I'm currently using fio to simulate user reads and writes to a NAS service, and am testing the performance of different shared filesystems (NFS, Samba, Glusterfs etc). In order to properly load the box I'm using a pool of clients to connect to the fs (fio --server) and then launching the same job on multiple clients. (fio --client=a --client=b job.fio) The resulting jobs on the different servers all seem to use the same file - the filename seems to be <name>.<thread>.<fileno> - in this case 'job.1.0': [global] directory=/mnt/gluster rw=randrw size=10M ioengine=libaio [job] Is there any way to have the server include some personal identifier (such as hostname) in the filename so that there isn't this conflict? Thanks, Matthew -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.
Attachment:
signature.asc
Description: OpenPGP digital signature
