[00:01:33] *** wdp has joined #postfix
[00:04:23] *** wdp has left #postfix
[00:08:18] *** cmdln has quit IRC
[00:09:28] *** Haris__ has joined #postfix
[00:15:17] <kithpom> what is a good way to deal with faked sender bounced emails?
[00:15:35] <robtone_> !backscatter
[00:15:36] <knoba> robtone_: "backscatter" : http://www.postfix.org/BACKSCATTER_README.html
[00:19:06] *** scout has quit IRC
[00:20:41] <kithpom> robtone: thank you.
[00:20:55] <kithpom> robtone_:  ^
[00:21:22] *** alex12 has quit IRC
[00:24:03] *** stickystyle has quit IRC
[00:24:11] *** Brownoxford has quit IRC
[00:24:23] *** scout has joined #Postfix
[00:24:52] *** stickystyle has joined #postfix
[00:25:08] *** stickystyle has quit IRC
[00:25:59] *** flIr_ has joined #postfix
[00:31:07] *** flIr_ has left #postfix
[00:32:13] *** Draecos has left #postfix
[00:38:41] *** action09 has quit IRC
[00:39:13] *** Haris has quit IRC
[00:45:29] *** jpalmer has joined #postfix
[00:48:41] *** diogo_79 has quit IRC
[00:59:32] *** stickystyle has joined #postfix
[01:00:49] *** war9407 has quit IRC
[01:06:21] *** JoseUK has joined #postfix
[01:06:54] <JoseUK> any chance of some clavam help please? Can't locate object method "check_clamav" via package "Mail::SpamAssassin::PerMsgStatus" at /usr/lib/perl5/vendor_perl/5.8.8/Mail/SpamAssassin/PerMsgStatus.pm line 2647
[01:07:08] <JoseUK> thats coming from spamd
[01:08:21] <shasta> your spamassassin installation seems to be broken; note that it's not spamassassin support channel
[01:08:29] * shasta points at the channel name
[01:09:16] <JoseUK> its more active here :P taking my chances while im awake
[01:10:39] <shasta> we so don't care :>
[01:22:47] *** [diablo] has quit IRC
[01:27:39] <JoseUK> <g> thanks :P
[01:29:35] *** suuuper has quit IRC
[01:51:05] *** madrescher has quit IRC
[01:59:50] *** Motoko-chan has quit IRC
[02:20:01] *** ftp3 has left #postfix
[02:20:42] *** FWP^^^ has joined #Postfix
[02:21:07] *** fwp has quit IRC
[02:22:46] *** mwalling has quit IRC
[02:32:14] *** mwalling has joined #postfix
[02:37:32] *** Zblakany has quit IRC
[02:42:13] *** kithpom has quit IRC
[02:46:11] *** JoaoCarneiro has joined #postfix
[02:46:50] <mwalling> shasta: fun spoiler
[02:47:23] <shasta> better? ;)
[02:48:17] <mwalling> eh
[02:48:35] <mwalling> thats an eh performance
[02:54:06] *** Tjikkun has quit IRC
[02:59:23] *** makerc has joined #postfix
[03:04:52] *** Motoko-chan has joined #postfix
[03:10:07] *** ming_zym has joined #postfix
[03:10:17] *** Tachy has joined #postfix
[03:11:17] *** pitakill has joined #postfix
[03:20:14] <adaptr> is that equivalent to a so-so waggle fingers performance ? or closer to his "we don't care" ?
[03:20:44] <adaptr> I know the expression, just curious as to your views
[03:20:50] <adaptr> it won't happen again, I promise
[03:21:04] *** tshine has quit IRC
[03:23:24] <mwalling> 20:41 < Invert314> now the problem i am facing is my PC won't start
[03:23:24] <mwalling> 20:41 < Invert314> i upgraded the PSU and it still won't start
[03:23:24] <mwalling> 20:41 < mwalling> plug it in.
[03:23:33] <mwalling> 20:46 < shasta> suspecting overheated cpu
[03:23:38] <mwalling> shasta ruined it
[03:23:48] *** Tachy_ has quit IRC
[03:23:55] <mwalling> he tried to redeem himself with: 20:47 < shasta> or your pc is just offended by running win32 for a few days straight. :-P
[03:29:16] *** tshine has joined #postfix
[03:33:37] <adaptr> too little, too late
[03:33:51] <adaptr> He should at least have mentioned Vista and Sunspots
[03:34:23] <mwalling> shasta: yeah, you hear that?
[03:41:09] *** githogori has quit IRC
[03:41:13] *** makerc has quit IRC
[03:51:50] <mwalling> !cheatsheet
[03:51:52] <knoba> mwalling: "cheatsheet" : http://jimsun.linxnet.com/misc/postfix-anti-UCE.txt : A HOWTO for pre-DATA spam control.
[03:57:39] *** Southron has left #Postfix
[03:59:34] *** aarcane has joined #postfix
[03:59:44] <aarcane> hello, I have a few questions about postfix.
[03:59:52] <aarcane> can postfix store delivered mail in a database ?
[04:01:17] <rob0> no, but Postfix can deliver to a 3rd-party message store called dbmail, which does this.
[04:01:25] <adaptr> dbmail is nice
[04:01:36] <adaptr> has built-in imap, too
[04:02:11] <aarcane> hrrm
[04:03:28] <aarcane> also, can I configure a small postfix cluster to provide mail services for a single domain distributed across several hosts with different users mail delivered to different hosts ?
[04:03:45] <adaptr> again, postfix does not STORE mail
[04:03:55] <adaptr> so you would not need any type of cluster for that
[04:04:09] <aarcane> adaptr, no, but it delivers it.  it has to know where to deliver it.
[04:04:14] <adaptr> yes
[04:04:19] <adaptr> that's what maps and transports are for
[04:04:44] <adaptr> so if you run 3 websites, you need 3 servers ?
[04:04:46] <adaptr> silly
[04:05:51] *** nfi|ermes has quit IRC
[04:05:56] <aarcane> is there some sort of guide to configuring that sort of cluster ?  where it would be thousands of users at a single domain get their mail delivered to, say..  1 of 2 machines based on their username or some other chriteria ?
[04:06:37] <adaptr> as I already told you, there is no reason to use a cluster-anything to deliver mali to different mailstores
[04:07:05] <aarcane> adaptr, well then what would you call it when you have your userbase for a single domain spread across several physical machines ?
[04:07:08] <adaptr> and "cluster" is a very vague term, with little meaning for postfix
[04:07:22] <aarcane> a cluster is any group of computers working together on a common goal
[04:07:33] <adaptr> as postfix cannot be made aware of any other cluster nodes, it is not, strictly speaking, capable of it
[04:07:49] <adaptr> well, that's simply too broad, and not a working definition ANYWHERE
[04:08:12] <aarcane> anyway, lets simplify the question with an example.
[04:08:12] <adaptr> a cluster is a group of systems that is each AWARE that it belongs to that cluster
[04:08:29] *** xpoint has quit IRC
[04:08:32] <adaptr> failover capability and load sharing are the usual examples
[04:11:02] <aarcane> assume I have 1000 users, all named with a number user000 through user999 where the numbers are sequential, a domain example.com, and user mail addresses are username at example dot com  if I want users user000 through user499 delivered to mx1.example.com, and users user500 to user999 to be delivered to mx2.example.com, what's the best way to accomplish this sort of goal ?
[04:11:41] <adaptr> again, mail is not delivered TO an MX - it is deliverd BY an MX
[04:11:58] *** pirho has quit IRC
[04:12:02] <adaptr> if the domain has multiple MXen, you have little influence in which one is chosen for any particular user
[04:12:22] <adaptr> other than that it'll normally be the one with the highest priority, unless that goes down
[04:12:53] <aarcane> adaptr, but if mail for user700 is sent to mx1, I'd like it to be transported to mx2 for final delivery, and vice versa.
