uvirtbotNew bug: #519567 in euca2ools (main) "euca2ools does not correctly specify port when $http_proxy is set" [Undecided,New] https://launchpad.net/bugs/51956700:06
SpaceBasshad a software raid5 array fail yesterday - was a bad disk connection. Fixed the connector and its been rebuilding for over 24 hours - does that sound right?00:09
uvirtbotNew bug: #519163 in samba (main) "package samba-common 2:3.3.2-1ubuntu3.3 failed to install/upgrade: subprocess post-installation script returned error exit status 3" [Undecided,New] https://launchpad.net/bugs/51916300:09
=== dendro-afk is now known as dendrobates
sponzorhi is there any kdump in ubuntu?03:03
sponzorhi what to use for crash dump? that it captures kernel panic etc..?03:35
Roxyhart0hi somebody know what is automonth for ldap, what is does?05:53
ja660khello, what directory is the php pear modules in?05:59
smosersoren, erichammond anyone else who cares: "Query Interface" at https://wiki.ubuntu.com/EC2UpgradesSpec is updated. tree at http://uec-images.ubuntu.com/query/ has info06:01
smoseri'm going to bed now.06:01
uvirtbotNew bug: #519653 in eucalyptus (main) "missing directories on NC after install" [Undecided,New] https://launchpad.net/bugs/51965306:31
uvirtbotNew bug: #505482 in euca2ools (main) "euca-bundle-vol dies with "Invalid cert"" [High,New] https://launchpad.net/bugs/50548206:51
uvirtbotNew bug: #519664 in squid (main) "squid FTBFS: dh_installinit: command not found" [Undecided,Confirmed] https://launchpad.net/bugs/51966407:06
=== ttx_ is now known as ttx
fluvvelli want to display photos on the screen of a server, and use eog -s -f --display= {second monitor} but don't know the syntax07:49
fluvvellany ideas?07:50
pts_anyone know if hb_gui is available for CRM (pacemaker-openais CIB manager) in any ubuntu repos?08:28
_rubenpts_: according to packages.ubuntu.com, no08:40
pts_I hoped it maybe was hidden somewhere :) Anyone knows why it hasnt been added? Using the pacemaker-openais package and wondering even if its compatible in a case a find the sources and want to try to compile it08:43
_rubenpts_: wouldnt know, i hadnt had a chance to dive into the "new" pacemaker stuff, hell, i even havent had a change to even look into the crm stuff :p09:02
c0dem4gnetic_what would a reasonable textmode mail client to use be?09:28
c0dem4gnetic_mutt it is then, thanks :)09:31
pts_could suggest pine too, but no idea how it compares to mutt09:32
rossouwaphi, can anyone give me some info on the KVM -incoming command?10:05
TeTeTttx: is the cloud installation from CD already quite stable, or will it be changed? If not I can start to update the training materials10:24
ttxTeTeT: it should be quite stable now.10:28
TeTeTttx: thanks!10:28
=== ogra_ is now known as ogra
AnAntHello, why did Ubuntu use w3m instead of elinks-lite for building doc for mutt package ?10:57
jpdsAnAnt: apt-cache showsrc mutt - shows that elinks is used?11:04
AnAntjpds: look at changelog of -5ubuntu111:04
jpdsAnAnt: Ask mathias when he's online.11:07
pts_is there any repos with updated pacemaker/corosync and DRBD?11:09
RoyKisn't that in the default repo?11:10
RoyKor is that just heartbeat?11:10
pts_pacemaker 1.0.5, but 1.0.6 is newest I think11:11
pts_and corosync is also outdated11:11
RoyKoutdated as in doesn't work?11:15
* RoyK still has a Solaris 7 running11:17
RoyKs/run/box run/11:18
pts_well, maybe. I'm totally new with clustering and it makes it a bit harder when CRM crashes while I follow a howto. When running the command sequence crm configure <enter> property no-quorum-policy=ignore <enter> commit <enter> crm crashes. CIB havent been changed at all before this, just made sure it exists as I expect11:19
uvirtbotNew bug: #495505 in tomcat6 "server.number reports not 6.0.X.0" [Medium,Triaged] https://launchpad.net/bugs/49550511:27
uvirtbotNew bug: #427701 in openvpn "OpenVPN client always runs" [Wishlist,Triaged] https://launchpad.net/bugs/42770111:32
Sorellhey guys,11:33
Sorelldoes anyone on here know anything about how LTSP uses it's DHCP server11:34
SorellI'm having a bear of a time getting this setup.11:34
pts_If I remmeber correctly it uses /etc/ltsp/dhcpd.conf. The "special" options you need are filename "/ltsp/i386/pxelinux.0";11:40
pts_option root-path "/opt/ltsp/i386";11:40
pts_you could try #ltsp if you need more help11:40
SorellI will give it a try ty11:43
sabgentonI want to use cup-lpd12:08
sabgentonanyone know the package name?12:09
dholbachis something with php pear known to be broken at the moment?12:09
dholbachbug 513765 looks like it12:09
uvirtbotLaunchpad bug 513765 in php-versioncontrol-svn "Sync php-versioncontrol-svn 0.3.3-2 (universe) from Debian unstable (main)" [Undecided,New] https://launchpad.net/bugs/51376512:09
sabgentonoks cups-bsd12:12
=== sabgenton_ is now known as sabgenton
kaushali have a issue about Maximum Capacity: 4 GB while running dmidecode,but the actual memory is 8GB on san box. please suggest ?12:47
sherringhamkaushal: a) Check your BIOS settings (RAM/memory mapping) b) What arch - 32bit or 64 bit kernel? uname -a12:53
kaushal64 bit12:53
kaushalsherringham: any clue ?12:57
sherringhamkaushal: Did you check your BIOS settings?12:57
sherringhamBIOS sees 8 GB? Some chipsets don't support > 4GB RAM - see if you can find out for sure. Maybe try switching RAM around in the slots ...13:00
kaushalsherringham: i was interested in how much RAM my san box would support ?13:00
kaushalI mean the Max Total RAM Support13:01
kaushalwith out trying to contact the manufacturer13:01
kaushalis there a way to find out that information ?13:01
kaushalfrom the OS Side13:01
sherringhamDon't think so - you need hardware/system specs e.g. chipset. If it's Intel or AMD, should be easy to find though - or google.13:02
sherringhamNormally, the manufacturer will have this information.13:03
pts_anyone know why I cant start corosync via the init.d script? Have to run corosync -f to start it13:27
pmatulisanyone heard of apache2 on 8.04 to suddenly stop logging?  this might be happening because of logrotate config13:50
zula new release is a paper cut?13:50
ivoksis meeting today or tomorrow?13:50
zulpmatulis: ive seen it before the large file becomes too large13:51
pmatuliszul: this is the logrotate config if it matters:13:51
pmatuliszul: maybe i should specify a log file size?13:52
zulpmatulis: no it should be ok, what if they do the default one?13:52
* pmatulis looking for default one...13:53
pmatuliszul: i'll have them try it13:54
incorrecti've had a request to upgrade a 8.04 to 9.1013:59
incorrecti am guessing i will have to go via 8.10 and 9.0414:00
RoAkSoAxivoks, today i believe14:00
Piciincorrect: yes, you will14:00
