[00:01] <cane> Hello
[00:02] <cane> I need help converting the default LVM install to RAID0
[00:04] <cane> no one here today?
[00:07] <funkyHat> Hello people!
[00:08] <cane> funkyHat, hey
[00:08] <funkyHat> Has anyone set up 2 exim servers with one configured as a "satellite"?
[00:09] <persia> I think most folks recommend postfix for new setup.
[00:10] <funkyHat> I already have a complex exim setup on the "server" side (i.e. not the satellite)
[00:10] <funkyHat> And gmail is managing to authenticate with that using the same transport and auth as I've set up for my satellite
[00:10] <funkyHat> But for some reason the satellite fails to auth correctly
[00:10] <persia> Ah, then the "most folks" situation doesn't apply to you :)
[00:11] <cane> Anyone know how to migrate from LVM to RAID0
[00:18] <Sp3c1alK_> If anyone has a static ip setup could you copy and paste your /etc/network/interfaces file onto pastebin for an  example?
[00:20] <lil_cain> http://pastebin.org/157951
[00:20] <lil_cain> you don't need the post-up bit
[00:21] <cane> Anyone know how to migrate from LVM to RAID0
[00:21] <Sp3c1alK_> ah, I see you have "network" listed but not "gateway"
[00:23] <lil_cain> Sp3c1alK_: You probably want to add a gateway
[00:23] <lil_cain> that network doesn't have a default gateway
[00:24] <Sp3c1alK> what's the differnce between network and gateway?
[00:24] <lil_cain> Gateway gives you a default route.
[00:24] <Sp3c1alK> right, but what abou8t network
[00:25] <lil_cain> Specifies the start address of the network, I'd imagine. I think it's more useful for people reading that anything else.
[00:25] <lil_cain> ah, network is required, apparetly if you're using a 2.0 kernel
[00:26] <lil_cain> It's probably legacy cruft in our case.
[00:51] <cane> Anyone know how to migrate from LVM to RAID0
[03:48] <swift> hi guys, I have a router with no snmp agent on it... ive installed mrtg on my ubuntu server...also installed snmpd.. how can i monitor this router internet traffic?
[03:49] <swift> this line doesn'
[03:49] <swift> thie line doesn't have a static IP.. that's the problem
[03:50] <swift> i have a virtual ppp1 interface representing this line... can i use the interface name in 'cfgmaker' to monitor it?
[04:23] <arrrghhh> hey all, does anyone use rtorrent with xmlrpc-c support?  i'm having trouble compiling it, on the surface it seemed like everything compiled correctly but when i try to run the "rtorrent" command I get an error that makes it look like libtorrent didn't compile correctly... but i can't see why.
[04:25] <arrrghhh> nvm, sudo ldconfig fixed my issue.  i'd like to know why... but i'm glad it's fixed
[04:46] <swift> guys, if a router doesn't have an snmp agent, can i use snmpd to create snmp traffic from that interface?.. that way mrtg can monitor it
[04:47] <swift> also, this internet line doesnt have a static IP
[04:47] <swift> please advise
[04:51] <darkk^> swift, is it ubuntu-based router or some "hardware router" ?
[04:52] <swift> the router is a separate hardware...
[04:52] <swift> but it is represented by a virtual interface on ubuntu...
[04:52] <swift> forgive my terms.. but it's just a modem... all routing is done via the ubuntu server
[04:53] <swift> routing and authentication.... basically there are 2 internet lines coming from two modems... and ubuntu is the router... so, all traffice passes via ubuntu outside
[04:55] <darkk^> I assume, it's possible to use interface name in snmpd. I don't know how (I've not deployed snmp myself yet), but I see nothing wrong with that.
[04:55] <flyback> ok
[04:55] <flyback> ubuntu server box
[04:55] <flyback> 8.x lts whatever it is
[04:55] <flyback> console only
[04:55] <flyback> upsmon, perl, mabye some bare bones compiling tools
[04:55] <flyback> what roughly should I be looking at disk space wise
[04:56] <flyback> basically it monitors a ups and then ssh logs into 30 hosts and executes a vmware esx/esxi or server shutdown script
[04:56] <flyback> I think I am around 700 meg or so so I could almost get away with 1gb flash
[04:57] <flyback> but need some for swap
[04:57] <flyback> so I dunno
[11:08] <KristianDK> I'm sorry if its a silly question, but if i have an UEC and one of the nodes crashes entirely, what will happen to the VM's on the server?
[11:21] <binBASH> KristianDK: Btw. I managed to get it fully working at hetzner ;)
[11:22] <KristianDK> binBASH, How!? :D
[11:22] <binBASH> KristianDK: Just configure Eucalyptus Network setup to SYSTEM
[11:22] <KristianDK> i even spoke to them today, and asked different stuff and they just said it wasnt possible
[11:22] <binBASH> and then I installed a dhcpd on each node
[11:22] <binBASH> which provides the 3 additional ips
[11:23] <KristianDK> oh, then it can distribute everything?
[11:23] <binBASH> yup
[11:23] <binBASH> I also have vnc running for the vms
[11:23] <KristianDK> but still, if that server crashes - all of the VMs with IP from it will die, right?
[11:23] <KristianDK> nice :D
[11:23] <binBASH> yup
[11:23] <binBASH> all vms will get los
[11:23] <binBASH> t
[11:24] <KristianDK> really? Isn't there any backup?
[11:24] <binBASH> no
[11:24] <binBASH> it's not persistent image
[11:24] <KristianDK> but why not store the data on walrus? that's distributed, right?
