[02:33] <brad100> does anyone here use thunderbird with postfix-dovecot? for some reason thunderbird cant verify my account and i cant figure out why
[02:36] <Patrickdk> well, only one of two problems
[02:36] <Patrickdk> setup thunderbird wrong
[02:36] <Patrickdk> setup dovecot wrong
[02:36] <Patrickdk> you do have logs right?
[02:37] <brad100> yeah
[02:38] <brad100> how would i show them to you? xD im a complete noob im sorry
[02:38] <Patrickdk> !pastebin
[02:39] <brad100> which logs do i show you
[02:39] <brad100> it would have to be dovecot
[02:43] <brad100> @patrickdk
[02:51] <chaz68> http://imagebin.org/261604 - Anybody got some ideas here.  I'm running out of ideas...
[02:52] <chaz68> Using Apache
[02:53] <brad100> @Patrickdk
[03:09] <brad100> anyone?
[03:47] <TimRe> does anybody know how I can setup domain.com/~username just for webpages and be able have my users just login into there directory an noting else?
[03:51] <qman__> TimRe, apache mod_userdir
[04:00] <Austneal> Hello!
[04:00] <Austneal> I'm trying to set up postfix / dovecot to use virtual users / multiple domains... would anyone like to help me? :S
[04:02] <Austneal> have looked at tons of guides and tutorials, but none of them seem to turn out for me
[04:03] <Austneal> I mean, I can get postfix / dovecot installed and working with system accounts... but when I try to do anything with virtual users / domains is where I have a ton of trouble
[04:07] <qman__> Austneal, https://help.ubuntu.com/community/PostfixVirtualMailBoxClamSmtpHowto
[04:11] <Austneal> qman__ if that works, Ill love you :P
[04:15] <Austneal> Nope, already not working
[04:15] <Austneal> its telling me to remove a line that doesnt exist
[04:17] <Austneal> "Comment the line  mailbox_command = procmail -a "$EXTENSION". It will not be used."
[04:17] <Austneal> that line isnt in the file
[04:42] <Austneal> continued on.... now I've hit another dead end
[04:43] <Austneal> sigh... should have stuck with windows server
[04:44] <Austneal> telling me to go to /home/vmail/domain.com/
[04:44] <Austneal> doesnt exist....
[04:55] <qman__> create it
[04:56] <qman__> it's not meant to be taken verbatim, you have to substitute your actual setup for it
[04:56] <Austneal> that kind of defeats the purpose of setting this up, doesnt it?
[04:57] <Austneal> alright though... I'll try it
[04:58] <Austneal> nope, still failed
[04:59] <qman__> that directory must be writable by postfix
[05:01] <Austneal_> that was strange
[05:04] <Austneal> how would I see if its writable by postfix?
[05:04] <qman__> postfix run as the user postfix
[05:04] <qman__> that user needs write permission, so you either need to chown it or chmod it appropriately
[05:04] <Austneal> uh...
[05:04] <qman__> postfix can't deliver mail there if it can't write to it
[05:05] <Austneal> tutorial made me chown it for a different user
[05:05] <Austneal> er
[05:05] <Austneal> it set that as the home directory to a user I created
[05:06] <qman__> oh, didn't see that
[05:06] <qman__> it has you create a user dedicated to this
[05:06] <qman__> ok
[05:07] <qman__> so what failed
[05:07] <Austneal> it didnt make the mailbox folder
[05:07] <Austneal> when I sent it mail, it said the delivery permanently failed
[05:07] <qman__> and /home/vmail existed, and was owned by the vmail user?
[05:08] <Austneal> correct
[05:08] <Austneal> actually... wait
[05:08] <Austneal> how would I see who owns that directory?
[05:08] <qman__> ls -l
[05:09] <qman__> also, it doesn't automatically create the domain directories, the user add script does
[05:09] <Austneal>  drwxr-xr-x  3 vmail  vmail  4096 Jun 17 04:57 vmail
[05:09] <qman__> the script under "Create Dovecot Users"
[05:09] <Austneal> Ok... what but I manually created them like you said to earlier :/
[05:09] <Austneal> let me restart service and try again
[05:10] <qman__> did you follow through that part with that script?
[05:10] <Austneal> not yet
[05:10] <Austneal> it told me to test it with the mail command first
[05:10] <Austneal> which also didnt work
[05:11] <qman__> ok
[05:11] <Austneal> should I try to create the script first?
[05:11] <qman__> looks like an oversight in the guide
[05:12] <qman__> while the domain.com may or may not be auto generated, and the users' directories are auto generated, /home/vmail is not necessarily auto generated
[05:13] <Austneal> has been there anyway
[05:13] <Austneal> ....
[05:13] <Austneal> message didnt go through
[05:13] <Austneal> "/home/vmail" is there
[05:13] <qman__> show me the exact error line
[05:13] <qman__> tail /var/log/mail.err
[05:14] <Austneal> screen just flooded with lines....
[05:14] <qman__> should have only grabbed the last 5
[05:14] <Austneal> which one would you like? :/
[05:14] <Austneal> this looks relevant:
[05:14] <qman__> just pastebin it all
[05:14] <Austneal> Jun 17 01:45:51 domU-12-31-39-0E-16-55 postfix/postmap[18340]: fatal: open /etc/postfix/transport: No such file or directory
[05:15] <Austneal> ok
[05:15] <Austneal> one sec
[05:15] <qman__> ah, that's definitely important
[05:15] <qman__> was that from some other guide? it's not in this one
[05:16] <Austneal> um... Im not sure
[05:16] <Austneal> http://pastebin.com/SLdEEseB
[05:16] <Austneal> It very well could be >.<
[05:16] <qman__> ok, what I think you need to do is start fresh
[05:16] <qman__> and by that I mean, apt-get purge all the mail packages
[05:17] <qman__> and reinstall, getting fresh, stock config files to work with
[05:17] <Austneal> I did apt-get purge
[05:17] <Austneal> and autoremove
[05:17] <Austneal> there was no /etc/postfix directory when it was done :/
[05:17] <qman__> did you use --purge when you did autoremove?
