Hi, I am using gluster 3.5.3 via packages from the gluster repository and am seeing some incidents where applications (mainly PHP/perl) when accessing NFS shared gluster filestore are having timeout issues when a flock() system call is used.
It occurs to me that the obvious thing here is to change my NFS mount options to forcibly tell the NFS client not to use locking and presumably to let the Gluster service deal with this at the server end, but does raise the following questions: 1. is forcibly disabling client side locking the 'right' thing to do? 2. if the above is true then why does the gluster documentation not recommend this? (this is not intended as snarky dig - honest) Thanks Paul Paul Osborne Senior Systems Engineer Infrastructure Services IT Department Canterbury Christ Church University +44 1227 782751 _______________________________________________ Gluster-users mailing list [email protected] http://supercolony.gluster.org/mailman/listinfo/gluster-users