Davieyzul: My main concern with 5.3 was upstream php switching to default of short tags off.  Just had a look at debian experiment 5.3, and there is a debian/patch that re-enables it.  This will reduce stuff broken, so i am happy.14:08
zulDaviey: good14:08
incorrectoh i have to edit something to allow me to upgrade from 8.04 to .1014:09
smosercommand line completion is mostly fixed, kirkland ?14:28
kirklandsmoser: well, chris added the code upstream, but it's not working just yet14:29
smoserlike bash completion ?14:29
kirklandsmoser: yes14:29
kirklandsmoser: see /etc/bash_completion.d/14:29
=== dholbach_ is now known as dholbach
=== LyonJT_ is now known as LyonJT
kirklandsmoser: euca_conf with bash completion just uploaded to lucid!15:00
stgraberkirkland: do we have anything against running tftpd-hpa as a standalone daemon ?15:01
stgraberthat's what would cause the biggest delta with Debian15:01
kirklandstgraber: as opposed to under inetd?15:02
stgraberthe rest is basically a two line diff to change the default path and debconf priority15:02
stgraberkirkland: yes15:02
* Daviey tends to deploy tftpd-hpa as a daemon15:02
=== RoAk is now known as RoAkSoAx
stgraberpushing an updated package to my PPA15:05
kirklandstgraber: i don't care much either way, as long as the damn thing starts on boot ;-)15:08
kirklandstgraber: most of the documentation in the Ubuntu wiki and server guide talks to running it under inetd15:08
kirklandstgraber: we'd just need to ensure that those docs are updated (not hard to do, certainly not a blocker)15:08
ttxI agree that changing the location of the tftpboot is a nice killer :)15:11
ttxkirkland: btw, I didn't get a copy of your mirror rsync script, could you email/pastebin it to me ?15:12
kirklandoh, sure15:12
kirklandttx: http://pastebin.com/f26ee31bd15:14
kirklandttx: you may want to tweak it, but you should get the idea15:14
ttxsure, I'll use it as a starting point15:14
kirklandttx: btw, one paper cut down ;-)15:15
kirklandttx: just uploaded euca_conf with bash completion15:16
ttxI'll make sure we count it :)15:16
kirklandttx: okay, uploaded iptables/db-deadlock fix for karmic/eucalyptus to ~ubuntu-virt ppa15:27
kirklandttx: updated the bug15:27
kirklandnijaba: could you talk to TeTeT and/or laclasse about perhaps helping test?15:28
nijabakirkland: for euca ?15:29
nijabakirkland: url?15:29
kirklandnijaba: https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/503180/comments/1115:29
uvirtbotLaunchpad bug 503180 in eucalyptus "eucalyptus-cloud doesn't reply to requests (eucalyptus doesn't work after reboot or services restart issues due to upstart networking behavior)" [High,In progress]15:29
nijabakirkland: ok, will ask them to test15:29
kirklandnijaba: ack, thanks!15:31
kirklandnijaba: hopefully this will put the worst karmic/euca bug to bed ;-)15:31
* nijaba crossing his finger15:31
nijabamy fingerS too15:32
TeTeTnijaba + kirkland : booting my karmic UEC :)15:32
kirklandTeTeT: the package is building in the PPA15:33
kirklandTeTeT: should be published any minute now15:33
TeTeTkirkland: need to update the systems anyway, they haven't been touched for a while15:33
kirklandTeTeT: great, please try to reproduce the bug before you upgrade to the ppa version15:34
kirklandTeTeT: i think you can reproduce it easily by rebooting, and seeing that you can't talk to the CLC for a long time (like 10-20 minutes)15:34
kirklandTeTeT: at least, that's how I experienced the bug15:34
kirklandTeTeT: then, I upgrade the packages, reboot again, and the CLC is up and communicative almost immediately (within 1 minute of login)15:35
TeTeTkirkland: will do15:35
TeTeTkirkland: I can reproduce it on the training cloud in the datacenter, but I don't have access to it right now15:35
uvirtbotNew bug: #224828 in nis (main) "NIS demon fails during startup if the roaming mode is turned off" [Low,Confirmed] https://launchpad.net/bugs/22482815:37
* dholbach tries asking again: is something with php pear known to be broken at the moment? bug 513765 looks like it15:39
uvirtbotLaunchpad bug 513765 in php-versioncontrol-svn "Sync php-versioncontrol-svn 0.3.3-2 (universe) from Debian unstable (main)" [Undecided,New] https://launchpad.net/bugs/51376515:39
dholbachmathiaz, zul: ^15:39
zuldholbach: looking15:41
sarthorHi, My pppoe-server is according this url "  http://ubuntuforums.org/showthread.php?t=291926  " and is working, but same username and passwords can connect from several computers, I to solve this problem. example, if username is "username" and password is "passwdd" So serveral clients can connect this  username from same time from several computers, My clients are on my lan, HELP,15:42
kirklandTeTeT: okay, amd64 is built15:44
kirklandTeTeT: and published15:44
TeTeTkirkland: bummer, I only have a 32bit cloud at home15:44
kirklandTeTeT: 32 is built too now15:45
nijabakirkland: reproduce 1 --> done, ow updating to current packages (no ppa) to reproduce again, will add ppa afterward15:47
TeTeTkirkland: reproducing problem now, will wait for 15 minutes15:50
kaushali am still not convinced about the response i got from ubuntu-server mailing list about dmidecode. please suggest15:57
kaushalso what i understand from that email thread is to consult the manufacturer and not dmidecode to know the Maximum Capacity a san box can support ?15:59
kaushalam i understanding it correctly ?15:59
nijabakirkland: :( can't reproduce after doing a normal update (without your ppa)15:59
kirklandnijaba: hrm?16:00
nijabakirkland: running euca 1.6~bzr931-0ubuntu7.416:00
kirklandTeTeT: you should be able to just wait ~5 minutes16:00
kirklandTeTeT: if it's not up by then, your db is wedged16:01
uvirtbotNew bug: #519904 in samba (main) "package samba-common 2:3.4.0-3ubuntu5.3 failed to install/upgrade: subprocess dpkg-deb --control returned error exit status 2" [Undecided,New] https://launchpad.net/bugs/51990416:01
TeTeTkirkland: after 15 minutes it's there16:02
TeTeTkirkland: should I do another reboot and test after 5?16:02
kirklandTeTeT: right, it takes ~15 to get unwedged16:02
kirklandTeTeT: sure, if you like16:02
kirklandTeTeT: basically, eucalyptus loses connection to the database16:02
mathiazkirkland: hm bug 519653 - seems that running an NC and a front-end on the same system is working16:02
uvirtbotLaunchpad bug 519653 in eucalyptus "missing directories on NC after install" [Undecided,New] https://launchpad.net/bugs/51965316:02
kirklandTeTeT: due to the changing network configuration16:02
kirklandTeTeT: after ~15 minutes or so, the connection gets reset16:03
kirklandTeTeT: with the fix, the db connection should be ready immediately on boot16:03
kirklandTeTeT: were you able to confirm that you could not talk to the CLC for a feew minutes after boot?16:03