[05:18] <Austneal> apt-get autoremove --purge
[05:18] <Austneal> ?
[05:18] <qman__> yes
[05:18] <Austneal> ugh
[05:18] <Austneal> so... I need to remove postfix and dovecot, yeah?
[05:19] <qman__> yeah, purge those packages, then apt-get autoremove --purge
[05:19] <Austneal> dovecot didnt remove
[05:19] <qman__> that will blow away all the config and data relating to those packages so you can start fresh
[05:19] <qman__> might be dovecot-common
[05:20] <Austneal> it didnt
[05:20] <Austneal> all of the dovecot config files are still there
[05:20] <Austneal> no again :/
[05:22] <qman__> ok
[05:22] <qman__> try also dovecot-imapd and dovecot-pop3d
[05:22] <qman__> and then, do dpkg -l | grep dovecot
[05:22] <qman__> if any are still installed, purge them
[05:23] <qman__> and, after that, if the directory still exists, remove it
[05:23] <Austneal> this is strange
[05:23] <qman__> same with postfix
[05:23] <Austneal> ah, ok
[05:24] <Austneal> 4 packages left
[05:24] <qman__> the thing is, the default modes are intentionally non-destructive, they will keep your old config and databases in case you wanted to recover it
[05:25] <qman__> but in this case we want to blow it away and start fresh
[05:25] <qman__> you won't have to remove and re-add the vmail user though
[05:25] <qman__> a user is a user
[05:25] <Austneal> alright, all done
[05:25] <Austneal> now... start over on the guide? >.<
[05:25] <qman__> yes
[05:26] <qman__> this time you should have a stock default config to work with and you should get a working mail server out of it
[05:26] <Austneal> alrighty... here goes
[05:26] <qman__> if not, pastebin the errors and your postfix configs
[05:27] <Austneal> rgr
[05:28] <Austneal> ...still missing that first step
[05:28] <Austneal> it says to comment out a line that doesnt exist... that shouldnt be a problem though, should it?
[05:29] <qman__> no
[05:29] <qman__> it's telling you to remove it, so if it's not there, you're already set
[05:36] <Austneal> Ok, im to the part of sending a test email
[05:37] <Austneal> failed again...
[05:38] <Austneal> http://pastebin.com/LDFFTSUP
[05:39] <Austneal> qman__, theres the log... its still giving that same error
[05:40] <qman__> ok
[05:40] <qman__> do service postfix restart
[05:40] <Austneal> done
[05:40] <qman__> in that log, it's showing that postfix was called with no options, and then it was attempted to start while running
[05:40] <qman__> try the mail again and post again
[05:40] <Austneal> rgr, one sec
[05:41] <Austneal> message failed again
[05:42] <qman__> same errors?
[05:42] <qman__> try to find any references to that file, /etc/postfix/transport, in the postfix config
[05:42] <qman__> just grep for it
[05:43] <Austneal> nothing in the main.cf
[05:44] <Austneal> and Im not seeing anything in master.cf
[05:45] <qman__> ok
[05:46]  * Austneal is stumped
[05:47] <qman__> just as a test
[05:47] <qman__> touch /etc/postfix/transport; postmap; service postfix restart
[05:53] <Austneal_> Sorry, webchat froze up on me
[05:53] <Austneal_> what about that error? :/
 touch /etc/postfix/transport; postmap; service postfix restart
[05:53] <qman__> then try the mail again
[05:53] <Austneal_> fatal: usage: postmap [-Nfinoprsvw] [-c config_dir] [-d key] [-q key] [map_type:]file...
[05:53] <qman__> oh, postmap transport
[05:53] <qman__> sorry
[05:53] <Austneal_> kk
[05:54] <Austneal_> ...transport?
[05:54] <Austneal_> or vmaps?
[05:54] <qman__> actually I don't think that's right, hold on
[05:55] <qman__> postmap /etc/aliases; postmap /etc/postfix/vmaps
[05:55] <Austneal_> run that?
[05:55] <qman__> yes
[05:56] <qman__> then restart postfix, then mail again
[05:56] <Austneal_>  postmap: warning: /etc/aliases, line 2: record is in "key: value" format; is this an alias file?
[05:56] <qman__> ok, what about vmaps
[05:57] <Austneal_> failed again
[05:57] <Austneal_> um... idk, thats all it gave me
[05:57] <qman__> ok, so vmaps worked
[05:57] <qman__> post the new log
[05:57] <Austneal_> looks like it
[05:58] <Austneal_> http://pastebin.com/BYyPmrnp
[05:59] <qman__> ok, so the only new line is that error
[05:59] <qman__> what about mail.log, anything in there?
[06:01] <Austneal_> there is...
[06:01] <Austneal_> http://pastebin.com/kYwENYxu
[06:01] <qman__> there we go, now we have something
[06:02] <qman__> user unknown
[06:03] <qman__> is that email address in vmaps?
[06:03] <Austneal_> it is
[06:03] <Austneal_> vmaps looks like this:
[06:03] <qman__> oh, I know what's up
[06:03] <qman__> see that earlier line about mydestination
[06:03] <Austneal_> test@domain.com      domain.com/test
[06:03] <Austneal_> aaah
[06:04] <Austneal_> I forgot the trailing slash
[06:04] <qman__> mydestination should be a local domain name, not your internet domain
[06:04] <qman__> it's conflicting with the virtual domain
[06:05] <Austneal_> ?
[06:05] <qman__> in main.cf
[06:05] <qman__> for example, mine is "mydestination = hilltopsrv.home.hilltop.local"
[06:06] <Austneal_> so I should remove my domain name from it?
