[00:17] <adam_g> RoAkSoAx: if those cobbler changes get uploaded, a quick fix in orchestra-import-isos should keep the cache updated over time: http://paste.ubuntu.com/707071/
[00:33] <ruben23> hi guys how do i loadbalance two apache web server only..? any idea guys please
[01:47] <twb> I'm in the market for a build bot.  Any opinions/suggestions?  My rough functional requirements are http://paste.debian.net/136152/
[03:25] <socomm> Anyone got recommendation to any good tutorials covering mail server on Ubuntu-server?
[03:28] <twb> socomm: the ubuntu server guide
[03:52] <socomm> twb: thx I will start there
[04:25] <RoAkSoAx> adam_g: lol ok
[05:28] <crmccreary> Anybody free to help diagnose a dns problem?
[05:28] <twb> !anybody
[05:36] <crmccreary> OK. First time irc user. I have been using a server for DHCP and dynamic DNS with BIND for some time. I "demoted" the server to a dhcp client when I purchased a new server. The new server works great, al the clients are resolveable *except* the old dhcp server. I thought that I've scrubbed it clean but  when restarting networking, I keep getting "DHCPREQUEST for 192.168.59.3 (192.168.59.2) from 00:06:4f:4f:59:2f via eth1"
[05:36] <crmccreary> where 192.168.59.2 was the old static ip of the old server and is now the new static ip of the new server. I've assigned 192.168.59.3 in dhcpd.conf:
[05:36] <crmccreary> host pe4400.crmeng.lan {
[05:36] <crmccreary>   hardware ethernet 00:06:4f:4f:59:2f;
[05:36] <crmccreary>   fixed-address 192.168.59.3;
[05:36] <crmccreary>   default-lease-time 86400;
[05:36] <crmccreary>   max-lease-time 86400;
[05:36] <crmccreary> }
[05:37] <twb> So what's the problem
[05:37] <crmccreary> The old server's name pe4400  is not resolveable anywhere except on itself.
[05:39] <crmccreary> It gets the right ip address, but bind ignores it. I've added "send host-name "<hostname>";" to the old server's dhclient.conf to no avail
[05:46] <twb> You understand that host names provided by DHCP clients, don't necessarily propagate to the DNS server?
[05:52] <crmccreary> Yes, but the server is set up for dhcp/dns server with dynamic dns (see http://ubuntuforums.org/showthread.php?t=730717 for an example set up). All of the dhcp clients are resolveable by their FQDN's *except* the old server. I'm not using caching, so where is the "memory" of the old server's MAC coming from?
[05:54] <twb> Is the old server acting as a DHCP client?  Is it an *ubuntu* (not debian) server?
[05:54] <crmccreary> Yes
[05:54] <twb> The format "<hostname>" is an ubuntu-specific patch
[05:54] <twb> crmccreary: OK, not sure then
[05:55] <twb> I use dnsmasq, not bind, so I'm not sure how to do much more diagnostics
[05:55] <crmccreary> Ubuntu - DISTRIB_DESCRIPTION="Ubuntu 10.04.3 LTS"
[07:38] <josePhoenix> Hello all
[07:39] <josePhoenix> I just did 'service mysql stop' and 'service mysql start... start hung so I ctrl-Ced out of it
[07:39] <josePhoenix> Then I did 'service mysql start' and it worked
[07:39] <josePhoenix> but none of the php sites on this server can establish a db connection
[07:40] <josePhoenix> ... okay, weird.
[07:40] <josePhoenix> stopped and started the service again, now it works
[08:03] <spiekey> Hello!
[08:03] <spiekey> i need some ideas for a (Config)-File Deployment strategie for multiple ubuntu/linux server.
[08:04] <jamespage> morning all
[08:04] <spiekey> has anyone an idea or a nice project?
[08:05] <jamespage> spiekey: there are a few options - puppet or chef provide good configuration management toolsets for individual servers
[08:05] <jamespage> if you want todo something at a higher level with co-ordination across servers you might like to take a look a juju
[08:06] <RoyK> #873198
[08:06] <RoyK> bug #873198
[08:06] <RoyK> that's a rather annoying one
[08:13] <RoyK> how long does it take from a bug is filed till it becomes public?
[08:14] <jamespage> RoyK, not normally that long - was it specifically marked private?
[08:17] <RoyK> ah. dunno why it was
[08:18] <RoyK> bug #873198
[08:26] <soren> RoyK: It was marked private because it was marked security.
[08:27] <RoyK> ah
[08:27] <RoyK> it's security-related, but not a big hole, unless you consider running a 1YO kernel a major security issue - it might be...
[08:51] <soren> RoyK: Heck, running an up-to-date kernel is probably a major security issue.
[08:57] <ivoks> :(
[08:58] <ivoks> we should devote this release to Dennis Ritchie
[08:58] <soren> Yeah, I just heards about it, too :(
[09:07] <chris-> hey all
[09:09] <chris-> i have the following problem: i have installed another ubuntu server 10.04 LTS , i had a problem connecting via putty over ssh ( always said access denied ). then i set protocol to v2 only and it worked but after a while i get a network error message and putty disconnects, then i cant connect anymore no matter what protocl version i chose i always get a access denied error after auth
[09:13] <soren> chris-: If I had to guess (and I guess I do), you have another host with the same IP on the network.
[09:13] <soren> chris-: ..and the fact that changing to v2 fixed it is a coincidence.
[09:14] <soren> chris-: The disconnection is caused by the other host suddenly getting the network packets and sending back an RST (because it doesn't recognise the connection).
[09:14] <soren> chris-: ..and now you're back to not being able to auth because you're talking to a host different from the one you're expeting.
[09:14] <soren> expecting, even.
[09:15] <soren> I can't quite explain why you're not getting SSH host key errors, though.
[09:15] <soren> Are these machince built from an image?
[09:15] <soren> chris-: So they could potentially have the same SSH host key?
[09:17] <chris-> soren, no its a fresh install
[09:18] <chris-> soren, when i reboot the server i can connect to it again until i get disconnected for some reason ( maye after about 30mins )
[09:23] <chris-> soren, funny u seem to be right
[09:24] <chris-> i am installing the machine remote and have been given the ip configuration i should make by the admins
[09:24] <chris-> soren, seems like they made a mistake , i shut down the server and can still ping the ip adress, so theres probably another device here
[09:29] <chris-> soren, u were right just called the admin and he made a mistake - theres another device with this ip haha. funny i didnt think of this myself. thanks a lot mate ;)
[09:29]  * soren pats his crystal ball
[09:29] <soren> Comes in handy every so often.
[09:29] <soren> chris-: You're welcome.
[09:30] <soren> I still wonder about those ssh host keys, though.
[09:35] <chris-> hehehe :)
[10:04] <koolhead17> morning everybody!!
[10:43] <lynxman> koolhead17: morning
[10:45] <koolhead17> hello lynxman :)
[10:53] <Daviey> jamespage: around?
[10:53] <Daviey> jamespage: when you get this, how did bug 870244 go?