kirklandmathiaz: with system mode network, yes, Ng told me as much16:04
mathiazkirkland: oh ok.16:04
kirklandmathiaz: i have not tried it16:04
kirklandmathiaz: and none of the other, advanced networking modes will work, though16:04
kirklandmathiaz: but Ng has documented it on the topologies page in the wiki16:04
TeTeTkirkland: yes, could not take to it immediately after boot16:05
TeTeTkirkland: bummer, now it is available :(16:06
TeTeTright away16:06
kaushalchecking in again for my query ?16:09
TeTeTkirkland: ok, can't reproduce the bug anymore on my UEC here. 2 out of 3 reboots the cloud is responding to 'euca-describe-availability-zones' right awya16:11
TeTeTkaushal:  sudo dmidecode | grep "Maximum Total Memory Size"  works for me16:14
kaushalTeTeT: yeah i have already done it16:15
kaushalTeTeT: shall i pastebin the output ?16:16
TeTeTkaushal: afraid I'm not of any more help, that's all I had to offer.16:17
zuldholbach: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=54616416:18
uvirtbotDebian bug 546164 in php-pear "pear download directory is inherited from the build" [Important,Open]16:18
dholbachzul: ah ok, I'll see what it's about16:18
kklimondaany idea whenever django 1.2 is going to make it to lucid?16:26
rbrunhuber9729Can init/upstart start every (cli) application or just daemonized applications?16:44
kreuterhowdy #ubuntu-server.  can upstart run a service as a non-root user?16:44
karlhegkreuter: dpkg --listfiles upstart | less16:48
karlhegkreuter, that will show you what files it installed... see if it uses shell scripts to start the services.16:48
karlhegkreuter, ps wwwauxf | less16:48
karlhegkreuter, You'll see many services running as non-root users.16:49
kreuteram I reading you as saying that it's necessary to indirect through a shell script to run as non-root (i.e., the upstart configuration files don't have any special support for it?)16:50
mathiazkreuter: correct - upstart doesn't natively support running programs as a non-root user16:50
uvirtbotNew bug: #519940 in keepalived (main) "Please merge keepalived 1.1.17-2 from debian testing" [Low,In progress] https://launchpad.net/bugs/51994016:51
karlheg... but it runs as root so it can do anything you program it to do via a shell script. The script can call 'su'.16:51
karlhegSo, there's no need to put that kind of functionality into upstart itself.16:52
kklimondazul, what is your opinion about django 1.2 in Lucid? I'm asking as you were the one requesting MIR for it. It's still likely to be released in march.16:53
kreuterkarlheg: sure, of course.  I just couldn't find anything about this by googling, so I figured I'd ask.16:53
kreuterkarlheg: while I'm at it, which FM did you have in mind?16:53
zulkklimonda: i dont have a problem with it it probably wont get into main for lucid though16:54
karlhegdpkg --listfiles upstart | grep '/usr/share/\(man\|doc\)'16:55
kklimondazul, so MIR was a one time only? I guess there is no problem then as universe is known not to be perfect anyway :)16:56
karlhegman 5 init16:56
karlhegman 8 init16:56
zulkklimonda: it was rejected for main because it has a poor security track record but no problems with universe16:56
karlhegman 8 pizza16:57
kklimondazul, it was? hmm.. was it discussed somewhere after MIR was accepted? I must have missed it.16:58
zulkklimonda: MIR was never accepted and wasnt filed because of the security history review done while considering the MIR16:59
kirklandTeTeT: nijaba: so are you suggesting that the fix is not necessary for SRU to karmic?17:01
nijabakirkland: no, I am not suggesting this.  I am just saying that I am havig trouble reproducing on my setup today, but that does not mean it does not occur at all17:01
kklimondazul, Are we both talking about python-django? It was promoted to main in karmic.. I'm confused :)17:02
nijabazul: me too am confused...17:02
zulkklimonda: yeah i was confused :)17:03
nijabazul: python-django | 1.1.1-1ubuntu1 |        karmic | source, all17:03
nijabapython-django | 1.1.1-2ubuntu1 |         lucid | source, all17:03
zulkklimonda: march is cutting a bit close for me but submit a bug to launchpad and we'll have a look closer to the day17:04
nijabakklimonda: AFAIK it won't be demoted for lucid, and I don't know of anyone suggesting that we should17:04
kklimondazul, there is already a bug 505440 - the problem is during 1.1 cycle django developers have decided to postpone RC for two or three months after the planned release. The number of opened bugs with milestone set to 1.2 on their tracker is also quite big. Can we just wait and request an exception from FeatureFreeze?17:08
uvirtbotLaunchpad bug 505440 in python-django "django 1.2 for lucid" [Undecided,New] https://launchpad.net/bugs/50544017:08
zulkklimonda: probably ill have to look at it17:09
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
smosermathiaz, ping.17:42
mathiazsmoser: hi!17:45
smosermathiaz, i got an answer elsewhere. you were too slow :)17:45
TeTeTkirkland: same here, the latest I saw the problem happening was on the cloud training. I will try to reproduce it there tomorrow17:50
tolikПривет всем.18:26
smoserjjohansen, ping18:33
jjohansensmoser: whats up18:33
* smoser knows that jjhohansen really hates it when i piing18:34
smoseri'm looking at a mail to ec2ubuntu18:34
jjohansensmoser: only if it involves more work :)18:34
smosererr, wait18:34
smosernot that list.18:34
smoserlooking for mail18:35
smoserit looks like there are some broken package dependencies in karmic right now18:35
smoseri've just booted the ami in question (ubuntu-karmic-9.10-amd64-server-20091027.1.manifest.xml)18:36
smoserthen done apt-get update18:36
smosersudo apt-get install linux-ec2 --dry-run18:36
smoserThe following packages have unmet dependencies:18:36
smoser  linux-ec2: Depends: linux-image-ec2 (= but is to be installed18:36
smoserE: Broken packages18:36
jjohansenhrmm, yep that looks like broken deps18:37
smoserinterestingly, though, the broken packages "save" him18:39
smoserwhile nothign bad would happen, installing those new abi kernels is useless as he can't use them.18:39
smoserjjohansen, bug 52001518:55
uvirtbotLaunchpad bug 520015 in linux-meta-ec2 "bad dependencies on karmic linux-ec2, linux-image-ec2" [Undecided,New] https://launchpad.net/bugs/52001518:55
sponzorhttp://pastebin.com/m75d80fca what should be wrong? after this only established connections work. if i try to connect to server on 80 ssh etc.. everyting is denied. and ping from server also? any ideas?18:56
sponzornevermind i got it19:04