[06:06] <qman__> because yours is set to your actual domain name, postfix sees that first and tries to deliver to a local user that doesn't exist
[06:06] <qman__> set it to a local, internal-only name
[06:06] <Austneal_> "localhost"
[06:06] <Austneal_> ?
[06:07] <qman__> no
[06:07] <qman__> localhost has special meaning
[06:07] <qman__> it should be your servers' hostname .something
[06:08] <Austneal_> .something? :/
[06:08] <qman__> you would follow whatever convention you're already using on your network
[06:08] <qman__> mine is hilltop.local
[06:08] <Austneal_> ...this is the only server on the network
[06:09] <Austneal_> im not following >.<
[06:09] <qman__> then you need to start one, do you have a workgroup or something?
[06:09] <Austneal_> no....
[06:09] <Austneal_> this is an aws server
[06:09] <qman__> ok, then I would do it like this
[06:10] <qman__> hostname.aws.yourdomain.com
[06:10] <qman__> you don't actually have to make it resolve beyond your own server
[06:10] <qman__> it's also a good idea to add that to the hosts file so that it points at itself
[06:12] <Austneal_> that sounds like a good idea >.<
[06:13] <Austneal_> ok... I think I got it
[06:14] <Austneal_> added that name to the hosts file, and pointed it to localhost... yeah?
[06:14] <qman__> well, point it to your IP, but yeah
[06:14] <qman__> not a big deal as it will only affect locally-bound mail
[06:15] <Austneal_> I think it worked >.>
[06:15] <Austneal_> ...it didnt work
[06:15] <Austneal_> the mail sent, but its not showing
[06:16] <Austneal_> oh...
[06:16] <Austneal_> one sec
[06:16] <Austneal_> bleh
[06:16] <Austneal_> still now showing up
[06:17] <qman__> ok, what's mail.log say now?
[06:18] <Austneal_> ...wheres the log at again?
[06:18] <qman__>  /var/log/mail.log
[06:18] <Austneal_> http://pastebin.com/wad00TL0
[06:19] <qman__> ok, this is the relevant error: maildir delivery failed: create maildir file /home/vmail/arachtis.com/test/tmp/1371449797.P12689.arachtis.com: Not a directory
[06:21] <qman__> so, I would do this: rm -rf /home/vmail/arachtis.com
[06:21] <qman__> and see if it recreates it
[06:21] <Austneal_> ok, resend the mail?
[06:22] <qman__> if it doesn't, I would do: mkdir -p /home/vmail/arachtis.com/test; chown vmail.vmail -R /home/vmail
[06:22] <qman__> don't need to resend, it should keep trying
[06:22] <qman__> you can send more if you like
[06:23] <Austneal_> it made a bunch of directories
[06:24] <qman__> ok, did the mail get delivered now?
[06:25] <Austneal_> i think so...
[06:25] <Austneal_> there are 3 files in the "new" folder
[06:25] <qman__> ok, then it did
[06:25] <qman__> postfix is working
[06:26] <Austneal_> \o/
[06:26] <Austneal_> now on to the dovecot mess xD
[06:26] <qman__> pretty sure the myhostname was the root cause of the problem and the rest was caused by poking at it
[06:26] <Austneal_> cd /
[06:26] <Austneal_> ...that was supposed to go into the terminal >.<
[06:26] <Austneal_> lol
[06:29] <Austneal_> ...will a self signed cert work well with dovecot?
[06:31] <qman__> it will work
[06:31] <qman__> you have to accept it at the client
[06:31] <qman__> if you want peoples' mail clients to 'just work' without giving them a warning, though, you will need to buy one
[06:32] <qman__> there's also startssl which offers a free, trusted SSL cert
[06:36] <Austneal_> I bought an SSL cert
[06:36] <Austneal_> Guess I just need to set it up with dovecot
[06:36] <qman__> dovecot takes it in pem format out of the box
[06:36] <qman__> just swap it in place of the existing snakeoil pem
[06:36] <Austneal_> it works with apache at least :/
[06:40] <austneal> irc froze again...
[06:43] <Austneal> Im a bit confused...
[06:43] <Austneal> at the bottom where its making the script... it says to add this code segment
[06:43] <Austneal> ...add it to what and where at? >.<
[06:45] <Austneal> derp, nevermind
[06:54] <Austneal> qman__ which config file was it that set the SSL cert?
[06:55] <qman__> both the postfix and dovecot config files point at one
[06:55] <qman__> just search for SSL
[07:04] <Austneal> email is a nightmare....
[07:04] <Austneal> getting an error "cant connect to server" when I use SSL
[07:12] <Austneal> #$&@YT#YU
[07:12] <Austneal> dovecot is broke
[07:12] <Austneal> jesus christ
[07:13] <Austneal> qman__ you still there?
[11:38] <hackeron> hey, how would I detect and automatically mount a plugged in hard drive? - It used to be dbus/hal but I believe it changed in the last few years?
[11:54] <halvors> Hi! I have following mailbox_command in my main.cf configuration file for Postfix: mailbox_command = /usr/lib/dovecot/deliver -c /etc/dovecot/conf.d/01-mail-stack-delivery.conf -m "${EXTENSION}"
[11:54] <halvors> But Procmail needs to be set in the mailbox_command in my main.cf file? How do i do that?
[11:55] <mardraum> vi main.cf ?
[11:56] <halvors> No i mean, what should my new mailbox_command look like?
[11:57] <mardraum> http://www.postfix.org/postconf.5.html#mailbox_command
[11:58] <mardraum> there is this sweet new thing out there, it's called google
[11:58] <mardraum> it saves you from asking basic and well documented question on irc
[12:00] <halvors> Seriously, do you think i'd be asking here if i'd found anything that helps me on google? :P
[12:00] <halvors> So why not shut down the irc channel since google exists then?