[10:54] <jamespage> Daviey: sure am
[10:54] <jamespage> hmm - not much progress
[10:54] <jamespage> lemme just see where I got to
[10:57] <Daviey> cool
[11:08] <jamespage> Daviey: sorry - dovecot fell of my list for some reason
[11:09] <jamespage> I hit an issue with restarting dovecot - just testing again now
[11:23] <jamespage> Daviey: see my comment on bug 870244 - ATM restarting or stop/starting is a bit unreliable
[11:31] <Daviey> jamespage: I think we need to add this to the release notes for Oneiric.. and fix it in P..
[11:31] <Daviey> not sure it's worth SRU'ing.. agree?
[11:38] <koolhead17> jamespage: hey. how did you associated set of floating IP to a project? Currently i been manully adding it in DB
[11:38] <koolhead17> in diablo
[12:21] <smoser> jamespage, ping
[12:33] <jamespage> blimey - I pop out for lunch and see what happens
[12:33] <jamespage> Daviey: agree on the SRU
[12:33] <jamespage> the fix is not reliable
[12:34] <jamespage> koolhead17: hrm - not sure - I setup a set of floating IP addresses as detailed in https://wiki.ubuntu.com/ServerTeam/Oneiric/OpenStackTestPlan
[12:34] <jamespage> I then used euca-allocate-address and euca-associated-address to fix them to instances
[12:34] <jamespage> smoser: pong
[12:35] <smoser> how can i run a tiny little images test ? with ubuntu-server-ec2-testing
[12:35] <smoser> i made a merge proposal last night, but completley untested and i want to attempt to test
[12:35] <smoser> :)
[12:36] <jamespage> smoser: you should be able to run it from the command line
[12:36] <jamespage> from within the bzr branch
[12:36]  * jamespage refreshes his memory
[12:37] <smoser> jamespage, i see execute_all_ec2_tests.py, run_ubuntu_ec2_multi_test.py, execute_ubuntu_ec2_test.py, run_ubuntu_ec2_test.py
[12:37] <smoser> all of those looked like good candidates :)
[12:37] <jamespage> smoser: execute_ubuntu_ec2_test.py
[12:37] <jamespage> is the one
[12:38] <jamespage> --help has good advice
[12:39] <jamespage> smoser: you will need a branch of  lp:~ubuntu-server-ec2-testing-dev/+junk/ec2-automated-tests
[12:39] <smoser> yeah, and thats expected to be ~/tests by dfault ?
[12:39] <jamespage> and to setup a ~/.ec2.yaml file containing access credentials
[12:39] <jamespage> yes
[12:40] <smoser> ok. i'll run this test and then write a README or doc/ something on how to do it.
[12:40] <jamespage> just has to keys - AWS_ACCESS_KEY_ID & AWS_SECRET_ACCESS_KEY
[12:41] <jamespage> smoser: that would be great - there is the start of one in docs
[12:45] <jamespage> smoser: the test names are in lp:~ubuntu-server-ec2-testing-dev/+junk/ec2-automated-tests global/tests.yaml
[12:45] <jamespage> that file pulls together metadata about the test and the actual detail of the test
[12:46] <zul> morning
[12:47] <jamespage> morning zul
[12:47] <kainore> hey i got 3 servers i want to connect remotely from 1 public ip but they got 3 diffrend local ip's anyone know how i can set that up ?
[12:49] <zul> kainore: google nat
[12:49] <kainore> network access translater ?
[12:49] <kainore> address sorry mybad
[12:50] <kainore> but thanks
[12:50] <jamespage> kainore, how exactly do you want to access them? SSH?
[12:51] <kainore> server01.domain.dk server02.domain.dk server03.domain.dk they just all get to server01
[12:51] <Daviey> Does anyone have a fresh oneiric server install to hand?
[12:51] <kainore> and ssh need ofc to get around as well
[12:53] <zul> Daviey: i can fire up a kvm if that helps
[12:54] <Daviey> zul: nah, no worries.
[12:57] <jamespage> Daviey: I have quite a fresh one
[12:57] <jamespage> as in later yesterday
[12:58] <jamespage> Daviey: are you writing the release note for mail-stack-delivery?
[13:00] <Daviey> jamespage: I'm not sure it's worth noting is it?
[13:00] <Daviey> it was in Natty, and easy to work around.
[13:00] <Daviey> think we should include it?
[13:00] <jamespage> meh
[13:00] <jamespage> probably not - its a minor niggle
[13:01] <jamespage> I just noted it had a release notes task that was all
[13:01] <Daviey> If you think we should, sure.
[13:01] <Daviey> i think i added that task as a maybe.
[13:02] <jamespage> as its such a minor issue lets not add it - if we get lots of post release bugs (which we won't - as you state its like this in natty) then we can consider an SRU
[13:02] <jamespage> Daviey: ^^ - I'll pickup the bug to resolve early next cycle
[13:03] <jamespage> we also need to fix the racey restart issue
[13:03] <jamespage> but that prob requires some upstream co-op
[13:03] <Daviey> jamespage: rocking!
[13:03] <Daviey> jamespage: is the task assigned to you and targetted? :)
[13:04] <jamespage> Daviey: it is now :-)
[13:04] <Daviey> awesome!
[13:04] <jamespage> I marked the oneiric task won't fix as well
[13:04] <Daviey> sounds good.
[13:23] <zul> Daviey: im going to push out an SRU for nova
[13:24] <nijaba> hello.  Do you know where we should be pointing users that looks for instruction to deploy Ubuntu Cloud Infra ?
[13:25] <koolhead17> a
[13:25] <zul> the nova docs maybe?
[13:25] <nijaba> zul: which certainly exlain how to use orchetra and juju, I am sure!
[13:26] <lynxman> nijaba: hm, don't think there's complete docs yet :/
[13:26] <zul> nijaba: right i think RoAkSoAx was suppose to do orchestra and juju but hasnt finished yet
[13:34] <koolhead17> i wasted so many hours with role based access control :(
[13:36] <w00> ooh it's on!
[13:37] <koolhead17> lynxman: the secret is out :P
[13:37] <lynxman> koolhead17: yeah, preparing my blog post :)
[13:37]  * koolhead17 pokes Daviey 
[13:38] <koolhead17> lynxman: i have many things to write on my blog if am allowed too, rather many failures i encountered with :P
[13:38] <koolhead17> daiblo/old auth :P
[13:48] <lynxman> koolhead17: I'd rather try to fix those for P ;)
[13:48] <zul> lynxman: its ok to say precise ;)
[13:49] <lynxman> zul: I so much preferred the name I proposed :)
[13:49] <zul> lynxman: i dont want to even know
[13:50] <lynxman> zul: you sure? :D
[13:50] <medberry> Precise is going to have more name collisions than Oneiric.
[13:50] <medberry> context collisions, etc.
[13:50] <medberry> lynxman, P?