[11:24] <binBASH> you can do that
[11:24] <binBASH> that data will be persistent
[11:25] <binBASH> Though it's slow with hetzner ;)
[11:25] <binBASH> only 100 Mbit
[11:25] <KristianDK> yeah, but you can actually get a 1gbit router :)
[11:26] <binBASH> yup
[11:26] <binBASH> you just need to pay a lot more ;)
[11:26] <KristianDK> they didnt provide me with the price, however :P
[11:26] <binBASH> I have it
[11:26] <KristianDK> how much is it?
[11:26] <binBASH> You need flexipack for each server which is additionally 15 Eur / Month
[11:26] <binBASH> and a 2nd nic
[11:27] <binBASH> then you have to let them move all nodes in hetzner datacenter so they are located at each other
[11:27] <binBASH> costs 69 Eur / Machine
[11:27] <KristianDK> lol
[11:27] <KristianDK> and then how much for the router itself?
[11:29] <binBASH> sec, I'm searching this
[11:29] <KristianDK> ok :P
[11:29] <binBASH> for 8 Port Gbit it's 59 Eur once
[11:29] <binBASH> and for 24 Port it's 199 Eur once
[11:30] <KristianDK> well, its not that bad
[11:30] <binBASH> just the relocation is bad ;)
[11:30] <KristianDK> btw, as you are mentioning the flexipack for the router - you don't use the raid controller?
[11:30] <KristianDK> yeah, but if you order the new ones together, you can request them in the same datacenter, they said :)
[11:31] <binBASH> yup
[11:31] <KristianDK> and the rest of my servers are in RZ10 anyway
[11:31] <KristianDK> you don't use raid?
[11:31] <binBASH> I don't have flexipack at all
[11:31] <binBASH> I have all systems software raid 0
[11:32] <binBASH> each node has a secondary one for backup ;)
[11:32] <binBASH> at least for the storage partitions ;)
[11:32] <KristianDK> hehe, ok :D
[11:32] <binBASH> glusterfs does this
[11:33] <binBASH> KristianDK: http://78.46.23.157/
[11:33] <KristianDK> well, i guess software raid can do the trick too
[11:33] <binBASH> See cherokee webserver runs on a vm ;)
[11:33] <KristianDK> nice :D
[11:34] <binBASH> If you want details about network config per node just tell me
[11:34] <KristianDK> I'd love to have that, because I'm certainly gonna try it out :D
[11:35] <binBASH> ok sec, I'll pastie it
[11:36] <KristianDK> awesome :)
[11:36] <binBASH> KristianDK: http://www.pastie.org/926849
[11:37] <binBASH> you have to set net.ipv4.ip_forward=1 in /etc/sysctl.conf as well
[11:40] <KristianDK> ok :) btw, what is the different IPs in terms of hetzner? The .129 gateway, is this your own machine?
[11:41] <KristianDK> i gues .151 is the normal ip of the server?
[11:41] <binBASH> yup
[11:41] <binBASH> 160-162 are the 3 additional ones
[11:42] <binBASH> it's possible to use those in the vms without special nat setup
[11:42] <KristianDK> ok, so you have only 4 ips in total, right?
[11:42] <binBASH> 4 ips / Nodes
[11:42] <binBASH> the main ip is not usable within the vm
[11:43] <KristianDK> the .151, correct?
[11:45] <binBASH> yup
[11:46] <binBASH> I will use it though
[11:46] <binBASH> it's better to run varnish not in the vm for speed :)
[11:47] <KristianDK> i can understand that you are using the cloud for web projects too :D
[11:48] <KristianDK> btw - this line - what does the .129 and .129 mean? up route add -net 78.46.22.128 netmask 255.255.255.192 gw 78.46.22.129 eth0
[11:48] <binBASH> it's the default route which goes through the hetzner router
[11:48] <KristianDK> how would i translate this into my own setup?
[11:48] <binBASH> but you have this line already in the config
[11:49] <KristianDK> it will be there from the beginning?
[11:49] <binBASH> you just need to add routes for the additional ips
[11:49] <binBASH> yup
[11:49] <KristianDK> ahh, cool :D
[11:49] <binBASH> the br0 routes you need to add
[11:49] <KristianDK> yeah
[11:51] <KristianDK> in the br0, why is the netmask .255 in the end?
[11:51] <KristianDK> and not 192
[11:52] <binBASH> Well I took that myself from another tutorial KristianDK :)
[11:52] <KristianDK> hehe, ok :D just seemed weird to me :P
[11:53] <binBASH> just can tell you it works for me like this ;)
[11:53] <KristianDK> hehe :P
[11:53] <KristianDK> what kind of web apps are you going to host btw?
[11:53] <binBASH> gettyimages.com
[11:53] <binBASH> and some subsites
[11:54] <binBASH> things like gettyimageslatam.com
[11:54] <binBASH> our company provides software for image agencies and press content providers.
[11:55] <KristianDK> ahh, cool :)
[11:55] <KristianDK> why does it say .aspx in the end of some of the files? :P
[11:55] <binBASH> KristianDK: Well currently we host only gettyimageslatam.com from Getty Images
[11:55] <binBASH> others will follow
[11:56] <KristianDK> ahh, ok - but the varnish is for php then?
[11:56] <binBASH> varnish is a caching proxy
[11:56] <binBASH> it can cache all kind of files
[11:57] <binBASH> the goal is to setup a geoip based system which distributes content from serverclusters which are spread world wide
[11:57] <KristianDK> yeah, i know - but sometimes you can interact with it from the scripts - at some point i saw some tutorials with it
[11:57] <KristianDK> thats cool :-D
[11:58] <binBASH> yeah, you can purge the objects from scripts for example
[11:58] <binBASH> so it will reload the cached objects from the source
[11:58] <binBASH> Must have for image changes ;)
[11:59] <KristianDK> hehe, yeah, i guess they would normally cache for quite some time
[11:59] <binBASH> Don't want to serve old thumbnails when user uploaded new pic for example ;
[11:59] <KristianDK> true :P
[12:00] <KristianDK> anyway, my macbook pro is running out of battery - i'll see you around later ;)
[12:00] <binBASH> see ya ;)
[12:00] <binBASH> I have free day now.