[12:01] <mardraum> you couldn't even find the postfix documentation?
[12:01] <greppy> halvors: you would be amazed at the number of people that ask very well documented questions here.
[12:02] <mardraum> and halvors, yes, I think you would ask here, because I think you are probably lazy and don't reallt want to understand things, just make something work, or think it's working until it breaks and you are back at square one.
[12:02] <halvors> Well, since i'm using the mail-stack-delivery package, i don't know how tat works with procmail...
[12:02] <mardraum> 1. install procmail 2. configure postfix
[12:03] <halvors> I'm asking here since i simply don't find anything useful on google for that issue. If you know the answer please help me instead of bugging me about google :)
[12:03] <halvors> mardraum: No shit!
[12:06] <halvors> Could i add 2 mailbox_command lines?
[12:06] <halvors> Or merge both into one?
[12:13] <zul> jamespage/yolanda: https://code.launchpad.net/~zulcss/nova/nova-ftbfs-jun17/+merge/169770
[12:29] <yolanda> zul, why my changelog entries are appearing in the diff?
[12:33] <TimRe> does anybody know how I can setup domain.com/~username just for webpages and be able have my users just login into there directory an noting else?
[12:34] <zul> yolanda:  because i merged them
[12:34] <yolanda> mm, wasn't it already merged?
[12:34] <yolanda> anyway, it looks good
[12:35] <pmatulis_> TimRe: research mod_userdir
[12:35] <TimRe> ok
[12:36] <TimRe> anything else I need there pmatulis_
[12:38] <pmatulis_> TimRe: no, just enable the mod and do a little configuring of apache, it's not much
[12:38] <TimRe> alright thx for your help
[12:39] <pmatulis_> TimRe: not quite sure what you mean by 'login there directory and nothing else'.  you probably want chroot ftp access, if so, that would be a separate thing, investigate 'vsftpd chroot'
[12:39] <TimRe> alright
[12:40] <TimRe> thx again for that info
[12:46] <thyrant> hi guys, I need to remove some malicious code from my ubuntu server 10.04.
[12:46] <thyrant> sudo rm -R /dev/
[12:46] <thyrant> rm: cannot remove directory `/dev/shm': Device or resource busy
[12:46] <thyrant> rm: cannot remove `/dev/pts/2': Operation not permitted
[12:46] <thyrant> rm: cannot remove `/dev/pts/0': Operation not permitted
[12:46] <thyrant> rm: cannot remove `/dev/pts/1': Operation not permitted
[12:46] <thyrant> rm: cannot remove `/dev/pts/ptmx': Operation not permitted
[12:46] <thyrant> how do I stop it from running?
[12:48] <hallyn> smb: hey, you mentioned a libvirt bug with 1.0.6 in saucy?  can you give some more details for zul and I?
[12:49] <hallyn> thyrant: what is malicious - the thing pinning/dev/pts/2, or the thing trying to rm -rf /dev?
[12:49] <hallyn> thyrant: lsof /dev/pts/0 will show you who is pinning that file
[12:49] <smb> hallyn, sure... I was wondering whether I actually should better open a lp bug report so it is simpler to track
[12:49] <hallyn> smb: sounds good, thanks
[12:49] <thyrant> I am the one trying to RM the folder, it won't let me as something there is running
[12:50] <thyrant> no respons from lsof /dev/pts/0
[12:51] <thyrant> also it has closed my ssh .. ssh_exchange_identification: Connection closed by remote host
[12:51] <thyrant> but I still have a ssh window open
[12:55] <smb> hallyn, zul bug 1191782
[12:56] <hallyn> smb: thanks
[12:57] <thyrant> I am running a ssh session in nautilus which is currently deleting as much as possible of the dir
[13:06] <Pici> thyrant: what makes you think that there is 'malicious code' in /dev/ ?
[13:09] <thyrant> well, many of my sites where hacked this friday and when I searched for recently changed files I found a php.ini hidden in the joomla 1.5 filesystem. This joomla install was from way back three years ago and grabbed from a shady site.
[13:12] <thyrant> so nautilus managed to remove the directory without any errors
[13:31] <thyrant> I need to make ssh working again, I currently cannot connect
[13:32] <thyrant> the ssh session i currently have says slots@warcamp:/sites$ cd /
[13:32] <thyrant> -bash: /dev/null: Permission denied
[13:32] <thyrant> -bash: /dev/null: Permission denied
[13:32] <thyrant> -bash: /dev/null: Permission denied
[13:32] <thyrant> -bash: /dev/null: Permission denied
[13:32] <thyrant> -bash: /dev/null: Permission denied
[13:32] <thyrant> slots@warcamp:/$
[13:38] <airtonix> !pastebin
[13:39] <thyrant> sorry..
[13:39] <greppy> thyrant: I have a feeling the permissions on /dev/null got hosed.
[13:39] <greppy> which will break all kinds of things.
[13:40] <thyrant> oh so it's not because of the malicious code?
[13:43] <jpds> thyrant: Compare "ls -l /dev/null" on a working system and yours.
[13:44] <thyrant> ls -l /dev/null
[13:44] <thyrant> -rw-r--r-- 1 root root 0 2013-06-17 16:39 /dev/null
[13:45] <jpds> thyrant: Well, something changed that.
[13:46] <thyrant> it's supposed to say crw-rw-rw- 1 root root 1, 3 Jun 17 09:06 /dev/null
[13:54] <thyrant> anything I can so except for reinstall my OS?
[14:03] <rbasak> Restore from backup?
[14:04] <thyrant> which program should I use for this?
[14:06] <pmatulis_> thyrant: regain confidence in your system by re-installing it.  how big a problem is that for you?