ZimCSHey guys.  I am not really new to linux, but usually have the installer partition the drives for me.  This time I would like to do it myself using a software RAID 1 setup.  If I make a boot partition and and apply it to RAID, if disk 1 fails and is removed from the system and the system is rebooted, will drive to be able to boot the system using GRUB?19:17
alkosmurfHi. How do i setup MRTG for Apache2 on Ubuntu-Server edition?19:17
lifelessI hear and obey19:19
kirklandlifeless: yo19:21
lifelessso I had a few teething problems I've filed as bugs19:21
lifelessmissing directories19:21
kirklandlifeless: okay, so i gather you're running everything on a single machine?19:21
lifelessmissing files19:21
kirklandlifeless: yup, yup, i was looking into those19:22
lifelessits a karmic desktop (started life as jaunty)19:22
kirklandlifeless: i'm surprised you're the first to hit that19:22
lifelesskirkland: I'm not19:22
lifelesskirkland: there are a bunch of random 'help me' on the web that appear to have the same symptoms I saw at different stages19:22
kirklandlifeless: oh, interesting; i wonder how network manager will play with eucalyptus' crazy networking and bridging stuff19:22
lifelessNM is disabled19:22
kirklandlifeless: nice19:22
kirklandlifeless: i'll nail those missing dirs in the packaging19:23
lifelesswell, I say disabled - told not to touch etho0/br019:23
kirklandlifeless: i just need to figure out which package they belong in, and why others don't hit that problem19:23
lifelessI'm happy to purge and retry tests for you once we get it all working ;)19:24
lifelessnow, networking19:24
lifelessiface eth0 inet manual19:24
lifelessauto br019:24
lifelessiface br0 inet dhcp19:24
lifeless    bridge_ports eth019:24
lifeless    bridge_fd 919:24
lifeless    bridge_hello 219:25
lifeless    bridge_maxage 1219:25
lifeless    bridge_stp off19:25
lifeless(just under my pastebin threshold :P)19:25
lifelessthe iface eth0 stanza tells NM to go jump :)19:25
kirklandlifeless: okay19:25
kirklandlifeless: looks okay to me19:25
lifelessmy home server is, it does DNS and DHCP19:26
kirklandlifeless: cool19:26
kirklandlifeless: can you set static dhcp leases?19:26
lifelesswhen I installed UEC I told it
kirklandlifeless: okay, and you told your dhcp server not to touch that range?19:26
lifelessas 'public addresses', but that seemed silly in hindsight: nothing could get to it19:26
lifelesskirkland: yah19:26
lifelessit allocates down in the 20's19:26
kirklandlifeless: yeah, we should work on that wording19:26
kirklandlifeless: sabdfl said he didn't like the wording either19:27
lifelessso, given I have at home19:27
lifelessand complete control19:27
lifelessand I don't care what happens above (say) 8019:27
kirklandlifeless: okay19:27
lifelesswhat should I have put it ?19:27
kirklandlifeless: well, just any range of addresses that eucalyptus can assign to your vm's19:28
kirklandlifeless: and not conflict with ips that your dhcp server will dole out19:28
lifelessand by assign, it has them dhcp initially on virbr0, and then later on br0 ?19:28
lifelessbecause they seem to end up with my main dhcp server issuing ip addresses to them19:29
lifelessthey get   for example19:29
kirklandlifeless: grep VNET_MODE /etc/eucalyptus/eucalyptus.conf19:30
glphvgacshi, apt-get install openssh-server says pkg is not avail. (samba karmic) /etc/apt/sources.list looks fine, I think19:30
kirklandlifeless: https://help.ubuntu.com/community/UEC/Topologies19:30
kirklandlifeless: your topology is "1 Physical System"19:30
kirklandlifeless: see the notes there19:30
lifelessok, so I need to use SYSTEM ?19:31
kirklandlifeless: i have never personally installed this topology; it was actually Ng who told me this was actually possible, and added the entry to that wiki page19:31
kirklandlifeless: yes, according to Ng;  we might need his assistance to move you along much further19:31
kirklandlifeless: otherwise, you're welcome to attempt my suppositions19:31
jjohansensmoser: ogasawara updated the linux-ec2 meta package and it should be uploaded soon19:31
lifelessok, well I can switch to SYSTEM in a second now that I know of the 'CLEAN=' upstart parameter19:32
smoserjjohansen, so what happened ?19:32
smoserit appeared that the meta depended on a newer version that was available?19:32
kirklandlifeless: yes, CLEAN=119:32
kirklandlifeless: also ....19:32
lifelesskirkland: I'm guessing that the CC is expecting to issue 172.19.1.x addresses to the VM's19:33
kirklandlifeless: have you hit a problem where it takes a *really* long time to get your CLC to start responding to requests?19:33
lifelesskirkland: god yes19:33
kirklandlifeless: yes; the 172.* is the internal addresses that nodes can use to communicate with one another19:33
lifelesskirkland: and massive long waits for 'ec2 metadata server' on the console19:33
kirklandlifeless: and the 192.* is how external network users talk to the vms19:33
lifelesskirkland: and finally when it does come up the key doesn't seem to be on the vm19:33
kirklandlifeless: okay, I think have a fix for that19:33
kirklandlifeless: it's in the ~ubuntu-virt ppa, or I can give you a 2-line patch to add to your /etc/init/eucalyptus.conf19:34
kirklandlifeless: i'd like to SRU this fix, but I'm looking for someone to verify it ;-)19:34
lifeless2-line me baby19:34
kirklandlifeless: http://paste.ubuntu.com/373451/19:35
lifelessactually what do you mean by CLC to start.. just so that I know we're measuring the same thing.19:35
kirklandlifeless: just apply those two lines by hand19:35
kirklandlifeless: that's the full debdiff19:35
jjohansensmoser: I'm not sure.  This was the first round of updates to ec2 and the arm topic branches, there was a series of updates and the final abi is 304.10 so there was more than one thing missing19:35
kirklandlifeless: well, there's a boatload of related problems...  basically, takes a long time for the web interface on 8443 to respond19:35
kirklandlifeless: which stems from the fact that the CLC can't talk to the database19:36
lifelessah, no I see the API service timing out19:36
lifelessgiving permission errors19:36
kirklandlifeless: right, i think that's related19:36
lifelessand then it will come good19:36
lifelessso thats applied19:36
lifelessrebooting - clean test for you19:36
kirklandlifeless: yeah, so the problem is that the ip connection tracker from iptables is not getting loaded before eucalyptus services come up19:36
kirklandlifeless: so network connections established before the network settles get dropped19:37
lifelessmakes sense19:37
kirklandlifeless: eventually (10-20 minutes later), everything resets, and it starts working again19:37
kirklandlifeless: those two iptables commands ensure that iptables comes up in time19:37
lifelessit'll be a sec before that machine comes back :)19:37
kirklandlifeless: no problem19:37
lifelessso to fix one-machine installs19:40
