Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 31 Oct 1995 13:51:14 -0500 (EST)
From:      John Capo <jc@irbs.com>
To:        adam@veda.is (Adam David)
Cc:        freebsd-questions@freebsd.org
Subject:   Re: clnttcp_create: RPC: Program not registered
Message-ID:  <199510311851.NAA19082@irbs.irbs.com>
In-Reply-To: <199510311745.RAA14436@veda.is> from "Adam David" at Oct 31, 95 05:45:17 pm

index | next in thread | previous in thread | raw e-mail

Adam David writes:
> 
> Various people wrote:
> >>> I've looked everywhere and I still don't have any idea what this message
> >>> means.  It comes out on the console four or five times when the system
> >>> first boots up, and once every time I log in or su.  Aside from the
> >>> error messages, everything else seems to be working OK, until this
> >>> evening when I finally got another machine up on FreeBSD on my ethernet,
> >>> and discovered rcp won't work:
> >>> 
> >>> rcp: clnttcp_create: RPC: Program not registered
> 
> ypbind is not running, and you have 'nis' support enabled in /etc/host.conf
> 
> >> Hmmmm... Since I've set NIS up, I see this a few times when my system is 
> >> booting, but never any other time.   Everything seems to work for me, so 
> >> I haven't tried to figure it out.
> 
> > I believe that the cause of this message is that rpc-based services are
> > used before portmap is loaded.
> 
> clnttcp_create: RPC: Port mapper failure - RPC: Timed out
> 
> It waits one minute trying to resolve each hostname arg by NIS while setting
> up the static routes, and portmap is not loaded (nor can it be loaded from an
> NFS /usr at this stage).
> 
> The real fix would probably be to change pmap_getport() to not wait 60 seconds
> for a reply from the local portmapper if it is not even running. Couldn't it
> check whether the remote machine is localhost or the current host, and if so
> just see if tcp PMAPPORT is bound (and listening) on the local machine, in
> order to abort early if appropriate?
> 
> The temporary workaround is to put 'mv /etc/host.conf /etc/host.conf.tmp'
> before setting up the static routes in /etc/netstart, and mv it back again
> afterwards (a few lines down).
> 
> --
> Adam David <adam@veda.is>
> 

Run rpcinfo -p and you will see the real problem.  There are a
zillion attempts to connect to ypbind whether or not nis is enabled
in /etc/host.conf.

Here is where the message comes from:

clnttcp_create: RPC: Program not registered
265     }
(gdb) where
#0  clnt_pcreateerror (s=0x8cd2 "clnttcp_create")
    at /usr/src/lib/libc/rpc/clnt_perror.c:265
#1  0x8fe2 in _yp_dobind (dom=0x23d70 "davsys.com", ypdb=0x0)
    at /usr/src/lib/libc/yp/yplib.c:299
#2  0x922e in yp_bind (dom=0x23d70 "davsys.com")
    at /usr/src/lib/libc/yp/yplib.c:376
#3  0xa307 in _yp_check (dom=0x374a4) at /usr/src/lib/libc/yp/yplib.c:868
#4  0x3489 in getrpcbynumber (number=100000)
    at /usr/src/lib/libc/rpc/getrpcent.c:102
#5  0x2520 in pmapdump (argc=0, argv=0xefbfdca8)
    at /usr/src/usr.bin/rpcinfo/rpcinfo.c:530
#6  0x18f5 in main (argc=2, argv=0xefbfdca0)
    at /usr/src/usr.bin/rpcinfo/rpcinfo.c:157

Not sure what the best fix is.  I guess looking at /etc/host.conf
would be a start.

John Capo
IRBS Engineering                      High performance FreeBSD systems


help

Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199510311851.NAA19082>