[01:23] <irx> hello. i'm having problems with 2TB drive attached to 8.04 server x64. does show up as /dev/sdb, but fdisk can't handle it, nor dows it show any partitions
[01:24] <twb> Define "can't handle it"
[01:24] <irx> Unable to read /dev/sdb
[01:26] <irx> lshw does show some info, as does dmesg
[01:27] <irx> http://pastebin.com/TS7swJwP
[01:29] <irx> any ideas?
[01:29] <twb> Are you running fdisk as root?
[01:29] <irx> yes
[01:29] <twb> Dunno, then
[01:30] <twb> Could be a non-512b block device, and that is pissing it off
[01:30] <twb> You can use MBR disk labels up to 2TiB, so a 2TB disk will be fine
[01:32] <irx> so ... no luck on getting this to work?
[01:35] <twb> I have no idea
[01:35] <twb> I haven't experienced that issue before
[01:39] <irx> according to manual it uses 4096 bytes per sector, is that what you meant by non-512b block device?
[01:45] <irx> i'm assuming that this drive uses the new 4k format, so could that be unsupported by the 8.04 kernel?
[01:48] <twb> Could be
[01:48] <twb> I don't know if 8.04 handles those
[01:48] <qman__> could be, I don't remember when that support was added
[01:48] <qman__> however I also recall problems with fdisk in particular, have you tried parted?
[01:49] <qman__> or cfdisk, or anything else
[01:50] <twb> Mmm, IME they're all stupidly picky in different ways
[01:50] <irx> "parted /dev/sdb" just returns, nothing ...
[01:51] <irx> cfdisk also can't read the drive
[01:51] <qman__> probably too old a kernel then
[01:52] <irx> i'd love to update, but i'm using vmware server ...
[01:52] <irx> and that beast just requires me to use 8.04
[01:53] <qman__> well, provided you verify the drive works in a more recent system, there's not really anything else you can do
[01:54] <qman__> get a newer kernel, or get a drive that doesn't have 4k sectors
[01:54] <irx> drive came straight from my other machine, running windows server 2008 r2, no problems there ...
[01:55] <qman__> if the drive is using GPT, that may also be an issue
[01:56] <irx> don't think it's the case, dmesg shows it has 0 512b sectors
[01:56] <irx> and if i'd use gpt, fdisk would at least be able to access the drive
[01:57] <qman__> yeah
[01:58] <irx> so it's just that my 2.6.24 can't handle 4k drives and vmware won't allow me to update it. f**l
[01:59] <irx> guess i'll have to find another virtualization method then ...
[01:59] <cjwatson> IIRC the point where WD sent me a 4k sector drive so I could make Ubuntu support it better was 9.10 or 10.04 or so
[02:00] <cjwatson> yeah, changelogs suggest 10.04
[02:00] <cjwatson> before that parted definitely didn't understand it properly
[02:01] <qman__> yeah, that's about right
[02:01] <irx> so out of luck with this drive and this kernel?
[02:01] <qman__> bear in mind that 4k sectors didn't show up until the 3TB drives did
[02:01] <qman__> even though many 1TB+ drives now use them
[02:01] <cjwatson> I don't recall the kernel state, but the userspace was definitely not up to it in 8.04
[02:02] <qman__> I can't say without any doubt, but that's my best guess
[02:02] <qman__> you'd have to read through changelogs to know for sure
[02:02] <cjwatson> I'm pretty certain
[02:03] <cjwatson> sorry - it was a pretty invasive set of changes, not easily backportable
[02:04] <irx> well yeah, who would have thought that you'd need such an old kernel with those drives ...
[02:04] <irx> i guess it's time to get comfortable with KVM
[02:06] <irx> thanks guys :)
[02:09] <patdk-wk> hmm, I ran 4k sectore disks in 9.10 without any issues
[02:09] <patdk-wk> currently running 10.04 on it though
[02:13] <cjwatson> It might have worked by then if you could get it installed, ye.
[02:13] <cjwatson> *yes
[02:14]  * patdk-wk notes 3 out of 3 of them died a horrible death, so currently have 0 4k sector disks
[02:15] <cjwatson> I have two, but one is in a USB enclosure and that doesn't pass through the topology information, and one is in a currently-dead machine.  Sigh.
[02:15] <cjwatson> Must carve out time for some home sysadmin.
[02:50] <SpamapS> cjwatson: EC2 has pretty much relieved me of all home sysadmin duties.. the only "server" I have is my WRT54G :)
[02:50] <SpamapS> all media is stored on USB drives that I plug into whichever device needs to access said media.
[02:51] <SpamapS> Actually lately I don't even use those.. I just use Amazon/Netflix
[02:51] <cjwatson> You have a lot more bandwidth than I do
[02:51] <SpamapS> True, 12Mb down, 1Mb up
[02:51] <cjwatson> That, and I go over my charging thresholds as it is; it would be unaffordable to do everything over the network
[02:52] <SpamapS> metered bandwidth would definitely be a problem for me too
[02:52] <cjwatson> And yes, at 2Mb down it's not very exciting to put everything in the cloud
[02:52] <SpamapS> I dunno, I got rid of my home server when I got my first 1.5Mbit down/up connection
[02:52] <SpamapS> But I wasn't doing OS development at the time
[02:55] <cjwatson> SpamapS: http://www.phdcomics.com/comics.php?f=1456
[02:56] <SpamapS> :-D
[02:59] <qman__> eh, I could never do it that wy
[03:00] <qman__> I get mildy annoyed with gigabit
[03:00] <qman__> waiting for 10g consumer devices
[03:10] <SpamapS> qman__: what do you find yourself transferring?