[04:13:08] <adaptr> if you want multiple postfix instances to receive the same network mail, you need IP load balancing, not a cluster
[04:13:33] <adaptr> aarcane: then those are mailstores, and not Mail eXchangers
[04:13:44] <adaptr> put one postfix in front and use that to route
[04:14:16] <adaptr> you can DO it, but you'd have to keep full lists on both near-identical postfix instances, for very little advantage
[04:14:27] <adaptr> VERY little advantage
[04:14:33] <rob0> And vice-versa?
[04:14:49] <adaptr> what does he have to do with it ? :D
[04:15:00] <rob0> That's what I'm wondering.
[04:17:14] <aarcane> adaptr, the point is to reduce the load on each machine by dividing a userbase amongst multiple machines which recieve and store mail for later retrieval via pop or imap.
[04:17:39] <rob0> !virtual
[04:17:41] <knoba> rob0: "virtual" : a way to configure additional domains and user accounts (that do not need to exist in your /etc/passwd). See: http://www.postfix.org/VIRTUAL_README.html
[04:17:45] <mjoseph> aarcane: it's not impossible, but it doesn't reduce the load
[04:17:56] <mjoseph> the problem is, let's say that you get 50% to each mx
[04:18:06] <mjoseph> of that
[04:18:11] <mjoseph> 50% will be mailstore-misses
[04:18:18] <mjoseph> and will have to be forwarded to the other mx
[04:18:20] <rob0> the virtual_mailbox_maps distribute half to one mailstore, half to the other
[04:18:28] <rob0> trivial
[04:18:41] * adaptr trivializes rob0
[04:18:45] <mjoseph> therefore, the other mx (both mx's) now handle 25%
[04:18:57] <mjoseph> so effectively, each mx is still handling, at minimum, 75% of your load
[04:19:05] <mjoseph> and that implies perfect distribution, which is unlikely
[04:19:14] <adaptr> which is 50% more than with a single postfix instance
[04:19:29] <adaptr> because there are no misses
[04:19:37] <mjoseph> well, not exactly
[04:19:42] <mjoseph> i mean
[04:19:42] <adaptr> ssshh
[04:19:44] <aarcane> the thought I had initially was a single "master" mxer which is configured to recieve ALL inbound mail, and forward it to an internal mx/mail store without ever touching disk unless the final destination is "unavailable" temporarily.
[04:19:58] <adaptr> mail always touches disk
[04:20:00] <adaptr> always
[04:20:02] <mjoseph> in the front-end model, yes, the front end handles 100%, and the backends handle 50% each (or 1/n)
[04:20:28] <mjoseph> but the front end has less work to do, since it doesn't do local delivery (which costs about 2x relay)
[04:20:36] <adaptr> mjoseph: we're talking about postfix; since you have multpile mailstores, you can assume they will always be able to handle the load from one frontend
[04:21:04] <mjoseph> not necessarily
[04:21:11] <mjoseph> local-delivery is actually somewhat expensive
[04:21:31] <adaptr> mjoseph: no, no, and no - even using smtp you would not run postfix again on a backend - and I would not, I would run dbmail or somesuch for direct injection
[04:21:39] <aarcane> adaptr, with 4+ GB of RAM being commonplace and inexpensive, any disk IO can be eliminated by using tmpfs mounts if they can't be eliminated by application configuration.
[04:21:47] <adaptr> so postfix *delivers* - to a backend
[04:21:56] <adaptr> aarcane:  that's not the point
[04:22:21] <mjoseph> aarcane: bad idea, you lose mail
[04:23:30] <mjoseph> adaptr: i'm not that convinced that there is universal consensus to use some other MDA on the backends
[04:23:47] <aarcane> mjoseph, with proper battery backing, and configuring it properly to merge RAM to DISK using, say, unionfs when RAM is near full or power is lost, those potential hazards can be reduced to the same likelihood as losing data in the event of a power failure with an unclean journal in any journaling filesystem.
[04:24:16] <mjoseph> aarcane: and if your kernel panics?
[04:27:58] <aarcane> mjoseph, highly unlikely.  but it has the same chance of leaving a disk unclean as if the power fails.
[04:28:33] <mjoseph> not if said disk is in memory, but clearly you seem convinced of the wisdom of this...so i'll leave it at that :)
[04:29:23] <aarcane> mjoseph, I just suggest it as an alternative to enhance performance of a dedicated mail relay which is not intended for final delivery of anything.
[04:31:38] <aarcane> what about using aoe of iscsi to create several different mail stores, each of which is on a completely different disk storage system ?
[04:32:41] <aarcane> s/aoe of iscsi. aoe or iscsi/
[04:32:45] <aarcane> erm..
[04:32:49] <aarcane> s/aoe of iscsi/aoe or iscsi/
[04:33:27] <mjoseph> you seem to be tossing technologies out there, but what fundamental problem are you really trying to solve?
[04:34:47] *** EpochWolf has joined #postfix
[04:35:58] <aarcane> mjoseph, the fundamental problem is disk conjestion, and in some cases CPU/RAM usage caused by a high volume of mail to a large number of users
[04:36:15] <aarcane> mjoseph, the CPU and RAM usage can have money thrown at them, but disks don't get signficantly faster.
[04:36:44] <Motoko-chan> What FS are the disks running, and what options?
[04:36:46] <Motoko-chan> Also, what OS?
[04:37:17] <EpochWolf> I don't know anything about mail servers and I would like to learn :) I'm just looking for some good tutorials because I'm buried in google search results.
[04:37:37] <Motoko-chan> EpochWolf, the Postfix website has some good info.
[04:37:48] <aarcane> Motoko-chan, the filesystem will only help so much, the operating system is requisitely linux.
[04:37:50] *** ming_zym has quit IRC
[04:38:02] <Motoko-chan> aarcane, what FS and what options on the mount?
[04:38:12] <EpochWolf> Motoko-chan, thanks, I'll start there.
[04:38:38] <mjoseph> aarcane: and you are using a database to store the mail?
[04:39:24] *** AcTiVaTe has joined #postfix
[04:39:39] <aarcane> Motoko-chan, lets posit that the disks are raid 5 with 5 SSD disks, and an aditional raid1  array of 15K SAS disks for an external journal.
[04:39:55] <Motoko-chan> Filesystem and mount options?
[04:40:27] <aarcane> xfs, and since I don't have the man pages for xfs handy, lets assume optimal configuration.
[04:40:38] <Motoko-chan> Optimal?
[04:40:48] <Motoko-chan> So, you have noatime and all that jazz?
[04:41:01] <aarcane> furthermore, lets assume that the server is constantly at 90+% disk IO performance.
[04:41:04] <Motoko-chan> Because that one mount option can do some nice speed boost on the read side.
[04:41:12] <aarcane> Motoko-chan, lets assume that, yes.
[04:41:26] <aarcane> remember, I've not got this theoretical configuration yet.
[04:41:28] <Motoko-chan> Hmmm
[04:41:42] <Motoko-chan> And the spool is on a non-journaled FS on different disks?
[04:41:53] <Motoko-chan> And no RAID.
[04:41:56] <aarcane> no, this is the spool
[04:42:11] <Motoko-chan> Spool will speed up if you aren't journaling it.
[04:42:41] <Motoko-chan> And aren't doing a heavy-IO RAID config.
[04:42:43] <aarcane> the operating system and all applications would be booted with the toram boot option which loads the entire operating system into RAM.
[04:43:17] <aarcane> Motoko-chan, journal is external, and the raid is raid 5 with 5 disks to spread the IO evenly.
[04:43:45] <mjoseph> raid1 is faster
[04:43:47] <mjoseph> err
[04:43:47] <mjoseph> sorry
[04:43:49] <mjoseph> raid0
[04:44:28] <aarcane> mjoseph, raid0 provides no failure protection, and raid1 provides only read speed increases.
[04:45:02] <Motoko-chan> If this is just a mail spool, it doesn't need insane protection unless you are holding spooled mail for hours waiting for delivery to local boxes.