[12:29] <e-DIO-t> yo!
[12:42] <e-DIO-t> !landscape
[13:25] <Italian_Plumber> how much of a threat, really, are viruses contained in Powerpoint presentations, i.e. ones passed around in email forwards?
[13:26] <RoyK> for the mail server?
[13:26] <Italian_Plumber> for the pc
[13:26]  * RoyK points to channel name
[13:28] <Italian_Plumber> oh sorry wrong channel
[13:43] <lau> hello, I am running 8.04 and have an issue with apparmor
[13:43] <lau>  $ sudo /etc/init.d/apparmor restart
[13:43] <lau> Reloading AppArmor profiles Error: Could not allocate temporary file. Profile /etc/apparmor.d/usr.sbin.mysqld failed to load
[13:43] <lau> : Failed.
[13:44] <lau> http://paste.ubuntu.com/418567/ any idea ?
[13:44] <zul> jdstrand: ^^^
[13:47] <ivoks> brb
[13:49] <ScottK> zul: Would you have a chance to upload ivok's dovecot-postfix changes today?
[13:49] <ScottK> If not, I'll do it tonight.
[13:49] <zul> ScottK: not yet
[13:50] <zul> ScottK: but I was going to look at it today
[13:50] <ScottK> zul: That'd be great.  Thanks.
[13:50]  * ScottK is busy with $WORK, car repair, and fixing clamav today.
[13:52] <ijhhonuibjhn> has anyone had issues with Ubuntu Server running on hyper-v won't show more than ~139GB for hard drive size?
[13:53] <ijhhonuibjhn> I have a 500GB VHD attached but Ubuntu sees is as around 139GB and can't find anyone else seeing a similar situation
[14:16] <zul> smoser: ping can yo have a look at #564355?
[14:18] <smoser> bug 564335
[14:18] <zul> smoser: crappers https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/564355
[14:44] <ivoks> zul: could you push this one?
[14:44] <ivoks> zul: https://bugs.edge.launchpad.net/ubuntu/+source/drbd8/+bug/562832
[14:45] <zul> ivoks: i have it open in my web browser already,,,once I get through the daily traige Ill get to it
[14:45] <ivoks> zul: awesome, thanks
[14:55] <flyback> flyback> ubuntu server box
 8.x lts whatever it is
 console only
 upsmon, perl, mabye some bare bones compiling tools
 what roughly should I be looking at disk space wise
 basically it monitors a ups and then ssh logs into 30 hosts and executes a vmware esx/esxi or server shutdown script
 I think I am around 700 meg or so so I could almost get away with 1gb flash
[14:55] <smoser> kirkland,
[14:55] <smoser> https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/564355
[14:57] <da65> anyone using liferay?
[15:03] <lau> I found a work around in order to set up a mysql multi instance on my hardy server
[15:03] <lau> But I had to disable apparmor mysql :(
[15:04] <lil_cain> You should just have to modify the policy.
[15:04] <lau> I tried https://wiki.ubuntu.com/DebuggingApparmor but was not able to set the policy properly
[15:04] <lau> any idea, suggestion appreciated :) http://paste.ubuntu.com/418567/
[15:05] <jdstrand> lau: we need the denied messages from dmesg/kern.log
[15:05] <_ruben> flyback: 1gb flash is pushing it with a default -server install, 2gb should be fine .. when using 1gb you might wanna strip it down after install, or use alternate install cd to do minimal install (not sure how much smaller that'd become)
[15:06] <flyback> no X btw
[15:06] <lau> jdstrand: ok, something like ? kernel: [1146644.138980] type=1503 audit(1271684949.075:317): operation="mknod" pid=7594 parent=7568 profile="/usr/sbin/mysqld" requested_mask="w::" denied_mask="w::" fsuid=113 ouid=113 name="/var/lib/mysql2/mysql/db.frm"
[15:06] <flyback> yeah I was thinking 2 gig might be the safe point
[15:06] <_ruben> default install is around ~700 megs indeed, but you want enough spare space to be able to run upgrade and stuff like that
[15:07] <_ruben> flyback: my fileserver at home runs from 2gb flash
[15:07] <flyback> nice
[15:07] <flyback> this is just a little mini-itx box to monitor a ups
[15:07] <flyback> and send shutdown scripts to some vmware esx/esxi hosts
[15:07] <flyback> and some others
[15:07] <lau> the problem always occured when sudo mysql_install_db --user=mysql --datadir=/var/lib/mysql2/
[15:08] <flyback> thx guys
[15:08] <flyback> oh here's a tip
[15:08] <flyback> they are cheap on ebay also
[15:08] <flyback> http://webdevsys.com/lightsOut.htm
[15:09] <flyback> I already found use for 2 at home and 2-3 at work
[15:09] <lau> my assumption was to add a line like /var/lib/mysql2/ in /etc/apparmor.d/usr.sbin.mysqld
[15:09] <flyback> and if you read carefully they work with non-hp's (minus a few features)
[15:10] <jdstrand> lau: just add the following:
[15:10] <jdstrand>   /var/lib/mysql/ r,
[15:10] <jdstrand> err
[15:10] <jdstrand>   /var/lib/mysql2/ r,
[15:10] <jdstrand>   /var/lib/mysql2/** rwk,
[15:10] <lil_cain> what does the 'k' mean?
