| |
| |
| |
|
Comments:
<0> is there another forward application besides the standard forward one ?
<1> alias, virtusertable
<2> hey guys just a quicky.. wonder if anyone can explain the syntax of the maillog file to me. ie what do the different sections (retr, time, etc) of this mean: Feb 20 18:24:30 server1 pop3d: 1140423870.154837 LOGOUT, user=paul, ip=[60.240.44.44], top=0, retr=0, time=1, rcvd=18, sent=792, maildir=/var/qmail/mailnames/website.com/paul/Maildir
<3> consult your pop3 server's documentation.
<4> I've yet to set a mail server up for use on my lan, and am now on several blacklitsts.
<4> So, I need to find out how to block access to all but a few mail servers so that I can get off the blacklists.
<4> Can someone help me with this?
<4> I'm thinking that a few iptables rules will do the trick.
<4> But, I need to set up a mail server for our LAN. I have a Fedora Core4 machine that has Sendmail installed and can move it to another alias IP for use, and then direct all mail thorugh it, and this would solve our problem. But... need some pointers on how to do this. Would anyone be willing to point me in the right direction for getting started with this project.
<4> ?
<4> Any help on this issue would be greatly appreciated.
<3> which blacklists?
<3> normally speaking sendmail won't allow any relaying, provided it wasn't misconfigured.
<4> how can I get sendmail to start using a new alias address, ifcfg-eth0:1 ?
<5> hi all
<5> whats the best log reporting tool for sendmail ?
<5> can anyone suggest a good log reporting software
<6> hi all
<6> sendmail uses only mbox format for messages, isn t it ?
<1> only mail.local, often other LDAs are used
<1> which can stored differently
<7> Hey, wanted to bandy some discussion about. Hopefully someone is in a talkative mood. :-)
<7> We use LDAP for primarily our virtuser but all maps are actually in LDAP. This centralizes management and works very well.
<7> The one drawback is the w cl***. Sendmail only loads new w cl*** entries when restarted or given signal HUP.
<7> Doing a 'killall -HUP sendmail' reloads it, but also kills all inbound and outbound sessions. We do that every 2 hours.
<7> The big problem is when a new domain is added to the system. We consider 2 hours to be excessive to have to make the users wait.
<7> I was looking to see if there was some kind of signal I could send to sendmail that would be kind of like the graceful signal to apache (which is USR1 I think).
<8> that is a good idea, but I don't think it is supported
<7> It would tell sendmail to let all children live until their current session finishes. The root daemon would reload the w cl*** and any new children spawned would have that newer w cl***.
<7> IYHO, do you think it's impractical? Or is it just that nobody has sat down and coded it?
<8> it would likely entail locking, or something like RCU for cl*** pointers
<8> It makes good sense, to me
<7> I'm going to put on my thinking cap and ask what is the obvious question (at least to me).
<7> My method has always been 'killall -HUP sendmail'. Could I effectively achieve the same thing I'm wanting by doing 'kill -HUP `head -n 1 /var/run/sendmail.pid`' ?
<7> That only HUPs the root daemon, but I can't tell what it's doing exactly. It doesn't kill all the children, so I'm pretty sure that they continue unchanged.
<7> But I am not sure about new children.
<8> that is what my reload script does (signal only the daemon)
<7> Ok, I'll approach it from that angle and do some testing. It might solve the problem for me outright.
<8> I know it doesn't affect children - on bad days I've had to manually do killall -HUP sendmail to nail them
<7> If there was only a way to get the w cl*** of the running daemon and/or children, I could easily test :-)
<8> any new children should automatically have the new cl***es
<7> Your commentary is very re***uring :-)
<8> interesting, it looks like it does trap USR1, now to see what it does with it
<7> According to RELEASENOTES, it dumps $j and a few other little debug items.
<8> ah, dump state - might get you the info you're looking fore
<9> "cannonball" at 216.35.188.120 pasted "kill -USR1 `head -n 1 /var/run/sendmail.pid`" (25 lines, 2.1K) at http://sial.org/pbot/15956
<7> Doesn't really give much in the way of info.
<8> yeah, just ran it here :(
<7> It probably wouldn't take too much coding to add a USR2 signal handle and make it dump the w cl***.
<8> I was excited when grep showed a test in the code for $j and $w, but it was just a sanity check :(
<8> however, you can add $w to the help text, or hello text and test by telnetting
<7> Alright, have a patch and it builds. Now to get it into gentoo's build system.
<8> patch to do what ?
<7> Cowboy: my patch doesn't work. I made it so that it accepts SIGUSR2 but I can't figure out the sequence of commands to dump the w cl***.
<7> I got the signal handlers all spread throughout the programs the way it needs to be. Taht part works.
<7> I got it printing strings to the logs, that part works too.
<7> But I can't figure out how to tell it to "print out the whole w cl***". Oh so close...
<7> Right now I am able to print out the $w macro, which is the short hostname. Just have to figure out how to make it dump the whole w cl***.
<7> When I find the code that handles echo '$=w' | sendmail -bt I hope to be able to figure out.
<10> hey guys
<10> why would my smtp connections would start timing out all of the sudden?
<10> log is at 15, im not seeing anything..
<11> exit
<11> heh
Return to
#sendmail
or
Go to some related
logs:
how to change screen resolution wiht ubuntu
ipp2p dapper
udp always firewalled fc5
#openssh
disable local transport vpopmail
Net::SSH::Perl getpwuid is unimplemented
#kde
mapbender mbstring
ubuntu switch resolution
#perl
|
|