From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0) Description of problem: A server with multiple ethernet cards has problems with RPC and NFS Server mounts through the an ethernet card not defined as the GATEWAY device. If a client tries to mount an nfs drive using an IP address for an ethernet device that is not defined as the gateway, it gets an RPC timeout. The /usr/sbin/rpcinfo -p nfsserver.notgateway.name will give a list of services, regardless of the gateway device, however /usr/sbin/rpcinfo -u nfsserver.notgateway.name portmapper will hang. The client mount command will get an RPC timeout when trying to mount using the non-gateway device. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Setup two NIC cards on two separate subnets, using the eth0 as the GATEWAY device 2. then try either (from a second system on the second subnet) of the following: a) /usr/sbin/rpcinfo -u ipaddr.of.eth1 portmapper b) mount -t nfs ipaddr.of.eth1:/tmp/foo /mnt/foo Actual Results: for testcase 2.a) the rpcinfo command will hang 2.b) the client will return with an RPC timeout error Expected Results: for testcase: 2.a) should return with program 100003 version 2 ready and waiting 2.b) should mount the nfs share under /mnt/foo Additional info: This only happens with RedHat 7.1 using NFS server, but works correctly with RedHat 6.2 NFS servers.
This seems to be fixed in later kernels.