Skip to content

Data migration to Cirrus NCR

If you want any of your ARCHER2 data on Cirrus, you must transfer this yourself using standard data transfer methods as documented in the ARCHER2 documentation on data transfer.

Important

Globus Online is not yet available on Cirrus for data transfer so you cannot use this method at the moment.

Use EPCCFS/NSCDS to move data from ARCHER2 to Cirrus

The EPCCFS/NSCDS storage is shared between ARCHER2 and Cirrus so copying data to this storage system on ARCHER2 (mounted at /nscds) is the simplest way to move your data to Cirrus (where it is mounted at /epccfs). More details on transferring data in this way are given below.

/home and /work file systems

While Cirrus has /home and /work file systems like ARCHER2, these are completely different file systems to the ARCHER2 ones so none of your ARCHER2 data will be automatically available on Cirrus.

If you have questions about data transfer, please contact the Cirrus service desk.

EPCCFS/NSCDS

The EPCCFS/NSCDS storage is shared between ARCHER2 and Cirrus:

  • On ARCHER2 it is mounted at /nscds
  • On Cirrus it is mounted at /epccfs

Copying data to EPCCFS/NSCDS from ARCHER2 file systems (on ARCHER2)

You can use the standard Linux cp command to copy data from ARCHER2 file systems to EPCCFS/NSCDS when you are logged into ARCHER2. For example, to transfer the file important-data.tar.gz from the ARCHER2 /work file system to EPCCFS/NSCDS you would use the following command on ARCHER2 (assuming you are user auser in project e05):

cp /work/e05/e05/auser/important-data.tar.gz /nscds/e05/e05/auser/

(remember to replace the project code and username with your own username and project code).

Use rclone parallel local data transfers for large datasets

If you are transferring a large amount of data to EPCCFS/NSCDS, you should consider using parallel rclone local data transfer (perhaps in a serial job submission script) rather than using the basic cp command.