[03:11]  * SpamapS asks, while waiting for erlang's 32MB source file to upload to the ubuntu archive
[03:11] <qman__> oh, everything
[03:12] <qman__> all the normal media, and now with HD video it's a lot bigger
[03:12] <qman__> but one of the more annoying tasks is backups
[03:13] <qman__> in truth my array is far from optimal and actually the limiting factor
[03:13] <qman__> but the main point stands, no way I could deal with my stuff on 10/100, much less go without my stuff on-site
[03:14] <SpamapS> qman__: but, what actual media are you transferring?
[03:15] <qman__> like I said, everything, TV and movies, music, games, I don't use discs anymore
[03:16] <SpamapS> I don't either.. but I just stream it in off the net
[03:16] <qman__> keep all my music in FLAC, and an HD show is usually over 1GB
[03:17] <SpamapS> I do only have 8GB of music.. so I am I think odd
[03:17] <SpamapS> I went through and purged all the music I hadn't paid for about 3 years ago.
[03:18] <qman__> most things I use it for, 10/100 would be acceptable, but just enough
[03:18] <SpamapS> qman__: I'm convinced that 90% of the hard drives sold to the consumer market are used primarily to house pirated content. :)
[03:19] <qman__> but I also do a lot of fixing computers, and that means many gigs of backups
[03:19] <qman__> I won't disagree
[03:19] <SpamapS> Yeah backups definitely aren't tolerable w/o Gigabit. :)
[03:20] <qman__> and I have about 6 users actually storing data here
[03:20] <SpamapS> which is why I have no data anymore.. my backup is U1 syncing my Photos/Documents
[03:21] <kirkland> hallyn: have you tried byobu's new feature to save and restore custom layouts?
[03:22] <kirkland> hallyn: configure your tmux however you want, with vertical and horizontal splits, etc.
[03:22] <kirkland> hallyn: and then ctrl-shift-f8, you'll be prompted to name your layout
[03:22]  * SpamapS wanders off
[03:22] <kirkland> hallyn: and then shift-f8, you'll see a list of layouts and you'll choose one
[03:22]  * kirkland waves at SpamapS 
[03:22] <kirkland> hallyn: it was pitti's idea
[03:23]  * SpamapS waves back on his way to find an opener for the last Leffe Blonde Ale he just found in the fridge
[03:23] <hallyn> kirkland: hm, that actually sounds good
[03:23] <hallyn> kirkland: and what's more, you know who might be swayed by that?
[03:23]  * kirkland tries not to concentrate on the word "actually"
[03:23] <hallyn> paulmck
[03:24] <kirkland> hallyn: heh :-)  he and i had a chat about something like that at uds
[03:24] <hallyn> well!  my first reaction was "why would i need that".  i'm quite the sceptic
[03:24] <hallyn> i know i was there
[03:24] <hallyn> and he likes to fire off his xterms in certain layouts.  byobu layouts might help him
[03:25] <kirkland> hallyn: cool, i'll reach out to him
[03:25] <kirkland> hallyn: i'm trying to get mouse-mode working well, but i'm not there yet
[03:25] <hallyn> eh, maybe best to just wait and mention it in person :)
[03:25] <kirkland> hallyn: good point ;-)
[03:25] <kirkland> hallyn: though i don't know when i'll see him again with the new job
[03:25] <hallyn> you know there's a guy (paul t) working on something that might compete with byobu
[03:25] <hallyn> oh, right
[03:25] <kirkland> hallyn: neat
[03:26] <kirkland> hallyn: who's paul t?
[03:26] <twb> Does byobu speak tmux now?
[03:26] <kirkland> twb: yes
[03:26] <twb> huh
[03:26]  * kirkland waits for twb's tirade...
[03:26] <hallyn> tagliamonte
[03:26] <twb> Why not just fix tmux
[03:26] <twb> Seeing as how it has an active upstream
[03:27] <twb> Incidentally, my netbook has no f8 key ;-)
[03:27] <hallyn> after all i would prefer dvtm to tmux's panes if only it would do scrollbar
[03:27] <twb> And can't type shift+<function key> in fbcon anyway
[03:27] <kirkland> hallyn: what's it called?
[03:27] <kirkland> twb: yeah, tmux is very active, that's nice
[03:27] <hallyn> i forget, he keeps renaming it, but look through planet.ubuntu.com, he keeps mentioning it there
[03:28] <hallyn> ah yes rokkaku
[03:28] <hallyn> rolling his own all the way down to an ansi escape library
[03:30] <kirkland> hallyn: i only see one post, http://blog.pault.ag/tagged/rokkaku
[03:30] <hallyn> he had an older one when it was all called shibuya i think
[03:31] <hallyn> http://blog.pault.ag/post/12642018666/silly-little-project-activity-for-this-week
[03:31] <kirkland> twb: what are your blocking issues for tmux?  I know you were complaining about it earlier this week...
[03:32]  * hallyn out
[03:33] <twb> hallyn: cough
[03:34] <twb> kirkland: the main ones are in http://cyber.com.au/~twb/.tmux.conf
[03:34] <twb> kirkland: bottom line is I couldn't make it do everything I do in screen, in an afternoon of effort
[03:34] <twb> i.e. inertia
[03:35] <twb> If I spent a couple of weekends I probably could, with some C patching, but ICBF and screen is already installed on every box I have to babysit, so it would be a bit like learning zsh or dvorak
[03:42] <kirkland> twb: interesting;  I had to solve a few of those same problems in byobu
[03:42] <kirkland> twb: particularly the lack of separate refresh frequencies for status items
[03:42] <twb> How did you do it?
