Attention: Confluence is not suitable for the storage of highly confidential data. Please ensure that any data classified as Highly Protected is stored using a more secure platform.
If you have any questions, please refer to the University's data classification guide or contact ict.askcyber@sydney.edu.au
Troubleshooting access to Artemis
- Chengtao Liang (Unlicensed)
1. How to access Artemis and RDS via terminal?
Refer to Artemis HPC guide and RDS guide.
2. Where are my project folders in Artemis and RDS?
For Artemis, your project folder is within scratch:
ssh abcd1234@hpc.sydney.edu.au
cd /scratch/PRJ-###
For RDS, you can access:
i) via the hpc:
ssh abcd1234@hpc.sydney.edu.au
cd /rds/PRJ-###ii) via rcos-int:
ssh abcd1234@rcos-int.sydney.edu.au
cd /rds/PRJ-###
3. How can I transfer data from my local computer OR an RDS location (e.g. PRJ-Frontier) into /scratch on Artemis?
Note: submitting a job on Artemis via command line is the FASTEST way to complete this
a) If your data is local on your computer, you can use a tool like Filezilla or Cyberduck to transfer files to /scratch. Instructions for how to do this are here. These are the methods supported by USYD. If you run into problems using these, please contact Stephen Kolman and copy Ramon in so he can keep an eye on it.
b) If your data is on RDS, you can access the data from HPC directly via terminal at /rds/PRJ-SHORTCODE - see here for more info.
c) If you do not want to use command line, and prefer to mount /scratch visually, you can use MountainDuck. However, MountainDuck is a paid program, you will need to purchase licenses for your team. This method is not officially supported by USYD.
4. How can I log into a virtual desktop environment to visualise my analyses?
a) Using Nomachine. Nomachine logs into a server location (like the HPC graphical node) as if it was a virtual desktop. This is for checking on data or running some light processing/visualisation, not for running large jobs or simply for moving data around. Instructions for using nomachine to access the HPC graphical node are here.
b) The Argus virtual desktop is also a good option for this. Argus allows you to run graphical programs. See here for more info on Argus. There are some important things to consider when using Argus:
i) While HPC has many nodes, allowing for many users to access and submit jobs efficiently, Argus is by definition a single node shared among many users. There is a single “Shared” VRD node that can be used by many users. Having many people logging in simultaneously (and/or not properly logging out when you’re finished) will adversely affect the speed for you and others. To access this node, please contact Stephen Kolman, CCing Ramon.
There are also 3 separate “private” VRD nodes that are organised by supervisors. These private nodes can only be accessed by one person at a time - to use these nodes, please organise your access with your supervisor and remember to log out after each session. Using your respective large node will speed up performance for you and your colleague's jobs.
ii) Much like the HPC graphical node, Argus is not for running large jobs where you can submit via command line and let it run. Doing this on argus will be slower than on HPC and will slow down jobs for your colleagues using Argus as well. You should run the bulk of your analysis via HPC, then use argus just for minor analyses, post-processing or visualisation.