[14:08] <thyrant> it's something I planned with my hosting guy last Friday when the problems began. I have it booked tomorrow. Backing up my stuff, just wanted to know of my options
[14:10] <thyrant> for future reference how often to you guys do complete backups of webhosting servers? twice a year? and which program do you use?
[14:15] <psivaa> hallyn: out of lxc and ceph test hangs that you were investigating, we now have only ceph test hangs. The lxc tests appear to complete fine starting from the images of yesterday
[14:16] <psivaa> hallyn: i've reported a bug for the ceph test hangs, bug 1191800
[14:17] <hallyn> psivaa: can you re-run the lxc tests once again?  half expect them to fail again...
[14:18] <hallyn> psivaa: when the tests fail, does the utah vm then stay up?
[14:18] <hallyn> I ran the tests twice on my own box - using utah to setup and start a vm - and it exited both times though with some weird log msgs
[14:19] <psivaa> hallyn: ok, when the tests hang, then utah vms would be active since utah does not timeout for some reason and destroy the machine.
[14:19] <psivaa> hallyn: but when the tests complete (even with failure) the vm usually gets destroyed, but we could add a flag -n to keep the vm after the tests
[14:20] <hallyn> psivaa: you mean utah supports -n now, or you mean that's something that could be added to utah?
[14:22] <psivaa> hallyn: yes, that support was there from the beginning but we do not use it by default in order not to clutter our servers in the lab
[14:28] <hallyn> ok, thanks
[14:28] <hallyn> (right now the box is doing other things :)
[14:50] <genii> I'm just curious if anyone knows why for mailutils, has to be set in PAM and theres no conf file like /etc/mailutils.conf ... was pissing me off because i have everything else set for that (Dovecot/Exim4/Prayer setup)
[14:51] <patdk-wk> has to be set in pam?
[14:52] <genii> patdk-wk: Yes, to use Maildir instead of /var/mail
[14:54] <genii> eg: So the webmail works but when you ssh/console login, always said "No Mail". Finally got it sorted but finding the info on how to set it took some time, not well documented.
[14:56] <zul> jamespage/yolanda/roaksoax: https://code.launchpad.net/~zulcss/ceilometer/ceilometer-fbfs-jun17/+merge/169820
[14:56] <jamespage> zul, ack
[14:56] <jamespage> zul, I could do with your help on a quantum ftbfs I'm seeing
[14:57] <jamespage> I had a stab at fixing up the udev issues we see in PPA's
[14:57] <zul> jamespage:  thats next
[14:57] <jamespage> but I hit some other weird problems
[14:57] <zul> openvswitch?
[14:57] <jamespage> zul, bah - OK - I'll leave it with you then
[14:57] <zul> jamespage:  i was just going to skip the test if it raises and exception
[14:58] <jamespage> zul, for udev?
[14:58] <zul> jamespage:  for the udev race in the test
[14:58] <jamespage> zul, I don't see a udev race
[14:58] <jamespage> udev is not supported in the PPA's right?
[14:58] <jamespage> we already have a skip for one of the tests - I just added it to the two new ones in the same test file
[14:59] <psivaa> hallyn: a failed lxc vm that's in our server is utah-10304-saucy-server-amd64, if you'd like to have a look
[14:59] <zul> jamespage:  dont think so
[14:59] <zul> jamespage:  are you seeing something different?
[14:59] <zul> jamespage:  EnvironmentError: Could not create udev monitor
[15:00] <zul> https://launchpadlibrarian.net/142655662/buildlog_ubuntu-saucy-i386.quantum_1%3A2013.2%2Bgit201306170931~saucy-0ubuntu1_FAILEDTOBUILD.txt.gz
[15:00] <jamespage> zul, yeah - its not a race, its just not accessible in the PPA
[15:00] <hallyn> psivaa: i just wanted to know that it would fail again.  sounds very sporadic.
[15:00] <jamespage> zul, but I hit some other issues after that one
[15:00] <zul> jamespage:  okies ill take a look
[15:00] <patdk-wk> genii, mailutils does have it's own config file, it's right in the manual
[15:00] <zul> jamespage:  like what?
[15:00] <patdk-wk> http://mailutils.org/manual/html_node/configuration.html
[15:00] <genii> patdk-wk: Deprecated
[15:00] <hallyn> jamespage: the ceph utah testcases, do they use lxc at all?  (psivaa is getting failures in both ceph and lxc test cases, but sporadic)
[15:00] <patdk-wk> heh? I tested it, it works on my system
[15:01] <jamespage> hallyn, nope
[15:01] <genii> The mailituls.rc doesn't get parsed
[15:01] <patdk-wk> genii, it does on ubuntu 12.04
[15:01] <hallyn> jamespage: would you mind taking a look at the ceph failures to see if you can see a root cause?
[15:01] <patdk-wk> but mailituls.rc won't be, mailutils.rc will be
[15:01] <hallyn> really I'm thinking it must be environmental...  bad proxy or something
[15:01] <genii> patdk-wk: Hm, odd. I readt that documentation first and tried setting it that way. I'm using 13.04, maybe it's changed
[15:02] <genii> Typos aside ;)
[15:02] <patdk-wk> I have never used mailutils till now :)
[15:03] <jamespage> psivaa, I thought we got that test disabled as it was not reliable?
[15:04] <jamespage> hallyn, ^^
[15:04] <hallyn> huh
[15:04] <hallyn> :)
[15:06] <psivaa> jamespage: yea, i think we disabled one of the ceph tests (i386 or amd64 which i could not remember) during the raring cycle but since saucy they were somehow re-enabled
[15:06] <jamespage> psivaa, just disable them again please
[15:06] <psivaa> jamespage: both ceph tests?