[04:45:16] <aarcane> mjoseph, raid5 provides the best balance of performance and failure protection in smaller arrays.
[04:46:10] <aarcane> Motoko-chan, this is the theoritcal final store.  I posited a relay spool backed entirely in RAM, but someone refuted the idea on the grounds that a kernel panic would destroy the temporary spool.
[04:46:41] <Motoko-chan> Then it isn't a spool, it's a mailstore. It is the user's boxes, right?
[04:47:28] <aarcane> Motoko-chan, it's the machine on which postfix is running, and is accessed via webmail, imap, pop3, etc.
[04:48:46] <mjoseph> then it is a mailstore
[04:48:53] <mjoseph> i thought you were talking about a spool
[04:49:32] <mjoseph> but eitehr way, i still suggest that even a single disk with no raid at all is more resilient than a tmpfs filesystem w/ battery backup and some clever powerfail scripts
[04:49:44] <aarcane> mjoseph, I've suggested a single relay with a ram backed spool relaying to several mxen each of which is a mail store.  I've also suggested using aoe or iscsi to provide multiple mail stores over the network.
[04:50:05] <mjoseph> aarcane: yeah, i get the idea, it's not a bad one except for the ram-backed spool
[04:50:52] <aarcane> mjoseph, the point of the ram-backed spool is that this machine is expected to relay every single incoming mail in an environment where the diskio is the biggest problem.
[04:51:29] <mjoseph> okay, here's how you solve that problem: shard the frontends, and teh backends, until you are no-longer capacity constrained
[04:51:32] <mjoseph> you use load-balanced frontends
[04:51:40] <mjoseph> and multiple backends
[04:51:40] *** grigora has joined #postfix
[04:51:57] <mjoseph> you just increase the sharding at whichever layer is experiencing congestion
[04:52:17] <grigora> Hi, I am wondering if it's possible to set up Postfix to relay off of another server that requires SSL authentication? Thanks
[04:52:34] <mjoseph> making huge sacrifices to increase the capacity of a single node isn't usually a good idea (since eventually, that node will still hit a limit)...it just doesn't scale
[04:52:54] <mjoseph> however, i really, really question how much mail you are handling that you will need more than one front end host
[04:52:57] <aarcane> mjoseph, you mean like multiple mx entries with the same value and a dns server capable of proper round-robin balancing ?
[04:53:18] <mjoseph> well, one _or_ the other :)
[04:53:36] <mjoseph> and if you really need better balancing, an ip loadbalancer as someone else suggested
[04:55:08] <Motoko-chan> grigora, yes.
[04:55:16] <aarcane> so several frontends can still route the incoming mail to the proper mail store based on the username and some arbitrary chriteria ?
[04:55:27] <mjoseph> sure
[04:55:43] <mjoseph> personally, i'd use some sort of LDAP database or the like to index the users to backends
[04:55:51] *** Zeit|awy_ has joined #postfix
[04:55:59] <grigora> Motoko-chan: thanks, any pointers to web resources describing how to actually do that?
[04:56:24] <mjoseph> (you take a speed hit on the ldap lookup, but that too can be sharded, or you can dump the user table to a hash periodically)
[04:56:38] <mjoseph> you can also do it deterministically based on the username as you propose
[04:56:41] <grigora> Motoko-chan: most of what I come across described how to set up Postfix to accept authentication through TLS/SSL
[04:56:43] <mjoseph> though, honestly, that seems a bit limited
[04:56:48] <aarcane> mjoseph, as long as it's possible to say that usera goes to arbitrary server a, and user b goes to arbitrary server q for final delivery, even if the users are in the same single domain.
[04:57:04] <mjoseph> aarcane: yes
[04:57:49] <mjoseph> i'm pretty sure that rob0 and adaptr actually described this method a while ago :)
[04:57:55] <aarcane> mjoseph, the leading principal behind this is that there is too much mail coming in to handle final delivery on a single system due to disk speed constraints.
[04:58:03] <mjoseph> yes
[04:58:14] <aarcane> mjoseph, it's been danced around for the past half hour or so.
[04:58:14] <mjoseph> i really, really, really understand that is your _theoretical_ concern
[04:58:37] <mjoseph> there are lots of ways around that, and sharding is always the most scalable
[04:58:56] <aarcane> mjoseph, I contend that sharding sounds exactly like clustering by any other name.
[04:59:18] * mjoseph sighs
[04:59:54] <mjoseph> as adaptr pointed out, clustering does not have a very clear definition
[04:59:59] *** EpochWolf has quit IRC
[05:01:07] <mjoseph> i don't doubt that you could (and many would) call a group of mail servers handling a single domain jointly a "cluster", but just saying a mail cluster doesn't imply anything about how that group of machiens is setup or operates
[05:01:25] <mjoseph> so you can call it a cluster if you wish, and you'd not be wrong outright, but it doesn't help us to understand what you are looking for
[05:01:34] <mjoseph> and i think that was adaptr's point (it certainly is mine)
[05:01:57] *** Zeit|awy has quit IRC
[05:02:10] <aarcane> mjoseph, it makes sense.  I think some diagrams would be the best way to explain it, but my ascii art skills are severely limited.
[05:02:36] * mjoseph sighs
[05:02:40] <mjoseph> okay, but lacking the diagrams
[05:02:48] <mjoseph> you get what all three of us have basically been suggesting?
[05:03:16] <aarcane> the concepts are fairly clear, I'm jumbling them around in my muddled brain now.
[05:03:18] <mjoseph> one or more front end machines which receive mail, lookup a user->mailhost somehow, and then forward it on to the proper mailhost
[05:03:27] <mjoseph> and btw, start with one frontend
[05:03:33] <JoaoCarneiro> adaptr, remember me? i'm back since yesterday. i have been reading an excellent doc on postfix http://workaround.org/articles/ispmail-etch/
[05:03:33] <mjoseph> and see if you actually _need_ more than one
[05:03:35] <JoaoCarneiro> pretty cool
[05:03:49] <JoaoCarneiro> just want to thank you for your patience yesterday
[05:03:57] <aarcane> mjoseph, obviously.
[05:04:05] <mjoseph> (or potentialy two, for redundancy, but probably not more than that)
[05:04:13] <aarcane> mjoseph, when a single postfix instance functions as both mta and mda, does it write mail to disk twice ?
[05:04:16] <JoaoCarneiro> anyway, still don't know what i want, but i'll get there :)
[05:04:24] <mjoseph> aarcane: yes
[05:04:40] <aarcane> mjoseph, so only spooling it for delivery would cut the disk writes in half then, correct ?
[05:04:45] <mjoseph> yes
[05:04:59] <mjoseph> 22:20 <mjoseph> but the front end has less work to do, since it doesn't do
[05:04:59] <mjoseph>                 local delivery (which costs about 2x relay)
[05:05:15] <mjoseph> that was 45 minutes ago, btw
[05:05:29] <aarcane> mjoseph, but each back end machine would have to write 1x to spool and 1x to final delivery ?
[05:05:37] <mjoseph> yes
[05:05:39] <mjoseph> but you can easily shard that
[05:05:44] <aarcane> but only for a fraction of users.
[05:05:49] <mjoseph> right
[05:06:10] <aarcane> and it can scale almost indefinately.
[05:06:10] <mjoseph> adding additional layers isn't uncommon either
[05:06:22] <mjoseph> like one for fitlering/virus detection, etc
[05:06:30] <mjoseph> yes, it can scale almost infinitely
[05:06:44] <mjoseph> where is that article floating around, about scaling mail systems
[05:06:46] <mjoseph> ?
[05:06:49] * mjoseph looks
[05:07:19] <aarcane> mjoseph, is there some way that postfix can be configured to skip the spool and cut directly to delivery ?
