The clients that this is occurring on are Sun Solaris client OS level 5.8,
client version 5.1.5.0.

The idletimeout value is set to 60 and I'm not quite sure where to find the
tcpwindowsize.

I have looked through the server activity log and I have found nothing out
of the ordinary.  It appears that the server that I have looked at was
archiving data, but there were no failures to transmit or even a connection
issue.

I have noticed that there are several archives starting at the same time,
but this isn't anything abnormal.  We do have mpthreading set to yes on
this test server. (Not quite sure if that makes a difference)

06/21/2003 17:18:22   ANR0403I Session 94347 ended for node FJSU001 (SUN

                       SOLARIS).

06/21/2003 17:18:23   ANR0406I Session 94348 started for node FJSU001 (SUN

                       SOLARIS) (BPX-Tcp/Ip 157.154.43.107(-23745)).

06/21/2003 17:18:24   ANR0406I Session 94349 started for node FJSU001 (SUN

                       SOLARIS) (BPX-Tcp/Ip 157.154.43.107(-23744)).

06/21/2003 17:18:28   ANR0406I Session 94350 started for node FJSU001 (SUN

                       SOLARIS) (BPX-Tcp/Ip 157.154.43.107(-23734)).

06/21/2003 17:18:29   ANR0403I Session 94350 ended for node FJSU001 (SUN

                       SOLARIS).

06/21/2003 17:18:31   ANR0403I Session 94349 ended for node FJSU001 (SUN

                       SOLARIS).

Thanks!

Joni Moyer
Systems Programmer
joni.moyer AT highmark DOT com
(717)975-8338



                      Richard Sims
                      <rbs AT BU DOT EDU>             To:       ADSM-L AT 
VM.MARIST DOT EDU
                      Sent by: "ADSM:          cc:
                      Dist Stor                Subject:  Re: TCP/IP socket error
                      Manager"
                      <[email protected]
                      .EDU>


                      06/23/2003 08:04
                      AM
                      Please respond to
                      "ADSM: Dist Stor
                      Manager"






>Our TSM server was upgraded from 4.1.5 to 5.1.6.2 on os/390 and ever since
>then I have been noticing the following errors on the clients.  Has anyone
>seen this?  Is there a known fix?
>
>06/21/03   16:57:57 ANS1005E TCP/IP read error on socket = 5, errno = 131,
>reason : 'Connection reset by peer'.

Joni - Two pieces of information are important here:

1. Your client type and level.

2. The TSM server is the "peer", and so because it terminated the session,
   there should be a reason in the server Activity Log.  There may be
   configuration issues like server timeout values and TCPWindowsize.

 Richard Sims, BU