Losing telnet and ftp functions on RS6000 2

[marcmcconnell]

On an RS6000 43P Model 150 some time after the system is booted, users are unable to telnet or ftp onto the machine.

You can still ping the RS, however it takes 40ms to send 64 bytes rather than the 1 second it should do.

Errpt reports problem with parallel port. Diagnostics point this to a sysplanar fault.

Could this cause the network problems or is it something else ?

 

[PSD]

Marc,

I assume that on the 170 you do not have a PCI ethernet card and are using the on-board? If you are using the on-board if you have sysplanar problems as the IP port is built in it could have repercussions of this nature. I would run diags. on the ethernet controller with a wrap plug if you have one. Also, is there nothing in the error report about ethernet errors.

Can you log onto the console? and you say after a while it happens, does it work as normal to begin with....also, when the network is playing up is the box slow when logged on locally?? Could it be a network problem? can you ping from the box to other servers within normal repsonse times?

What OS and patch level are you on?

Hope that helps



PSD
IBM Certified Specialist - AIX V4.3 Systems Support
IBM Certified Specialist - AIX V4 HACMP

 

[christos]

check the portmapper
eg put in a cron job to reset it every couple of mins -

if the problem goes then there is a port mapper problem

 

[MikeLacey]

Marc,

I would get the reported sysplanar problem investigated first; look at s/w problems after the h/w appears OK. Mike
michael.j.lacey@ntlworld.com
Email welcome if you're in a hurry or something -- but post in tek-tips as well please, and I will post my reply here as well.

 

[marcmcconnell]

Hello PSD,

The sysplanar has now been changed. Same problem exists.

OSLEVEL is 4.2.1 in all my machines.

You can log on normally to the machine but the application will hang from time to time because of this network like problem.

If you try and telnet from another machine and it fails then trying to telnet back from the faulty machine causes the ping to hang (100% packet loss).

I have eliminated the network by changing cables, ports etc.

Engineer thinks something is corrupt in AIX.

Any more ideas ?

Thanks for your time,

Marc.

 

[aixqueen]

Just my two cents...

What patch level are you at for 4.2.1
There were a lot of patches that had to do with hangs in the ethernet in
4.2.1

Including
U468649 - bos.net.tcp.client.4.2.1.35
with co and pre reqs....

If you arent at the latest level of AIX for 4.2, I suggest getting the maintenance
level at least.........

instfix -i | grep ML
AIX 4.2.1:
IY08272 February 2000
IY03101 RECOMMENDED Maintenance Level(4.2.1-06)

Number two mbufs can cause it to hang as well.....you didn't say if you
were running out of mbufs.... netstat -m

Also is it a 10/100 or 10...if 10/100 make sure it is not set on auto, but set to
10 or 100..........

entstat -d ent0
netstat -v | egrep "STATISTICS|Media Speed"
lsattr -E -l en0

Some network performance ideas:

http://www.rs6000.ibm.com/doc_link/en_US/a_doc_lib/aixbman/prftungd/2365ch9.htm#23142

 

[marcmcconnell]

Problem solved. The problem was that the network had several ghost IP addresses still active and not flushed from the network. This unit had been allocated one of these IP addresses. Allocated a different IP address then rebooted switch to flush old IP addresses.

Thanks for all who helped.