[03:42] <kirkland> twb: which i solved with a cache based approach
[03:42] <twb> Yeah, figured :-)
[03:43] <kirkland> twb: not many other options
[03:43] <twb> HTFS
[03:43] <kirkland> twb: but i retrofited that to work with both screen and tmux
[03:43] <kirkland> twb: there is a nice upshot though
[03:43] <kirkland> twb: which is that if you have multiple byobu/screen or byobu/tmux sessions
[03:44] <kirkland> twb: they can all share the same cache file
[03:44] <kirkland> twb: and each cache is updated at its own defined frequency
[03:44] <twb> Meh
[03:44] <kirkland> twb: it's quite functional and comparatively efficient
[03:44] <twb> Incidentally that's that I did as a workaround too
[03:45] <twb> imapbiff (using IMAP IDLE) writes to a file; screen tails -f it, byobbu poll's it with tail -1
[03:45] <twb> Whereas before screen just ran imapbiff directly
[03:45] <kirkland> twb: the two main parts of the profile are:
[03:46] <kirkland> twb: http://bazaar.launchpad.net/~kirkland/byobu/trunk/view/head:/usr/share/byobu/profiles/tmux
[03:46] <kirkland> twb: and
[03:46] <kirkland> twb: http://bazaar.launchpad.net/~kirkland/byobu/trunk/view/head:/usr/share/byobu/keybindings/f-keys.tmux
[03:46] <kirkland> (excuse LP's bad color coding)
[03:47] <twb> I don't get colour anyway
[03:49] <twb> I think tmux would've been better off just using liblua for its config files, tho
[03:49] <twb> There's a screen git branch that does that
[09:10] <angelete2> hi
[09:11] <angelete2> i have my server with 2 hd, and 2 raid partitions (/boot and /)
[09:11] <angelete2> my mysql didn't started, and searching for more info, i found there was a problem with permissions in my filesystem
[09:12] <angelete2> when i looked the filesystem i found some files with ?????????? info instead of permissions, inode and so on
[09:12] <angelete2> so i made a fsck /dev/md2
[09:12] <angelete2> no my system does not reboot
[09:12] <angelete2> and says that there are no partition on /dev/md2
[09:13] <angelete2> what can i do?
[09:13] <twb> pastebin /proc/mdstat
[09:14] <angelete2> twb: it's fine, both mds are UU
[09:14] <angelete2> i would like to know what could cause this problem
[09:14] <twb> both as in md0 and md1, or md2 ?
[09:14] <angelete2> in both md1 and md2
[09:15] <angelete2> there wasn't md0
[09:15] <twb> Dunno why fsck is sad then
[09:15] <angelete2> now when i boot my server it shows only a initramfs console
[09:15] <angelete2> but i cannot access to fsck
[09:15] <twb> Did you read the error message at that console?
[09:16] <angelete2> no, i was working using ssh console
[09:16] <angelete2> but now i have to work using phisical console
[09:17] <angelete2> i'm gonna try to use a usb linux, give me 10mins,
[09:17] <angelete2> i'll be back
[09:34] <matti> http://i.imgur.com/2I6LQ.jpg :)
[09:42] <oCean> matti: please don't post offtopic here
[09:44] <matti> oCean: Heh, sorry.
[09:54] <koolhead11> hi all
[10:02] <uksysadmin> morning koolhead11
[10:03] <koolhead11> hey uksysadmin .
[10:05] <uksysadmin> so I got an answer from the OS team meeting regarding updates... there's gonna be a 2011.3.1 release of Diablo soon
[10:06] <twb> uksysadmin: uh, are you in the wrong channel?
[10:07] <uksysadmin> not really - discussing OpenStack Ubuntu packages on my 11.10 server is valid isn't it?
[10:07] <koolhead11> uksysadmin: sounds good. i am not touching OS till essex comes, not interested in getting frustrated anymore
[10:07] <koolhead11> twb: hello there
[10:07] <twb> uksysadmin: fair enough
[10:08] <twb> I guessed "OS Diablo" was some plan9 knockoff or something :P
[10:08] <koolhead11> twb: :P
[10:08] <koolhead11> lynxman: hola
[10:17] <Daviey> rbasak: Please can you sign up for some WI's regarding ARM, at the bottom of:
[10:17] <Daviey> http://status.ubuntu.com/ubuntu-precise/group/topic-precise-servercloud-service-orchestration.html
[10:18] <Daviey> hey uksysadmin o/
[10:18] <uksysadmin> Hi Daviey
[10:19] <uksysadmin> how goes it?
[10:19] <Daviey> uksysadmin: Not too bad.. How closely are you following the stable branches?
[10:19]  * rbasak looks
[10:20] <uksysadmin> quite - as we're not insane enouugh to touch Diablo in production, its not much of a priority - but its good for my sanity doing stuff that should work
[10:20] <uksysadmin> what's the plan for a bug fix set of packages coming to 11.10?
[10:21] <rbasak> Daviey: how many would be sensible?
[10:22] <Daviey> uksysadmin: Well.. zul and myself are on the openstack stable team.  We are currently looking at how best to backport it to Oneiric.
[10:22] <Daviey> We are working on some CI testing, but it's not quite ready.
[10:22] <Daviey> (almost!)
[10:23] <Daviey> rbasak: *shrug* :), for a release team perspective - i was just concerned so many are in unassigned state, and wanted to shrink the list :)
[10:25] <uksysadmin> awesome Daviey - I'll be lurking until its out ;-)
[10:25] <koolhead11> uksysadmin: :P
[10:26] <rbasak> Daviey: OK, I've taken three that I think would be most useful to do
[10:27] <koolhead11> rbasak: hey there
[10:27] <rbasak> hey koolhead11!
[10:31] <querier> Hi there: what is normal way to add/remove/start jobs onubuntu desktop running as server?
[10:31] <Daviey> uksysadmin: super :)
[10:31] <koolhead11> Daviey: sir. what is there 4 my plate :D
[10:31]  * koolhead11 needs some TODO :D
[10:32] <Daviey> koolhead11: Well Essex-2 is landing in Precise today. :)
[10:32]  * uksysadmin has a list of work as long as his arm for koolhead11 if he's that keen - shame not Ubuntu related
[10:32]  * uksysadmin is just firing up a Precise VM in prep for that!
[10:32] <koolhead11> uksysadmin: :P
[10:33] <koolhead11> Daviey: cool i will have  2 machine setup of precious plus openstack
[10:34] <koolhead11> uksysadmin: your yet to get me the cobbler manual partition preseed thing.