[15:07] <jamespage> psivaa, yes please
[15:08] <hallyn> jamespage: well then thanks for nothing :)
[15:08] <hallyn> j/k
[15:08] <hallyn> psivaa: all right i'll look at this after some sru stuff is done
[15:08] <jamespage> hallyn, we test ceph in a different way now
[15:10] <hallyn> jamespage: yeah, i was just hoping you'd find a root cause that explained the lxc failures too :)
[15:10] <psivaa> jamespage: ok will do. it would be better to remove them from the branch. because if they stay there the automatic jenkins smoke test creation script runs it will re-enable the ceph tests again.
[15:10] <jamespage> psivaa, OK - I'll remove them then
[15:10] <psivaa> jamespage: hallyn: thank you :)
[15:10] <jamespage> psivaa, next question - where is the branch?
[15:11] <psivaa> jamespage: https://code.launchpad.net/ubuntu-test-cases
[15:35] <progre55> Hi guys. When I run "grep -a <pattern> /dev/sda" I get lots of matches, but now I need to find where those files are located. Tried recursive grep on root, as in "grep -a <pattern> -r --exclude-dir=dev --exclude-dir=proc --exclude-dir=sys /" but didn't get any matches. Any suggestions, please?
[15:44] <Austneal> hi qman__ :P
[15:44] <alibama> hi all - i want to set it up so that when users are added to the system their new home directory is /some-path/another-path/email@address
[15:44] <alibama> where do I configure this?
[15:44] <alibama> oh yeah, all user names are email@address.edu
[15:45] <alibama> so it's really just /some/path/username
[15:45] <Austneal> useradd -b /some/path/username email@address.edu
[15:45] <Austneal> I think....
[15:45] <patdk-wk> progre55, why do you assume matches on /dev/sda == matches in files?
[15:46] <patdk-wk> that would only be true if you where at 100% full filesystem
[15:46] <PryMar56> useradd -D
[15:46] <alibama> is there some way to make this a default?
[15:46] <Austneal> what PryMar56 said
[15:46] <Austneal> :P
[15:46] <genii> alibama: DHOME= in /etc/adduser.conf
[15:47] <alibama> genii: thank you!
[15:47] <progre55> patdk-wk: well okay, I have split up my /dev/sda and set up lvm. And /dev/mapper/vg-root is mounted to /. When I grep /dev/mapper/vg-root, I get matches, but not when I grep /
[15:47] <genii> alibama: You're welcome
[15:48] <genii> alibama: It will automatically make a directory with their username underneath whatever directories you specify in there.
[15:48] <patdk-wk> progre55, same answer
[15:49] <alibama> genii: this is perfect - thanks!
[15:49] <progre55> patdk-wk: thanks :)
[16:57] <alibama> hey folks - have fixed where the home directory is (thanks genii ) now need to allow the period and @ symbol in the regex
[16:58] <alibama> basically user names need to be email addresses and so I'm working on editing name_regex
[16:58] <alibama> "^[a-z][-a-z0-9]*\$" is what I started with - i've tried "^[a-z][-a-z0-9\.\@]*\$" "^[a-z][-a-z0-9.@]*\$" and some other variations, but i'm not getting it right
[16:59] <sarnold> alibama: no need to backslash-escape . in a [] character class, no need to escape @ at all..
[17:00] <alibama> sarnold: thanks - let me test - do i need to restart any services?
[17:00] <alibama> NAME_REGEX="^[a-z][-a-z0-9.@]*\$" is what I've got and it's throwing errors
[17:01] <sarnold> alibama: depends on the service and where you're setting it...
[17:01] <alibama> sarnold: it's the adduser command?
[17:01] <sarnold> alibama: probably don't escape the $ either
[17:02] <alibama> sarnold: Thanks!
[17:02] <alibama> that fixed it :)
[17:02] <sarnold> alibama: nice :)
[17:02] <sarnold> alibama: oh, maybe change the * to + -- you want at least one character in the username :)
[17:02] <alibama> ;)
[17:19] <alibama> where do most folks put their .key and .pem files?
[17:19] <Corey> alibama: For what, exactly?
[17:19] <alibama> well just for safe keeping?
[17:20] <alibama> i just didn't know if there were some best practice or common place to put them
[17:20] <xsl> hello all ... can anyone help with a bridge configuration and LXC? i need to give a container an external ip from my server
[17:20] <Corey> alibama: No, I mean which .key and which .pem files?
[17:20] <alibama> Corey: doh :)  these are just some openssl self signed files
[17:21] <Corey> alibama: You're still not answering the question. I *get* that they're parts of an OpenSSL keypair, what are they *for*?
[17:22] <Corey> alibama: Web server? Jabber OTR creds? Mail server? Something else?
[17:22] <alibama> Corey: sorry man -- they are for sftp access to the server
[17:22] <sarnold> sftp with x509 certs? o_O
[17:22] <alibama> i'm going through this process here https://help.ubuntu.com/12.04/serverguide/ftp-server.html
[17:22] <Corey> Yeah, that's a little nutty, generally one uses SSH keys for that.
[17:23] <sarnold> ah, FTPS. :)
[17:23] <alibama> so i guess /etc/ssl/certs is my answer
[17:23] <Corey> FTPS is the term you're looking for.
[17:23] <alibama> yeah - sorry for the newb
[17:23] <sarnold> alibama: actually, can you use sftp instead of ftps? ftp is an old and cranky protocol.
[17:24] <alibama> sarnold: i think i probably should
[17:25] <Corey> alibama: But I either stash them in /etc/ssl or /etc/pki; but *watch* your permissions, you don't want those world readable.
[17:25] <alibama> Corey: cool - i think i have made this vastly more complex than need be
[17:26] <alibama> so sftp would run over ssh and i don't really need to set up an ftps server at all....
[17:26] <alibama> that makes pretty good sense, i was confused about sftp vs ftps
[17:28] <alibama> i think with this in mind I"m going to follow http://kpdirection.com/technology/setting-up-sftp-on-ubuntu/
[17:30] <Corey> !tutorial
[17:30] <Corey> alibama: What's the actual use case you're trying to get done?
