Re: [cobalt-users] Virtusertable again

[Jeff Lasman <jblists@xxxxxxxxxxxxx>]

Stuart Low wrote:
> /etc/rc.d/init.d/sendmail restart
> /etc/rc.d/init.d/poprelayd restart

While this works it actually completely stops then restarts the server. 
That's much slower than just telling the server to restart itself. 
Don't ask me why but I've measured, and it is...

# kill -HUP <pid#-of-sendmail>

You get the pid# of sendmail by looking at the first line of
/var/run/sendmail.pid

# head -n 1 /var/run/sendmail.pid

which in my case just gave me 24334.

Then do:

# kill -HUP 24334

and sendmail will silently and promptly restart itself.

Personally, I automate this:  I've got a file at /root/sendmail.restart

with the following two lines:

#!/bin/bash
kill -HUP `head -n 1 /var/run/sendmail.pid`

which will restart the server promptly.

The original poster may need to be reminded as well that before
restarting sendmail he must rebuild the access database thusly:

# makemap hash access < access

I'm taking the liberty of answering the original poster by replying to
your reply...

> > 2)Are the options REJECT and DISCARD accepted while editing the access
> file?
> 
> Not sure about what 'access file' you are talking about here? It is a stock
> mail setup so I guess if it works with a standard sendmail setup it will
> work.

Yes.  From the www.sendmail.or faq:

> Values can be REJECT to refuse connections from this host,
> DISCARD to accept the message but silently discard it (the
> sender will think it has been accepted), OK to allow access
> (overriding other built-in checks), RELAY to allow access
> including relaying SMTP through your machine, or an arbitrary
> message to reject the mail with the customized message. 
> 
> For example, a database might contain: 
> 
> cyberpromo.com  REJECT
> sendmail.org    RELAY
> spam@xxxxxxxxx  550 Spammers shan't see sunlight here

> 
> > 3)If i have all the updates to my Raq3, what is the maximum size of a
> > username with and without the . (dot in the middle)?
> 
> 255  characters I THINK.

I just did some tests.  The RaQ4 gui allows only 12 characters.  Using
SSH into the command prompt I was able to create a user
"antidisestablishmentarianism" but I couldn't send it mail; the mail
appears to have disappeared <frown>.  I didn't get it back and it
doesn't seem to be anywhere on the RaQ <frown>.

> > 4)Has someone found a reasonable way around the following situation: 2000
> > users to be distributed among 150 virtual sites, each of these virtual
> sites
> > are vanity email domains which have surnames after @ like:
> >    user1@xxxxxxxxxx   user2@xxxxxxxxxx  user3@xxxxxxxx
> >     It happens that all of them want a short first name as john, david,
> dan
> > ;-) but of course there are at least 20 john among the 2000. The webmail
> > program is not hackable for them to authenticate exclusively via email
> > address. There has to be a login name but i would prefer not to let them
> > notice they are john155 or john1023. Also, the Webmail program  (although
> it
> > is possible to configure a "reply to" address of john@xxxxxxxxxx for
> > everybody) would always "confess" to the recipient that the email was from
> > john155@xxxxxxxxxxx

I don't know what webmail program you're running; that doesn't seem to
be a RaQ issue, though.  You can't escape the "different" login with the
RaQ; you can by some fancy configuring on various other Linux and W2K
solutions, as you've already been advised in other threads.

> If someone has an easy way around this mess, then
> > prepare to fase 2: What if in some of the cases there are more than two
> john
> > with surname crosby, for instance? Should i seriously consider two or more

No matter what solution you use, you'll never be able to have more than
one "john@xxxxxxxxxx".  Can't happen.  There's no way the systems could
ever tell the difference.  Not even a Cray <wry grin>.

Jeff
-- 
Jeff Lasman <jblists@xxxxxxxxxxxxx>
Linux and Cobalt/Sun/RaQ Consulting
nobaloney.net
P. O. Box 52672, Riverside, CA  92517
voice: (909) 787-8589  *  fax: (909) 782-0205