Jesse Off wrote:
Is there any way to disable, manually flush, or change the timeout settings of RPC responses held in the duplicate request cache of NFS v2 over UDP?
no such interface exists. A workaround could be to overflow the DRC with dummy requests from another unrelated client. The size of the UDP DRC is 1024. Mahesh
My problem is I use NFS on a lot of constantly rebooting NFS clients and these clients have no decent random number generator to seed the RPC transaction id. (They are embedded systems) As a result, they all use the exact same XID and when they reboot they get the same IP from DHCP. As a result, the non-idempotent requests are not performed correctly since the server responds with the response it gave from the board's last bootup. //Jesse Off
_______________________________________________ nfs-discuss mailing list [email protected]