[17:31] <alibama> some folks uploading a bunch of ginormous files to their home directory for use in a bioinformatics app - they're the only ones that need to see their files, and no one else will ever see them
[17:31] <Corey> alibama: sftp is your friend.
[17:32] <sarnold> alibama: hrm. That tutorial would allow the web server to write to any user's files. I also don't see any way for users to actually add their own files to those directories. it looks odd.
[17:32] <Corey> (It also supports things like scp and rsync)
[17:32] <sarnold> alibama: skip that tutorial.
[17:32] <Corey> Yeah, that tutorial is nutty.
[17:32] <alibama> Corey: thanks :)  I think this is going to be much better - they'll have limited shell access with this too?
[17:37] <Corey> alibama: If you install scponly and set that to their shell, yes.
[17:37] <alibama> perfect
[17:40] <brad100> does anyone here use dovecot with thunderbird? i set up dovecot and thunderbird cant verify my account and i dont know why
[17:48] <mgw> What is the d-i option for setting an alternative ntp server?
[17:54] <JanC> brad100: check that you configure SSL or TLS correctly if you use that
[17:54] <JanC> (I hope you use that :p )
[18:40] <adam_g_> Daviey, any chance you can wave  openstack 2013.1.2 thru the raring upload queue? the original keystone got trumped by security, so there is a new one there (0ubuntu2) to include fixes
[18:44] <Daviey> adam_g_: I'll review after i've eaten, deal?
[18:45] <adam_g_> Daviey, sure. but don't forget to eat your vegetables
[18:46] <Daviey> Yes sir!
[18:48] <six_storm> hello everyone.  can anyone assist me on a BIND9 setup?
[19:14] <Daviey> adam_g_: why are there 2 of everything? https://launchpad.net/ubuntu/raring/+queue?queue_state=1
[19:15] <adam_g_> Daviey, uh i dont see duplicates, with the exception of keystone
[19:15] <adam_g_> but one supercedes the other
[19:18] <zul> SpamapS:  is anything happening with MariaDB?
[19:22] <SpamapS> zul: It has been packaged for Debian and is awaiting somebody to look at the packages.
[19:22] <zul> SpamapS:  like an archive admin?
[19:23] <SpamapS> no
[19:23] <SpamapS> like, a DD
[19:24] <SpamapS> For some reason I'm the only DD who actually gives a rats ass about MySQL and its demonspawn.
[19:24] <zul> SpamapS:  hah ok
[19:25] <Daviey> adam_g_: Ah, i thought it was simply a rebase for an already inflight SRU.  As it's a fresh one, i'd rather look again in the morning now.  Sorry.
[19:25] <adam_g_> ok
[19:43] <Corey> mgw: "d-i option?" Multiple "server" lines in ntp.conf isn't working for your use case?
[22:05] <kobrien> ahoy. everytime I log into my server with encrypted home I have to run "ecryptfs-mount-private". Any idea how I can have it automatically unlock at login?
[22:07] <tyhicks> kobrien: does the file /home/.ecryptfs/kobrien/.ecryptfs/auto-mount exist?
[22:08] <kobrien> no. if I touch it, will that sort it?
[22:08] <tyhicks> kobrien: yes, it should
[22:08] <kobrien> thanks, I'll try that
[22:10] <sarnold> tyhicks: is Dustin's answer here no longer The Best Answer? http://askubuntu.com/a/116198/33812
[22:12] <tyhicks> sarnold: his answer is the Best Answer for that specific question (How do I get an automatic password prompt if I use SSH key authentication?)
[22:13] <tyhicks> sarnold: kobrien didn't mention anything about using SSH key authentication and the file that indicates that an auto-mount should occur doesn't exist, so I think that creating the auto-mount file should work for kobrien
[22:15] <sarnold> tyhicks: aha, thanks :)
[22:15] <tyhicks> np :)
[22:19] <sh_t> hey folks. I'm trying to create a package with checkinstall that I can use on a few identical debian/ubuntu servers. I've added a few extra lines into my Makefile/install area to create some directories in addition to some similar commands already there (mkdir -p). When I checkinstall on the server that I'm creating the package on (in the source tree) everything in my Makefile is done
[22:19] <sh_t> correctly. When I copy the package to another server, only some of these commands are actually ran. I've tried --fstrans=no to no avail. any ideas what might cause this?
[22:23] <mgw1> so, I have this in my preseed — d-i     mirror/http/proxy string http://x.x.x.x:8000/
[22:23] <mgw1> But tcpdump on host x.x.x.x does not show any traffic on port 8000
[22:24] <mgw1> On the machine where the installer is running, I can jump to another console and wget via that proxy, but the installer itself is not using the proxy.
[22:24] <mgw1> Any ideas?
[22:30] <Corey> mgw1: To my understanding the syntax that kickstart at least is looking for is http_proxy
[22:30] <Corey> mgw1: Not familiar with "d-i" as you're using it, can you expand it for me?
[22:30] <sarnold> Corey: iirc, that'd be for the http_proxy environment variable; I don't think apt-get respects that, preferring its own http proxy configuration. iirc. :)
[22:31] <mgw1> Corey, sarnold: the d-i I believe are debian specific installer options in the preseed
[22:31] <Corey> sarnold: Right, but that's not during preseed generally. Once the box is up that belongs in etc/apt/apt.conf.d/00proxy
[22:32] <sarnold> everything I know about it (little) comes from section b.4.4 here: http://www.debian.org/releases/stable/amd64/apbs04.html.en
[22:32] <mgw1> the mirror/http/proxy string …. in the kickstart is along with "d-i mirror/http/hostname string archive.ubuntu.com"
[22:33] <mgw1> that part is working fine
[22:33] <mgw1> B.4.4. Mirror settings
[22:38] <Corey> mgw1: And your goal here is to get the proxy used for packages, or something else?