[10:34] <koolhead11> :p
[10:36] <uksysadmin> ahh, yes - I promised that didn't I?
[10:36] <uksysadmin> when I work it out - I'll let you know :p
[10:37]  * uksysadmin thinks installing everything under / is underrated
[10:37]  * koolhead11 wants to write blog on cobbler/ubuntu. 
[10:38] <koolhead11> uksysadmin: but what i have realized is this kickstart file its very perticular about syntax and spaces
[10:39] <uksysadmin> is that Python based too!?
[10:40] <koolhead11> uksysadmin: preseed in debian term. no it has notthing to do with python :P
[10:41] <uksysadmin> ;-)
[11:08] <querier> Is there a difference between ubuntu dekstop and server edition regarding the "startup" behaviour ?
[11:28] <jamespage> querier, there are some differences, yes
[11:29] <jamespage> querier, specifically which jobs do you want to add?
[11:30] <querier> jamespage: for example , does ubuntu server use /etc/inittab ? Im having problems with /etc/init/<job>.confs
[11:30] <jamespage> querier: no - both desktop and server use upstart as the init system
[11:31] <jamespage> you can still install regular init scripts /etc/init.d
[11:31] <jamespage> what issues are you having creating upstart configurations?
[11:31] <querier> creating is no problem...let me exaplain
[11:33] <querier> jamespage: on a (ubuntu desktop running apache et.c) livecd, that I produced, the initctl start <job> does not seem to work.
[11:33] <querier> jamespage: on the host  system which is equal it works jsut fine. (I used relinux to make the live cd)
[11:34] <jamespage> querier, do you get anything in /var/log/syslog?
[11:34] <uksysadmin> I'm just installing Essex-2 on Precise and doing my usual: apt-get install -y rabbitmq-server nova-api nova-objectstore nova-scheduler nova-network nova-compute glance qemu unzip it is pulling in X11 deps... what uses X11 - what is the easiest way of finding that out? (It never used to...)
[11:35] <jamespage> querier: you can also up the logging level by running sudo initctl log-priority debug
[11:39] <querier> jamespage: I just looked into the /var/log/syslog it does not tell me anthing about my attempts to start things.
[11:42] <querier> jamespage: now this was added to. (root) CMD (  [ -x /usr/lib/php5/maxlifetime ] && [ -d /var/lib/php5 ] && find /var/lib/php5/ -depth -mindepth 1 -maxdepth 1 -type f -cmin +$(/usr/lib/php5/maxlifetime) ! -execdir fuser -s {} 2>/dev/null \; -delete)
[11:42] <querier> jamespage: I dont understand what the maxdepth means?
[11:43] <jamespage> querier, only descend to one level of subdirectories when executing find
[11:44] <querier> jamespage: ic.
[11:50] <tewea> iam system admin one osf serevr iam using ubuntu to my server and in my serevr i cant acess it the page remotly ?it works somtime and after short of time it says"ERROR The requested URL could not be retrieved" and second thing i installed ssh allradey but i cant aces it remotly to the server
[11:50] <tewea> is there any body who can asssist me?
[11:55] <TeTeT> tewea: do you have console access to the system if ssh doesn't work?
[11:56] <jamespage> Daviey: minimal virtual grew again today
[11:57] <tewea> TeTeT:console means? by the way i can aces it even graphically
[11:58] <TeTeT> tewea: graphically is as good as console :)
[11:58] <tewea> TeTeT:ok
[11:58] <TeTeT> tewea: is there a firewall installed on the system? can you ssh localhost from it?
[12:00] <tewea> TeTeT:yeah it works
[12:00] <TeTeT> tewea: with /sbin/ifconfig you can find out your ip address, try that instead of localhost
[12:02] <tewea> TeTeT:yeah it works even when iam typing by it ipadress
[12:03] <TeTeT> tewea: hmm, what happens when you ssh from the outside to your server? can you ping it?
[12:05] <tewea> TeTeT:even this one is working proporly
[12:06] <querier> jamespage: the output php: :trying[sudo /sbin/initctl start id_._name] NOTHING. the job is /etc/init/id_._name.conf
[12:07] <TeTeT> tewea: ok, that's good. now with that being ok, onto the webpage
[12:07] <TeTeT> tewea: but wait a second - if something over the network works one time and not the other - is the assigned ip maybe taken already? E.g. two systems share the same ip address
[12:08] <tewea> TeTeT:but i can't accces it to the web page
[12:08] <TeTeT> tewea: is there some sort of network monitor running that could detect that?
[12:08] <TeTeT> tewea: can you access from the server via 'w3m http://localhost'
[12:09] <tewea> TeTeT:but i can't accces it to the web page it says"ERROR The requested URL could not be retrieved"
[12:10] <TeTeT> tewea: is there a webserver running on the server at all? what's the output of netstat -ntaup | grep 80
[12:12] <jamespage> querier, I need some more context to help - is php trying to run the start command?
[12:12] <querier> jamespage: yes it is. and as the correct user.
[12:13] <querier> jamespage: expected is " <job> start/running process_id"
[12:14] <jamespage> querier, anything when you up the logging level with sudo initctl log-priority debug?
[12:14] <jamespage> in (/var/log/syslog) that is
[12:20] <querier> jamespage: Ihave sent the only output I found regarding the certain job.
[12:21] <jamespage> querier, and you get nothing else when you run sudo /sbin/initctl start id_._name ?
[12:22] <querier> jamespage: if I do it by hand (user) everything is fine.
[12:22] <tewea> TeTeT: it says it established
[12:23] <TeTeT> tewea: so, what does w3m http://localhost do from the server commandline?
[12:25] <querier> jamespage: oh. when I do it now i am asked the password, which should not be, according to sudoers file.
