Re: [gentoo-server] qmail virtual domains + virtual users

[Ben Munat]

Well, that appears to have been one of my problems. I think it was 
because of the "start qmail processes" button in webmin. This is how my 
process tree looked right after I read your email:

      |-pagebufd
      |-qmail-send-+-qmail-clean
      |            |-qmail-lspawn
      |            `-qmail-rspawn
      |-sshd---sshd---bash---pstree
      |-svscan-+-2*[supervise---tcpserver]
      |        |-3*[supervise---multilog]
      |        `-supervise---qmail-send
      |-syslog-ng

I stopped the qmail processes from within webmin and then, just to be 
safe, restarted svscan with /etc/init.d/svscan. Now I get this:

      |-sshd---sshd---bash---pstree
      |-svscan-+-2*[supervise---tcpserver]
      |        |-3*[supervise---multilog]
      |        `-supervise---qmail-send-+-qmail-clean
      |                                 |-qmail-lspawn
      |                                 `-qmail-rspawn
      |-syslog-ng

My question now is, is it okay that my process tree says 2* and 3* for 
tcpserver and multilog?

And, unfortunately, it still doesn't seem to be working. Well, at least 
I don't get the "cannot start" messages in the log... but now they look 
like this:

@40000000409669710e4d4334 status: local 0/10 remote 0/20
@40000000409669710eec55dc bounce msg 248578 qp 3022
@40000000409669710eec7134 end msg 248578
@40000000409669710f065dec new msg 248577
@40000000409669710f067944 info msg 248577: bytes 1368 from <#@[]> qp 
3022 uid 206
@40000000409669710f5099d4 starting delivery 9: msg 248577 to remote 
postmaster@×××××.org
@40000000409669710f50bcfc status: local 0/10 remote 1/20
@40000000409669710fc3c24c delivery 9: failure: 
Sorry._Although_I'm_listed_as_a_best-preference_MX_or_A_for_that_host,/it_isn't_in_my_control/locals_file,_so_I_don't_treat_it_as_local._(#5.4.6)/
@40000000409669710fc3e574 status: local 0/10 remote 0/20
@40000000409669710fc3f12c triple bounce: discarding bounce/248577
@40000000409669710fc3f8fc end msg 248577

BTW, I'm trying to do the qmail-inject example in the gentoo docs... and 
I still have a problem with mutt. How else can I check to see if the 
mail has gone through? Should it just be a file under ~/.maildir/new?

Thanks for sticking with me!

Ben



Andy Dustman wrote:
> On Mon, 2004-05-03 at 11:41, Ben Munat wrote:
>
>>Thanks (you too, Dadi) for the ideas. My /var/log/qmail-send/current
>>contains over 500k of lines like this:
>>
>>
>>@400000004095ef010bac1efc alert: cannot start: qmail-send is already running
>>
>>I scrolled all the way through it... that's all there is. Doesn't sound
>>right.
>
>
> It isn't.
>
> In /service, there should be qmail-send, which is a symbolic link to
> /etc/qmail/supervise/qmail-send. Make sure there aren't other links
> pointing there. But it sounds more like qmail-send is being started
> somehow other that by svscan (which monitors /service).
>
> You can check the status of qmail-send like this:
>
> # svstat /service/qmail-send
> /service/qmail-send: up (pid 7162) 1206573 seconds
>
> If you get 0 seconds, something is wrong. I find that pstree is a good
> way to get a system overview:
>
> # pstree
> init-+-6*[agetty]
> ...
>      |-sshd---sshd---sshd---bash
>      |-svscan-+-supervise---qmail-send-+-qmail-clean
>      |        |                        |-qmail-lspawn
>      |        |                        `-qmail-rspawn
>      |        `-supervise---multilog
>      |-syslog-ng
> ...
>
> If your system doesn't look like this, something is wrong. Could you
> have possibly manually started qmail-send? If so, you need to kill it.
> It may very likely also be running with different default delivery
> instructions if started manually.