[13:50] <lynxman> medberry: Precise Pangolin of course
[13:50] <zul> medberry: at least i dont have to go to dictionary.conf to figure out how to pronounce it ;)
[13:50] <medberry> lynxman, that's what you wanted?
[13:51] <medberry> s/wanted/proposed/
[13:51] <lynxman> medberry: nah, mine was a bit more accurate :o)
[13:51] <medberry> lynxman, nod.
[13:52] <smoser> jamespage, ping
[13:52] <smoser> so, ec2-automated-tests/global/tests.yaml has 'simple-user-data'
[13:54] <jamespage> smoser: yep
[13:54] <smoser> and i can probably change 'instance_type' there.
[13:54] <smoser> i want to run t1.micro, just for testing
[13:54] <koolhead17> jamespage: thanks. :d
[13:54] <smoser> buti dont know how it knows arch and such
[13:55] <koolhead17> the !@#$ is_admin table
[13:55] <smoser> oh, i see. i can pass that in.
[13:55] <smoser> never mind. i'm a dolt.
[13:56] <jamespage> yes - the framework will then pick the right instance size of the arch from archs.yaml
[13:56] <jamespage> in this case 'default'
[13:57]  * jamespage thinks he might have gone a bit over the top with yaml in this framework
[14:00] <koolhead17> why ubuntu.com shows laptop with bloody "Dell" on it :P
[14:00] <lynxman> koolhead17: you don't like Dell?
[14:00] <koolhead17> lynxman: nopes. the morons sell with MSFT in india.
[14:01] <koolhead17> Acer is the only company which gives me linux installed lappy
[14:01] <koolhead17> :P
[14:01] <koolhead17> and then i replace limpus to ubuntu :D
[14:03] <Martyn> koolhead17 : That might be because Dell and Canonical are partners
[14:04] <Martyn> and that Dell will ship machines with Ubuntu pre-installed
[14:04] <kirkland> RoAkSoAx: yo
[14:05] <koolhead17> Martyn: i doubt it happens in reality of at all. The showroom shows only 7 :D
[14:05] <Martyn> koolhead17 : #awholenewworld :)   It didn't count down to the release of Oneiric :)
[14:05] <Martyn> koolhead17 : Could be something else :)
[14:07] <koolhead17> Martyn: i give damm to an OEM like dell, who licks MSFT ass  :)
[14:08] <patdk-wk> heh
[14:08] <Martyn> Well, how do you _really_ feel P)
[14:08]  * patdk-wk is pretty much stuck to hp and supermicro
[14:09] <koolhead17> anyways am happy if canonical is making something from them in partnership !!! :D
[14:13] <wip> only 1 day before my hardy server get hacked again... i really need to do-release-upgrade but i fear very much
[14:13] <wip> why hardy doesn't offer PHP 5.3
[14:13] <matthew> hi
[14:13] <matthew> anyone here good with SendMail?
[14:14] <lynxman> matthew: I'm kind of good with it :)
[14:14] <matthew> ok….
[14:14] <matthew> I am using Rapidweaver, and I have a "ContactMe" web page….
[14:14] <matthew> I published my site, and on the contact page, it says mail is being sent, but … guess what… its a no go….
[14:15] <matthew> I don't know if its a sendmail configuration issue or PHP5 issue…
[14:15] <lynxman> matthew: I'd say that's not even remotely related to sendmail, you're several layers above :)
[14:15] <lynxman> matthew: start by checking logs, writing some php script to check, etc
[14:16] <matthew> sadf
[14:16] <w00> do-release-upgrade, y u error fatally :|
[14:18] <wip> w00, you don't recommand doing a do-release-upgrade from hardy LTS?
[14:19] <lynxman> wip: if you're running in production I'd wait for the next LTS
[14:19] <rbasak> zul: re bug 873243, the consensus here seemed to be that it doesn't make sense to add xen-hypervisor-4.1-{amd64,i386} since that won't assure that the system is booted into dom0
[14:19] <w00> ^
[14:19] <rbasak> zul: I don't suppose it matters that much though
[14:19] <wip> lynxman, i cannot wait my server is running PHP 5.2 (hardy version) and this version is easily hacked
[14:19] <lynxman> wip: oops :/
[14:20] <lynxman> wip: maybe getting a backport?
[14:20] <zul> rbasak: well ideally we would have an xen-meta package that does everything for you and just depend on that
[14:20] <wip> i have 2 options: upgrade PHP to PHP 5.3 or do-release-upgrade
[14:20] <wip> lynxman, can you explain how to get a backport? is it possible to upgrade PHP5.3 from hardy?
[14:20] <w00> Where can i report the release-upgrade err?
[14:21] <lynxman> wip: I'd jump from hardy to lucid for starters, you want to be in the latest LTS
[14:21] <rbasak> zul: the argument (not mine but it makes sense to me) was that Depends: cannot describe a dependency on what is booted so it's the wrong place to define it and the sysadmin should know to do it himself (with an added helpful message if he tries to start a domU without being dom0)
[14:21] <wip> lynxman, i fear the worst, never been lucky with do-release-upgrade...
[14:22] <lynxman> wip: well, be ready, do a backup beforehand, usual precautions :)
[14:22] <yakster> ok, I was matthew… had issue with username….
[14:22] <yakster> ok
[14:22] <zul> rbasak: yeah
[14:22] <yakster> so sendmail…
[14:22] <yakster> I don't see any errors…. really I dont
[14:22] <rbasak> zul: so Daviey asked me to request an SRU but you seem to have beaten me to it :)
[14:23] <zul> rbasak: heh
[14:24] <yakster> this is the only thing that says fail… version=TLSv1/SSLv3, verify=FAIL
[14:24] <wip> anyone did HARDY to new LTS (lucid)?
[14:24] <wip> (do-release-upgrade)
[14:24] <lynxman> wip: I've done it in the past, it works :)
[14:24] <Pici> wip: Yes, whats the issue?
[14:24] <wip> with apache and php mysql ftp pgsql dovecot
[14:24] <lynxman> wip: and more stuff even
[14:24] <lynxman> wip: do a backup and go for it ;)
[14:25] <wip> like i said i never been lucky when doing a do-release-upgrade
[14:25] <wip> lynxman, what is the "good" way for making a back-up. right now i have all my sites, database, but not sure if i have all my apache conf, ftp user conf etc...
[14:26] <lynxman> wip: I'd rsync the whole structure to another disc, also mysqldump your databases, keep a separate copy of /etc as well and of your user data
[14:26] <lynxman> wip: as a paranoid that's the kind of backup I'd do
[14:27] <wip> lynxman, sadly the other disk is full (the back-up disk)
[14:28] <yakster> what address do I need to have on my hostile? just the local private IP and the name associated with it correct?
[14:28] <lynxman> wip: then connect another one? :)
[14:29] <wip> lynxman, good advice sadly i don't want to paid my hosting cie to install a new disk
[14:30] <wip> would it be possible to backup an image of my entire system within the same hd (i have plenty of space)
[14:30] <lynxman> wip: I'd not recommend that unless you have no other choice
[14:31] <yakster> nm I got it….