[05:08:01] <mjoseph> yes
[05:08:06] <mjoseph> but it's (1) a bad idea
[05:08:17] <mjoseph> and (2) i don't know of anyway to do a lookup first
[05:08:28] <mjoseph> you can use the before-queue content filter mechanism to just have postfix proxy blindly
[05:08:43] <mjoseph> but i don't know of anyway to use that to balance
[05:08:45] <grigora> why am i getting the following error? - (Host or domain name not found. Name service error for name=hotmail.com type=MX: Host not found, try again)
[05:09:04] *** githogori has joined #postfix
[05:09:32] <grigora> dig hotmail.com MX works fine ... same issue with gmail.com
[05:11:11] <mjoseph> aarcane: http://mysqldump.azundris.com/archives/69-Changing-everything.html
[05:11:30] * mjoseph goes AFK for a while
[05:13:34] <rob0> grigora, chroot
[05:18:23] <grigora> rob0: I don't think I have chroot set up ... where would it be? It's been years since I set up one up
[05:21:28] <rob0> !debug
[05:21:28] <knoba> rob0: "debug" : http://www.postfix.org/DEBUG_README.html : a good starting point for how to deal with problems and to report information to those who might help. Post your information in a pastebin such as http://pastebin.ca/ or http://rafb.net/paste/ .
[05:21:37] <rob0> #no_chroot anchor in that
[05:23:58] <grigora> rob0: I can't find another resolv.conf file on my system, I am fairly sure it's not chroot
[05:24:39] <grigora> rob0: spoke too soon, that's what it is, thank you
[05:30:10] *** xnixan has quit IRC
[05:30:41] *** xnixan has joined #postfix
[05:33:51] *** grigora has quit IRC
[05:41:10] *** saurabhb has joined #postfix
[05:42:03] *** stickystyle has quit IRC
[05:54:11] *** Haris_ has quit IRC
[05:56:19] *** c00l2sv has quit IRC
[06:06:15] *** hparker has quit IRC
[06:07:55] *** bhagat has joined #postfix
[06:08:11] *** AcTiVaTe has quit IRC
[06:22:26] *** kk_CHN has joined #postfix
[06:29:47] *** JoaoCarneiro has quit IRC
[06:32:19] *** ming_zym has joined #postfix
[06:45:11] *** pitakill has quit IRC
[06:52:48] *** loddafnir has joined #postfix
[07:05:27] *** fabounio has joined #postfix
[07:06:13] *** Zblakany has joined #postfix
[07:20:55] *** kk_CHN has quit IRC
[07:22:25] *** bhagat has quit IRC
[07:27:21] *** seekwill has joined #postfix
[07:34:43] *** Fallenou has joined #postfix
[07:45:05] *** aarcane has quit IRC
[07:57:02] *** seekwill has quit IRC
[07:59:13] *** fabounio has quit IRC
[08:05:24] *** fabounio has joined #postfix
[08:09:37] *** fabounio has quit IRC
[08:19:42] *** SniZ has quit IRC
[08:20:16] *** Gokee2_Work has joined #postfix
[08:22:31] *** Gokee2_Work is now known as Gokee2_Main
[08:26:47] <Gokee2_Main> Hello all,  I am trying to get mailman working.  I have a setup (mostly) following http://workaround.org/articles/ispmail-etch/#mailing-lists-with-mailman (very good article!).  But I am trying to get mailman working.  Last time I did mailman was with only one domain(well no real ones at all).  But I was able to get mailman to put out a file that postfix picked up on so I could add new mailing lists right from mailman and have them work.
[08:26:47] <Gokee2_Main>   Is there any way to do this with multiple domain names in a virtual setup?  Thanks
[08:31:08] *** carl- has joined #postfix
[08:32:17] *** crab has joined #postfix
[08:32:52] *** stony_ has joined #postfix
[08:33:16] <crab> hi. is it possible to say "allow clients that authenticated with the username $x to use only 'mail from:<$x>'"?
[08:34:24] <Gokee2_Main> crab, Hmm I have that problem too, I have not looked into fixing it yet.
[08:37:36] <crab> reject_authenticated_sender_login_mismatch, maybe.
[08:38:38] <crab> yes. smtpd_sender_login_maps etc.
[08:38:41] <crab> cool.
[08:40:39] <Gokee2_Main> Hmm I will have to play around with that :)
[08:44:06] <f3ew> crab yes
[08:45:04] *** stony__ has quit IRC
[08:46:33] *** phnord has joined #postfix
[08:47:26] *** CrazyFoam has joined #postfix
[08:48:31] <crab> crab yes?
[08:49:22] *** Filbert has joined #postfix
[08:50:40] <f3ew> In answer to your question
[08:50:47] <f3ew> but I see you found TFM
[08:55:58] *** amrit|wrk is now known as amrit|zzz
[09:01:06] *** glad_work has joined #postfix
[09:06:25] *** syneus has joined #postfix
[09:06:50] *** Tjikkun has joined #postfix
[09:14:05] *** Motoko-chan has quit IRC
[09:15:07] *** m_p has joined #postfix
[09:30:21] *** harobed has joined #postfix
[09:37:33] *** madrescher has joined #postfix
[09:42:46] *** denis has joined #postfix
[09:57:02] *** war9407 has joined #postfix
[09:57:25] *** carl__ has joined #postfix
[09:58:44] *** tsauter has joined #postfix
[10:02:13] <tsauter> hi all
[10:03:28] <tsauter> I am trying to forward all outgoing emails over two relay servers. I have now added an entry to my transport table "* smtp:myrelay.test.com" and everything works find. But can I specified more then one server here? To make the connection more robust in the case of one of the relay servers is down?
[10:13:38] *** suuuper has joined #postfix
[10:14:04] *** carl- has quit IRC
[10:19:03] *** tm-30740-exa has quit IRC
[10:21:25] *** carl__ has quit IRC
[10:32:33] *** ircmojo has quit IRC
[10:32:46] *** [diablo] has joined #postfix
[10:44:08] *** af_ has joined #postfix
[10:46:05] *** crab has left #postfix
[10:48:32] *** dragonheart has quit IRC
[10:49:09] *** dragonheart has joined #postfix
[10:52:35] *** ircmojo has joined #postfix
[10:56:08] *** visf_ has joined #postfix
[11:03:31] *** Daviey has quit IRC
[11:04:15] *** madrescher has quit IRC
[11:05:04] *** zoldar has joined #postfix
[11:08:29] <zoldar> hello, is there any way to check when was the last connection to given mailbox through pop3/imap (courier)? other than parsing logs?
[11:10:54] <sysmonk> only parsing the logs...
[11:11:05] <sysmonk> atleast as far as i know :)
[11:11:11] <sysmonk> but i don't use courier
[11:11:15] <sysmonk> and this is not #courier, it's #postfix
[11:12:38] <zoldar> i know, sorry, but courier is almost empty most of the time
[11:17:19] *** robboplus_ has quit IRC
[11:21:40] *** Draecos has joined #postfix
[11:25:42] *** muecke77 has joined #postfix
[11:27:37] *** havvg has joined #postfix
[11:31:04] *** muecke77 has quit IRC
[11:38:15] *** Haris_______ has joined #postfix
[11:41:35] *** CrazyFoam has quit IRC
[11:48:06] *** xpoint has joined #postfix
[11:50:24] *** lodder has joined #postfix
[11:50:40] *** Sypher has joined #postfix
[11:50:48] *** af_ has quit IRC
[11:51:31] <lodder> Hi, is there a way if you have a 550 the user in is not found in the virtual table to relay to an other server ?
[11:52:42] <lodder> if possible only for one domain?
[11:59:02] *** ming_zym has quit IRC
[12:02:34] *** tm-30740-exa has joined #postfix
[12:06:31] *** Haris__ has quit IRC
[12:12:42] <lodder> Hi, is there a way if you have a 550 the user in is not found in the virtual table to relay to an other server ?
[12:12:56] *** wdp has joined #postfix
[12:16:11] *** CrummyGummy has joined #postfix
[12:18:34] <CrummyGummy> Hi all, I'm writing a script that receives input from postfix, processes it and sends an optional reply. Would a content filter work for this or would I need to use pipe on its own?
