[Info-vax] HP OpenVMS TCPIP SMTP gives error after change of node name
Rich Jordan
jordan at ccs4vms.com
Tue Oct 30 15:55:42 EDT 2012
On Oct 30, 11:46 am, RobertsonEricW <robertsoner... at netzero.net>
wrote:
> On Tuesday, October 30, 2012 10:51:09 AM UTC-4, Stephen Hoffman wrote:
> > On 2012-10-30 14:21:23 +0000, RobertsonEricW said:
>
> > > All,
>
> > > I went through the node rename items described at
>
> > >http://labs.hoffmanlabs.com/node/589
>
> > > (original node name of IQDEV1, new node name of DEMO1) and ended up
>
> > > with everything working as expected except for the TCP/IP SMTP
>
> > > service. The situation seems a little schizophrenic because the TCPIP
>
> > > configuration seems to recognize it is now operating on the node with
>
> > > the new name, but when you attempt to start the service it issues an
>
> > > error and attempts to use a queue name with the old node name. Anyone
>
> > > got any ideas on why this might be? What follows is what I think is
>
> > > necessary to know. But if some other information would be useful, let
>
> > > me know and I will happily provide it.
>
> > ...
>
> > > SMTP Configuration
>
> > > Options
>
> > > Initial interval: 0 00:30:00.00 Address_max: 16
>
> > > NOEIGHT_BIT
>
> > > Retry interval: 0 01:00:00.00 Hop_count_max: 16
>
> > > NORELAY
>
> > > Maximum interval: 3 00:00:00.00
>
> > > TOP_HEADERS
>
> > > Timeout Initial Mail Receipt Data
>
> > > Terminate
>
> > > Send: 5 5 5 3
>
> > > 10
>
> > > Receive: 5
>
> > > Alternate gateway: not defined
>
> > > General gateway: not defined
>
> > > Substitute domain: not defined
>
> > > Zone: not defined
>
> > > Postmaster: TCPIP$SMTP
>
> > > Log file: SYS$SPECIFIC:[TCPIP$SMTP]TCPIP$SMTP_LOGFILE.LOG
>
> > > Generic queue Queues Participating nodes
>
> > > TCPIP$SMTP_DEMO1_00 1 DEMO1
>
> > > $ @SYS$STARTUP:TCPIP$SMTP_STARTUP.COM
>
> > > %TCPIP-I-INFO, image SYS$SHARE:TCPIP$SMTP_MAILSHR.EXE installed
>
> > > %TCPIP-I-INFO, image SYS$SYSTEM:TCPIP$SMTP_RECEIVER.EXE installed
>
> > > %TCPIP-I-INFO, logical names created
>
> > > Initializing TCPIP$SMTP_IQDEV1_00 queue got status %SYSTEM-F-BADPARAM,
>
> > > bad parameter value
>
> > > %SYSTEM-F-BADPARAM, bad parameter value
>
> > > %TCPIP-I-INFO, service queues started
>
> > > %TCPIP-I-INFO, service enabled
>
> > > %TCPIP-S-STARTDONE, TCPIP$SMTP startup completed
>
> > As for the queues, disable the SMTP server via the TCPIP$CONFIG menus,
>
> > and then enable it and restart it, and see if you get a new set of SMTP
>
> > server queues created. If not, you could manually recreate the SMTP
>
> > queues, matching the settings of the old SMTP queues. I just checked
>
> > a test box that I'd renamed, and I still have the old queues hanging
>
> > around; off to delete those. IIRC, I stopped and disabled and
>
> > re-enabled and restarted most of the TCP/IP Services servers as part of
>
> > the rename.
>
> > If that works, let me know and I'll update the article.
>
> > This whole networking area of OpenVMS clearly wasn't ever actually
>
> > designed, and it's become an aggregation of hacks stretching back
>
> > decades. ...Though resolving this mess would have required the source
>
> > code equivalents of a flame thrower and a grenade launcher as part of
>
> > the requisite "design renovations".
>
> > FWIW and IIRC from the release notes, there have been changes to SMTP
>
> > in a recent ECO; they went to a Unix-style configuration file for some
>
> > features, rather than integrating the updates into the database. Check
>
> > the files and see if they've picked up a host name. While not related
>
> > to the queues, also check for host names in the TCPIP configuration
>
> > files.
>
> > --
>
> > Pure Personal Opinion | HoffmanLabs LLC
>
> Yes, I too went through the entire TCP/IP configuration using SYS$MANAGER:TCPIP$CONFIG.COM and still had this problem. The interesting thing is that I have renamed nodes before but don't remember having this problem. However, there may have been a subtle difference in how I went about renaming the node this time around as compared to the last time I did it.
>
> Your suspicion regarding the UNIX-like configuration file may be correct. I found the file SYS$SYSDEVICE:[SYS0.TCPIP$SMTP]TCPIP$SMTP.CONF had an entry:
>
> Queue-Name : TCPIP$SMTP_IQDEV1_00
>
> within it. After changing this entry to read:
>
> Queue-Name : TCPIP$SMTP_DEMO1_00
>
> and saving it, I was able to restart the SMTP service without incident and the expected queues once again appeared in the system's list of queues.
>
> However, before you go to the trouble of changing the article, I will once again rename the node paying more close attention to the procedure I used last time to see if I obtain a better result.
>
> Thanks again for your help.
>
> Eric
You have been bit by the same thing that got me. TCPIP V5.7 is
TOTALLY different.
The TCPIP SET CONFIG SMTP command (and SHOW CONFIG SMTP), most of the
TCPIP$SMTP* logicals, and the SMTP.CONFIG file method of SMTP
management were all replaced in TCPIP V5.7 with the TCPIP$SMTP.CONF
file you noted above.
The release notes are the only docs I've found about this change; none
of the main TCPIP manuals have been updated (at least online at the HP
website) so they still show the SET CONFIG SMTP and old config files
as the way to go.
When a system is upgraded from V5.6 to V5.7 the generic and execution
queue names get changed (they still incorporate the nodename) and new
ones get created with those names, but the OLD queue names do NOT go
away. You'll have stopped/dead queues hanging around until and unless
you clean them up. The upgrade is supposed to transfer part of the
config from the old method/files to the new TCPIP$SMTP.CONF file.
Coincidentally both V8.4 systems I have access to (both upgraded from
VMS V8.3 and TCPIP V5.6) also no longer stop the SMTP execution queues
when the SYS$STARTUP:TCPIP$SMTP_SHUTDOWN command is issued; they have
to be stopped manually. I haven't had the chance to do a clean V8.4
install with TCPIP V5.7 to see if there's any difference.
More information about the Info-vax
mailing list