[14:31] <yakster> thanks all
[14:31] <wip> i have no other choice... sadly - i'm a poor fake sysadmin
[14:36] <wip> is this line ok to backup before do-release-upgrade? tar cvpzf backup.tgz --exclude=/proc --exclude=/lost+found --exclude=/backup.tgz --exclude=/mnt --exclude=/sys /
[14:37] <lynxman> wip: do that, then also a separate tar of /etc another of your web content and a mysqldump of your BBDDs (or pg_dump in your case)
[14:41] <RoAkSoAx> kirkland: yo
[14:41] <RoAkSoAx> kirkland: etckeeper is in -proposed
[14:41] <kirkland> RoAkSoAx: cool, thanks!
[14:41] <kirkland> RoAkSoAx: we probably need to talk to the upstream etckeeper devs and see if they'll take this simple patch upstream
[14:41] <RoAkSoAx> kirkland: the ping was about that bte
[14:41] <RoAkSoAx> btw*
[14:41] <wip> lynxman, should i change release-upgrades - Prompt=lts?
[14:42] <RoAkSoAx> kirkland: alright
[14:42] <kirkland> RoAkSoAx: i predict a bikeshedding discussion about *where* this should be fixed
[14:42] <kirkland> RoAkSoAx: they're going to say "fix it in bzr", and bzr is saying "fix this in etckeeper"
[14:42] <lynxman> wip: I'd follow this doc, it's quite good (I followed it myself) https://help.ubuntu.com/community/LucidUpgrades#Upgrade_from_8.04_LTS_to_10.04_LTS
[14:43] <wip> lynxman, THX!!
[14:43] <w00> Anyone did a release-upgrade from natty to oneiric yet?
[14:43] <RoAkSoAx> kirkland: agreed, but the simplicity of the patch gives us good chances, at it makes sense to set defaults as well
[14:44] <kirkland> RoAkSoAx: heh, i wish i just had to convince you :-)
[14:45] <lynxman> kirkland: hey o/
[14:45] <wip> lynxman, you may need to edit /boot/grub/menu.lst and change the default boot kernel to the newly installed 10.04 kernel: did you have to? i don't have access to the console...
[14:46] <lynxman> wip: I didn't need to, but your mileage may vary
[14:46] <kirkland> lynxman: howdy
[14:50] <RoAkSoAx> kirkland: lol xD
[14:50] <wip> anyone have any recommandation before doing a do-release-upgrade on hardy: i will backup using: tar cvpzf backup.tgz --exclude=/proc --exclude=/lost+found --exclude=/backup.tgz --exclude=/mnt | dump my mysql and pgsql database | bk /etc | bk manywebsites | ask for a kvm/ip (you may need to edit /boot/grub/menu.lst - If this step is not performed your server may attempt to boot into the 8.04 LTS kernel and will hang.)?
[14:52] <wip> and i am using dovecot on hardy...
[14:58] <koolhead17> lynxman: i am testing most of the nova commands and all of them are working , what would you suggest :P
[14:59] <lynxman> koolhead17: keep testing? :o)
[15:00] <w00> Please report this as a bug and include the files /var/log/dist-upgrade/main.log and /var/log/dist-upgrade/apt.log in your report. The upgrade has aborted.
[15:00] <w00> :[
[15:01] <koolhead17> lynxman: ok. sir :D
[15:05] <w00> Where should i report the bug more exactly?
[15:28] <w00> launchpad timeout, awesome, no upgrade for now then
[15:35] <___MAX> hi, i have toshiba disk drive with usb cable i wont to install ubuntu on it with persistence
[15:40] <kirkland> SpamapS: ping
[15:40] <___MAX>    can i install persistence ubuntu on 80 disk drive  with usb cable
[15:40] <SpamapS> kirkland: pong!
[15:41] <___MAX> any idea !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!11
[15:45] <smoser> jamespage, still around ?
[15:46] <jamespage> smoser: yep
[15:47] <smoser> i ran a test like : execute_ubuntu_ec2_test.py simple-user-data --tests=./ec2-automated-tests --release=oneiric --storage=ebs --region=us-east-1 --arch=amd64 --debug
[15:47] <smoser> having modified global/archs.yaml to say 't1.micro' for default size
[15:48] <smoser> the instance went to started, then to stopped, then started, but then back to stopped
[15:48] <smoser> it did not go to terminated
[15:49] <jamespage> logfile?
[15:49] <jamespage> ah - are you running that on natty or oneiric?
[15:49] <jamespage> started-stopped-started-terminated would be normal
[15:49] <smoser> http://paste.ubuntu.com/707408/
[15:50] <smoser> running a test of oneiric on my oneiric machine
[15:50] <smoser> that log is cleaned a bit... i tried to strip out boto debug
[15:51] <smoser> and it just now happened again
[15:51] <jamespage> smoser: did boto get an upgrade this release? I remember having todo something hacky with terminate/stop before
[15:51] <jamespage> infact I know it did
[15:52] <jamespage> fwiw I've only ever run the ec2-tests from natty
[15:52] <jamespage> so that might be making a difference.
[15:53] <smoser> boto did upgrade 1.9b-1ubuntu5 in natty, and 2.0-0ubuntu1 in oneiric
[15:55] <jamespage> smoser: it looks like that part of the api has changed - there is now a terminate method which did not exist before
[15:55] <SlimG> PHP complains that it cannot find the imagecreate() function I'm using, phpinfo() says GD is enabled, ubuntu server 10.04 with apache2 and php5
[15:55] <smoser> right.
[15:56] <smoser> jamespage, i'll try to monkey patch the test suite to account for that
[15:56] <SlimG> What do I need to do/install to get the php function imagecreate() ?
[15:56] <jamespage> smoser: ack
[15:56] <smoser> jamespage, doc merge proposal at https://code.launchpad.net/~smoser/ubuntu-server-ec2-testing/doc-updates/+merge/79304
[15:57] <jamespage> smoser: thanks for that - it really needed doing
[15:57] <jamespage> smoser: the stop command for ebs instances in PHASE_3 could also be updated as well
[15:57] <jamespage> I'm happy to move the minimum release requirement forwards BTW
[15:58] <jamespage> its not hassle
[16:05] <smoser> jamespage, so we think that boto.ec2.instance.stop() on natty would terminate ?
[16:05] <smoser> and on oneiric it 'stop'
[16:05] <smoser> right?
[16:14] <donspaulding> Can I have upstart spawn/respawn multiple processes from a single .conf?
[16:23] <smoser> jamespage, ok. i think we're good now.
[16:23] <smoser> i have 3 things i'd like merged
[16:23] <smoser>  https://code.launchpad.net/~smoser/+junk/ec2-automated-test/
[16:23] <smoser> err.. lets number them.
[16:23] <smoser> 1.) https://code.launchpad.net/~smoser/+junk/ec2-automated-test/ . I can't seem to submit a merge proposal for that at all, so you'll just have to review, and pull it in.