[12:25] <jamespage> querier, thats probably your issue; it may be failing silently under php
[12:25] <tewea> TeTeT: it says it works
[12:25] <jamespage> querier: hence no syslog output
[12:26] <TeTeT> tewea: ok, so from the remote computer you can ssh to the server, but not access the webpage?
[12:27] <querier> jamespage: hmm yes seems so. sudoeres:[ user   ALL = NOPASSWD:  ... /sbin/initctl ]
[12:28] <tewea> TeTeT: from the rmote computer i can acees the web page but after some minute when i refresh the browser it says"ERROR The requested URL could not be retrieved" and complutly  i cant ssh it
[12:30] <TeTeT> tewea: you sure the server has a unique ip address?
[12:31] <tewea> TeTeT: yeah
[12:31] <tewea> TeTeT: but how can i make it shure it have unique ip adress
[12:32] <TeTeT> tewea: do you have access to another system in the same LAN?
[12:32] <tewea> TeTeT: yeah
[12:33] <TeTeT> tewea: ping the server, check sudo arp -a for the mac to ip assignment, repeat that every 10 mins and see if the mac sometimes changes
[12:33] <TeTeT> tewea: there should be more sophisticated tools, but I'm not aware of any
[12:45] <caribou> lynxman: remember when I told you that grub wasn't able to boot with /boot as XFS ?
[12:46] <lynxman> caribou: I do I do :)
[12:46] <caribou> lynxman: well, I was wrong; it's the way I moved /boot from ext2 to XFS that was wrong
[12:47] <caribou> lynxman: rebuilding a brand new VM with a native XFS /boot works
[12:47] <lynxman> caribou: I was quite certain that grub supported xfs, not so sure about jfs
[12:47] <querier> jamespage: hmm just now I retryied and it worked doing it by hand without being asked for password.
[12:47] <lynxman> caribou: cool :)
[12:47] <caribou> lynxman: I'll tell  you in a minute, I'm testing that now
[12:47] <lynxman> caribou: cool
[12:47] <jamespage> querier: you sudo session remains intact for a bit - so you will not be asked everytime
[12:47] <caribou> lynxman: but I just can't figure out how to move from on FS type on /boot to another without rebuilding the whole VM
[12:48] <caribou> lynxman: there must be hooks in grub to tell it that the FS has changed and that grub.cfg needs to be changed, but I haven't found yet
[12:49] <lynxman> caribou: maybe reinstalling grub might help?
[12:49] <lynxman> caribou: that's what I would do a least
[12:49] <caribou> lynxman: well update-grub doesn't seem to work
[12:51] <querier> Jamespage:But I did not enter the user password. I did root$ su user . user$ sudo /sbininitctl start job. now
[13:12] <angelete2> hi
[13:12] <angelete2> i've just installed my server, and i get a black screen
[13:12] <angelete2> but i have ssh access
[13:12] <angelete2> does getty use any kind of video driver??
[13:21] <patdk-wk> vt100?
[13:36] <esde> Hello all, quick question, I am running ubuntu server 10.04 for some small projects. Right now I'm uploading all my files to my webserver than WGET'ing them into my server box. This sucks. I'd like to set up an SMB share for the /var/www directory (at least) so that I can simply drag and drop the files I'm working with. I have installed the required components but I'm in need of assistance with the .conf file, specifically configuring the WORKGROUP
[13:36] <esde>  section, please help :)
[13:37] <Sander^work> Anyone uses pureftpd here?
[13:37] <Jasonn> yeah
[13:37] <Jasonn> I have
[13:37] <Jasonn> Sander^work: ^
[13:39] <Daviey> Jasonn: it grew again?!
[13:39] <Daviey> jamespage: ^^
[13:39] <jamespage> Daviey: yes - I think the base install for amd64 is now around 600MB
[13:39] <Daviey> jamespage: I think next week we'll have to dig deep to find out what is being pulled in.
[13:39] <jamespage> agreed
[13:39] <Daviey> crikey
[13:41] <zul> morning
[13:41] <Sander^work> Jasonn, I'm wondring how I can debug what's going wrong there.. I've tried to turn on the -D -D option for debug mode..
[13:43] <Jasonn> Sander^work: whats the problem?
[13:45] <Sander^work> Jasonn, i'm using pure-ftpd-ldap.. and I'm unable to log into the server.. just getting Login authentication failed. and it takes like 5 seconds to give that answer.
[13:59] <Sander^work> Jasonn, wondring what could be wrong with the pure ftpd server.
[13:59] <Jasonn> Sander^work: whats wrong with it...
[13:59] <Jasonn> whats not working
[13:59] <_ruben> Sander^work: sounds like your ldap server is timing out or something
[14:00] <Jasonn> what is the connection you have on that server?
[14:00] <Sander^work> _ruben, I tried to put in another ldap IP in the pureftpd config.. and then it answered pretty quick..
[14:00] <Sander^work> Jasonn, what you mean with connection ?
[14:00] <Jasonn> speed
[14:00] <Jasonn> sorry, should have specified that
[14:00] <Jasonn> and the uptime/quality
[14:01] <Jasonn> Honestly, I would just screw pure, and go with vsftpd
[14:01] <Sander^work> It's been quite stable actually..
[14:02] <Sander^work> Jasonn, but not lately..
[14:02] <Sander^work> Jasonn, it's 100Mbit, why?
[14:02] <Jasonn> I havent used it recently
[14:02] <Jasonn> just wondering whether the timeout would have to do with the quality of the connection
[14:02] <Jasonn> that is, if you are running a home server or something
[14:05] <Sander^work> _ruben, Do you know how I can debug that, what is wrong?
[14:06] <Sander^work> Jasonn, does vsftpd comes with an ldap module?
[14:06] <Jasonn> hummmmm
[14:06] <Daviey> hallyn: Around?
