Printer system mail messages are coming from user root and get sent to root, as that's the only default member of
the printq group able to run the qdaemon.  They also unhelpfully have a blank subject in my experience.

There are three things which may have an effect on the messages:

From "man chvirprt", note the "si" attribute - e.g. for queue "asc" to a remote device "@lpd", to send mail to "nobody":
chvirprt -d @lpd -q asc -a si=nobody

smit chvirprt
- enter queue name from lpstat,
- choose option 2 Default Print Job Attributes
- go to the end ESC->,
------------- Messages/Diagnostics ------------                             
DIAGNOSTIC LEVEL                                    (normal) - print job; > +
- select "ignore" from the F4 list.

From "man qdaemon", check if you have this file:
/var/spool/lpd/pio/@local/fullmsg    # not present by default I think.

However, none of these things stop messages to root on my system 

So the only other approach seems to be some kind of mail filtering,
which I cannot really recommended, and will not by default work with the root user
as sendmail will block it as "unsafe".  So tough basically.

Oh, and BTW, to set up to simulate problems, generating messages...
If you have an existing remote print queue, make sure it is "enable"d READY,
unplug the physical printer and print something to it - after a timeout you'll
get the root Message from qdaemon as the queue is taken DOWN when no
contact can be made.

I have had problems in the past with sendmail resolving names. But that was especially because I was using my own server as nameserver.

If you are usig your own nameserver, try to get it use named9, rather than named8 (and certainly not named4 - there are MANY resolvers that refuse to talk to named4 resolvers).

The fact that you can nslookup the hostname is not enough. You must also be able to get the MX record.

Consider using dig <DNSname> for additional testing.

I'll have to think about this a bit. Will get back to it on Monday.

However, AIX level, NIM setup basics, etc..

Yes, you will almost certainly need to make some changes.

No guarantee on any of these approaches:

Save the nim database, and find the entry for master (if1 attribute iirc) and make the change, and restore the database.

Two: create a new interface with the new name - might even work with hostname in same nim network. It is about routing... If that starts working, you can delete the oirginal interface, and all installs will run via the new interface. I used to have three interfaces on my master (if1, if2, if3). The if chosen was determined by most direct route to client.

Last: ODM commands (odmget(s), followed by odmchanges once you have the records figured out. Start with saving the current definition.

Lastly - stop using eznim. At least when it came out several years ago, I tried to use it and just hated how it got in the way all the time.

Worst case: you have to rebuild your spots (i.e. delete nim database, and recreate resources from files you already have in place. You do not have to reread all the cds again, etc. But the spot will need to be rebuilt.

I tend to think more towards routing problem than name resolution. The relay server in sendmail.cf is resolvable.

can you do an nslookup of marksandspencer.com, and is the output correct?

did you try to mail another mail address, and did that work?

did you allready searched via google on sendmail and Transient parse error ?
i just did and got quite some tips?

I would like to know, wether it is possible to configure printing from AIX in such a way, we do NOT receive mails about failed prints/ printer problems.

Currently we get messages like:

Every 5 minutes we check the printqueues, and if they have gone down, we reset them. That most likely gives the job deleted message. I agree in principle we should try to tackle the problem at the source, unfortunately that will be quite hard. As the printers are placed in different sites, and simulating the problem is quite hard.

I tried to find more info about this via google, but was unable.

Or is it maybe possible to filter the message out of /var/spool/mail <user> file?

Hi

I have a NIM server on AIX 5.3.08.02. I had to change the host name and I now have nimsh connection issues to the NIM clients. All host files and DNS entries are correct with the new host name. I have tried editing the niminfo files on the NIM server and NIM clinet. I have tried removing the the NIM client from NIM and then recreating it through "smitty eznim" on the client, but still no luck. I get the following error in /var/adm/ras/nimsh.log

Tue Oct 14 15:52:34 2008    passing OpenSSL setting of 0
Tue Oct 14 15:52:34 2008        error: remote value passed, "new_host_name", does not match environment value "old_host_name"

Does anybody have any ideas? Could this require an ODM edit to change the NIM server host name entry?

Thanx

we have configured nameserver in /etc/resolv.conf and able to ping nameserver and we have specify smart relay host in /etc/sendmail.cf as well.

It looks like you did not configure an dns.
so it does not know where to deliver the mail?

Do you have an nameserver mentioned in /etc/resolv.conf?

And is your sendmail config (/etc/sendmail.cf) correctly configured? Maybe you need to specify an Smart relay host?