Hello gluster folks Glusterfs seems to be the wonderful solution we need, in fact the only one we still consider after some research of existing distributed fs. Using standard filesystem under glusterfs is very reassuring for us. Some of our applications are I/O bounded and exhaust our nfs servers (search lots of sparses small data amongst 10th GB data, and in the end have read all the datas). Are there some sample install script for simple configuration? I'd prefer to not manually edit 20 conf files. We would like to set up a glusterfs for a common /workspace in our small (SSI) cluster: - ~20 nodes - each with 500 GB of local disk - each node act as client and server of /workspace - everything replicated once - so far we are either cpu bounded or i/o bounded but far from both at the same time. Best regards, and congratulations for this wonderful job. Alain