[12:20:12] *** cpm has joined #postfix
[12:21:22] *** visf_ has quit IRC
[12:30:09] *** Haris1 has joined #postfix
[12:42:17] *** madrescher has joined #postfix
[12:43:24] <lodder> Hi, is there a way if you have a 550 the user in is not found in the virtual table to relay to an other server ?
[12:57:44] *** UQlev has joined #postfix
[12:59:41] *** Lap_64 has joined #postfix
[13:01:06] <f3ew> lodder relocated(5)
[13:10:11] *** stickystyle has joined #postfix
[13:19:47] *** UQlev has quit IRC
[13:26:52] <cite> lodder: You might want to have a look at the luser_relay statement in main.cf.
[13:28:27] <lodder> cite: canyou than do luser_relay = *@domain ?
[13:28:34] <lodder> f3ew: what do you mean
[13:30:39] <cite> lodder: man 5 postconf | less +/^luser_relay
[13:30:58] <lodder> cite: thx I'll have a look
[13:31:12] <cite> lodder: This will only work for the local(8) delivery agent.
[13:31:35] <lodder> cite: they have to be relay to external server
[13:32:19] <cite> lodder: You can try smtp_fallback_relay (fallback_relay with Postfix < 2.3).
[13:32:43] <cite> lodder: Still, this would require tempering with recipient validation and therefore is a "very bad thing[tm]".
[13:35:21] <f3ew> http://www.postfix.org/STANDARD_CONFIGURATION_README.html#some_local
[13:36:47] *** mehulved is now known as bhurji
[13:37:15] <lodder> cite: I know thats dangerous but I need it badly
[13:43:10] <lodder> cite: I have now configured it that all mails for that domain should be forwarded but now I need to make an exception for one email addres could that be possible?
[13:43:42] *** stickystyle has joined #postfix
[13:49:25] <lodder> So is there a way the user isn't found in the virtual table it's being relayed to a external server ;)
[13:49:40] <lodder> that is now the case i'm working on
[13:50:46] *** ramoni has joined #postfix
[13:52:53] <cite> lodder: Sure. Just rewrite the address using virtual_alias_maps.
[13:56:00] <lodder> cite: oke
[14:05:55] *** havvg has quit IRC
[14:06:00] *** reiner has joined #postfix
[14:06:02] *** havvg has joined #postfix
[14:13:06] *** reiner has quit IRC
[14:15:30] *** denis has quit IRC
[14:19:56] *** Nockian has quit IRC
[14:22:24] *** Lap_64 has quit IRC
[14:36:45] *** saurabhb has quit IRC
[14:37:11] *** c00l2sv has joined #postfix
[14:54:42] *** tsauter has quit IRC
[14:56:02] *** cmdln has joined #postfix
[15:03:38] *** User855 has joined #postfix
[15:04:32] <User855> anyone here?
[15:04:56] <User855> ich habe ein postfix problem
[15:06:42] *** marku5_ has joined #postfix
[15:06:55] *** User855 has quit IRC
[15:07:29] <marku5_> Hi  ich habe ein postfix problem
[15:08:04] <marku5_> hat gerade einer zeit?
[15:08:24] <wdp> marku5_, we're talking english in here, not german.
[15:08:29] <marku5_> ok
[15:08:31] *** spybsd has joined #postfix
[15:08:39] <wdp> marku5_, anyway, as long as you don't tell us about your problem, no one will have time.
[15:10:23] <spybsd> Hi there
[15:10:55] <spybsd> i'm running an antispam gateway on Openbsd with postgrey/amavis/spamassassin
[15:10:57] <marku5_> yes   i would use a accesslist for postfix  for incomming mails with any existing user  all the other non existing should  postfix block
[15:11:09] <spybsd> and i'm looking for some performance improvement
[15:11:25] <marku5_> anyone idea?
[15:11:30] <spybsd> currently an incoming is fully scan and forward within 15s
[15:12:09] <spybsd> <marku5_> why don't know simply use access_recipients file?
[15:12:19] <spybsd> why don't you simply use, sorry
[15:12:59] *** morpheus08 has joined #postfix
[15:13:11] <morpheus08> salve
[15:13:13] <morpheus08> helo
[15:13:19] <marku5_> i try check_recipient_access but it doesnt to the block    i am a postfix beginner
[15:13:32] <morpheus08> ther are sam one for help my
[15:13:55] <marku5_> i am not shure that my conf is correct
[15:13:59] <spybsd> <marku5_> you have few choice:
[15:14:14] <morpheus08> i have a veri problem with mi mail server
[15:14:18] <spybsd> <marku5_>use virtual_domains is you forward to a backend server
[15:14:30] <spybsd> or use access_recipients file
[15:14:34] <morpheus08> the port 25 il anrecable
[15:14:42] <morpheus08> the port 25 il unrecable
[15:14:59] <spybsd> <morpheus08> did you check if smtp port is open to incoming connection
[15:15:07] <spybsd> <morpheus08>netstat -an |grep :25
[15:15:13] <morpheus08> yes is open
[15:15:45] <marku5_> the server runs many virtual domains and one main domain
[15:16:03] <marku5_> the accesslist is only for the main domain
[15:17:00] <marku5_> in main.cf i have configured the param check_recipient_acces  and in the configured file i have the users  but this does not work
[15:17:02] <morpheus08> in netstat -an |grep 25 ther are more syn req ed any acept, but the clent wrote dont are possible send mail
[15:17:39] <spybsd> <marku5_> did you run postmap your_access_recipients_file?
[15:17:44] <marku5_> yes
[15:18:08] <marku5_> the .db file is existing
[15:18:13] <spybsd> <marku5_>ok hold on a sec
[15:18:20] *** glad_work has quit IRC
[15:19:43] <morpheus08> posibile is that too many requests for blocking the exit door?
[15:20:52] <morpheus08> posibile is that blocking the door too many requests for exit?
[15:21:57] <morpheus08> in mail.inf ther are mani message the thi type: ...from<> ... to
[15:22:07] <spybsd> <marku5_>http://www.postfix.org/ADDRESS_VERIFICATION_README.html
[15:22:18] <morpheus08> :p01.nome.doman.it
[15:22:39] <morpheus08> wher p01.nome.somain.it dont exist
[15:23:26] <spybsd> http://www.postfix.org/postconf.5.html -> check_recipient_access Syntax ;)
[15:23:55] <spybsd> <morpheus08>are you able to telnet the server locally?
[15:24:03] <spybsd> <morpheus08>telnet 127.0.0.1 25
[15:26:14] <spybsd> Anyone for tunning performance?
[15:27:23] <morpheus08> yes telnet localhost is ok bat don't work telnet ipremoto
[15:27:58] <marku5_> <spybsd> thanx for the  sites     they are known   i have exactly do that but all other mailaccounts are also accepted
[15:30:59] <spybsd> <marku5_>did you try to use a wild card to reject every addresses but the correct ones?
[15:31:16] <spybsd> <marku5_>for example * REJECT
[15:31:31] <marku5_> yes
[15:31:38] <spybsd> i'm not using this feature since i'm working only with virtual_domains
[15:32:54] <marku5_> ok we have users a at doain dot tld and b at doain dot tld      in the access file is following:     a@     OK     and    @domain.tld   REJECT
[15:33:19] <marku5_> but b at domain dot tld is accepted
[15:33:35] <spybsd> :s
[15:33:50] <marku5_> me too
[15:33:52] <spybsd> i'm looking to my conf at the moment
[15:33:57] <spybsd> hope i can find something revelant
[15:34:48] *** gladier has joined #postfix
[15:35:44] <marku5_> maybe you have another idea for the prob
[15:36:01] *** slashtom has left #postfix
[15:36:03] <spybsd> Take a look to this feature may be
[15:36:18] <spybsd> smtpd_reject_unlisted_recipient (default: yes)
[15:36:18] <spybsd>     Request that the Postfix SMTP server rejects mail for unknown recipient addresses, even when no explicit reject_unlisted_recipient access restriction is specified. This prevents the Postfix queue from filling up with undeliverable MAILER-DAEMON messages.