[16:24] <smoser> 2.) doc fixes : https://code.launchpad.net/~smoser/ubuntu-server-ec2-testing/doc-updates/+merge/79304
[16:24] <smoser> 3.) other fixes : https://code.launchpad.net/~smoser/ubuntu-server-ec2-testing/fixes/+merge/79174
[16:27] <donspaulding> I see where I can run a separate pre-start stanza to run a process, but can I exec a command in that stanza and have upstart manage it?  Will Upstart respawn the commands executed in the pre-start stanza if they terminate abnormally?
[16:41] <smoser> donspaulding, http://paste.ubuntu.com/707441/
[16:42] <smoser> i put that in /etc/init, and then ran 'sudo start testme'.  that wrote to /tmp/testme.log.  I got the pid that was in /tmp/testme.log and killed it (sudo kill).
[16:43] <smoser> and after the kill, post-stop, pre-start, start, and post-start all get called
[16:43] <donspaulding> smoser: my question regards exec'ing a daemon within the pre-start.
[16:44] <smoser> oh.. i see. i thoguht you were just asking if it would run pre-start on 'respawn'
[16:45] <smoser> i would doubt that it would handle the respawn of jobs started in a pre-start.
[16:45] <donspaulding> smoser: yeah, I can see where my question could look like that.  Unfortunately, I'm trying to screw with a hammer, I think.
[16:45] <donspaulding> smoser: I think I'm just going to "do it the Upstart way" and put the two daemons in different .confs
[16:57] <koolhead17> why all sudden channel went in silence :P
[16:59] <josePhoenix> Shhhh.
[16:59]  * donspaulding grabs the popcorn
[17:00]  * koolhead17 just had nice dinner!!
[17:01] <donspaulding> we're watching a movie
[17:01] <donspaulding> http://i.imgur.com/tCp90.gif
[17:01] <storrgie> how can I get a script to run at startup? Its daemonized... I have /etc/rc.d/init.d/mpdscribble ready to go
[17:02] <josePhoenix> symlink into the runlevel you want it, like old-style initscripts
[17:02] <josePhoenix> or write an upstart script for it
[17:03] <storrgie> josePhoenix, I remember there was a program i could call if I had an initscript, something like: program myscript defaults
[17:03] <storrgie> i forget what its called though
[17:03] <koolhead17> donspaulding: scary movie!! :P
[17:03] <donspaulding> storrgie: update-rc.d
[17:03] <josePhoenix> storrgie: update-rc.d? I don't remember if ubuntu installs it by default these days
[17:03] <storrgie> I'm actually in fedora right now, I just usually ask questions in here
[17:04] <storrgie> i believe you're right it is update-rc.d
[17:16] <josePhoenix> storrgie: you may get different answers here ;]
[17:19] <storrgie> josePhoenix, right. I did find it out though. I needed to use systemctl
[17:19] <storrgie> I'm not sure if thats in ubuntu, I think ubuntu uses upstart right?
[17:21] <josePhoenix> storrgie: it's kind of a confusing landscape in ubuntu :\
[17:22] <josePhoenix> upstart scripts are installed with services, but there's no "right" way to disable services without removing them altogether
[17:22] <josePhoenix> and to further complicate things, traditional rc#.d/ folders are still supported
[17:25] <storrgie> josePhoenix, this is why I shifted over to fedora... I feel really detached from whats going on in ubuntu. Its frustrating
[17:29] <josePhoenix> Aside from that weirdness I've been very happy with Ubuntu
[17:30] <josePhoenix> even when the packagers make nonstandard decisions, they usually aren't bad ones
[17:31] <koolhead17> storrgie: change is good!! :D
[17:33] <RoyK> koolhead17: I guess that depends what sort of change, the slow migrate to upstart while keeping the sysv scripts isn't too easy to handle for many sysadmins
[17:34] <RoyK> IMHO the jump to upstart was a wee bit hurried
[17:34] <koolhead17> RoyK: things will get much better in LTS :)
[17:34] <storrgie> koolhead17, change is great... arbitrary change is rough though
[17:34] <RoyK> koolhead17: I'm talking about LTS, lucid
[17:35] <koolhead17> RoyK: i meant coming one!! :)
[17:35] <RoyK> things should be _stable_ in an LTS
[17:35] <RoyK> in Hardy it was all fine
[17:35] <RoyK> in Lucid, things got a bit complicated
[17:35] <patdk-wk> ya, too earily upstart migration :(
[17:35] <RoyK> indeed
[17:36]  * RoyK just has to try 11.10 on his home server to get Xen - talking about cutting edge.....
[17:36] <patdk-wk> royk, I'm bad
[17:37] <patdk-wk> I haven't used anything other than maverick and lucid currently
[17:37] <KM0201> i'm gonna put 11.10 on my server in a little bit.
[17:37] <patdk-wk> only lucid on servers
[17:37] <KM0201> i'l go to LTS w/ 12.04, and stay there a while
[17:37] <patdk-wk> but I do test all iso's for all ubutu versions
[17:37] <patdk-wk> just don't use them, just test
[17:37] <patdk-wk> things will get fun when 12.04 comes out
[17:38] <patdk-wk> I'll have to start building 10.04 and 12.04 packages for all my custom stuff
[17:38] <patdk-wk> hoping 12.04 is nice, I'll upgrade to it quickly
[17:38] <RoyK> patdk-wk: for workstations, we upgrade rather often. that may change after 12.04, but then, I beleive we said that before 10.04 came out :P
[17:38] <patdk-wk> I'm still using 10.04 for workstations, (except my netbook, cause I tried maverick on it, just haven't bothered hardly use it)
[17:38] <koolhead17> RoyK: it takes time, after all we are rapidly going there :P
[17:39] <koolhead17> patience :D
[17:39] <RoyK> and then, when 12.10 comes, the users are bound to request an update
[17:39] <patdk-wk> yuk
[17:39] <shennyg> how to I list available versions (including past versions) of a specific program using apt-get?
[17:40] <patdk-wk> I can't stand to upgrade every 6 months, too quick, just geting settled in :)
[17:40] <patdk-wk> shennyg, apt-get doesn't do that
[17:40] <RoyK> still trying to find an easy way to integrate ubuntu into AD, single signon etc
[17:40] <patdk-wk> royk, that is easy
[17:40] <patdk-wk> been doing that since atleast hardy
[17:40] <RoyK> shennyg: apt-cache search will show you the list of current versions
[17:41] <RoyK> patdk-wk: really? the howtos I've come across weren't exactly simple - local user needed and whatnot
[17:41] <shennyg> thanks RoyK
[17:41] <patdk-wk> royk, local user is not needed at all
[17:41] <patdk-wk> atleast the way I do it, bind it into pam
[17:41] <RoyK> then the howto is wrong :P
[17:42] <patdk-wk> I also don't use that *simple* AD intergration howto
[17:42] <patdk-wk> or whatever that package is called
[17:42] <RoyK> I don't think there is one
[17:42] <RoyK> and with a bunch of workstations, you really want a simple one
[17:42] <shennyg> RoyK: that didn't give me versions numbers... just the name.