[14:06] <Jasonn> http://devzero.wikidot.com/debian
[14:06] <Jasonn> Sander^work: ^^
[14:07] <Jasonn> its a debian guide, but should be almost the same
[14:43] <zul> ttx: still around?
[14:48] <ttx> zul: yes
[14:48] <zul> ttx: have you had the ubuntu security team look at your rootwrapper yet?
[14:48] <zul> ttx: because i want to start using it
[14:49] <ttx> zul: yes, they reviewed it
[14:49] <zul> ttx: ok so we should be able to start using it next week
[14:49] <ttx> zul: you should ask them to cc you on the result...
[14:49] <zul> ttx: yeah ill talk to them
[14:49] <ttx> you still have my packaging patch ?
[14:49] <zul> ttx: no :(
[14:50] <ttx> My patch basically implemented : http://wiki.openstack.org/Packager/Rootwrap
[14:50] <zul> ttx: ok ill push somethin up for review then after i get the ubuntu packaging out the door
[14:50] <ttx> http://paste.ubuntu.com/772265/
[14:51] <ttx> zul: you might need a fuzz a bit to apply it now
[14:51] <zul> ttx: yep yep
[14:51] <ttx> A bit of extra testing couldn't hurt, as well :)
[14:51] <zul> ttx: dont worry ill be testing it :)
[15:31] <SpamapS> oh wow.. starting with MySQL 5.5.18 they're putting out native .debs
[15:33] <zul> oracle?
[15:34] <querier> Hi. where to get a list of packages the ubuntu server edition installs ?
[15:40] <jpds> querier: http://gb.releases.ubuntu.com/10.04/ubuntu-10.04.3-server-amd64.list
[15:42] <querier> jpds: Thank you.
[15:42] <SpamapS> zul: yeah
[15:42] <jpds> querier: But you probably want to look at the preseed/ubuntu-server.seed file.
[15:43] <zul> SpamapS: interesting are they providing the source package as well?
[15:45] <SpamapS> zul: no
[15:45] <SpamapS> :-P
[15:45] <SpamapS> fail
[15:46] <zul> yep
[15:46] <zul> that would have made things alot easier
[16:29] <pythonirc1011> does anyone run ubuntu in production environment here? especially web applications/servers.  I need to run a webserver on a cloud-- very small slice of a xeon box (512Mb RAM / 600 Mhz slice). Any suggestions for configuration?
[16:57] <pmatulis> pythonirc1011: yes, ubuntu is used in production all over the world
[17:05] <powergod> QZ87MYЩJӨP4
[17:05]  * powergod |;
[17:06]  * powergod MΛƧƬΣЯ ΣᄃӨПӨMIΣ ΣƬ ƧƬΛƬΣGIΣ D ΣПƬЯΣPЯIƧΣ
[17:07]  * powergod ๓гร Ŧคгђคภ ๏๒รเєђ ๒๏ยђ 		35 54 83/82 32 85 	๔เקlô๓є ยภเשєгรเtคเгє ๔є tєςђภ๏l๏ﻮเє 	Ŧгคภςคเร,คภﻮlคเร,คгค๒є
[17:07]  * powergod ,|;?
[17:07] <powergod> uke47*ikpex;
[17:07]  * powergod lo9 cxz ;
[17:07]  * powergod ADM!N
[17:07]  * powergod uk
[17:08] <powergod> uke47*ikp!qz
[17:13] <pmatulis> i think we were just visited by an unidentifed flying object
[17:14] <pmatulis> *unidentified
[18:13] <zul> Daviey/smoser: Can you have a quick look: http://paste.ubuntu.com/772493/
[18:14] <Daviey> zul: why does it include prior versions?
[18:15] <zul> Daviey: because i did bzr diff -r251..245
[18:15] <Daviey> zul: why? :)
[18:15] <Daviey> bzr diff -c-1, would be more useful :)
[18:16] <zul> because i suck
[18:20] <zul> Daviey: still looking?
[18:21] <Daviey> zul: can i have an updated diff?
[18:21] <adam_g> zul: typo in changelog (nova-console.intall)
[18:22] <zul> adam_g: thanks
[18:24] <Daviey> zul: wait, the diff includes debian/changelog old references but not the changes it references?
[18:24] <zul> Daviey: right because the ubuntu/nova and ubuntu-server-dev is out of sync
[18:25] <zul> so i just pulled in the debian/changelog from ubuntu/nova
[18:25] <Daviey> ahh
[18:26] <zul> minus the tyop that adam saw thats what is going to be uploaded
[18:27] <Daviey> zul: should fix-traversal-via-image-register.patch be refered to being dropped, fixed upstream?
[18:27] <zul> Daviey: yes
[18:28] <Daviey> i'm not convinced we are not abusing pydist-overides TBH, but if it works.. :)
[18:28] <zul> fixed
[18:28] <zul> Daviey: once we get nosexcover and coverage into main it will be dropped
[18:31] <zul> Daviey: anything else?
[18:34] <zul> adam_g: did i miss anything else?
[18:35] <gary_poster> hi hallyn and SpamapS.  I have another lxc problem I'd like some help with.  The instance I created the yesterday is now hanging when I try to start it.  This is the output.  http://pastebin.ubuntu.com/772509/  What can I do to fix or even diagnose?  This is after an apt update that was partially broken (me trying to see if bug #902190 were fixed) so this may simply be a hosed image, but this didn't happen before when I had the same problem.
[18:35] <Daviey> zul: nah, looks good i thik :)
[18:35] <zul> alright uploading
[18:37] <zul> alrighty glance is next
[18:37] <adam_g> zul: thats the only thing that popped out at me. is there a branch around? ill build and test if you haven't
[18:38] <zul> adam_g: already have....tested it yesterday
[18:38] <zul> its in the ubuntu-server-dev branch now
[18:38] <adam_g> zul: cool
[18:40] <adam_g> zul: ive got a script ive been working on that lets us define openstack juju deploymenst like this: http://paste.ubuntu.com/772517/   maybe next upload, we can prep everything in a testing PPA, point each service to it and test everything at once before uploading
[18:41] <zul> adam_g: is that the jenkins ci stuff jamespage is doing in the testlab?