[15:36:18] <spybsd>         * The recipient domain matches $mydestination, $inet_interfaces or $proxy_interfaces, but the recipient is not listed in $local_recipient_maps, and $local_recipient_maps is not null.
[15:36:18] <spybsd>         * The recipient domain matches $virtual_alias_domains but the recipient is not listed in $virtual_alias_maps.
[15:36:22] <spybsd>         * The recipient domain matches $virtual_mailbox_domains but the recipient is not listed in $virtual_mailbox_maps, and $virtual_mailbox_maps is not null.
[15:36:25] <spybsd>         * The recipient domain matches $relay_domains but the recipient is not listed in $relay_recipient_maps, and $relay_recipient_maps is not null.
[15:36:30] <spybsd>     This feature is available in Postfix 2.1 and later.
[15:37:02] <morpheus08> <spybsd> whai i dont telnet remotip 25?
[15:37:14] *** scout is now known as masterkiller
[15:37:21] <spybsd> <morpheus08>you can try both
[15:37:34] <spybsd> <morpheus08>but you need to be sure that can access smtp locally first ;)
[15:38:18] <morpheus08> yes
[15:38:45] <marku5_> we have 30 mail accounts   but all the spam for *@domain is accepted and relayed to internal server     do you another reason for the prob
[15:39:05] <morpheus08> i locally have access smtp
[15:39:27] <morpheus08> is connect to localhost port 25
[15:39:29] <spybsd> <marku5_>The recipient domain matches $relay_domains but the recipient is not listed in $relay_recipient_maps, and $relay_recipient_maps is not null.
[15:39:35] <spybsd> so mail is not rejected
[15:40:12] <spybsd> so you need to list all your user un relay_recipient_maps
[15:40:20] <spybsd> to reject all non-existing email
[15:40:29] <spybsd> i think that's the way it works
[15:40:43] <spybsd> <morpheus08>that's a good point
[15:40:56] <marku5_> ok fine     many thanx  to you spybsd
[15:41:04] <spybsd> can you paste an error msg from you log
[15:41:13] <spybsd> <marku5_>you're welcom ;)
[15:45:37] <morpheus08> ok and nao
[15:46:34] <spybsd> an you paste an error msg from you log?
[15:50:42] *** keffer has quit IRC
[15:51:46] *** keffer has joined #postfix
[16:00:26] *** madrescher has quit IRC
[16:02:05] *** madrescher has joined #postfix
[16:05:23] *** wdp has quit IRC
[16:06:36] *** madrescher has quit IRC
[16:08:30] *** netcrash has joined #postfix
[16:16:39] *** madrescher has joined #postfix
[16:26:48] *** marku5_ has quit IRC
[16:29:59] *** hever has joined #postfix
[16:32:04] *** xemacs3 has joined #postfix
[16:37:11] *** madrescher has quit IRC
[16:37:31] *** xemacs has quit IRC
[16:38:06] *** nik_ has joined #postfix
[16:39:08] *** pitakill has joined #postfix
[16:40:49] *** stickystyle has quit IRC
[16:41:23] *** xemacs3 is now known as xemacs
[16:44:06] *** madrescher has joined #postfix
[16:48:20] *** Haris_______ is now known as Haris
[16:52:20] *** zoldar has quit IRC
[16:55:19] *** CrummyGummy has quit IRC
[16:55:29] *** wdp has joined #postfix
[16:59:02] *** bhurji is now known as mehulved
[17:01:06] *** wdp has left #postfix
[17:01:48] *** SniZ has joined #postfix
[17:13:46] *** af_ has joined #postfix
[17:22:39] *** Samonoske_ has joined #postfix
[17:22:42] <Samonoske_> hey guyz
[17:35:40] *** j_s has joined #postfix
[17:38:05] *** phnord has quit IRC
[17:40:33] *** nik_ has quit IRC
[17:45:46] *** tshine has quit IRC
[17:53:34] *** Jense has joined #postfix
[17:57:31] *** madrescher has quit IRC
[18:02:54] *** jelly has quit IRC
[18:05:46] *** zbrown has left #postfix
[18:06:15] *** Sypher has quit IRC
[18:06:29] *** jelly has joined #postfix
[18:12:24] *** jellis-real has joined #postfix
[18:16:53] *** madrescher has joined #postfix
[18:18:29] *** suuuper has quit IRC
[18:20:37] *** Samonoske_ is now known as Samonoske
[18:22:31] *** PhilKC has quit IRC
[18:24:43] *** madrescher1 has joined #postfix
[18:33:15] *** tshine has joined #postfix
[18:33:45] *** syneus has quit IRC
[18:36:26] *** madrescher2 has joined #postfix
[18:36:54] *** denis_ has joined #postfix
[18:38:07] *** madrescher has quit IRC
[18:43:38] *** madrescher1 has quit IRC
[18:44:03] <linkslice> what's the esiest way to block messages bound for itself?  i.e. mx2.mydomain.com has been accepting a lot of mail sent from 'somespammer at mx2 dot mydomain.com'
[18:44:11] <linkslice> is it safe to reject these with a header check?
[18:48:41] <mwalling> ...
[18:48:46] <mwalling> !mydestination
[18:48:46] <knoba> mwalling: "mydestination" : a configuration parameter in the main.cf: The list of domains that Postfix delivers via the $local_transport mail delivery transport. By default, mail is given to the Postfix local(8) delivery agent that looks up all recipients in /etc/passwd and /etc/aliases, or their equivalents.
[18:49:01] <mwalling> dont tell me thats set to $myhostname
[18:49:09] <mwalling> because i am sure you already checked that
[18:50:07] <linkslice> mwalling, actually....it is....hrrrrm
[18:50:19] <linkslice> actually it's set to 'localhost'
[18:50:21] <mwalling> postfix is doing nothing wrong then
[18:50:50] <linkslice> mwalling, never said it was, just want to know how to tell it to do something else that is correct  :-)
[18:51:20] *** harobed has quit IRC
[18:51:20] <mwalling> postfix will accept mail for domains that it thinks it is the final destination for.
[18:51:20] <rob0> I think linkslice means that he's getting spam from outside with envelope sender set to $myhostname, but more information is needed.
[18:51:22] <linkslice> should that be set to 'mydomain.com'?
[18:51:30] <linkslice> rob0, yes that's correct
[18:51:38] <mwalling> oh
[18:51:40] <mwalling> sent from
[18:51:50] * mwalling read "sent to"
[18:51:55] * mwalling walks away
[18:52:19] <linkslice> is a header check in order or do I have some config problem?
[18:52:39] * cpm hands mwalling his hat and coat
[18:52:41] <linkslice> er, do I *likely* have a config problem
[18:52:47] *** githogori has quit IRC
[18:53:13] <rob0> "More information is needed", pastebin a sample and "postconf -n".
[18:53:40] * cpm hands rob0 more information, more than he can handle
[18:54:27] <linkslice> http://pastebin.org/40360
[19:00:46] *** af_ has quit IRC
[19:04:05] *** madrescher2 has quit IRC
[19:14:21] *** zaggy has joined #postfix
[19:14:34] <zaggy> cool, a postfix channel ;)
[19:14:52] <JoseUK> :)
[19:16:09] <lunaphyte_> it's not what you think.
[19:16:33] <lunaphyte_> darn, the topic gives it away.
[19:16:58] <zaggy> heh
[19:17:11] <zaggy> ok been trying to RTFM on this one
[19:17:52] <zaggy> if I want postfix to send mail from a specific interface, I should use smtp_bind_address right?
[19:18:41] <zaggy> simply be smtp_bind_address=1.2.3.4 on main.cf?