lifelesskirkland: which service embeds the crazy dhcp server?19:40
kirklandlifeless: the eucalyptus-cc19:40
kirklandlifeless: this is the "cluster controller"19:40
Mackcan anyone help me install the ioncube loaders?19:40
kirklandlifeless: nodes are associated with -cc's19:40
kirklandlifeless: well, nodes are associated with clusters19:41
kirklandlifeless: clusters are controlled by the cc19:41
kirklandlifeless: the cc doles out the dhcp addresses to the VMs19:41
kirklandlifeless: i *think* this should be disabled, though, when you're in SYSTEM mode19:41
kirklandlifeless: i *think* your networking is pretty much on your own, if you're in SYSTEM mode19:41
lifelesskirkland: so, to make MANAGED_NOVLAN work with a single machine19:42
kirklandlifeless: don't think it can be done19:42
lifelesskirkland: I don't see why not, iptables is sex19:42
lifelesskirkland: filter dhcp replies that are not from the cc and heading to nodes19:43
* Mack can't get the ioncube loaders installed no matter what i try19:43
kirklandlifeless: i defer to those with more iptables notches in their bedpost19:43
lifelesskirkland: I've made a note to try19:44
lifelessif I get it happy with something crude I'll let the devs polish it ;)19:45
kirklandlifeless: ack19:45
kirklandlifeless: okay, so your system is back up now?19:45
RoAkSoAxkirkland, Maybe you can help me out. What network management tools are good to monitor networks that implement the UEC and, by any chance, do you have any document on how network have changed to be used in UEC?19:45
lifelesskirkland: yes seeing if it hangs like before19:45
lifelesskirkland: the ugly hanging I was seeing appears gone19:46
lifelesskirkland: \o/19:46
kirklandlifeless: nice19:46
kirklandRoAkSoAx: docs are all under: https://help.ubuntu.com/community/UEC?action=fullsearch&context=180&value=UEC%2F&titlesearch=Titles19:46
lifelesskirkland: so please SRU that :)19:46
RoAkSoAxkirkland, thanks :)19:47
lifelesswhats the bug number19:47
kirklandlifeless: okay, i'll upload to -proposed19:47
kirklandlifeless: if you could update to that package and ack it at your earliest convenience, that would be shiny19:47
lifelesskirkland: sure, drop me a ping when its built in -proposed and I'll do that19:48
Mackkirkland: do you know how to install the ioncube loaders with php5?19:49
lifelessjust on the bug wil be fine, I've subscribed19:49
kirklandlifeless: uploaded19:49
kirklandMack: no idea whatsoever19:49
Mack:( i cant get this to wrk19:49
lifelesskirkland: ok, now I get stuck on Waiting for EC2 meta-data service19:50
lifelesskirkland: which is I presume another ip routing fail19:50
kirklandlifeless: what's sticking?19:50
kirklandlifeless: is this running a VM?19:50
kirklandlifeless: okay, we need smoser for that19:50
kirklandsmoser: ping19:50
lifelessINSTANCEi-40670700emi-E027107D192.168.2.101172.19.1.3running mykey ...19:50
lifelessbah copy paste _fail_19:50
lifelesssmoser: I have a new all-on-one machine UEC19:51
kirklandsmoser: lifeless is having trouble starting an instance on 9.1019:51
kirklandsmoser: hanging at ec2 meta-data19:51
smoseris the metadata service running?19:51
kirklandsmoser: what's the bug on that one, and is there a fix, or workaround for Karmic19:51
lifelessmy nodes are coming up ok, and being given a public ip like and a private lke
sarthorHi, My pppoe-server is according this url "  http://ubuntuforums.org/showthread.php?t=291926  " and is working, but username  and passwords in /etc/ppp/pap_secrete "sarthor9"      *               "99999999" can be used by several clients on the same time from the different computers, how to solve this problem? help19:51
smoserwhat odes that mean?19:52
lifelesssmoser: I have one physical machine running sc, cc, nc, clc19:52
smoserif there is no metadata service (and i think som eof the Eucalyptus modes have no metadata service) than images in karmic will hang waiting for it.19:52
lifelesssmoser: its a test environment so that I can check a code change I'm making to a EC2 using project to UEC enable it works19:52
lifelesssmoser: its running in MANAGED_NOVLAN at the moment19:53
lifelesshow can I tell if the metadata service is running19:54
smoserwell, reading about that mode http://open.eucalyptus.com/wiki/EucalyptusNetworking_v1.5 and http://open.eucalyptus.com/forum/startup-scripts-vms , it "should" be running19:54
smoserbut it would appear that it is not.19:54
smoserlifeless, what i would suggest, for debugging is:19:54
smoser- mount image loopback19:55
Macksmoser: you know how to install ioncube?19:55
smoser(this is karmic, right)?19:55
lifelesskirkland: note that I will need to get MANAGED_NOVLAN working :P19:55
lifelesssmoser: yes19:55
smoserkarmic instance19:55
smosermount image loopback19:55
lifelesskarmic on karmic19:55
lifelessemi-E027107D specifically19:56
smosermodify  /etc/ec2-init/is-compat-env (or remove it -- i think thats the file, i have19:57
smoserand put your public ssh keys manually into /root/.ssh/authorized_keys (or /home/ubuntu/.ssh/authorized_keys)19:57
smoserumount image19:57
smoserupload image, register, boot19:58
smoserthen, i think that what you will find is that there is no metadata service.19:58
lifelessso, the metadata service is 8773 ?19:58
smosermetadata service is a web service that runs on forom the guests point of view19:58
smosershouldn't be 8773, should be 8019:59
smoserand its per-instance data19:59
smoserso each instance has a different world view of that web service19:59
lifelesssmoser: its dnatted to 877319:59
lifelessI totally agree it should be port 80 :)19:59
lifelessDNAT       tcp  --  *      *     tcp dpt:80 to:
smoserthat might be magic20:00
smoseridont know.20:00
lifelessI'll follow this thread20:00
smoseryou'd have to see how its configured elsewhere.20:00
lifelessI suspect insufficient mangling in iptables20:00
smoserbut seriously, mount hte image and modify it as i suggested.20:00
lifelesswill do20:00
smoserthen you should be able to ssh to it and poke around from inside20:00
smoserfwiw, there is a package now https://launchpad.net/ubuntu/+source/cloud-utils that has some scripts to make publishing an image easier.20:01
smosermainly 'uec-publish-image' might be of use. its one command for bundle upload register20:01
kirklandsmoser: saaaaweeet20:02
kirklandsmoser: i so need that20:02
kirklandlifeless: send along patches, if you have any20:02
kirklandlifeless: and, of course, I'll note that the lucid packages are ***far*** better than the karmic ones :-)20:03
lifelesskirkland: do they depend on other changes in karmic ?20:04
lifelesskirkland: or could I reasonably sensibly snarf eucalyptus-* from lucid?20:04
kirklandlifeless: hrm ...  well, there's a few jars and libraries that have been updated20:05
