I have set up two linux boxes using the 2.1.78 kernel:
-- orion is a UDB (alpha cpu) and is a nfs server using the kernel nfsd and
the rpc.nfsd and rpc.mountd from linux-nfs-0.4.22.
-- nostromo is a 6x86MX and is the nfs client.
For both machines, nfs and nfs server stuff is modularized and loaded
cleanly by kerneld.
orion exports its root directory with the following line in /etc/exports:
/ 129.199.131.98/255.255.255.255(rw,no_root_squash,async)
(129.199.131.98 is nostromo's IP)
When I mount this directory on nostromo using the following command:
mount -t nfs -o rw,async,rsize=8192,wsize=8192 orion:/ /mnt
It first spends a dozen second putting the following messages in
/var/log/messages on nostromo:
Jan 12 16:44:43 nostromo kernel: portmap: RPC call returned error 111
Jan 12 16:44:43 nostromo kernel: RPC: task of released request still queued!
Jan 12 16:44:43 nostromo kernel: RPC: (task is on xprt_pending)
Jan 12 16:44:48 nostromo kernel: portmap: RPC call returned error 111
Jan 12 16:44:48 nostromo kernel: RPC: task of released request still queued!
Jan 12 16:44:48 nostromo kernel: RPC: (task is on xprt_pending)
Jan 12 16:44:48 nostromo kernel: lockd_up: makesock failed, error=-111
Jan 12 16:44:53 nostromo kernel: portmap: RPC call returned error 111
Jan 12 16:44:53 nostromo kernel: RPC: task of released request still queued!
Jan 12 16:44:53 nostromo kernel: RPC: (task is on xprt_pending)
And then mount ends, and the directory seems to be perfectly mounted, with
adequate performance (read and writes at more than 900 KB/s on a 10baseT
ethernet).
Are the messages and the delay perfectly normal, or is there something buggy
in my installation ?
Thanks in advance for any answer.
--Thomas Pornin