[15:10] <jdstrand> lau: then do: 'sudo apparmor-parser -r /etc/apparmor.d/usr.sbin.mysql
[15:10] <jdstrand> lil_cain: allows locking
[15:14] <lau> jdstrand: do you mean http://paste.ubuntu.com/418622/ ?
[15:15] <lau> jdstrand: sudo apparmor_parser -r /etc/apparmor.d/usr.sbin.mysqld returns nothing (echo $? is 0)
[15:16] <jdstrand> lau: that is correct and that is what apparmor_parser should return
[15:17] <lau> jdstrand: with you apparmor mysql config and after restarting apparmor service,
[15:18] <lau> now sudo mysql_install_db --user=mysql --datadir=/var/lib/mysql2/ is ok :)
[15:19] <lau> jdstrand: is it good to add lines like /var/run/mysqld/mysqld2.pid w, /etc/mysql/my.cnf r,
[15:19] <lau> err, jdstrand: is it good to add lines like /var/run/mysqld/mysqld2.pid w, /etc/mysql2/my2.cnf r,
[15:19] <jdstrand> lau: if mysqld needs to use those files-- absolutely
[15:20] <jdstrand> lau: you can also check to make sure it is confined with 'sudo aa-status'
[15:20] <lau> jdstrand: ok, I tried aa_status previously, reading https://wiki.ubuntu.com/DebuggingApparmor
[15:21] <jdstrand> lau: cool, good luck! :)
[15:21] <lau> jdstrand: what is the relevant info in the kernel log above ?
[15:21] <lau> I was not ale to analyze it and extract any substance to solve the issue
[15:22] <fallback> hello, trying to set up a bridge performing NAT for several KVM guests as well as allowing several KVM guests to have external IP's; can anybody help?
[15:22] <lau> except not allowed to create db table
[15:22] <jdstrand> lau: profile="/usr/sbin/mysqld" requested_mask="w::" denied_mask="w::"... name="/var/lib/mysql2/mysql/db.frm"
[15:22] <lau> since it is comming from audit does this meqn it is an apparmor output ?
[15:23] <lau> mean
[15:23] <jdstrand> lau: that tells you what denied it (the profile), what access was needed (requested_mask), what was not provided by the profile (denied_mask) and the file in question (name)
[15:23] <jdstrand> lau: apparmor uses 'audit', yes
[15:25] <jdstrand> lau: in Ubuntu 10.04 LTS, there is a handy apparmor_notify command that can be used to summarize the logs in a more human readable form
[15:26] <jdstrand> lau: in earlier releases, you have to just look at the dmesg/kern.log or if using auditd, /var/log/audit/audit.log
[15:26] <lau> ok, jdstrand, can you point me to a kernel audit reading ? I have no clue how it works and why ? what ?
[15:26] <lau> ok audit is a connex service to dmesg ?
[15:27] <jdstrand> lau: the 'audit' message is generated by the kernel
[15:31] <lau> ok, apparmor uses the 'audit' level message in order to log into /var/log/messages
[15:31] <lau> other services may also use that kernel log level message
[15:31] <lau> that is why if I suspect apparmor to create pbs, I should probably grep audit /var/log/messages
[15:32] <lau> is this good a approach jdstrand ?
[15:33] <jdstrand> lau: kern.log generally-- if you log them to /var/log/messages, that would work too
[15:34] <jdstrand> lau: 'grep audit /var/log/kern.log' should help you a lot
[15:43] <fallback> hello, can i ask about ubuntu-server, running KVM?
[15:44] <lil_cain> Don't ask to ask, just ak.
[15:44] <lil_cain> *ask
[15:46] <fallback> Thanks; i'd like to double-check certain issues i'm having. I'm trying to set up a bridge that allows several guests to have external IP addresses, while also providing NAT for several guests on the internal network range
[15:47] <fallback> the bridge comes up fine, but i can't get internet access to the private guests; should i send the firewall script or something?
[15:50] <ivoks> no you can't, this is fedora channel
[15:50] <ivoks> oh, sorry, wrong channel :)
[15:50] <Pici> s/fedora/ubuntu/
[15:50] <ivoks> fallback: ok, so KVM should be bridged, while others should be nated?
[15:53] <fallback> several guests are in a private range in a load-balancing configuration; access should be granted from the outside using port forward to the virtual IP. Another guest running FreeBSD and DirectAdmin should be accessible directly
[15:53] <fallback> (because DirectAdmin demands a public IP)
[15:56] <zul> ivoks: can you update the changelog in your dovecot bzr branch
[15:56] <ivoks> changelog?
[15:57] <ivoks> zul: isn't that extracted from bzr? :)
[15:57] <zul> ivoks: bug number etc
[15:57] <ivoks> bug number... there isn't one
[15:57] <ivoks> the bug that was reported was on similar issue
[15:57] <ivoks> well, you might say this upload fixes it
[15:58] <ivoks> ok, let me update it
[15:58] <zul> thanks
[16:05] <ivoks> zul: like this? (commit 70)
[16:45] <Jeeves_> ARGH!
[16:54] <jpds> Jeeves_: My sentiments exactly.
[16:59] <e-DIO-t> well! TTL reached: it's time to leave!
[16:59] <e-DIO-t> bye!
[17:15] <gzmask> why am I keep getting "FinishedVerify: Not enough resources (VmTypeAvailability{type=VmType{name='c1.medium', cpu=1, disk=5, mem=256}, max=0, available=0} < 1: vm instances." after "euca-run-instances -k mykey -t c1.medium emi-DE8D1064" ??
