telnetd[9352]: getpid: peer died: Error 0

[Guettache]

Hello,
For about 30 mn, clients lost there connection to the server . The syslog reported this error
telnetd[9352]: getpid: peer died: Error 0
The server was running but lost its network connection.
lanscan does not report any problem (lan cards were up)
But pinging the IP of the server gives a timeout and
pinging from the server a client IP gives as well a timeout.
the neetl.LOG00 shows this error:
***********************************STREAMS/UX********
Timestamp : Wed Jun 19 METDST 2002 10:17:40.331160
Process ID : 9861 Subsystem : STREAMS
User ID ( UID ) : 105 Log Class : ERROR
Device ID : 0 Path ID : 0
Connection ID : 0 Log Instance : 0
Location : 00123
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2 10:17:40 189080965 1 T.. 5321 123 tcp_rput_other: case T_ERROR_ACK, ERROR_prim == 1

After half an hour the server network connections were back.
my system is an HP9000/rp5400 OS 11.0
Does anyone had such event. What could have caused the server to loose its network connection:

Thanks for your help
Nariman

[Rosli Ahmad]

Hi,
There could be number of reasons for this to happen and such message appearing in your syslog. try to extract something from this thread.
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xd6a479bffde7d4118fef0090279cd0f9,00.html

regards
Rosli

[U.SivaKumar_2]

Hi,
Is Your Server connected to a port of switch or
hub ?. IF yes , try changing the port of the server in the switch or hub and see whether the
problem repeats.

If possible , for isolating the problem , change the switch or hub to which server is connected with another one.

regards,
U.SivaKumar

[Stefan Farrelly · Accepted solution]

The chances that it was your HP server that caused the problem are very small.

It was very likely it was caused by network equpiment problems - switch or hub or router or cabling. Talk to your network people to resolve.

[Guettache]

Hi,

Thanks for your reply.

It seems indeed like a network
equipment problem rather than
a problem with the HP system.

Regards,

Narimane

[Michael Tully]

I agree with Stefan. This problem is definitely at the network end. Perhaps they are doing something they shouldn't be and didn't tell anyone. If they had performed an un-authorised change maybe bring out the big stick.

Michael

[sven verhaegen]

Some explaining about 'Reset by Peer'

Many mistakes exist on this subject , reset-by-peer mean literaly that the receiving system , in this case the HP9000 has received a signal form the other side it was unable to cope with the communication due to 'circumstances' , these may be timing delay , bad packets , out-of-sequence data streams etc.. , to better understand wath happen only tracing the event will tell you why your PC clients refused to continue , this can be HP9000 related but also derive from network devices causing errors , general network latency or even interferences and bad lancables , if lanscan and lanadmin show no error , look at any switches or routers inbetween for error logs and if that isn't possible trace afailing communication on h tHP and on the PC and observe the behaviour , that will tell you exactly what is happening

[Paula J Frazer-Campbell]

Hi

Set up a ping routine to monitor all devices in the network chain all the way to a user.

Monitor and note where in the chain the ping stops at when problems occur.


You might have to enable it if you have a firewall in your chain.


HTH

Paula

[Guettache]

Hi stepfan,

The cause of the problem reported by the network team:

"The problem was because a user set up his PC by mistake with the IP address of the default gateway for the subnet nn, leaving isolated the network. Once the problem and the user creating the problem were identified, it was fixed."

note: my system is on subnet nn

Cheers,

Narimane