[22:40] <mgw1> Corey: yes, during a pxe install
[22:40] <Corey> mgw1: Okay. There was a bug around this back in ~2010 that resulted in that string being applied to all connections (curl, wget, etc) mentioned in the preseed config.
[22:41] <mgw1> yeah, i saw that  bug
[22:41] <mgw1> but this is the opposite
[22:42] <Corey> mgw1: Dumb question, can the node resolve x.x.x.x (assuming hostname and not IP) and route to it?
[22:42] <mgw1> It's an ip
[22:42] <mgw1> and yes, it can route to it
[22:42] <mgw1> I ran the exact same wget mentioned in syslog, but with http_proxy set, and it ran
[22:43] <Corey> mgw1: This is kinda ridiculous, but... after d-i, is that a space or a tab?
[22:44]  * Corey distrusts whitespace variances
[22:46] <mgw1> spaces, same as the other lines
[22:47] <Corey> (Unfortunately I don't have a convenient environment in which I can play with preseed at the moment)
[22:50] <Corey> mgw1: Can you redact and paste your preseed.cfg in its entirity?
[22:54] <mgw1> Corey: https://gist.github.com/mgwilliams/744119549a195edd3bdc
[22:54] <Corey> mgw1: Sweet, checking.
[22:54] <mgw1> I just tried changing it to 'manual' — checking now if it works
[22:55] <mgw1> It's ignoring the directive. 'manual' does nothing either.
[22:56] <mgw1> actually it does seem to make a difference, in an odd way
[22:57] <mgw1> syslog complains about no route to the public ip when I changed it to 'manual'
[22:57] <mgw1> Whereas with it there it just complains that the archive is broken
[22:58] <Corey> mgw1: Debian wiki states: Be aware there is only one space in preseed files between subkey and value on "owner key/subkey value" lines.
[22:58] <Corey> Not sure if that's a hard and fast rule or not; if so I'd expect you to see more failures?
[22:59] <mgw1> yeah, the preseed was based on one from elsewhere, and it arranged everything into columns
[22:59] <mgw1> Everything else (up to that point) is working.
[23:00] <BAMbanda> does ubuntu-server essentially offer the full ubuntu terminal eith the only exception of no gui?
[23:03] <jcastro> yes
[23:12] <Corey> mgw1: Sorry about that, the bat-phone rang. :-)
[23:13] <Corey> mgw1: In all of the examples I see that use the proxy setting, it gets set after the rest of the apt/http config options are passed.
[23:14] <mgw1> Corey: I tried it both ways
[23:14] <mgw1> As I saw it both ways
[23:14] <Corey> mgw1: So I wonder if what's happening here is that your mirror/http/hostname is overriding mirror/http/proxy
[23:14] <Corey> (Isn't blind troubleshooting fun?)
[23:14] <mgw1> yes
[23:14] <Corey> This seems particularly poorly documented (the entire preseed functionality)...
[23:15] <mgw1> for the pxe part the right solution might be to use an apt-mirror anyway
[23:15] <Corey> Yeah, I'm a big fan of using a proxy for apt-mirror. What's the goal here, not to redownload the same thing over and over?
[23:16] <Corey> mgw1: Perhaps blow out that stanza and crib from https://help.ubuntu.com/lts/installation-guide/i386/preseed-contents.html
[23:17] <Corey> (I still haven't ruled out weird whitespace issue!)
[23:19] <mgw1> Corey: #1 goal — get access to packages from behind a firewall (no public ip)
[23:19] <sarnold> mgw1: flailing-blindly idea: drop the "http://" from the front of the address.
[23:19] <mgw1> #2 goal — don't re-download everything for every install
[23:19] <Corey> sarnold: Yeah, except everything I've seen uses the http:// string for the proxy.
[23:20] <sarnold> Corey: d'oh. there's no string at all at https://help.ubuntu.com/lts/installation-guide/i386/preseed-contents.html  so I got to wondering...
[23:20] <Corey> sarnold: Right, it's like they want that there even in the event that you're not using a proxy, which is odd to me.
[23:21] <Corey> It's definitely a strange problem, that's for sure.
[23:22] <Corey> Not having much success in trawling the codebase to see what it's actually doing programmatically.
[23:25] <Corey> mgw1: Oh jeez.
[23:25] <Corey> mgw1: This is a really, really, really stupid idea. But.. what happens if you specify the proxy via hostname?
[23:29] <mgw1> instead of ip?
[23:29] <Corey> mgw1: Yes.
[23:30] <mgw1> Corey: give me a few minutes and I'll test that
[23:30] <Corey> (bonus points if it actually lives in DNS...)
[23:30] <Corey> I'm also admitting defeat and asking Kyle. :-)
[23:44] <kobrien> I have an encrypted home. I deleted the wrapped-passphrase in ~/.ecryptfs/. How can I restore this file?
[23:44] <qman__> from your backup system
[23:44] <Corey> kobrien: It's a good thing you've got a robust backup policy in place and can restore from backups though!  A lot of people with insufficient experience and bad administrative habits would not be in such a good position.
[23:44] <kobrien> when I try to run "ecryptfs-mount-private", it says Signature not found in user keyring
[23:45] <Corey> mgw1: <kyle> Corey: hmm off the top of my head I don't know
[23:45] <Corey> mgw1: Okay, we've got an actually interesting problem here. :-)
[23:45] <Corey> mgw1: (Kyle wrote "The Official Ubuntu Server Book")
[23:47] <kobrien> Well the problem is, that having restored this file using ecryptfs-wrap-passphrase, this error persists.
[23:50] <mgw1> Corey: I think maybe the preseed is not getting used at all, after making some more changes to it. Maybe the wrong boot option.
[23:50] <mgw1> trying some other options now, like "preseed/url" instead of "ks"