Operating System - OpenVMS
cancel
Showing results for 
Search instead for 
Did you mean: 

SMTP receive sequence and status

SOLVED
Go to solution
Troodon
Frequent Advisor

SMTP receive sequence and status

I notice that image accounting shows LOGINOUT exiting with a status of %X0 at the beginning of every conversation with a sending node. Is this the expected behavior, or did someone forget that 1 is the standard exit on VMS and not 0.

First is LOGINOUT with an exit of %X0.
Next, SET with an exit of %X10000001
Then DELETE with an %X10000001
Next TCPIP$SMTP_RECEIVER.EXE with an %X1
Finally, LOGINOUT.EXE with %X10000001
3 REPLIES
Wim Van den Wyngaert
Honored Contributor
Solution

Re: SMTP receive sequence and status

May be because loginout is not run as an image (don't know how to call it) ?

When you do logout it runs as an image and exits normaly.

When you do run/det RUNDET.exe is acticated that will run loginout in the newly created process. Here it exits with 0 too.

When it runs in a process context to initialize the process (as requested by job control) it could conform to other standards.

Fwiw

Wim
Wim
Hoff
Honored Contributor

Re: SMTP receive sequence and status

There are two issues, and yes, that status does look a little weird though loginout is itself a seriously weird image; its goal is to set up the DCL environment. The rundown implicit in its image exit already has implications for temporary identifiers, too. There are other oddities. (My guess? Somebody missed an exit path.)

The second issue is why TCP/IP needs a DCL-based process to receive an SMTP connection. That looks to be a fairly heavyweight and more complex approach toward receiving a message. DECnet evolved away from this, and allowed direct image activation as part of its server task support.

Not sure I'd class either of these as bugs or as "yep, it does that" cases, and that decision is up to HP in any case. Having an SMTP server start up, and start sucking in all arriving mail would seem a more resource friendly approach -- and current behavior would seem to have a wider effect than an errant R0 exit.
Troodon
Frequent Advisor

Re: SMTP receive sequence and status

I agree that the entire TCP/IP stack ought to be revisited for performance and security.

Another peeve is how SSH creates you as a "INTERACTIVE" (i.e., local) user, and none of the utilities show the remote host information like LAT used to in $SHOW USER/FULL

My guess is that SSH ought to be an ACP-like process for the SSH protocol, so I do not get 2 processes per connection. I think a more full-featured pseudoterminal like the old YW/PY pair would solve a lot of evils here.