smoserlifeless, also, i'm sure that kirkland and i are not the only people that would love to see "all in one system" working20:05
kirklandlifeless: so i can't quite recommend that yet20:05
kirklandlifeless: and we haven't yet solved the karmic -> lucid euclayptus upgrade path yet20:05
lifelesskirkland: so you're going to tease me. Fffffiine20:05
kirklandlifeless: well, just saying :-)20:05
smoserbut then as soon as you get that someone is going to want you to do it with virtual-box, or no-hardware-virt ...20:05
lifelessseriously though, I don't care about upgrade20:05
kirklandlifeless: and warning not to live-upgrade to lucid yet and expect UEC to work20:06
lifelessI can purge the packages and start over, if it is likely to make a different20:06
smoseri suspect that there is a bug with the metadata service in this setup.20:06
lifelesssmoser: I'm sure there is, I have money on iptables20:06
Mackso nobody here knows how to install ioncube? :(20:06
lifelessMack: I don't even know what it is.20:06
Macklifeless: stupid encoder php devs use so you cant steal their code20:07
Mackbut the loader required to use encoded files is such a pain to install20:07
smoserMack, sorry , i had no idea what it was either.20:07
Macknobody knows what it is, and the ioncube support site sucks :(20:08
kirklandencoding php files sounds like an awful idea; i'll avoid ioncube20:08
kirklandlifeless: i can try to push eucalyptus to a karmic ppa and see if it builds, if you like20:09
kirklandlifeless: i just haven't even tested the build, so i don't know how deep the rabbit hole goes20:09
lifelesskirkland: please!20:10
lifelesskirkland: that would rock20:10
kirklandlifeless: pushed to ~kirkland ppa for karmic20:15
smosermathiaz, ping20:16
smoserhttps://blueprints.launchpad.net/ubuntu/+spec/server-lucid-aws-client-libraries . i'm not shooting for main inclusion on those, right?20:16
smoserie, its ok for me to build-depend or runtime depend on universe20:16
kirklandlifeless: you *might* need a newer euca2ools too20:17
kirklandsmoser: i doubt we're looking for main inclusions on those libraries20:17
lifelesskirkland: you know what to do :)20:17
kirklandlifeless: heh20:17
smoseryou "should not" need a newer euca2ools20:18
smoserthat is something that actually needs to function, as the api has not changed so we will need to support older euca2ools clients interacting with lucid euca20:18
lifelesswhats the poll frequency on the wait-for-metadata thing20:20
kirklandlifeless: euca2ools karmic build pushed to same ppa20:22
lifelesskirkland: thanks20:22
kirklandsmoser: oh20:22
kirklandlifeless: okay, you might not need newer euca2ools20:22
lifelessI won't install it until it appears obvious that its needed ;)20:22
uvirtbotNew bug: #520058 in mysql-dfsg-5.0 (universe) "package mysql-server 5.0.67-0ubuntu6 failed to install/upgrade: subprocess pre-installation script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/52005820:23
smoserlifeless, i think it tries, waits for 2*N seconds, adds one to N, tries again.20:23
smosersomething like that.20:23
smoserit will timeout in an hour or so, or maybe half a day20:23
smoserits painful20:23
kirklandsmoser: jeebus20:24
kirklandsmoser: that sucks20:24
smoseryeah. soren copied the boto author on that.20:24
lifelessso, ipv6 for uec ? :)20:25
kirklandlifeless: i think it supports both20:28
kirklandlifeless: kinda20:28
lifelessuses dnsmasq, which doesn't do ipv6  ;)20:28
kirklandlifeless: tries to, anyway20:28
lifelesskirkland: did you see my http_proxy bug ? kinda fun20:32
kirklandlifeless: not yet20:32
lifelesson euca2ools20:32
_lance_i seem to have hosed my system with a careless apt-get install. any easy way to revert a bunch of packages to hardy? i've been doing a few by dpkg -i, but there's dozens to go.20:39
bwallenMy machine sounds like it's constantly powering off and on my second hard drive. What causes this?20:41
lifelesskirkland: is there a document somewhere that describes /what/ euca does with networking as nodes come up or down?20:55
lifelessor should I just finish reverse engineering it?20:55
Sorrelhey guys21:16
SorrelI'm having an issue setting up dhcp21:16
Sorrelon my 9.10 box.21:16
SorrelI have installed dhcp3-server21:16
Sorrelbut it comes up and says that it can't start21:17
SorrelI'm setting this up on a small local network21:17
Sorrelany suggestions?21:17
Spanglish_7776sorrel, have you tried to configure it?21:18
SorrelI have edited the dhcpd.conf21:18
Sorrelstill fails21:18
Sorrelto start21:18
Sorreland I can't seem to find the log file21:18
Spanglish_7776did you check the error log? (possibly syslog)21:18
Spanglish_7776maybe syslog.21:18
Sorrelcd /var/log/syslog21:19
Spanglish_7776cd /var/log21:19
Spanglish_7776vi or nano syslog21:19
lifelessdon't use vi on syslog, has a tendancy to rename files if you edit by mistake :)21:20
lifelessless /var/log/syslog21:20
Spanglish_7776or tail ;)21:20
lifelessanyhow, dhcpd logs to daemon.log, I think21:20
Sorrelis there a place that I can upload this too?21:21
Sorrelnvr mind21:21
Sorrellooking at the file right now.21:21
Sorrelit's a little overwhelming :)21:22
Spanglish_7776try to start dhcpd and then check the end of the file21:22
lifelesskirkland: all working21:23
lifelesskirkland: starting documentation now21:23
SorrelNot configured to listen on any interfaces!21:28
Sorrelthat's what I'm getting in the tail end of the log file21:28
Spanglish_77761 sec21:28
Sorrelgoogleing now21:29
Spanglish_7776give that a go, ubuntugeek has been invaluable to me.21:29
Sorrelthanks I will take a look21:30
Spanglish_7776Sorrel: INTERFACES=”eth0″ may be your fix21:30
Spanglish_7776where eth0 specify your interface (usually eth0, but verify)21:31
SorrelI have a 1 and a 021:31
Sorrelit was complaining on the install that I only had one NIC21:31
Sorrelso I put in another and then did the install21:32
Sorrel( LTSP from the alt CD )21:32
Sorrelthe documentation is really ambiguious for LTSP21:33
Spanglish_7776which one is the active interface, eth0 or 1?21:33
Spanglish_7776you're trying to get LTSP going?21:34
SorrelI guess 1 would be21:34
Spanglish_7776what I mean is, which interface is plugged into the network where dhcp is going to be used.21:35
Spanglish_7776i wonder why you need two intefaces.21:37
Spanglish_7776Sorrel, you using ubuntu?21:38
Sorrelit doesn't say21:38
Sorrelwell it's from the alternatives CD21:38
Sorrelbut yeah21:38
Sorrel9.10 I think21:38
Spanglish_7776ltsp iso, gotcha21:38
Sorrelall sorts of issues21:38
Spanglish_7776yea, a few years ago, I got it going but didn't really utilize it. it was a pain, that is for sure.21:40
Spanglish_7776Sorrel: on the dhcpd.conf, you specified an INTERFACES="eth1" and tried again?21:41