[19:23:05] <linkslice> rob0 ever get a chance to take a peek at that poastebin?
[19:23:37] <rob0> zaggy: Offer void where taxed or prohibited by law, or where you have routing table issues with the smtp_bind_address interface.
[19:23:49] <rob0> 16:53 < rob0> "More information is needed", pastebin a sample and "postconf -n".
[19:24:30] <zaggy> rob0: heh
[19:24:38] <zaggy> rob0: I'll give it a try ;)
[19:25:56] <linkslice> rob0, ah, I parsed that as a sample of postconf -n ;)  http://pastebin.org/40366
[19:29:03] <zaggy> YAY it works.;
[19:29:06] <zaggy> thanks rob0
[19:29:20] <linkslice> rob0, that telnet session was from a system that is not listed in mynetworks or anything, it's basically a random system on the net to the mail server
[19:31:13] <rob0> There's a Big Picture that you need to understand, but if you do it's easy to block those.
[19:32:14] <rob0> after "permit_mynetworks, permit_sasl_authenticated," put in a check_sender_access lookup.
[19:32:16] *** zaggy has left #postfix
[19:33:03] <rob0> But, be aware that by doing so you will break some legitimate use of your domain as envelope sender from outside.
[19:33:58] <rob0> some Web services where your users have signed up might let them send mail from their site, using your domain address.
[19:35:00] <rob0> A small enough site can handle that with a policy, just tell your users that they can only send mail with your domain address FROM your server.
[19:42:35] *** amrit|zzz is now known as amrit|wrk
[19:49:37] *** xpoint has quit IRC
[19:51:59] *** keffer has quit IRC
[19:52:36] *** githogori has joined #postfix
[19:54:24] *** PhilKC has joined #Postfix
[19:56:08] *** keffer has joined #postfix
[19:58:33] *** robboplus_ has joined #postfix
[19:58:52] <morpheus08> ce qualcuno che parla italiano ?
[20:04:28] <vice-versa> linkslice: are the forged sender at yourdomain dot tld addresses invalid?
[20:12:59] *** muecke77 has joined #postfix
[20:16:32] *** denis_ has quit IRC
[20:17:27] *** m_p has quit IRC
[20:26:51] *** muecke77 has left #postfix
[20:29:46] <linkslice> vice-versa, yes
[20:29:59] <linkslice> they are made up names
[20:30:45] <vice-versa> !smtpd_reject_unlisted_sender
[20:30:47] <knoba> vice-versa: "smtpd_reject_unlisted_sender" : a configuration parameter in the main.cf: Request that the Postfix SMTP server always rejects mail from unknown sender addresses. This can slow down an explosion of forged mail from worms or viruses. This feature is not available prior to Postfix 2.1.
[20:31:07] <vice-versa> hmm, that's a little misleading
[20:31:11] <vice-versa> http://www.postfix.org/postconf.5.html#smtpd_reject_unlisted_sender
[20:31:53] *** devdas has joined #postfix
[20:41:44] *** Gokee2_Main has quit IRC
[20:43:34] *** Gokee2_Main has joined #postfix
[20:46:34] *** netcrash has quit IRC
[20:47:58] *** action09 has joined #postfix
[20:49:01] *** godlie has joined #postfix
[20:49:15] <godlie> hi
[20:50:36] <godlie> im just struggling with an complex configuration, is it possible to send mails to specific user over an specific relay allthough the mailserver is responsible for the domain
[20:51:19] *** UQlev has joined #postfix
[20:51:33] <godlie> this means in detail: my domain is test.mail user1 at test dot mail should be able to send an email to user2 at test dot mail but user2 at test dot mail is not an mailbox of the mailserver its an external account
[20:53:45] <mwalling> so its an alias
[20:54:03] <mwalling> otherwise user2 really isnt at test.mail, are they
[20:54:22] <godlie> mwalling thats an great idea with the alias :)
[20:54:24] <godlie> its that way
[20:54:35] <godlie> i have to pull all mails from external pop accounts
[20:54:56] <godlie> and i send all mails through external smtp
[20:55:32] <godlie> cause our internet connection is to low, so im not able to work as frontend-mta
[21:02:10] *** devdas has quit IRC
[21:11:58] *** bpgoldsb has quit IRC
[21:14:48] *** githogori has quit IRC
[21:17:38] *** githogori has joined #postfix
[21:20:56] *** [diablo] has quit IRC
[21:21:29] *** [diablo] has joined #postfix
[21:23:26] *** Motoko-chan has joined #postfix
[21:32:20] *** rootsvr has joined #postfix
[21:36:28] *** Zelest has joined #postfix
[21:36:59] *** donspaulding has joined #postfix
[21:38:29] <donspaulding> hi all, I've got two postfix mta's that relay mail for several thousand virtual domains...would postfix ever add/change the Message-Id headers?
[21:40:35] <cpm> huh?
[21:41:44] <adaptr> donspaulding: the message-id is set by the originating MUA, and not changed
[21:42:06] <cpm> yeah, that's an odd question, wonder what is behind the question
[21:42:35] * adaptr suspects a Mass Mailer instead of a MUA
[21:43:08] <donspaulding> no,  I'm actually seeing legit emails here and there being rejected by the destination because it refuses all mail without a proper message-id
[21:43:22] <donspaulding> (I understand that's how it's supposed to work)
[21:43:27] <adaptr> yes
[21:43:38] <cpm> and who is sending these legit mails without a message-id ?
[21:43:43] <adaptr> so what is in there, if not a message-id?
[21:43:50] <adaptr> is this Outhouse again ?
[21:43:53] <cpm> heh
[21:43:59] <donspaulding> I'm guessing so
[21:44:07] <cpm> starting to smell like an outhouse
[21:44:07] <donspaulding> just reading up on Outlook 2003's problems
[21:44:11] <adaptr> I know for a fact that even that produces valid SMTP
[21:44:15] <cpm> endless, they are.
[21:44:29] <donspaulding> cpm: :D
[21:44:40] <donspaulding> I'll try and ferret out if that's the one
[21:44:56] <donspaulding> 553 5.0.0 Message-Id header line format error (in reply to end of
[21:44:56] <donspaulding>     DATA command)
[21:45:21] <donspaulding> line format error?
[21:45:34] <adaptr> could be non-ASCII, or too long
[21:45:38] <donspaulding> (that's what the next-hop said)
[21:45:40] <adaptr> both are in the RFC
[21:45:40] *** yfoo has joined #postfix
[21:45:45] <rob0> is this from Postfix logs?
[21:45:53] <cpm> what mta is the next hop running?
[21:45:58] <cpm> that's where to look.
[21:46:00] <donspaulding> no, a bounce message from postfix
[21:46:07] <cpm> that's not a bounce, that's a reject
[21:46:10] <adaptr> my bets are firmly on the non-ASCII, as Outhouse is notorious for that
[21:46:21] <adaptr> donspaulding: 5xx is always REJECT
[21:46:38] *** Nockian has joined #postfix
[21:46:53] <donspaulding> yeah, yeah, sorry, not an mta nomenclature junkie
[21:47:01] <cpm> maybe some fancy header check
[21:47:09] <adaptr> cough SMTP standard cough
[21:47:22] <adaptr> if you run public MTAs, there is no excuse for not knowing the protocols
[21:47:24] <donspaulding> adaptr: alright, I get the point :-)
[21:48:20] <cpm> postmaster for thousands of domains?
[21:48:42] <adaptr> well, yes, that's what I thought
[21:48:46] <donspaulding> yeah, unfortunately.   I'm a web developer, not a postmaster :-(
[21:49:23] <adaptr> so.. let the postmaster sort it out (pun fully intended)
[21:49:33] <donspaulding> wow
[21:49:38] <adaptr> to each his own
[21:50:02] <cpm> ouch!
