cannot set a new password to your account, please contact us via Helpdesk.

Your surname and first name are the details that Users cannot change themselves in the Portal. If your first name or surname has changed, please contact us via Helpdesk.

If your affiliation is no longer active, it has probably not been extended on time. If you are still authorized to use PLGrid, but your affiliation has not been renewed in time, please contact us via Helpdesk.

Make sure you already have a user account in PLGrid and enter the correct login (with the "plg" prefix). If you have forgotten your password, you can reset it by entering the email address you used when registering to PLGrid.

If you are our new user, please wait approximately one hour before all your account information propagates on the cluster.

This error may be due to exceeded (exhausted) grant resources. In the case of Ares and Athena clusters, you can check this with the command

hpc-grants

If the allocation status of your grant is indicated here as exhausted, the grant resources have already been fully consumed. In this case you can apply for a new grant.

Grant the script/program the rights to be executed using the command:

chmod u+x <full_script_name>

This should solve the problem.

If the job has not yet been scheduled, you can query the Slurm system for an estimated time to run it by adding the --test-only argument to the sbatch command:

sbatch --test-only my_script.sh 

The system will then verify the correctness of the script and provide an estimate of the job start time. The estimation is calculated pessimistically and in most cases the job is going to start earlier.

The estimated start time of a task already scheduled can be checked with the command:

squeue -t pd -o %i,%S

This time will only be visible if the Slurm system has already scheduled the job. Due to queue optimization, the system can only complete this information when the job starts. The given start times are only estimates, and are not mandatory. There are many users on the cluster, which makes it very difficult to determine the exact start time of a job. The difficulty arises from the unpredictability of the actions of other users, who affect the state of the job queue all the time. The priority of a job, its position in the queue, depends on a number of factors, such as the amount of resources needed to start the job, the declared job execution time, the last activity of the user and other grant participants on the cluster, as well as the resources allocated in the grant.