1) Do I need to install all of GT 4.2.1 on each host or can I just install GridFTP on 4 hosts ( 2 in A and 2 in B) and RFT on the 5th ? I intend RFT to be on a separate machine from the GridFTP hosts, is this ok ?
that sounds right.
2) Recent GT documentation mentions that a separate database (PostGreSQL, MySql) is not required since GT uses an internal DB called "Derby". Does RFT need this internal database ?
you can use derby or an external database (postgres, mysql), it is your choice
3) If we wanted to transfer a large file from Site A to Site B, how can we setup the system so that the all the GridFTP servers in both sites can be used for transferring the file ? I assume that using more servers will result in a faster file transfer, is this correct ? if so how to set it up ?
this is called a striped transfer: http://www.globus.org/toolkit/docs/latest-stable/data/gridftp/admin/#gridftp-admin-striped
4) For authentication between the servers and certificate management, is MyProxy required on each server or is only one instance of it required ?
1 instance should suffice.
5) Is there any example of GridFTP setups anywhere, i.e. documentations on such setups etc that I can refer to ?
there is a bit of information here: http://www.globus.org/toolkit/docs/latest-stable/data/gridftp/admin/#gridftp-configuring and a tutorial here: http://www.globus.org/toolkit/data/gridftp/tutorials/
Thanks in advance Arn
