Configuration of cryoSPARC environment
License
Each user should have his/her own license obtained from https://cryosparc.com/download and apply for membership in plggcryospar
team in Portal PLGrid.
To get access to cryoSPARC instalation at Prometheus cluster
- Apply for apply for membership in
plggcryospar
team in Portal PLGrid and ask for registration in Cyfronet's internal cryoSPARC users database and dedicated port for access to cryoSPARC master through Helpdesk PLGrid. Log in to Prometheus login node
Log into Prometheus login nodessh <login>@pro.cyfronet.pl
Load cryoSPARC module using command
Set cryoSPARC environmentmodule add plgrid/apps/cryosparc/3.1
Run cryoSPARC configuration script. It will configure your cryoSPARC environment and create your user in cryoSPARC database and configure two lanes for external jobs -
prometheus-gpu
which is going to useplgrid-gpu
partition for GPU jobs andprometheus-gpu-v100
forplgrid-gpu-v100
partition. Both lanes are going to useplgrid
partition for CPU only jobs. As argument for script pass license id, your e-mail and password (they are going to be used to login to cryoSPARC webapp), your first and last name.Configure cryoSPARCcryosparc_configuration --license <XXXX> --email <your-email> --password <password> --firstname <Givenname> --lastname <Surname>
Access problems
In case of "
cryosparc_configuration: command not found
" error run in terminalnewgrp plggcryospar
to start new subshell with permissions of
plggcryospar
team.Access to GPU partitions
To use GPUs on Promehteus cluster you have to apply for GPU resources at Portal PLGrid.
To check whether you have an access to partition on Prometheus login node run below command and check whether your PLGrid computational grants are on AllowAccounts list
partition plgrid-gpu
scontrol show partition plgrid-gpu | grep Accounts | grep <PLGrid grant name>
partition plgrid-gpu-v100
scontrol show partition plgrid-gpu-v100 | grep Accounts | grep <PLGrid grant name>
In case that you do not have access to one or both of above partitions check your PLGrid computational grant details at Portal PLGrid. If your grant lists GPU resources, and access to required queue or queues is not possible please contact Helpdesk at https://helpdesk.plgrid.pl.
- Your cryoSPARC master setup already done. All succeeding crypoSPARC master instances should be run in batch jobs.
cryoSPARC master job
cryoSPARC master must not be run on login nodes of Prometheus cluster. It should be run in plgrid-servicies
trough SLURM job described below.
Automated cryoSPARK master in batch job
cryoSPARC master could be started trough batch job.
#!/bin/bash #SBATCH --partition plgrid-services #SBATCH --nodes 1 #SBATCH --ntasks-per-node 1 #SBATCH --time 14-0 #SBATCH -C localfs #SBATCH --dependency=singleton #SBATCH --job-name cryosparc-master #SBATCH --output cryosparc-master-log-%J.txt ## Load environment for cryoSPARC module add plgrid/apps/cryosparc/3.1 ## get tunneling info ipnport=$CRYOSPARC_BASE_PORT ipnip=$(hostname -i) user=$USER ## print tunneling instructions to cryosparc-master-log-<JobID>.txt echo -e " Copy/Paste this in your local terminal to ssh tunnel with remote ----------------------------------------------------------------- ssh -o ServerAliveInterval=300 -N -L $ipnport:$ipnip:$ipnport ${user}@pro.cyfronet.pl ----------------------------------------------------------------- Then open a browser on your local machine to the following address ------------------------------------------------------------------ localhost:$ipnport ------------------------------------------------------------------ " ## start a cryoSPARC master server cryosparcm restart ## loop which keep job running till scancel <JobID> by user or automatic kill by SLURM at end of requested walltime while true; do sleep 600; done
Above script is located at /net/software/local/cryosparc/3.1/cyfronet/cryosparc-master.slurm
. You could copy it your working folder
cp /net/software/local/cryosparc/3.1/cyfronet/cryosparc-master.slurm .
Submit job
job submisionsbatch cryosparc-master.slurm
cryoSPARC master job
There should be only one job which run cryoSPARC master in
plgrid-servicies
partition per user.Check whether job was started
jobs statussqueue -j <JobID>
Common states of jobs
PD - PENDING
- Job is awaiting resource allocation.R - RUNNING
- Job currently has an allocation and is running.CF - CONFIGURING
- Job has been allocated resources, but are waiting for them to become ready for use (e.g. booting). On PrometheusCF
state could last for up to 8 minutes in case when nodes that have been in power save mode.CG - COMPLETING
- Job is in the process of completing. Some processes on some nodes may still be active.
Make a tunnel
In your directory cat job log file:
Listing of job's logcat cryosparc-master-log-<JobID>.txt
where `
XXXXXXX
` is your sbatch job id which is displayed after you run it f.e. `cat cryosparc-master-log-49145683.txt
`It will show you something like this:
Example of job logCopy/Paste this in your local terminal to ssh tunnel with remote ----------------------------------------------------------------- ssh -o ServerAliveInterval=300 -N -L 40100:172.20.68.193:40100 plgusername@pro.cyfronet.pl ----------------------------------------------------------------- Then open a browser on your local machine to the following address ------------------------------------------------------------------ localhost:48511 ------------------------------------------------------------------
Exec in another shell at your local computer given command to make a tunnel:
Tunnelingssh -o ServerAliveInterval=300 -N -L 40100:172.20.68.193:40100 plgusername@pro.cyfronet.pl
- Log into cryoSPARK web application - open in browser: `
localhost:40100
`
Adding optional lanes
You could create additional lanes for other maximal duration of SLURM job:
Start interactive job using command
Interactive jobsrun -p plgrid-services --nodes=1 --ntasks=1 --time=0-1 --pty bash
Load cryoSPARC environment using modules
Load cryoSPARC environmentmodule add plgrid/apps/cryosparc/3.2
Copy cluster config
cluster_info.json
and script templatecluster_script.sh
from$CRYOSPARC_ADDITIONAL_FILES_DIR
directory to your working directoryCopy filescp $CRYOSPARC_ADDITIONAL_FILES_DIR/cluster_info.json . cp $CRYOSPARC_ADDITIONAL_FILES_DIR/cluster_script.sh .
- Modify files accordingly
- in config
cluster_info.json
change name of lane/cluster to avoid overwriting default prometheus* lanes - in
cluster_script.sh
change--time
,--partition
or other parts of script template accordingly
- in config
Start cryoSPARC master
run cryoSPARC mastercryosparcm restart
cryoSPARC master job
There should be only one job which run cryoSPARC master per user. Therefore you should stop job with cryoSPARC master before this step
run command
cryosparcm cluster connect <name-of-cluster-form-cluster_info.json>
to add lane/clusteradd linecryosparcm cluster connect <name-of-cluster-form-cluster_info.json>
- Repeat above points to create another lane if necessary
Stop cryoSPARC master
run cryoSPARC mastercryosparcm stop
End interactive job
end interactive jobexit