Sorrelyeah gives me21:41
SorrelNo subnet declaration for eth0 (
SorrelFeb 10 16:40:53 ubuntu dhcpd: ** Ignoring requests on eth0.  If this is not what21:41
SorrelFeb 10 16:40:53 ubuntu dhcpd:    you want, please write a subnet declaration21:41
SorrelFeb 10 16:40:53 ubuntu dhcpd:    in your dhcpd.conf file for the network segment21:41
SorrelFeb 10 16:40:53 ubuntu dhcpd:    to which interface eth0 is attached. **21:41
Sorrelno idea what's happening21:42
lifelessyou haven't configured it with a range that is to be served out on eth021:42
Spanglish_7776yep, looks that way.21:42
Sorreldo I have to specifiy that in /etc/dhcp3/dhcpd.conf ?21:43
Sorrelwhat's in there now21:43
Sorrelddns-update-style none;21:43
Sorrel#option domain-name-servers,;21:43
Sorreldefault-lease-time 86400;21:43
Sorrelmax-lease-time 604800;21:43
Sorrelsubnet netmask {21:43
Sorrel        range;21:43
Sorrel        option subnet-mask;21:43
Sorrel        option broadcast-address;21:44
Sorrel        option routers;21:44
mathiaz!paste | Sorrel21:44
ubottuSorrel: For posting multi-line texts into the channel, please use http://paste.ubuntu.com | To post !screenshots use http://tinyurl.com/imagebin | !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic.21:44
Sorrel( is it okay to cp that much into the chat window?21:44
Roxyhart0i i have a question. can i just move a ldap certificate form a server to another i mean without create again?21:45
Spanglish_7776Sorrel: ifconfig, is your eth0 on the subnet?21:46
Sorrelit seems to be completely ignoring that file21:47
Sorrelthere's my ifconfig21:48
Spanglish_7776ah, I don't think dhcpd likes that.21:49
Spanglish_7776you're specifying and you have a different subnet alltogether21:50
Spanglish_7776do you know how to configure the IP for the second NIC? eth121:50
Sorrelbeen trying to figure it out21:51
Spanglish_7776shows you how to set the ip via command line21:51
Spanglish_7776are you on the physical machine? (the one you're trying to configure ltsp on)21:52
Sorreldo I need to get off it?21:52
Spanglish_7776no  do this, plug in the second NIC(eth1) and21:53
Spanglish_7776follow the instructions and use eth1 instead of eth021:53
Sorrelthat page really doesn't want to pull up for me21:53
Spanglish_77761 sec21:55
guntbertSorrel: try again21:55
Spanglish_7776hope you're familiar with vi.21:56
Spanglish_7776you can substitute nano if you want to.21:56
SorrelI like nano a good bit better :)21:57
Spanglish_7776use nano (me too)21:58
Roxyhart0hi can i use a ldap certificate from a difirent server, i mean move it?22:01
kirklandbug 51891122:06
uvirtbotLaunchpad bug 518911 in cloud-utils "[MIR] cloud-utils" [Undecided,New] https://launchpad.net/bugs/51891122:06
SorrelRoxyhart0,  have you tried copying the file from the old server onto the new one22:06
uvirtbotNew bug: #520116 in mysql-dfsg-5.1 (main) "package mysql-server-5.1 5.1.37-1ubuntu5 failed to install/upgrade: subprocess new pre-removal script returned error exit status 1" [Undecided,New] https://launchpad.net/bugs/52011622:06
SorrelIDK if it will work22:06
Sorrelbut you could try it.22:06
Roxyhart0hi Sorrel...no yest, it is just copy that?22:07
Sorrelno idea22:07
Sorrelbut it seems like it could work22:07
Sorrelthat's going from the assumption that linux treats everything like a file22:08
Sorrel( which it does )22:08
Roxyhart0i will try22:08
Sorrelif you could get a server setup and put the file in the right place22:08
Sorrelyou may have to edit some config files as well.22:09
SorrelSpanglish_7776: I think I got it to work22:10
Sorrelbe back in a sec22:11
lifelesskirkland: success22:12
kirklandlifeless: \o/22:12
kirklandlifeless: do you have a list of hacks required?22:12
lifelesskirkland: no hacks22:12
lifelesskirkland: the PackageInstall instructions created a bad config22:12
lifelessbecause eucalyptus is a tad redundant with the live system config, changing that config means you need to also change eucalyptus - specifically putting stuff onto br0 not eth022:13
Nglifeless: did you get your one-machine setup going?22:14
lifelessNg: yes22:15
lifelesseasy once I finished reverse engineering what it was doing22:15
lifelessNg: I've rewritten the one machine topology notes22:16
kirklandlifeless: cheers, thanks22:17
kirklandNg: i'm curious why the Bug:123456 works on wiki.ubuntu.com, but not on wiki.ubuntu.com/community ?22:17
lifelesskirkland: no probs; let me know if anything is unclear. I'm happy to write more prose if you want.22:17
kirklandNg: are they different instances?22:17
kirklandlifeless: thanks for documenting; i might bust those config instructions out to a separate page, and link to it from the Topologies page22:18
lifelesskirkland: pretty sure they are, yeah.22:18
mathiazkirkland: are you still doing testing with tamarind?22:18
kirklandlifeless: just to keep the Topos page simple22:18
lifelesskirkland: sure22:18
kirklandmathiaz: i've been side tracked; i need to run the tests on the current Topo/install22:18
lifelesskirkland: I'm going to blog as well I think22:18
kirklandlifeless: thanks for doing in any case22:18
kirklandlifeless: please do ;-)22:19
mathiazkirkland: ok - do you plan to reinstall soon?22:19
kirklandmathiaz: let me kick the test ...22:19
* kirkland puts everyone else on pause22:19
mathiazkirkland: I'd like to test the new code I added for installation22:19
kirklandmathiaz: how long does your tests take to run?22:19
mathiazkirkland: I don't want to *reinstall* the UEC now22:19
kirklandmathiaz: oh22:19
mathiazkirkland: just test that the layout on tamarind is correct22:20
kirklandmathiaz: okay, i won't be reinstalling now22:20
mathiazkirkland: great - thanks22:20
kirklandmathiaz: http://pastebin.ubuntu.com/373535/22:21
mathiazkirkland: hmm - twisted is not installed on tamarind22:21
mathiazkirkland: hm wait22:21
mathiazkirkland: where did you run that command?22:21
lifelessis there a simple way to get the root user enabled in the UEC images?22:21
kirklandmathiaz: kirkland@tamarind:~/uec_testing/testing$  ./uec_test.py -c config_single.yaml22:22
mathiazkirkland: right - you can't run the test on tamarind22:22
kirklandmathiaz: oh, i need to be on the CLC22:22
mathiazkirkland: because the euca2ools commands are not there22:22
mathiazkirkland: yop22:22
mathiazlifeless: hm - we're really trying to avoid the use of the root account on UEC images22:23
Nglifeless: I'm curious now what I did to make it work. I wasn't really making notes at the time22:25
Ngkirkland: as far as I know they're the same wiki22:26
Nglifeless: ah yes, I set my bridge up identically to how you've done it. The only difference is I didn't bother even trying MANAGED_NOVLAN, I just went for SYSTEM so the nodes would appear as any other machine on my LAN22:28
