hp servers unreachability in network

[g durga prasad]

I am having one hp l1000 server, l2000 servers. These servers are frequently becoming unavailable in network. i have to disconnect the network cable and connect it again... to be available to the network. The servers are not giving any error messages. these servers are connected to a cisco switch and the link is still established even the servers are not reachable by ping.

[Paula J Frazer-Campbell]

Hi

Lots of things to check:-

1. cables
2. all devices which the servers use to connect to network.
3. Duplicated ip addresses.
Etc.

As you said "Servers" I would very strongly suspect the network and not the servers.


HTH

Paula

[Shahul]

Hi

Can U give some more inputs like whether both the servers are in same segment or not? Are U able to connect to other machines?..etc

If U are able to solve the problem by just removeing and inserting the network cable, It must be some loose connection or cable crimping problem. If U are using UTP cable please recrimp it once again. It may solve ur problem.

Are U able to ping to other systems from a server while it is not communicating to other server? Give me more inputs..

Best of luck
with rgds, Shahul

[John Waller]

Normally if it is a cable problem you will see a message on the Console to the effect of :

LAN DISASTER

If you are not seeing these messages, then I suspect that your problem may be with the switches. Unfortunatly I need a bit more info to diagnose further

[Vincenzo Restuccia]

I think the problem is with the cisco switch.

[KapilRaj]

hi ,

Just do a search on the forums. i hv seen a similiar prob few days back. HPUX 11 had utils which will disable the gateway if it cannot be ping.

Just give a try

kaps

[Glen Van Lehn]

hi - yes lots to check...

? are you using 10/100 Tx ports ? if so: following; else:?
- if you aren't getting 'Lan Disaster' msgs & you can ifconfig the interface down /up without errors, then you probably have carrier sense between server and the switch port

- verify speed/duplex settings on both switch & server
-- lanscan to get the NMID for your lan interface
if interface is 'down', that's a problem to focus on*
else
-- /usr/sbin/lanadmin -x tells the server setting
-- log onto switch & check status of the switch port the server is plugged into. I suggest the port NOT be in "Auto-detect" mode--nail it to server settings.
-- while on the switch, clear the port counts & ping out..
do the counters increment ?

- the gateway is config'd to not answer pings...
ran into this on a firewall -- vi /etc/rc.config.d/nddconf and turn off the default probe:
nddconf: add ndd_name[x]=ip_ire_gw_probe, \ ndd_value[x]=0, transport[x]=ip
[x={0..n} index number.. see comments in file ]

there are a lot more possibilities.. good luck