[17:23] <masu3701> i was able to create a samba share and access it from my other pc on the network...but how can the client pc drag and drop files to the share?
[17:28] <mirko_> c' è nessuno italiano ?
[17:41] <gzmask> My node controller is not talking to the CLC, what would I do? I checked the /etc/eucalyptus/eucalyptus-ipaddr.conf and manually set the ip addresses but still doesn't work.
[17:43] <gzmask> I even ran the "euca_conf --no-rsync --discover-nodes" and still nothing
[17:48] <gzmask> And my registration.log shows:
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3242 -> Calling walrus Walrus 142.3.31.158
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3243 -> Calling storage cluster1 storage 142.3.31.158
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3245 -> Calling node cluster1 node 142.3.31.158
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3239 -> Calling storage cluster1 storage 142.3.31.158
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3244 -> Calling cluster cluster1 142.3.31.158
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3238 -> Calling walrus Walrus 142.3.31.158
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3241 -> Node 192.168.122.1 is already registered.
[17:48] <gzmask> 2010-04-19 10:39:08-06:00 | 3245 -> Node 142.3.31.158 is already registered.
[17:48] <gzmask> 2010-04-19 10:39:09-06:00 | 3240 -> Cluster cluster1 is already registered.
[17:48] <gzmask> 2010-04-19 10:39:09-06:00 | 3238 -> Walrus 142.3.31.158 is already registered.
[17:48] <gzmask> 2010-04-19 10:39:09-06:00 | 3244 -> Cluster cluster1 is already registered.
[17:48] <gzmask> 2010-04-19 10:39:09-06:00 | 3239 -> SC for cluster1 is already registered.
[17:48] <gzmask> 2010-04-19 10:39:09-06:00 | 3243 -> SC for cluster1 is already registered.
[18:09] <pmatulis> gzmask: use pastebin please
[18:10] <zul> gzmask: please use pastebin
[18:37] <zul> nxvl: ^^^
[18:38] <zul> jib: mysql 5.0 totally gone from lucid
[18:41] <ScottK> \o/
[18:41] <Daviey> zul: The concern I have is that people who explicitly installed mysql-server-5.0 won't have an upgrade path.
[18:41] <ScottK> Thanks for seeing it totally out of the archive and not just out of Main.
[18:42] <nxvl> zul: working on it
[18:42] <nxvl> zul: wait augeas is in main
[18:42] <zul> nxvl: yeah it is ;)
[18:42] <nxvl> zul: ok, will work on it later today
[18:42] <zul> ScottK: dovecot is uploaded as well
[18:43] <ScottK> zul: Great.
[18:43] <ScottK> Just waiting for a Debian upload of clamav to merge and I think server stuff I'm worried about for Luicid is done.
[18:56] <gzmask> zul: sorry I was unaware of the long log file. anyways, I got the problem fixed with the help from nerko_ and he smells a potential bug. Here's how I got it fixed: http://ubuntuforums.org/showthread.php?p=9144875#post9144875
[18:57] <smoser> hggdh, is it possible for me to modify an image and try to recreate https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/566792 ?
[19:07] <jiboumans> zul++
[19:27] <libertiy> hi everyone i was reffered to here from ubuntu
[19:27] <libertiy> i wondered if it would be possible to have qmail on ubuntu with qmail toaster
[19:30] <RoAkSoAx> kirkland: what does '<number>!' with red backgroun in byobu means?
[19:33] <hggdh> smoser: yes
[19:33] <hggdh> go ahead, the environment is set up
[19:33] <smoser> hggdh, cempedak?
[19:33] <hggdh> cempedak is the CLC, yes
[19:33] <smoser> k
[19:33] <Daviey> RoAkSoAx: Pending updates
[19:34] <RoAkSoAx> Daviey: thanks! :)
[19:34] <Daviey> np
[19:34] <hggdh> smoser: topo2, cempedak=CLC, mabolo=walrus, marula=CC, santol=SC, (sapodilla|soncoya)=NC
[19:35] <smoser> hggdh, how often does this happen ?
[19:35] <libertiy> anyone knows of an easy way to install + manage postfix mailboxes?
[19:36] <RoyK> libertiy: well, it's linux after all, so you can run what you like, but see https://help.ubuntu.com/8.04/serverguide/C/dovecot-server.html for a good imap server
[19:36] <smoser> hggdh, other question, where did you get "20100419.1/UEC"
[19:36] <smoser> 20100419.1 is not available anywhere
[19:36] <hggdh> smoser: last 200 runs, at least 5 times
[19:36] <libertiy> i found out sudo apt-get install postfix it looks good  i found devocot indeed :P
[19:37] <hggdh> smoser do you want me to save the current logs before you start?
[19:37] <smoser> sure
[19:37] <RoyK> libertiy: postfix works well
[19:37] <smoser> hggdh, i'll need instruction on how to run also
[19:37] <smoser> give me 10 minutes to get an image registered
[19:37] <libertiy> i come from centos with qmail so this is something different
[19:37] <libertiy> i hope a small challange :)
[19:37] <RoyK> it shouldn't be very hard
[19:38] <libertiy> when installing postfix with sudo apt-get install postfix i dont get imap with it i guess right? i need to seperatly install devocot
[19:38] <RoyK> no, postfix is an mta, not a poa
[19:39] <RoyK> so yes, install dovecot for mailbox access
[19:39] <binBASH> Hi RoyK
[19:40] <RoyK> hi, bashy :)
[19:41] <libertiy> guys sudo apt-get install dovecot-postfix gives package not found
[19:41] <libertiy> do i need to set  a remote repository for packes in the ubuntu "rpm" manager?