lifelessmathiaz: you may be trying to avoid it, but umpteen million third party tools assume it is where the key is loaded when you use EC222:28
NgI hope you have something significantly more useful to do with you local cloud than I do. I only really set it up to see if I could ;)22:28
lifelessNg: making Hudson run with UEC22:29
lifelesswhich it looks like it would, if root was usable.22:29
mathiazlifeless: what do you mean by the key is loaded?22:29
lifelesseuca-run-instances -k mykey emi-E027107D22:29
lifelessinstalls a ssh key to use to log into the machine and $do $stuff22:29
mathiazlifeless: if you use the ubuntu account instead, you'll have passwordless sudo22:29
lifelessmathiaz: yes, but that means changing about 4 layers of code and writing a UI to say 'use the ubuntu account and do sudo'22:30
mathiazlifeless: right - the root account is disabled in the Ubuntu images - instead use the ubuntu account with sudo22:30
lifelessmathiaz: which is _not_ the same as EC222:30
dasunsrule32I am having a grub issue I believe, I cannot boot into the system. I have a seperate part for /boot on /dev/sda5, / is /dev/sda6. How can I go about getting this to boot again?22:30
mathiazlifeless: well - it's not the same as the images published in EC222:30
lifelessthus it gets in the way of 'seamless transition'22:30
* lifeless whinges some more22:31
TrickFinlay2ooops sorry22:31
mathiazlifeless: right - we had that discussion at the very begining22:31
lifelessmathiaz: can one take the AMI's from EC2 and just use them ?22:32
mathiazlifeless: on one hand there is the Ubuntu policy that says: not root account enabled by default22:32
mathiazlifeless: OTOH there is the fact that EC2 images are run with the root account by default22:32
mathiazlifeless: yes22:32
lifelessmathiaz: I can see the rationale, I'm just expressing my displeasure at the time and effort it is costing me; and the suspicion I have that it will add this cost and overhead to many other users too22:33
mathiazlifeless: it's an issue with the image itself and its sets itself up22:33
mathiazlifeless: running an unmodified EC2 EMI should work22:33
mathiazlifeless: *porting* an application to run on an UEC image requires some work22:34
lifelessmathiaz: This means that you can't seamless transition though: when you exceed UEC capacity you need to move to a different image to spill onto EC222:34
lifelessmathiaz: thats the bit I have trouble understanding why we did this22:34
mathiazlifeless: not really - the image that is running on UEC can be rebundled to EC222:35
mathiazlifeless: the part that sets the ssh key for the root|ubuntu account is done by the image itself22:35
lifelesssure, if you weren't using vanilla images22:35
lifelessmathiaz: yes, I am aware of that, it looks it up in the 16KB user data region or thereabouts.22:36
TrickFinlay2if I want to install JeOS from my 9.10 CD, is it F4 -> OEM install?22:36
lifelessmathiaz: given that UEC lives behind a firewall, I don't really see that avoiding root is really important.22:38
* lifeless adds this to his blog post22:39
mathiazlifeless: well - it's a policy in Ubuntu to not enable the root account by default22:41
lifelessthen the EC2 Ubuntu images should not enable it either ;)22:41
mathiazlifeless: whether you're behind a firewall or not is irrelevant here22:41
mathiazthe EC2 Ubuntu images don't have the root account enabled by default22:41
lifelessmathiaz: I think it is relevant, because the policy exists to solve specific problems, not as a fait accompli22:41
lifelessTrickFinlay2: its ubuntu-vm-builder these days22:42
lifelessTrickFinlay2: there is a wiki page about it.22:42
dindasommer: ping - you around?22:43
mathiazlifeless: hm - I'm not sure I get your point22:44
mathiazlifeless: none of the products in Ubuntu have the root account enabled by default22:44
mathiazlifeless: we decided that images published to EC2 and for UEC should behave the same way as any other Ubuntu product22:44
lifelessmathiaz: I thought based on your earlier comment that the Ubuntu official EC2 AMI's put the supplied key into the root account22:44
mathiazlifeless: oh no - I don't think so22:45
lifelessmathiaz: if they don't, then thats fine, its not a difference22:45
mathiazlifeless: I didn't not express myself correctly22:45
lifelessstill annoying to find bugs where folk assume it goes into root on EC2, but not unique to UEC in that case22:45
erichammondlifeless: http://alestic.com/2009/04/ubuntu-ec2-sudo-ssh-rsync22:45
mathiazlifeless: yes - that's one of the difference the Ubuntu images and other images22:46
erichammondlifeless: If you ssh to root@ on an official Ubuntu image on EC2, it prints out a message pointing you to use ubuntu@ instead.22:46
mathiazlifeless: you'd have to adapt your deployment/configuration scripts to use the ubuntu account when you deploy to an Ubuntu image22:46
lifelessmathiaz: yeah22:47
lifelesserichammond: thanks22:47
erichammondlifeless: In the above article I also describe one command which can enable the root ssh if you really want to use the EC2 standard instead of the Ubuntu  standard.22:47
erichammondI don't think I've been excommunicated yet for listing that.22:48
lifelessAs I need to change third party code anyhow22:48
jMyles_I'm interested in using subversion on my ubuntu-server, but I don't know where to begin.22:48
lifelessI'm just going to teach it how to use an arbitrary username22:48
lifelessand a command prefix when running admin code22:48
erichammondlifeless: The older community Ubuntu AMIs I published used ssh to root@ because I didn't think I could as an unknown individual sway the EC2 community to adopt the Ubuntu standard, but when Canonical started publishing AMIs, I encouraged them to do it the consistent (Ubuntu) way.  It did require a lot of back and forth and a compromise with passwordless sudo, but I am happy with the result.22:50
lifelessI was semi ranting because of the wrong end of the stick; I thought we were publishing UEC images differently to EC2 ones.22:52
lifeless*that* would be very hard to justify22:52
lifelessnow I'm totally happy (though still with code changes to make to get this thing working)22:53
erichammondlifeless: Not only is the behavior the same with EC2 and UEC, but the images themselves are exactly the same, except for the kernel.22:54
erichammondlifeless: Can you share what product you're working on?22:54
lifelessmoving the dx team test environment into the internal UEC instance once that comes online22:55
sherringhamjMyles_ : Google? https://help.ubuntu.com/9.10/serverguide/C/subversion.html22:57
=== rberger_ is now known as rberger

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!