[17:43] <RoyK> shennyg: man apt-cache :P
[17:43] <patdk-wk> royk, ya, I think it takes about 5min to do it, edit pam, use winbind I think, get kerberos token for machine
[17:43] <shennyg> thanks ;)
[17:43] <patdk-wk> then it should be good
[17:45] <patdk-wk> royk, likewise is the same of that package that is suppost to make it simple
[17:46] <patdk-wk> the one I used as a sample, though my config is alittle different is: https://help.ubuntu.com/community/ActiveDirectoryWinbindHowto
[17:46]  * RoyK took tomorrow off and really cba to worry about windoze integration atm
[17:47] <RoyK> but thanks - I'll bookmark it
[17:47] <dnmons> Hi all. Is it possible to do SSH attack throttling as explained here > http://my.opera.com/TMS/blog/show.dml/194002 using Ubuntu’s ‘Uncomplicated Firewall’?
[17:47] <aaabbbccc> hi everyone
[17:47] <RoyK> dnmons: I just use denyhosts
[17:48] <RoyK> dnmons: a bit more nazi, but works well
[17:48] <patdk-wk> dnsmons, there are many ways to do it using ufw, but they aren't very smart, as they don't know a good vs bad connection
[17:48] <RoyK> dnmons: and then, it's distributed, so the usual suspects are denied early
[17:48] <patdk-wk> therefor I normally use fail2ban
[17:48] <RoyK> patdk-wk: fail2ban < denyhosts
[17:49] <patdk-wk> denyhosts isn't dynamic
[17:49] <patdk-wk> or are you talking about some program?
[17:49] <patdk-wk> not hosts.deny
[17:49] <aaabbbccc> I have a question about dovecot-auth & postfix, after a recent update (I believe) no node is made at /var/spool/postfix/private/dovecot-auth
[17:49] <RoyK> patdk-wk: it uses hosts.deny, but it updates it dynamically for sure
[17:49] <RoyK> patdk-wk: apt-cache search denyhosts :P
[17:49] <dnmons> what are the advantages of fail2ban vs denyhost?
[17:50] <aaabbbccc> it seems 01-mail-stack-delivery.auth is not loaded anymore
[17:50] <dnmons> I liked the method explained in that blog post because it does’t block, just drop packets and slow things down
[17:50] <patdk-wk> who said you had to just block?
[17:50] <patdk-wk> you can do any number of things, only limited by your creativity and time
[17:51] <patdk-wk> but as that blog is not about linux iptables, not that useful
[17:51] <RoyK> patdk-wk: and yes, denyhosts != hosts.deny even though it uses it (as fail2ban may do if you want to)
[17:52] <patdk-wk> ya, I'm not a fan of hosts.deny
[17:52] <RoyK> why not?
[17:52] <patdk-wk> these days I use ipset
[17:52] <patdk-wk> and just have a custom rule chain that triggered off it
[17:53] <RoyK> hosts.deny does the same job only at a higher level
[17:53] <patdk-wk> yep
[17:53] <patdk-wk> higher level, depends on application support for it compiled in, ...
[17:53] <patdk-wk> also only works on the machine the app runs on
[17:53] <RoyK> patdk-wk: how many services do you really check for in fail2ban?
[17:54] <patdk-wk> about 12
[17:54] <patdk-wk> across 30+ servers
[17:54] <RoyK> I get your point....
[17:54] <necromancer> hi guys, i'm currently running a bunch of JeOS 8.04 VMs on this one VM server, and one app i have to deploy can't run on 8
[17:54] <necromancer> what version of Ubuntu is most like 8.04 in terms of footprint?
[17:54] <patdk-wk> royk, the only thing I don't check with it currently is web logs :)
[17:55] <RoyK> necromancer: with lucid (10.04) you have a "minimal virtual machine" choice in the install menu - that should be rather close to JeOS
[17:55] <necromancer> awesome
[17:55] <RoyK> might be a bit bigger, though
[17:55] <patdk-wk> min virtual machine = JeOS
[17:56] <RoyK> patdk-wk: in theory, yes :P
[17:56] <patdk-wk> atleast according to the testing I do every month :)
[17:56] <necromancer> i mean as long as it's really close
[17:56] <patdk-wk> it's suppost to be <500MB
[17:56] <patdk-wk> it's normally in the 450MB about size
[17:56] <RoyK> necromancer: it's the smallest one from ubuntu
[17:56] <RoyK> necromancer: a minimal debian install is likely to be half of that
[17:56] <necromancer> it's easier for my boss to keep track of the server's size and if it needs hardware upgrades
[17:56] <necromancer> RoyK: reeeally
[17:57] <RoyK> debian is very minimal
[17:57] <necromancer> now is debian difficult to configure to be very minimal?
[17:57] <RoyK> necromancer: but really, do you need to go that low?
[17:57] <RoyK> most - even cheap VMs - have a few gigs of space
[17:59] <necromancer> RoyK: our VMs are 128mb RAM, 8GB HD and using 1 vCPU
[18:00] <RoyK> necromancer: you can go a long way with 8GB
[18:00] <necromancer> yeah
[18:00] <RoyK> and 128MB should suffice for a small vm
[18:01] <necromancer> well as long as ubuntu 10 can run on those specs we're good i guess
[18:01] <RoyK> 10.04 will run well on that
[18:01] <RoyK> even a standard install
[18:01] <patdk-wk> smallest ram usage I could make ubuntu 10.04 use is 24megs
[18:04] <RoyK> necromancer: I have a few 64MB VMs
[18:04] <RoyK> works
[18:05] <RoyK> necromancer: and on my home server which has plenty of software installed, it's only using 1,6GB for the root
[18:06] <aaabbbccc> can someone help me with dovecot sasl / mail-stack-delivery
[18:06] <necromancer> RoyK: awesome
[18:06] <aaabbbccc> somehow my mail system stopped working, postfix cannot authenticat imap on dovecot is working fine
[18:07] <aaabbbccc> '/var/spool/postfix/private/dovecot-aut does not exist anymore it looks like 01-mail-stack-delivery.auth is not loaded, but i do not know why...
[18:08] <aibo> hola, where I can current software version list for 10.04 LTS
[18:08] <Pici> aibo: http://packages.ubuntu.com is probably the easiest way.
[18:09] <RoyK> aibo: man apt-cache
[18:09] <aibo> RoyK, i am not on 10.04, thinking about migrating from 11.04
[18:12] <RoyK> aibo: on which version are you?
[18:12] <aibo> I said, 11.04
[18:12] <RoyK> well, my first advice is: If everything works, don't fix it
[18:12] <pmatulis> an XFS filesystem was created on /dev/sdc instead of on /dev/sdc1.  does it matter?