[18:42] <adam_g> zul: well, my goal is to use it there as well, but right now im firing everything up in ec2. been using it to get everything up in one command so i can start working on keystone + horizon quickly
[18:42] <edgy> hallyn: ping
[18:42] <zul> adam_g: ah yeah that wouldnt be a problem
[18:43] <zul> adam_g: besides there probably isnt going to be an upload until the new year since im goin to be on holiday starting next wed ;)
[18:43] <edgy> with kvm or kvm-spice, shouldn't -enable-kvm be default?
[18:44] <zul> Daviey: you wanted me to use invoke-rc.d in glance right?
[18:49] <adam_g> where do bug reports for ubuntu cloud images go?
[18:49] <adam_g> smoser: ^
[18:50] <utlemming> ubuntu
[18:50] <utlemming> agam_g: and then assign it to me
[18:50] <utlemming> adam_g: what are you seeing?
[18:50] <zul> adam_g: you know the db_sync stuff for glance-api/glance-registry we were talking about last night i think it should go into the glance-common.postinst since its going to be installed anyways if you are going to install either api or registry
[18:51] <adam_g> utlemming: a friend is trying uec images + nova for the first time and his guests can't boot, no /dev/vda
[18:51] <adam_g> zul: the api server doesn't touch the database
[18:51] <utlemming> adam_g: humm...that doesn't sound right
[18:52] <utlemming> adam_g: oneiric or precise?
[18:52] <zul> adam_g: are you sure? so just glance-registry?
[18:52] <adam_g> utlemming: oneiric
[18:52] <adam_g> zul: almost certain, let me check
[18:53] <zul> adam_g: yeah you are right, so ill put it in glance-registry.postin
[18:53] <adam_g> zul: werd.
[18:54] <utlemming> adam_g: probably a configuration issue. Is he using "root=label=cloudimg-rootfs" as the commandline argument? Also, the cloud-images for KVM don't need a kernel, they should just boot.
[18:54] <adam_g> zul: db_sync is still weird issue with bug 883988 still around
[18:55] <adam_g> utlemming: *shrug* he used cloud-publish-tarball pointed at nova. ttylinux booted fine
[18:57] <zul> Daviey adam_g: glance http://paste.ubuntu.com/772531/
[18:57] <adam_g> zul: its not specific to glance packaging, but most seem to only call 'db_sync' if an sqlite database is in use. but most of the components automatically migrate the database if necessary when the service starts (not true of nova, where that postinst logic originated i think)
[18:58] <zul> adam_g: k
[19:00] <Madkiss> hi there
[19:00] <pmatulis> Madkiss: hi
[19:00] <adam_g> zul: dont see the glance-registry.postinst
[19:00] <adam_g> utlemming: meet madkiss. :)
[19:00] <Madkiss> I was told I could fine Ben Howard here? I'm having a problem with his ubuntu-uec-images ;-)
[19:00] <Madkiss> utlemming: hi! please to meet you :)
[19:01] <utlemming> hello Madkiss -- I take it your using KVM?
[19:02] <Madkiss> utlemming: that's correct
[19:03] <utlemming> Madkiss: can you file a bug report against ubuntu and assign it to me?
[19:03] <utlemming> I would like to see the details of the problem
[19:04] <Madkiss> bugs.launchpad.net?
[19:04] <utlemming> yes
[19:04] <pmatulis> against what package?
[19:04] <zul> adam_g: http://paste.ubuntu.com/772542/
[19:04] <utlemming> just against ubuntu
[19:05] <pmatulis> Madkiss: https://bugs.launchpad.net/ubuntu/+filebug
[19:06] <pmatulis> maybe https://bugs.launchpad.net/ubuntu/+filebug?noredirect
[19:07] <adam_g> zul: /var/log/glance /etc/glance should be 0700 glance:glance
[19:09] <Madkiss> utlemming: just a sec, I am having a terrible suspect
[19:09] <zul> adam_g: good catch
[19:09] <zul> fixed
[19:09] <adam_g> zul: also, not sure if its outside the context of those diffs, but shouldn't those directories be created in glance-common.postinst?
[19:09] <zul> adam_g: they are in glance-common.dirs
[19:11] <adam_g> zul: ah! wanna take this to #Openstack-packaging, ive got some other questions?
[19:11] <zul> adam_g: sure
[19:39] <Madkiss> utlemming: sorry, false alarm. It's just that I am too stupid (/var was full)
[19:39] <utlemming> Madkiss: :) no worries
[19:40] <utlemming> Madkiss: feel free to ping me here any time for image problems
[19:41] <Madkiss> sure, thanks a lot
[19:43] <Madkiss> oh,
[19:44] <Madkiss> utlemming: turns out it wasn't false alarm, i will report a bug. the 10.04 images works nicely interestingly
[19:44] <Madkiss> where's the web frontend for doing it?
[19:51] <Madkiss> utlemming: the thing wouldn't let me chose ubuntu, so I chose "I don't know"
[19:51] <Madkiss> utlemming: here it is: https://bugs.launchpad.net/ubuntu/+bug/905469
[19:54] <utlemming> Madkiss: do you have the entire console output?
[19:55] <Madkiss> sure
[19:55] <Madkiss> shall I paste it into the bug?
[19:56] <Madkiss> done
[19:59] <utlemming> Madkiss: have you tried the daily? I'm curious if that will work for you. http://cloud-images.ubuntu.com/oneiric/20111214/
[20:00] <utlemming> I think I know the problem and it is fixed in the daily, but let me confirm it
[20:00] <Madkiss> just a sec
[20:00] <Madkiss> will do so straight away
[20:02] <Madkiss> intresting
[20:02] <Madkiss> it started downloading the image with 500k, and now it's up to 8.5mb/s
[20:03] <Madkiss> utlemming: do you know offhand how I can remove the "old" image from the image store?