[19:42] <binBASH> RoyK: Damn, I have to migrate from qmail, to postfix at some day this week :)
[19:42] <libertiy> im sorry if in not into the correct terms here, just wonder , i guess basic principes are the same
[19:42] <libertiy> hehe i never had problems with qmail though binBASH
[19:42] <binBASH> libertiy: Here it crashes on mails comming from 1/4 of all mailservers
[19:43] <binBASH> really drives me nuts, but maybe it's caused by a plugin / patch.......
[19:43] <hggdh> smoser: I used the base install from 20100419.1, and the UEC from 20100419
[19:43] <libertiy> hmm thats a shame to hear, but sometimes it can be nice to try other solutions and why not keeps you sharp ;p
[19:43] <binBASH> libertiy: I'm using qmail since 2000
[19:43] <binBASH> so it's a shame for me to move :p
[19:43] <hggdh> smoser: running the single user test is set up. I would still have to create the accounts for the multi
[19:44] <smoser> ah. iso 20100419
[19:44] <hggdh> smoser: but the single test is where I got the last failure
[19:44] <smoser> that makes since
[19:44] <hggdh> why?
[19:45] <libertiy> i havent really looked into the differences on qmail and postfix, but i guess thats a discussion for another channel
[19:47] <smoser> hggdh, ok. i've registered emi-27101629
[19:47] <smoser> i'm sniff testing it to see that the debug output that i wanted gets printed
[19:47] <lamont> libertiy: qmail has a license that relegates it to multiverse
[19:48] <smoser> and then if so, run it in a test like yo uhave
[19:49] <binBASH> libertiy: Like I said, maybe it's caused by a qmail plugin. I'm using qmail-spp
[19:49] <binBASH> It has a plugin which looks up sender ips against dns blacklist
[19:49] <libertiy> hmm i havent used that i used it very basic to host mailboxes for projects and customers
[19:49] <binBASH> so spammers will be rejected and not accepted by the server
[19:49] <hggdh> smoser: OK. Tell me when you are ready
[19:50] <smoser> ready
[19:50] <hggdh> :-)
[19:50] <hggdh> smoser: let's do it as 200 runs, when the error happens we can stop it
[19:50] <binBASH> libertiy: Well I do not want to use it any longer. What I want is. Have Mails accepted by the server and put them into a SPAM folder
[19:50] <smoser> k
[19:50] <binBASH> just like gmail does for exmaple
[19:51] <binBASH> however I dunno how to configure spamassassin yet
[19:52] <hggdh> all you need to do: cd ~/uec-testing-scripts; ./uec_test.py -c config_single.yaml
[19:52] <hggdh> you may want to put it in the background
[19:52] <smoser> modify config_single for my ami
[19:52] <smoser> right?
[19:52] <hggdh> dammit
[19:52] <hggdh> yes, of course
[19:52] <hggdh> doing it
[19:52] <smoser> done
[19:52] <binBASH> smoser: What happens btw. if vms in uec are rebooted?
[19:53] <binBASH> Will they be deleted?
[19:53] <smoser> no.
[19:53] <smoser> reboot should come back up
[19:53] <binBASH> ok
[19:53] <smoser> binBASH, terminate is "go away"
[19:53] <hggdh> smoser: then just run it as I pointed above. Results are written to ./resutls/single*
[19:53] <smoser> and good bye data
[19:55] <smoser> hggdh, was there user data provided to the instance ?
[19:56] <smoser> hggdh, how do i know when i see it fail ?
[19:57] <smoser> (you can attached to screen session there: screen -x -r smoser)
[19:57] <hggdh> smoser: no consumer data
[19:58] <hggdh> attached. No real errors so far (apart from instances that seem to have failed to start)
[19:58] <binBASH> smoser: What is the best way to use local storage of the node inside the vm?
[19:58] <binBASH> ... if network speed is only 100 Mbit
[19:58] <binBASH> maybe local iscsi provider on the node?
[19:59] <smoser> binBASH, hm.. i dont know. its not really covered.
[19:59] <binBASH> I think EBS is too slow on 100 Mbit ;)
[19:59] <hggdh> smoser: all reschedules are instances to which we could not SSH. The script tries 4 times, I think, then considers the test a failure
[20:00] <hggdh> smoser: a string of failures, right now
[20:00] <smoser> yeah.
[20:01] <hggdh> if you want you can grep the log for the instance id, and this will return all lines written for it
[20:01] <smoser> hggdh, so. you see that screen, right?
[20:02] <smoser> (and you have a terminal larger than mine)
[20:04] <ARTSIOM> HI!
[20:05] <ARTSIOM> tring to install sun-java6-jdk package on 10.04 and getting no installation candiadte found error (multiverse is enebled in sources list, had also tried apt-get update). Any ideas?
[20:06] <smoser> hggdh, i can just ctrl-c ?
[20:07] <hggdh> smoser: yes, this should be trapped by the script, and it will terminate gracefully
[20:08] <smoser> k
[20:08] <smoser> i'll modify image for more debug and try again
[20:08] <smoser> i updated bug
[20:08] <hggdh> smoser: cool, thank you.
[20:09] <hggdh> smoser: I do see the screen and I have a very large terminal
[20:09] <smoser> :)
[20:09] <hggdh> :-)
[20:15] <kirkland> RoAkSoAx: man byobu
[20:16] <RoAkSoAx> kirkland: got it already :) thanks though
[20:17] <RoAkSoAx> kirkland: btw... you are perl programmer to right? Do you know anything about network programming in perl?
[20:19] <smoser> hggdh, this is easy to debug
[20:19] <smoser> happens reproducibly enough
[20:20] <kirkland> RoAkSoAx: a fair bit; did some socket programming in Perl long, long ago
[20:20] <hggdh> yes, fortunately. What worries me is why neither Thierry, Mathias, or Dustin don't get it
[20:21] <kirkland> hggdh: what don't i get?  the oops?