[18:13] <RoyK> pmatulis: hardly
[18:13] <RoyK> pmatulis: unless there were more partitions on the drive :P
[18:13] <aibo> RoyK, I know
[18:13] <RoyK> pmatulis: if there were, they are probably gone
[18:15] <pmatulis> RoyK: nah, there were no other
[18:15] <RoyK> pmatulis: then no harm done
[18:15] <RoyK> pmatulis: sdc now doesn't contain a partition table, but neither needs one
[18:16] <pmatulis> RoyK: ok, makes sense.  thanks
[18:18] <dnmons> another question, where are ufw rules stored? so I can back them up.
[18:19] <patdk-wk> pmatulis, never put that drive into an openfiler system though :)
[18:19] <patdk-wk> atleast the of I used would automatically put a gpt on every disk, even if you didn't want one
[18:19] <patdk-wk> dnmons, /etc/ufw
[18:19] <patdk-wk> you should backup /etc though
[18:20] <RoyK> patdk-wk: no reason for a partition table on drive unless you want to split it
[18:21] <patdk-wk> royk, correct, as long as no other programs think it should have one, and installs one :)
[18:21] <dnmons> patdk-wk: ufw allow 25
[18:21] <dnmons> I expect to se port 25 somewhere in that directry.
[18:23] <RoyK> patdk-wk: if you use a program that automatically installs a partition table on a drive, then either you or that program is pathologically incorrect :P
[18:24] <patdk-wk> royk, yep, that is why my tests with openfiler only lasts a day :)
[18:24] <RoyK> hehe
[18:26]  * koolhead17 is back homr
[18:26] <koolhead17> e
[18:26] <RoyK> omg - duck and cover
[18:34] <jdstrand> dnmons: rules added via the ufw cli are in /lib/ufw
[18:37] <patdk-wk> oh? they got moved
[18:38] <RoyK> shouldn't those to /etc somewhere?
[18:40] <koolhead17> jdstrand: hey
[18:43] <jdstrand> hello
[18:43] <dnmons> jdstrand: thanks.
[18:43] <dnmons> can confirm that they are there in 11.04
[18:51] <Pici> hrm.. my byobu status widgets are gone.
[19:06] <kirkland> smoser: hey, meat mjfork
[19:06] <kirkland> meet, even
[19:06] <kirkland> smoser: he has some questions about cloud-images.ubuntu.com
[19:07] <mjfork> they are the cloud-init based images that are used in EC2 as well as other envs?
[19:10] <mjfork> smoser: gotta step away, back after bit
[19:11] <marshall> hey ubuntu-server
[19:11] <RoyK> hi
[19:11] <marshall> free -m says I only have 18 mb of free memory. how do I see which processes are eating the most memory?
[19:12] <patdk-wk> that is normal
[19:12] <patdk-wk> what EXACTLY is the output of free -m
[19:12] <RoyK> marshall: free memory is memory not used by the system. you want the system to use memory for caching....
[19:13] <marshall> http://pastebin.com/2aEC6YdC
[19:13] <patdk-wk> 558megs free
[19:13] <marshall> huh...
[19:13] <marshall> okay, thanks
[19:13] <RoyK> marshall: free also gives you memory not allocated, used for buffers etc
[19:14] <RoyK> marshall: the kernel doesn't allocate that memory, it just uses it for caching/buffering, which is good
[19:14] <marshall> ah
[19:14] <marshall> thanks RoyK, patdk-wk
[19:15] <patdk-wk> basically, the ONLY thing that uses *free* memory, is interrupt drivers, to allocate memory.
[19:16] <patdk-wk> At any other time, it would dump some cache, to and use that memory instead
[19:16] <patdk-wk> as the cache is all unused memory, used temporarilly for speeding up disks
[19:18] <RoyK> damn - 25Mbps is slow - I should upgrade
[19:19] <Jasonn> RoyK: I have 100/1
[19:19] <Jasonn> \o/
[19:19] <RoyK> Jasonn: 25/25
[19:19] <RoyK> atm
[19:19] <Jasonn> where you from?>
[19:19] <RoyK> a phonecall and I'll get 60/60 for NOK 100 more
[19:19] <Jasonn> NOK?
[19:19] <KM0201> anybody know the package name for mt-daapd .... tried to install it, and i get a not found
[19:19] <RoyK> norwegian kroner
[19:20] <Jasonn> RoyK: in USD?>
[19:20] <RoyK> google it
[19:20] <Jasonn> 17.7073 U.S. dollars
[19:20] <Jasonn> oh
[19:20] <RoyK> about $100 for 60Mbps symmetric
[19:20] <Jasonn> go for it :D
[19:20] <Jasonn> :o
[19:20] <Jasonn> I pay $52ish for 100/1
[19:20] <Jasonn> and its dedicated to me
[19:20] <Jasonn> thats the difference
[19:20] <Jasonn> that line is not dedicated to you
[19:21] <Jasonn> meaning
[19:21]  * RoyK likes a good uplink
[19:21] <Jasonn> oh
[19:21] <Jasonn> I have a 10gbit dedi for that :p
[19:21] <RoyK> this is at home
[19:21] <RoyK> at work, we have more bandwidth
[19:22] <RoyK> 60Mbps for $90 or so isn't really bad
[19:22] <RoyK> _symmetric_
[19:31] <garo> I did a upgrade to 11.10 with 'do-release-upgrade'. When it was creating a initrd in /boot it crashed because my /boot was full
[19:31] <jeh> I'm attempting to provision natty via cobbler, but I can't seem to get a working preseed file.  Can anyone direct me to one?
[19:31] <garo> So i deleted some old kernels, did a 'aptitude full-upgrade' and this time it worked
[19:33] <garo> but was the creating of the initrd really the last step in the upgrade proces ? (if not, which other steps should i do)
[19:38]  * KM0201 sighs
[19:42] <RoyK> garo: bug #873198?
[19:43] <KM0201> man, i hate forked-daapd
[19:43] <garo> grub seems fine here
[19:44] <garo> it actually looks like everything works ok, but most bugs aren't immediately visible
[19:46] <garo> what i fear is that the upgrade proces was planning to write some configfiles after making /boot/initrd.img-3.0.0-12-server
[19:47] <garo> if those configfiles are related to things that are almost never used then it will take a long time before the bugs are visible
[20:04] <mtaylor> kirkland: ping
[20:06] <mtaylor> kirkland: orchestra server feature request: orchestra should be able to run an ldap server which acts as a debconf ldap database source ... so that you can preconfigure stuff systemically across everything you're spinning up (I got to this from looking at how devstack and juju both install mysql)
[20:06] <mtaylor> which is a little assinine
[20:10] <dkn> if i have an lvm on a single disk, unmount the lvm, and pull that disk from the hot swap bay, how can i tell the computer where the find the disk after it's reinserted?
[20:21] <KM0201> anybody know why rhythmbox wouldn't see a UPNP share?  i can see the share fine w/ other apps.
[20:21] <KM0201> how to install mt-daapd on 11.10... forked-daapd seems buggy
[20:26] <kudu> I'm trying to use mailman with postfix
[20:26] <kudu> I'm not receiveing emails on the mailman side
[20:38] <zul> mtaylor: database source?
