pinging OpenVMS servers

[Keith Whitwell]

A simple question.

Why can't I ping the server on which I have logged in? I can ping the other server in the cluster and other devices on the network.

[Volker Halle]

Keith,

welcome to the OpenVMS ITRC forum !

(assuming you're running TCP/IP services):

$ TCPIP PING localhost should always work. As well as TCPIP PING IP-adress-of-any-local interface (show them with TCPIP SHOW INT).

What is the error message ?

Volker.

[Phillip Thayer]

Keith,

If you do a TCPIP PING Localhost and get a response then your TCPIP on your system is probably configured properly. If you do a TCPIP PING IP-address-of-interface and it doesn't work then you it could be a number of problems.

1. If you cannot ping any systems outside of your VMS system then check your configured gatway.

2. If you can ping your gateway but nothing else then your gateway system may have the ping (ICMP) disabled for your system.

3. Can't think of anything else right now. Maybe someone else can add to what the possible problems could be.

Phil

[Arch_Muthiah]

Keith,

Welcome.

If the ping command for localhost does not respond correctly, try the ping command with the IP address 127.0.0.1 .

$tcpip ping 127.0.0.1
If this command displays correct output, the TCPIP database is missing a definition for localhost .

As you said, you can able to ping to other node, that means there is no problem in the upper (application) layer protocols (such as FTP, TELNET).

Anyway you can check your network interface and routing works fine using....
ifconfig , netstat , and arp commands.

Btw, what type of error do you get such as host is unreachable, connection timed out, and network is unreachable.

Archunan

[Steven Schweda]

What's your "ping" command?

What does it do? (Hint: "Not work" is not a
useful answer.)

Which TCP/IP software and version?

If there's a name in your "ping" command,
does nslookup resolve it?

With almost no useful info in the question,
it's difficult to offer a useful answer.

[Keith Whitwell]

Apolgies for vagueness of original post.
More flesh on the bone.
We're running TCP/IP services (not sure which version) on Open VMS 7.2-1

when I ping localhost or 127.0.0.1 explicitly
I get responses indicating o ms timing.(success)

When I ping DFG001 or 10.12.2.1 (the VMS server I'm logged onto) I get no response after 2 minutes, Ctrl/C gives

TCPIP> ping 10.12.2.1
PING 10.12.2.1 (10.12.2.1): 56 data bytes
Cancel



----10.12.2.1 PING Statistics----
204 packets transmitted, 0 packets received, 100% packet loss
%SYSTEM-F-TIMEOUT, device timeout


I can successfully ping devices outside of the VMS cluster and the other node in the cluster, DFG002 or 10.12.2.2. I can also ping both VMS servers from a DOS prompt on my PC using both the IP address and the server name.

The converse is also true, in that, when logged into to DFG002 a can ping dfg001 but not dfg002 or its IP address 10.12.2.2.

I also have a problem with timeout connectivity on a thin client (Gembase) application which I think is also related.
This application had been operational up until August last year, since when it has been moribund. I now want to start to use it again and have encountered these problems.

I should say that networking is not my forte so any further help with diagnosis would be gratefully recieved.

[Volker Halle]

Keith,

you can find out about your TCPIP version with:

$ TCPIP SHOW VERSION

How are the interfaces and default route defined:

$ TCPIP SHOW INT
$ TCPIP SHOW ROUTE

Volker.

[Keith Whitwell]

Volker,

TCPIP> sh version

DIGITAL TCP/IP Services for OpenVMS Alpha Version V5.0A
on a AlphaServer DS20 500 MHz running OpenVMS V7.2-1

TCPIP> sh int
Packets
Interface IP_Addr Network mask Receive Send MTU

WE1 10.12.2.1 255.255.0.0 256384383 218836594 1500
LO0 127.0.0.1 255.0.0.0 39 39 0
TCPIP> sh route

DYNAMIC

Type Destination Gateway

DN 0.0.0.0 10.12.100.1
AN 10.12.0.0/16 10.12.2.1
DH 10.12.2.1 10.12.10.254
AH 10.12.2.2 6.4.0.0
DH 127.0.0.1 127.0.0.1
TCPIP>

Hope that helps.

The servers allow Powerterm users access and can FTP to and from other other servers/PCs.

Keith

[Martin Vorlaender]

Keith,

>>>
TCPIP> sh int
...
WE1 10.12.2.1 255.255.0.0
...
TCPIP> sh route

DYNAMIC

Type Destination Gateway
...
DH 10.12.2.1 10.12.10.254
<<<

I think this routing entry is your problem.
Remove it please.

HTH,
Martin

[Volker Halle · Accepted solution]

Keith,

you seem to be running dynamic routing (ROUTED or GATED routing daemon).

The 'D' in DH and DN indicates a dynamic route created by dynamic routing.

Try to stop routing with TCPIP> STOP ROUTING/GATED and/or manually remove the routes with TCPIP> SET NOROUTE 10.12.2.1

You might also want to check the permanent routes with TCPIP> SHOW ROUTE/PERM

Volker.

[Wim Van den Wyngaert]

NOte that the MTU size of your local interface is 0.

Wim

[Galen Tackett]

Off topic a bit, but--

This morning I posted a reply saying pretty much the same thing as Martin's, but it never showed up here. My browser crashed perhaps a half hour later, but I can't see why that would matter.

[Wim Van den Wyngaert]

Galen,

Had that several times.

$ifconfig -a

will show your interfaces with MTU.

$ifconfig lo0 ipmtu 4096

will change it.

But I don't have your version of tcp. My version is refusing values 0 and 1. So, no idea where it came from.

Wim

[Keith Whitwell]

Martin,

Could you give me the intstructions to remove the route.
Do I need to shut down and start TCP/IP services?

Keith

[Steven Schweda]

TCPIP HELP SET ROUTE

SET NOROUTE

[Martin Vorlaender]

Keith,

Volker already mentioned the relevant commands (and explained it better than I did).

No, you don't have to restart TCP/IP. It should immediately start to work.

Keep an eye on those routing entries. It seems that some node (possibly 10.12.10.254) announced a route to 10.12.2.1 - which your VMS machine picked up.

HTH,
Martin

[Keith Whitwell]

Many thanks to all respondents.

I have removed the offending routes from both nodes, both dynamically and from the permanent database and am pleased to report that both the ping and the Gembase problems are now a thing of the past.

[Keith Whitwell]

see my last posting.

[Wim Van den Wyngaert]

For my information : is the MTU size in ucx sho int still zero for lo0 ?

Wim