[20:22] <smoser> hggdh, speed.
[20:22] <smoser> the node comes up before the metadata service is all the way up
[20:23] <zul> hggdh: you are a freaking lightning rod
[20:23] <smoser> hggdh, i suspect that the node controllers in this system have much better IO
[20:24] <smoser> than kirkland ttx or my crappy laptops
[20:24] <smoser> so the create instance happens much faster, starts, and the metadata service isn't up yet
[20:26] <hggdh> zul it seems I am :-(
[20:27] <hggdh> smoser: so it is a timing issue...
[20:27] <smoser> i think so yeah.
[20:27] <hggdh> and yes, these machines are fast
[20:27] <smoser> well, here i think its that IO is faster than cpu for this case
[20:27] <smoser> i'm guessing still
[20:27] <RoAkSoAx> kirkland: by any change.. do you know if it would be possible to capture a current SSH connection with a perl app and send data through that current connection, without having to create a connection through the scritp?
[20:28] <smoser> RoAkSoAx,
[20:28] <smoser>   ControlMaster auto
[20:28] <smoser>   ControlPath /tmp/.ssh-%r@%h:%p
[20:28] <smoser> ?
[20:29] <kirkland> RoAkSoAx: ^
[20:29] <RoAkSoAx> smoser: awesome, I'll look into it :)
[20:29] <smoser> RoAkSoAx, man ssh_config will tell you more about those
[20:29] <smoser> i have tha tin  my default .ssh/config
[20:31] <RoAkSoAx> sommer: ok cool. I'll look into it :) thanks.
[20:32] <RoAkSoAx> s/sommer/smoser
[20:35] <erimar77> ls
[20:35] <erimar77> whoops, lol.. sorry
[21:03] <kirkland> jdstrand: ping
[21:03] <jdstrand> kirkland: hi
[21:03] <kirkland> jdstrand: can you check my last comment to https://bugs.edge.launchpad.net/ubuntu/+source/mdadm/+bug/557429
[21:03]  * jdstrand reads
[21:05] <kirkland> jdstrand: i expect user "ceg" to come down hard on me. he always does.
[21:05] <jdstrand> kirkland: I haven't been following what it would take to fix it, but your comment seems reasonable
[21:06] <kirkland> jdstrand: there's one part that i left unwritten ....
[21:06] <kirkland> jdstrand: while it's easy to force this situation in a VM, i question the practicality with real hardware
[21:06] <kirkland> jdstrand: if a real disk "dies" and "disappears", they don't often just "come back", do they?
[21:08] <jdstrand> kirkland: I tried to convey something along those lines in the release meeting last week. I don't think it is something that most people would do, but if someone removes the disk to say test the array, or remove/move around some hardware or something, if the bug hits it is devastating
[21:08] <jdstrand> kirkland: so while an unlikely event, if it happens, ouchy
[21:08] <kirkland> jdstrand: fair enough
[21:09] <kirkland> jdstrand: i agree that it would be fixed in an ideal world
[21:09] <kirkland> jdstrand: but i don't think we can hack that into the mdadm's failure hooks easily
[21:09] <jdstrand> kirkland: not to mention, it is the test case in http://testcases.qa.ubuntu.com/Install/ServerRAID1. if someone does that on real hardware... again, ouchy
[21:10] <kirkland> jdstrand: how/why would someone do this on real hardware?  (i'm trying to clearly understand the test case, not being pedantic)
[21:10] <jdstrand> kirkland: I'm not suggesting that we try to rush something for lucid, but it seems it should be fixed in maverick
[21:10] <kirkland> jdstrand: excellent; if you care to +1 that analysis in the bug, that would be pretty cool!
[21:11] <sbeattie> well, no one should be doing iso testing on live machines that they have data that they care about; I do for some tests, but I recognize that I'm risking an installer bug by having it toast a wrong partition.
[21:11] <jdstrand> kirkland: boot with both disks. shutdown, remove one. boot. shutdown, put in the other one, boot. put in the first first. boom
[21:11] <kirkland> jdstrand: right, but why would someone realistically do that?
[21:11] <jdstrand> kirkland: it isn't that far fetched that someone would do this to verify their array-- there is a reason it ended up in the test case
[21:12] <kirkland> jdstrand: i could drop my laptop in the bathtub too, boom!
[21:12] <jdstrand> kirkland: I agree that it wouldn't happen often. I'm just saying if it did happen, it is devastating total data loss
[21:12] <sbeattie> welll, it can happen if you've got a squirrely controller.
[21:13] <jdstrand> kirkland: I think a user would expect the system to be hosed if it was dropped in a bathtub. I don't think total data loss is expected by adding back a drive
[21:13] <kirkland> jdstrand: sbeattie: alright, thanks
[21:13] <kirkland> jdstrand: should be fixed, Maverick issue/timeframe
[21:14] <kirkland> jdstrand: if the fix is "contained", an SRU is perhaps meritted
[21:14] <jdstrand> that is fine be me. afaict, the issue was in karmic too
[21:14] <jdstrand> it would be nice to not have it in an LTS, so hopefully it is SRU worthy
[21:14] <kirkland> jdstrand: yup
[21:14] <kirkland> jdstrand: okay, thanks a bunch
[21:14] <jdstrand> kirkland: sure! :)
[21:17] <kirkland> smoser: hggdh: okay ... ran 1000 instances on today's UEC, and uec-images:
[21:17] <kirkland> failed: 44
[21:17] <kirkland> success: 949
[21:17] <kirkland> success_rate: 0.94999999999999996
[21:17] <kirkland> i'll tar up the results
[21:17] <smoser> please.