[20:39] <koolhead17> zul: is nova shipped with oneiric is diablo-milestone?
[20:40] <zul> koolhead17: diablo final release+backported patches
[20:40] <koolhead17> zul: why is role based access control not working correctly then. :(
[20:41] <koolhead17> i have allready spent 48 hours on it wonder what is next :(
[20:41] <zul> koolhead17: no idea i never used it, you are talking about keystone right?
[20:41] <koolhead17> zul: nopes
[20:41] <zul> koolhead17: never used it
[20:41] <koolhead17> nova role based access control
[20:41] <koolhead17> okey. have to check it again for 20th time then
[20:42] <koolhead17> seems like something is there to look upon.
[20:49] <mtaylor> zul: yes. debconf can be told to get answers to its questions from an ldap server
[20:49] <kirkland> mtaylor: yo
[20:50] <mtaylor> kirkland: yo! I'm all making suggestions about shit again
[20:50] <kirkland> mtaylor: ooh, that's interesting
[20:50] <kirkland> mtaylor: orchestra ldap server
[20:50] <mtaylor> yup.
[20:50] <kirkland> mtaylor: you have a favorite ldap server in ubuntu?
[20:50] <mtaylor> could be helpful/useful
[20:51] <mtaylor> I do not ... I don't really know much about anything that isn't openldap
[21:15] <boxybrown> any pam experts?  I have a question about autocreating home directories
[21:15] <boxybrown> this has a pretty good description of it: https://help.ubuntu.com/community/LDAPClientAuthentication
[21:15] <boxybrown> but I have a question about what happens if the home directory is an NFS mount and it happens to not be mounted when a user logs in...
[21:34] <boxybrown> All the sudden I've started getting these errors when I try to access the GUI:
[21:34] <boxybrown> TGT NOT FORWARDABLE:
[21:34] <boxybrown> so the gui complains "did not receive kerberos credentials"
[21:43] <boxybrown> oops, just realized those last posts were to the wrong room... my bad
[21:44] <ersi> No harm done :)
[21:44] <aaabbbccc> him I am still struggeling with mail-stack-delivery, SMTP SASL authentication fails, IMAP is accessible. I believe after an update /var/spool/postfix/private/dovecot-auth disappeared /etc/dovecot/auth.d/01-mail-stack-delivery.auth  is not included anywhere I believe
[22:04] <aaabbbccc> I really do not understand why the socket at /var/spool/postfix/private/dovecot-auth suddenly disappeared, the /etc/dovecot/auth.d section seems to be includen nowhere, and that's the only place dovecot-auth is mentioned (except foor in postfix' config of coursr)
[22:10] <wip> back home, time to do a do-release-upgrade from hardy to lucid... wish me luck!
[22:14] <wip> oh... This session appears to be running under ssh. It is not recommended to perform a upgrade over ssh currently because in case of failure it is harder to recover. If you continue, an additional ssh daemon will be started at port  '9004'. - Should i wait for the kvm/ip?
[22:14] <wip> but tomorrow is friday night... i wish i could party not do-release-upgrade...
[22:16] <JanC> wip: the additional ssh daemon should help with recovering, and in any case I hope you're running the upgrade script in a screen session?
[22:16] <wip> JanC: what do you mean "screen session" - i am connected from my ubuntu box via ssh to my server located somewhere in the world
[22:17] <JanC> wip: do you use 'screen' or 'byobu' on the server?
[22:17] <aaabbbccc> I believe /etc/dovecot/auth.d/01-mail-stack-delivery.auth should be included in dovecot's auth default {....} section right?
[22:17] <JanC> (byobu being an andvanced config for screen, actually)
[22:18] <wip> JanC: i'm using Gnome Terminal... not sure if it's a stupid answer
[22:18] <JanC> wip: I mean server-side
[22:19] <wip> JanC: hardy default via ssh
[22:19] <wip> JanC: no gui just pure ssh
[22:19] <wip> what are the chance of successfully do-release-upgrade from hardy to lucid (i'm choking)
[22:20] <wip> > 60% or < 40%
[22:20] <JanC> not sure
[22:21] <JanC> but using screen on the server will certainly help a lot
[22:21] <wip> what i fear the most: you may need to edit /boot/grub/menu.lst and change the default boot kernel to the newly installed 10.04 kernel. If this step is not performed your server may attempt to boot into the 8.04 LTS kernel and will hang.
[22:22] <JanC> if you run do-release-upgrade from the plain ssh session, it will most likely be killed when the connection gets lost, leaving you with a halfway upgraded system...
[22:23] <JanC> wip: that sounds like something you can check before reboot
[22:24] <wip> JanC: yes
[22:24] <JanC> wip: and you should certainly read about screen if you don't know it yet  ☺
[22:24]  * wip is reading http://www.howtoforge.com/linux_screen
[22:25] <JanC> (and 'byobu' too, but I don't think it existed at hardy time)
[22:25] <JanC> 'byobu' makes 'screen' easier to use
[22:26]  * kirkland high fives JanC ;-)
[22:27] <aaabbbccc> sorry to be a bit of e nuisance, but I cannot fix my postfix/dovecot setup, /var/spool/postfix/private/dovecot-auth does not get created since today
[22:27] <aaabbbccc> so SASL does not wrk for postfix
[22:27] <JanC> wip: kirkland is byobu's main author  ;-)
[22:28]  * wip high fives kirkland
[22:28] <JanC> kirkland: I think byobu isn't available in the hardy repositories, right?
[22:29] <JanC> in any case, plain screen should be
[22:30] <JanC> 9.04
[22:31] <kirkland> JanC: wip: there is an older, backport of byobu available in ppa:byobu/ppa
[22:31] <kirkland> JanC: wip: it's not the latest/greatest, but it should be functional
[22:34] <JanC> kirkland: that reminds me, I have a little "script" in ~/.byobu/bin that might be useful for other people too; it just runs 'wget -q -O - http://www.myexternalip.com/raw', which fetches the "external" IP when you are behind a NAT (sometimes that's useful...)
[22:35] <JanC> not sure if that would be useful as one of the default status items?
[22:35] <JanC> (default = included)
[22:36] <kirkland> JanC: oh, neat
[22:36] <JanC> not sure if querying a third party without asking is a good idea anyway
[22:37] <JanC> and that site only supports IPv4
[22:37] <kirkland> JanC: i could make that a configuration option
[22:38] <kirkland> JanC: as currently, the ip_address item supports a few different methods to get that
[22:38] <aaabbbccc> for dovecot/postfix/sasl with mail-stack-delivery installed should I just include auth.d/*.auth in conf.d/10-master.conf?
[22:38] <kirkland> JanC: would you file a bug against byobu with that info?
[22:38] <kirkland> JanC: i know how we could do that cleanly
[22:40] <JanC> what do you want the bug report named?  ☺
[22:48] <JanC> kirkland: https://bugs.launchpad.net/byobu/+bug/873736
[22:51] <kirkland> JanC: cheers