Hi Craig,

What I do is create separate TCPMAINT userid's for each separate TCPIP
stack, and of course separate TCPIP and associated userids.  

Each TCPMAINT userid has it's own copy of the TCPMAINT 198 minidisk so
that changes you make to one stack effect ONLY that stack.  You could
also give each TCPMAINT it's own 592 minidisk, but the only file your
secondary TCPMAINTs and TCPIP userids really need is TCPIP DATA so I
keep that on the 198 instead of the 592 minidisk.  It also makes it
simple to have customized FTP, SSL, SMTP, POP, etc. configurations for
each stack.  So for example, I have 3 stacks:

Production Stack:
TCPMAINT
 - TCPIP
 - FTPSERVE
 - SSLSERV
 - NAMESRV
 - SMTPSRV
 - POPSRV

Stack A:
TCPMAINA
 - TCPIPA
 - FTPSERVA
 - SSLSERVA
 - NAMESRVA
 - SMTPSRVA
 - POPSRVA

Stack B:
TCPMAINB
 - TCPIPB
 - FTPSERVB
 - SSLSERVB
 - NAMESRVB
 - SMTPSRVB
 - POPSRVB

Etc.

Having each stack complete isolated from each other stack helps ensure
that you don't accidentally make a change intended for say Stack A to
your Production stack.  Whenever you want to use use commands like
NETSTAT against a particular stack, just link and access the appropriate
TCPMAINx 198 disk ahead of the production TCPMAINT 198 disk.

It's also VERY handy to have multiple running TCPIP stacks.  On those
rare occasions when something has happened that takes down our
production TCPIP stack, or I need to bounce the production TCPIP stack,
I have 2 "back doors" that I can use to get in and diagnose the problem
with via the Telnet servers on TCPIPA and TCPIPB.

PS:  With regards to SSLSERV keep in mind that it accesses a BFS that
contains the key data.  You'll need to either setup up a separate BFS
tree for EACH SSLSERV userid you intend to use, OR if you want to share
the key data across all of your SSLSERV's you'll need to give each
SSLSERVx userid SFS access to the BFS tree (I think R/O is OK for this).
Since we rarely change keys I just let all of the SSLSERVx's share the
one BFS tree and maintain it with one GSKADMIN (i.e. I don't have
GSKADMNA, GSKADMNB, etc.).

-Mike

-----Original Message-----
From: The IBM z/VM Operating System [mailto:[email protected]] On
Behalf Of Craig Dudley
Sent: Wednesday, March 25, 2009 4:06 PM
To: [email protected]
Subject: Question about separate TCP stacks and SSL server machines


Hi,
I am trying to set up a 2nd TCPIP/IP stack (TCPIPW2H) along with a 2nd
SSL server (SSLSRV1). When I start up SSLSRV1 manually I get these
messages:

DTCSSL002I main(): calling CQadminMain()... 
DTCSSL126E setibmsockopt() failed; rc: -1 errno: 1123 DTCSSL095E errno
detail: 1123 => EDC8123I Socket already connected. DTCSSL002I ckEPIPE():
Invoking shutdownNow() -- requested by caller DTCSSL127E Server shutdown
has commenced DTCSSL103E SSLADMIN communication error: QueueReceiveBlock
error; rc: 8 
reason: 221
DTCSSL128E Server shutdown is complete

I already have a TCP/IP stack running (TCPIP) and another SSL server
(SSLSERV) running.

My current guess is that SSLSRV1 is trying to use TCP stack TCPIP 
rather than stack TCPIPW2H.

Do I need a TCPIP DATA (with TCPIPUSERID TCPIPW2H) on SSLSRV1s A-disk 
for SSLSRV1 to use TCPIPW2H as its stack?

Thanks

-- 
Craig Dudley
Manager, Mainframe Technical Support Group
Department of Information Technology
State of New Hampshire
27 Hazen Drive
Concord, NH 03301
603-271-1506    Fax 603-271-1516

Reply via email to