network connectivity problem in Hp unix 10.20

[sacthivel.C]

Dear all,
we are using HP 9000 severs(hpunix 10.20 with oracle7).
Nearly 100 clients accessing this server.But from yesterday onwards some of the pc's only getting connectivity.
The remaining are not event pinging to server.
i tried by restarted the service(net),replaced the lan cable and restarted the server also.
So pls help me to solve this issue.

Regards,
C.Sacthivel

[Steven Schweda]

> [...] some of the pc's only getting
> connectivity.

If some PCs can talk to the HP-UX system,
then it would seem that the HP-UX system is
working, and it's connected to the network.
(Wouldn't it?)

It sounds to me like a network problem, and
we have no idea what's in your network
between the various PCs and the HP-UX system
(hubs, switches, cables, routers, ...).

[sacthivel.C]

Dear all,
I will try to explain my network setup.
We are having server in one place and the customers are accessing from another location (200 km's away)
We are having cisco routers on both the sides connected using Lease line.
On both sides we are connected using 10/100 mbbs switches.

regards,
C.Sacthivel

[Matti_Kurkela · Accepted solution]

If some clients at the remote site *can* access the server, then the leased line probably works. The server's NIC and LAN cable are obviously fine too.

Has anyone changed anything related to the routers or switches in the last two days?

Has there been any power outages, or anything else that might have caused a router or switch to restart? (A common error in configuring Cisco switches and routers is forgetting to save the configuration after the changes. When there is a power outage, the latest configuration changes are lost and the device returns to the last saved configuration.)

Without knowing the network topology in detail, I would guess that the problem would most likely be at the remote site, but that guess needs to be verified.
So, the first test would be to ping the router at the client site from the client PCs.

If one group of PCs can do this and the other can't (and the grouping is the same as with the server connectivity), you'll know for sure that the problem is somewhere inside the client site, before the router.

If all the PCs can ping the router, check any access restrictions in the routers and other network components in the server site.

If none of the PCs can ping the router at the client site (not even those that *can* access the server), you may need to plug in a network analyzer in various places at the client site to see what is going on. Or pick one of the PCs that is known to work and plug it into each of the switches of the remote site in turn. You may find a failing switch (or a failing module in a big switch), or a failed/disconnected interconnect between two switches.

MK

[rick jones]

You might try collecting "traceroute" output from the HP-UX server for the case of both a PC that can be reached, and one which cannot.

Do the same thing from each of those PCs - under Windows the utility is IIRC called tracert.

Post the results here.

[sacthivel.C]

Dear all,
When i analise the trace route output it first crosses the router then it shows 97 series ip address continuously.it goes on continuously.i terminate it after 55 traces.
The pc's which didn't get connection is not constant.
It keeps on changing.
According to the network support engineer,It may be virus problem.
Is their any possibilities for that?

regards,
C.Sacthivel

[Steven Schweda]

> When i analise the trace route output [...]

I might tell you what happens when _I_
analyze the traceroute output, but I can't
see the traceroute output. Perhaps one of
the Forum psychics can help you, but I'd need
to see the traceroute output, and I'd need to
know some of the IP addresses of these
systems.

> Is their any possibilities for that?

Many things are possible, but "PC" doesn't
really tell us what the OS on the PC is.
Assuming that it's some form of Windows, you
probably should be running some sort of
anti-virus software in any case. I'm not a
frequent Windows user, but I tend to doubt
that a virus would mess with your IP routing.
Most viruses want the networking to work
properly, so that they can propagate
themselves, and attack other systems.

[sacthivel.C]

Hello,
I attached here the output of the traceroute from server.

Note : pls open it in wordpad.

regards,
C.Sacthivel

[Matti_Kurkela]

These traceroutes were from the server towards the clients, right?

Looks like 97.97.cc.dd is the last common point. After that, one client is directly reachable without going through any other routers (as far as traceroute can detect), and the other just fails totally.

This is a possible confirmation that the problem is *not* within your server (without knowing the exact topology of the network and the network/netmask values at each subnet, I cannot be sure).

What is the role of 97.97.cc.dd? Is it the router at the server end of the leased line, is it the router at the client end, or is it a completely different box?

There seems to be some sort of a problem in connectivity in between 172.27.ee.ff and 97.97.cc.dd.

Try to run a traceroute (tracert) from 172.27.ee.ff towards the server. If it cannot reach 97.97.cc.dd at all, you've found a broken connection (either a configuration error or actual missing/broken hardware) somewhere between the client and 97.97.cc.dd.
The next question will be, what _can_ 172.27.ee.ff reach? The answer to that question may allow you to pinpoint the location of the problem:

- nothing at all: check the NIC and cable on 172.27.ee.ff, then the switch port 172.27.ee.ff is plugged into.

- other machines in the same network segment plugged into the same switch, but nothing beyond that: seems like a problem in the interconnect between the switches.

- only some machines in the same network segment plugged into the same switch, but not all of them: most likely a broken module/internal chip in the switch.

- other: more information about the physical network topology is required to analyze the situation. Give your network admin all the information you've collected about the problem so far, and *gently* ask him/her to stop sidetracking you with wild guesses about viruses and do his/her job :-)

If the traceroute from the client does reach 97.97.cc.dd but then displays only asterisks after that, it looks like the problem is on 97.97.cc.dd. In that case, it might be a routing error or an access restriction configuration error.

MK

[sacthivel.C]

It means their is no problem in the routing,right?

I sent all the log files to our network administrator and waiting for the reply.

Thank you for your reply.

regards,
C.Sacthivel

[sacthivel.C]

Hello,
I give the details which you want.
Pls find the attachment.


Note:Open it in wordpad.



Regards,
C.Sacthivel

[sacthivel.C]

Thanks to every body who replied me.

And special thanks to Mr.Matti Kurkela.As per his suggestion i asked my network admin to check the router configurations once again and he found the flaws and rectified it.

Now we are getting the connectivity.

Regards,
C.Sacthvel