Portmapper error

[blabelle]

Before Upgrading our Unix server I loaded the new server with SCO and our app program When I first installed Openserver 5.06 on this new server (Dell PowereEdge 4600) I configured it with the IP address of 192.168.0.11 the old had 192.168.0.1 also the domain was defaulted to scosysv.uucp.com. Once I removed the old server and install the new I changed the IP to 192.168.0.1, and removed the old server from the network. I also remove the domain of uucp.com from the new server. Now I am getting a portmapper error at boot up that says WARNING: portmapper on server 213.130.63.234 is not responding. If I ping this IP it sees it as scosysv.uucp.com. I tried removing the uucp.com in netconfig and rebooting but it keeps comming back. I've read the knowledgbase docs on sco.com but nothing there has helped so far.

 

[Wonboodoo]

Is that 213.130.63.234 machine a different machine than the one you are working on? If so it's odd that they have the same names.

portmapper is an NFS related process. Your SCO server is probably trying (and failing) to mount a drive from the other machine.

The domain name when you "ping" is probably received from your DNS server or if you are not using DNS then is probably an entry in /etc/hosts.

 

[drtrinidad]

I agree that the hosts file is a good place to look.

I'm not sure that NFS is trying to mount a drive. I've frequently had this message simply after changing a machine's ip address while keeping the name the same. Netconfig does not remove the old entry from /etc/hosts, just adds a new one at the bottom. When NFS starts up it looks up the ip address of the symbolic name (in your case scosysv.uucp.com) and tries to communicate with the local host using that ip. When it fails you get this message.

If you find duplicate entries for scosysv.uucp.com in your hosts file, I recommend you comment out or rename the bad ones.

FWIW if you don't need NFS all this message does is cause a delay at boot time. I lived with the problem for a while before I figured out how to fix it

 

[blabelle]

The 213.130.63.234 ip is linked somewhere in this system to this server. The hostname is scosysv and the domain name is uucp.com. No other machines with this ip or domain or hostname on network. In /etc/hosts this server has 192.168.0.1 as ip and scosysv as hostname and scosysv.uucp.com after the hostname. the command hostname comes back with scosysv.uucp.com, if I go into netconfig the hostname is scosysv and the domain name is uucp.com.

# SCCS IDENTIFICATION
127.0.0.1 localhost
192.168.0.1 scosysv scosysv.uucp.com

Local Host Name: |scosysv_____________|
IP Address: 192.168.0.1_________| *a
Netmask: |255.255.255.0_______| *a
Broadcast Address: |192.168.0.255_______| *a
Domain Name: |uucp.com____________| *a

 

[Annihilannic]

Did you check /etc/hosts like they suggested? grep -i scosysv /etc/hosts will give you a quick answer; if there are two entries remove the wrong one. I've had the same problem changing the IP address in netconfig. Annihilannic.

 

[drtrinidad]

If you don't have duplicate /etc/hosts entries, perhaps your /etc/resolv.conf is set to look up names by DNS first? Mine ends with these two lines:

# finally the order to use for resolution - use /etc/hosts first
hostresorder local bind

When I run 'host scosysv.uucp.com' I get

scosysv.uucp.com has address 213.130.63.234

so I'm not surprised you do. It's not clear to me: is uucp.com really your domain? Is it your intention for the public ip address for that machine to be 213.130.63.234 (presumably translated by NAT)?

FWIW, in the SCO article below it does seem to put all the emphasis on /etc/hosts

http://stage.caldera.com/cgi-bin/ssl_reference?102428

I'm guessing you didn't see that one. It's not at all obvious how to reach the legacy OpenServer knowledge base, but you'll find it at

http://stage.caldera.com/ta