[21:18] <jdstrand> sbeattie: I agree about the iso testing part on a live machine. I can imagine someone trying the non-install part of the test case, to be thorough and ensure raid is working properly. on real data without backups? pretty dumb, but it shouldn't be doing this...
[21:19] <kirkland> hggdh: where's that uec testing branch you had with some fixes?
[21:19] <kirkland> hggdh: did you give me that in irc or in email?
[21:19] <kirkland> hggdh: i forgot to merge that when i ran these tests
[21:21] <jdstrand> kirkland: added my +1 cent to the bug
[21:22] <hggdh> kirkland: it was in the email, hold on a sec
[21:23] <hggdh> kirkland: lp:~hggdh2/%2Bjunk/uec-testing-logged
[21:23] <sbeattie> kirkland|jdstrand: I have in fact suffered from a raid failure (under windows, alas) that was quite similar, due to a controller/electrical issues: disk 1 goes clicky, clicky, no spinnee; I replace disk 1 and try to sync it to disk 2; very shortly after sync starts, machine reboots; next reboot, system attempts to resync disks and decides the new disk 1 with nearly all zeros is the one to sync from.
[21:24] <jdstrand> ouchy
[21:42] <smoser> hggdh, can i tell it collect all console logs ?
[21:42] <smoser> even on success ?
[21:48] <hggdh> smoser not as written, but this is a small change. Give me 5
[21:48] <gzmask> Apple SUCK: kvm-ok
[21:48] <gzmask> INFO: Your CPU supports KVM extensions
[21:48] <gzmask> INFO: KVM is disabled by your BIOS
[21:48] <gzmask> HINT: Enter your BIOS setup and enable Virtualization Technology (VT),
[21:48] <gzmask>       and then hard poweroff/poweron your system
[21:48] <gzmask> KVM acceleration can NOT be used
[21:48] <gzmask> how the hell do I go into the bios on a MAC?
[21:49] <hggdh> smoser: I will change it to collect *all* consoles for all instance runs. This will be a local hack right now
[21:50] <hggdh> which is to say, do not rm uec_test.py from cempedak ;-)
[21:50] <smoser> yeah
[21:52] <smoser> hggdh, ok. i'm ready to run
[22:00] <smoser> hggdh, i have to run.
[22:00] <smoser> i would much appreciate seeing log output of this config_single run with emi-28171637
[22:00] <smoser> maybe run 200 times and just collect it all.
[22:00] <smoser> please ?
[22:00] <smoser> i'll check in later
[22:04] <libertiy> i need a simple pop imap mailserver with terminal administration to add accounts and domains and virtual hosts, its for webhosting
[22:04] <libertiy> anyone can advise me on something?
[22:04] <libertiy> seems to be harder then i thought and i have no php so i skipp the webinterface, a terminal session to add some accounts would be okay also
[22:08] <clrg> libertiy: postfix? Very good mail server, but not specially adapted for webhosting.
[22:08] <libertiy> how about exim?
[22:08] <hggdh> smoser: will do
[22:08] <libertiy> ive just red that seems a good choice
[22:08] <libertiy> else i would choose qmail but that i a bit harder on ubuntu server
[22:08] <libertiy> just need it for some clients like 15
[22:09] <libertiy> all info@theres.com and need some way to easily manage adding clients and passwords without messing with 1000s of config files
[22:09] <clrg> libertiy: I never used exim, can't advise you on that.
[22:09] <libertiy> what would you use then on ubuntu clrg?
[22:10] <clrg> libertiy: postfix
[22:10] <libertiy> for webhosting?
[22:10] <libertiy> else i hve to go back to centos to server but thats not handy i have to setup that all from scratch
[22:11] <libertiy> this ubuntu imafge im on now is very clean and memory safe and secure
[22:11] <clrg> libertiy: As e-mail-server. I work in the hosting business as well, but we only have very large customers, for which we make individual solutions. We use postfix for alarming per mail.
[22:11] <libertiy> just have to add the mail server to it somehow and im done
[22:11] <libertiy> okay then postfix will be it,do you know of a way to easily manage accounts from the terminal?
[22:11] <libertiy> since im all ruby on rails now i cannot run the postfix admin which is php
[22:12] <libertiy> and i dont feel like installing php to my server just for postfix admin
[22:12] <lamont> libertiy: mostly, I use vim
[22:12] <libertiy> just some config file editment thus ;p is that cost a lot of time?
[22:12] <libertiy> or is it just 1 config file with the client details
[22:13] <clrg> liberty: It doesn't matter. Do the procedure once, then write a script which does the work for you.
[22:14] <clrg> liberty: Is this what you are looking for? http://postfix.wiki.xs4all.nl/index.php?title=Virtual_Users_and_Domains_with_Courier-IMAP_and_MySQL
[22:15] <libertiy> something like that indeed but than more tailered for ubuntu with imap support ireally need that also
[22:15] <libertiy> which makes it a unpleasent task for someone tired and its late here ;p perhaps better check into it tomorrow
[22:17] <libertiy> thats what i need sortoff clrg thx altough i dont want to use postfix webadmin since i dont host php anymore
[22:17] <libertiy> but it could be something to look into
[22:18] <libertiy> or zimbra but thats heavey on the memory
[22:18] <lamont> zimbra is also heavy on the violation of APIs in postfix.
[22:34] <mok0> Anyone here familar with mailman w. postfix?
[22:34] <mok0> I can't understand the howto here: https://help.ubuntu.com/community/Mailman
[22:59]  * clrg wishes everyone a pleasant evening