[21:50:11] <adaptr> yes, that would be yours, cpm :P
[21:50:24] * cpm chuckles
[21:50:26] * adaptr quickly hides the LARTazer
[21:51:57] <cpm> thing is, postfix doesn't reject mail on headers, unless asked to do so, (last I checked)
[21:53:05] <cpm> so, *someone* ( a postmaster ) has customized something with intent, best to find out what.
[21:54:58] * adaptr blames cpm
[21:55:00] *** cpm has quit IRC
[21:55:55] <donspaulding> adaptr: yeah, this isn't postfix rejecting the mail, that came from a reject message postfix sent back to the MUA, saying that the next-hop had rejected because of a bad message-id
[21:56:17] <donspaulding> I'm not sure I should tell next hop to accept all non-message-id emails
[21:56:39] <adaptr> postfix doesn't send reject messages, those are bounces
[21:57:33] <donspaulding> <donspaulding> no, a bounce message from postfix
[21:57:44] <donspaulding> <cpm> that's not a bounce, that's a reject
[21:57:53] <adaptr> if it is in a LOG, then it is a REJECT
[21:58:06] *** j2^ has joined #postfix
[21:58:08] <adaptr> if it is in a mailed bounce message, then it is STILL a REJECT, but not from your MTA
[21:58:13] <j2^> 'evening all!
[21:58:23] <adaptr> you need to make these differences clear
[21:59:20] <adaptr> for example, I have by now deduced your actual question(s) to be: 1. does my MTA or its relayhost reject messages without a proper message-id, and/or 2. how can I solve this from my MTA, which is not involved in the rejection process
[21:59:33] <adaptr> the answer to 1. is, of course, Yes.
[21:59:47] <adaptr> and the answer to 2. is, why would you ?
[22:00:04] <donspaulding> adaptr: you have correctly deduced my questions
[22:00:10] <adaptr> apart from also being No, you can't, as you most probably have no control over the rejecting MTA
[22:00:21] <adaptr> so you need to do 3.: fix your clients
[22:00:37] <donspaulding> adaptr: right, which are the general public.
[22:00:48] <adaptr> postfix is behaving as designed, and very sane
[22:00:59] <j2^> I am trying to "duplicate" a trick i did for a client with Exim. Basically, i am using a transport map to split incoming mail between a number of servers. BUT i need to make alterations to the routing depending on with interface the mail is received on. Possible?
[22:01:01] <donspaulding> which works out perfectly for me, since I know everyone
[22:01:13] <adaptr> donspaulding: as I have never encountered this issue, I rather think something entirely different is happening
[22:01:15] *** PcPixel has joined #postfix
[22:01:45] <donspaulding> adaptr: really?  your public mtas *don't* ever send these bounce messages based on next-hop rejects?
[22:01:56] <adaptr> j2^: bind different smtpd to the different interfaces, and go from there.. (you can make rules per transport if you read the manual closely)
[22:02:04] <PcPixel> I am trying to ensure that senders on the inside only use approved from addresses. and conversely, that no mail attempts to come in claiming to be from me. the outside in is working, but inside out is not. MY files are at: http://pastebin.com/m579f056b . Thoughts?
[22:02:27] <adaptr> donspaulding: my public MTAs do not encounter SMTP traffic with actual INVALID message-ids
[22:02:31] <adaptr> not even from Outhouse
[22:02:39] <donspaulding> hrm
[22:02:47] <adaptr> the rejection issue never arises
[22:03:01] <donspaulding> adaptr: do they encounter traffic with missing ids?
[22:03:06] <adaptr> as cpm already "hinted": investigate header checks
[22:03:55] <donspaulding> adaptr: I know what header_checks are/do, but I'm not familiar with how that could fix my clients.
[22:03:57] <j2^> adaptr: Could you clue me in to where in the manual i might find this? I am feeling a bit overwhelmed right now.
[22:03:57] <adaptr> donspaulding: if a client connected without adding a proper message-id (as it would , say, when you telnet and test a server), postfix will add one which is guaranteed unique across the known galaxy
[22:04:32] <donspaulding> adaptr: but is that good practice? fixing bad clients silently
[22:04:34] <donspaulding> ?
[22:05:09] <j2^> donspaulding: Actually, a RFC says that the mta SHOULD supply a MSGID if one isnt present i think?
[22:05:17] <adaptr> j2^: you'll need to do some digging, sorry - it's not directly obvious, but if you have two smtpd listeners in master.cf, say mx1 and mx2, you can set transport-specific options in some cases by prefixing the equivalent global options with the name of the transport, so do_routing_magic_trick becomes mx1_do_routing_magic_trick
[22:05:36] <adaptr> and simlar for mx2
[22:05:59] <j2^> adaptr: So, I would definitely need two processes then?
[22:06:04] <adaptr> really, dig through the intermediate/advanced manual pages and you'll encounter them
[22:06:24] <adaptr> j2^: since it is logically impossible to listen on two distinct interfaces with one socket process, yes
[22:06:53] <adaptr> and if you don't separate the interfaces then you can't know which one it came in on
[22:07:04] <adaptr> so it's pretty much the only way
[22:07:10] <j2^> adaptr: Oh, sorry! Misphrasing from my part! I meant "route based on IP-adress of sending host".
[22:07:24] <adaptr> that's easy
[22:07:35] <adaptr> !sender_dependent_transport_maps
[22:07:35] <knoba> adaptr: Error: "sender_dependent_transport_maps" is not a valid command.
[22:07:40] <adaptr> yes it is !
[22:07:45] <adaptr> something like that, anyway
[22:07:59] <adaptr> postfix has too many too long options
[22:08:13] <adaptr> (although merely saying that is sacrilege)
[22:08:43] *** pirho has joined #postfix
[22:09:23] <j2^> adaptr: Hm, a fairly recent post says "run teo processes" http://www.irbs.net/internet/postfix/0803/0009.html
[22:10:18] *** UQlev has quit IRC
[22:12:25] <PcPixel> Any idea why internal senders can use other peoples sender domains? Config files: http://pastebin.com/d48abffa0
[22:13:52] <j2^> adaptr: Do you mean sender_dependent_relayhost_maps ?
[22:17:00] *** yfoo has quit IRC
[22:18:21] <vice-versa> PcPixel: ha, your main.cf looks like a page right out of the 'Book of Postfix' ;)
[22:20:54] <j2^> adaptr: Sorry, that does not seem to do what i need, as this would also trigger on mail froming from the other direction.
[22:21:14] <j2^> I need something to route based on the IP of the sending server, and it appears that it doesnt do that(?)
[22:23:12] *** godlie has quit IRC
[22:29:22] *** havvg has quit IRC
[22:31:41] *** jellis-real has quit IRC
[22:38:11] *** PcPixel has quit IRC
[22:38:54] *** Haris1 has quit IRC
[22:39:52] *** war9407 has quit IRC
[22:40:49] *** Haris_ has joined #postfix
[22:43:22] *** [diablo] has quit IRC
[22:48:20] *** Haris_ is now known as Haris1
[22:54:22] <donspaulding> adaptr: so if header_checks removes the message-id, cleanup will add it back in right?   How do I get it to detect that a message-id is invalid, do I handcraft my own regex?
[22:55:36] *** quieteyes has joined #postfix
[22:56:26] *** quieteyes has left #postfix
[22:58:14] *** rootsvr has quit IRC
[22:58:23] *** ramoni has quit IRC
[22:59:45] *** Haris has quit IRC
[22:59:50] *** Haris has joined #postfix
[23:11:58] *** Draecos_ has joined #postfix
[23:12:37] *** Haris1 has quit IRC
[23:12:49] *** havvg has joined #postfix
[23:16:03] *** Haris_ has joined #postfix
[23:16:42] *** tshine has quit IRC
[23:19:51] *** tshine has joined #postfix
[23:21:50] *** Draecos has quit IRC
[23:23:57] *** j_s has quit IRC
[23:28:40] *** Siegfried has joined #postfix
[23:40:47] *** cmdln has quit IRC
[23:56:32] *** madrescher has joined #postfix





top