![]() |
|||
|
File Transfer Method for Large Data Sets
03-Jul-2007 Introduction People who have accounts on the SSH Gateway system to access our shared computing clusters from off-campus are also able to use this service to transfer files to the clusters. This is a convenient solution if the data sets are small. If they are large, however, then disk space on the SSH Gateway and the need to copy the data twice (once to the SSH Gateway and then again to the cluster) can become a problem. This document will describe an advanced method for file transfers that will eliminate this problem. Note: This method will only work from Unix/Linux desktops. This method is not available on Windows systems. Create a FIFO The advanced method for file transfers will include the use of a FIFO. A FIFO is a special file type that permits independent processes 1. Login to the SSH Gateway Login to the SSH Gateway with SSH as you normally would. 2. Create and open the FIFO Create and open the FIFO on the SSH Gateway by running these commands:
The last step above will block and wait until data is written to the FIFO. When data is written to the FIFO, it will be copied via SSH to the cluster and untarred. 3. Write data to the FIFO On your Unix/Linux desktop, tar your file and pipe to ssh connected to the FIFO on the SSH Gateway with the following command:
This command will pipe the output of the tar command to an SSH connection which will in turn write the data onto the FIFO. This will result in the file being tarred from your desktop to the FIFO
on the SSH Gateway, which in turn writes it to the cluster via the FIFO where it is |
|||
|