[20:03] <utlemming> euca-deregister should do the job, if I recall. But then again, I'm not as farmiliar with NOVA as I'd like to be.
[20:05] <Madkiss> thanks a lot. I'm uploading the new image-
[20:12] <Madkiss> utlemming: yup, that's way better
[20:13] <utlemming> Madkiss: thanks for confirm that for me
[20:13] <utlemming> I'll be spitting out a new image on Monday or Tuesday then.
[20:13] <utlemming> what happened was there was a bug in the build process that put the wrong kernel in the .tar ball and it looks like you hit that
[20:14] <Madkiss> utlemming: could you make a short statement to the bug? i think people are going nuts already ,)
[20:14] <utlemming> yup, its a dup of Bug 901826
[20:14] <hallyn> edgy: unfortunately enable-kvm is not the default on qemu.git.  Apart from patching (which we could do) not much to be done, but libvirt should do it for you, and you can alias it...
[20:15] <Madkiss> utlemming: thanks for helping me out :)
[20:15] <hallyn> btw, i'm on holiday today.  so slow and few responses
[20:16] <hallyn> gary_poster: try 'lxc-start -n name -l debug -o outout' and look at outout
[20:16] <utlemming> Madkiss: no, thank you for taking the time to point out the problem; much appreciated
[20:16]  * hallyn slips quietly back away to holiday
[20:17] <gary_poster> hallyn, np and thank you.  I made a new container.  I'll record that debug spelling for the future though.  Have a great holiday.
[20:23] <Madkiss> utlemming: well, i'm in a rush of getting openstack knowledge, and interestingly enough, ubuntu provides the only images available not exclusively via AWS ... thanks for the work :)
[20:24] <jca1981> why does halt not power my box off but shutdown now -h does?
[20:24] <Madkiss> try halt -p?
[20:26] <jca1981> halt is easyer to type :D
[20:26] <jca1981> sorry thourght you typed why halt :P. Thanks
[20:28] <Madkiss> jca1981: ;)
[20:29] <jca1981> is the halt a wrong way of shutting down a server?
[20:30] <pythonirc1011> is there a way to benchmark a ubuntu server i'm root on? what tools could i use? i'm mostly looking for io/cpu benchmarks
[20:30] <Madkiss> jca1981: shutdown -h now doesn't do anything other than calling halt -p at the end, it just does numerous foo and magic around it.
[20:31] <jca1981> do i need the Magic?
[20:31] <smoser> Madkiss, 2 things regarding your 11.10 image
[20:31] <jca1981> does halt shutdown my services corectly?
[20:32] <smoser> a.) you should use the .img file rather than the .tar.gz
[20:32] <smoser> b.) if for some reason you want the .tar.gz file, you can use a "loader kernel" which would get you booting with an initramfs.
[20:33] <smoser>  "loader kernel" == http://people.canonical.com/~smoser/lucid-loaders/ which basically works as a bootloader to find kernel/ramdisk inside image.
[20:33] <Madkiss> smoser: AIUI, these .tar.gz files provided by utlemming are supposed to be installable with cloud-publish-tarball, aren't they?
[20:35] <smoser> Madkiss, yes, they are. its a bug, no doubt. but for 11.10 and 11.04, the .img file is superior.
[20:35] <smoser> and you can publish that with "cloud-publish-image" which has as-simple an interface.
[20:36] <Madkiss> great, thanks for the hint!
[20:43] <mattrobenolt> Hey, any advice for adjusting the boot message after all of the startup scripts have been run? I don't even really know what it's called, but I basically want what you see when you login with the /etc/motd to be displayed on boot without logging in first.
[20:43] <mattrobenolt> I'm bundling up a OVF and want some information displayed on boot, like the eth0 address and other misc things.
[20:44] <koolhead17> smoser: hello there
[20:44] <pythonirc1011> root@Mobile:~# sudo tasksel install lamp-server  ------> tasksel: aptitude failed (100)
[20:44] <pythonirc1011> what am i doing wrong?
[20:45] <smoser> koolhead17, hey.
[20:46] <pythonirc1011> I think it might have been apt-get update...checking again...
[20:46] <smoser> mattrobenolt, i think /etc/issue
[20:47] <mattrobenolt> smoser: Can I populate that with dynamic data on boot? Or I guess I could hook in a little startup script to write the file on boot, correct?
[20:47] <pythonirc1011> yup that was it
[20:47] <mattrobenolt> How exactly is the /etc/motd generated on login? This is new to me, coming from Debian land.
[20:48] <smoser> mattrobenolt, that is writtten every time getty runs (i think)
[20:48] <smoser> s/written/written to the getty/
[20:49] <smoser> i think you could update it on boot, yes.
[20:49] <smoser> /etc/motd is dynamic, see man 'update-motd'
[20:49] <mattrobenolt> Ok, so I could essentially write out to /etc/issue with a startup script, and it should be solid.
[20:49] <smoser> but it is only written on login
[20:50] <smoser> mattrobenolt, you might be able to tie into the starting of a getting via upstart (start on starting tty1) or something.
[20:50] <mattrobenolt> Oh my, there's a whole update-motd.d scripts folder, haha
[20:51] <mattrobenolt> smoser: That's pretty nice. Anything similar for /etc/issue ? Like an update-issue -esque script?
[20:58] <smoser> mattrobenolt, no.
[20:59] <mattrobenolt> smoser: I can work from this: http://humanreadable.nfshost.com/files/update-issue
[20:59] <mattrobenolt> All good, you pointed me at least in the right direction. Thanks smoser. :)
[21:02] <smoser> SpamapS, can you do 'start on starting tty*' ?
[21:03] <smoser> err... 'start on starting JOB=tty*' i guess.
[21:03] <SpamapS> smoser: I've never tried that, but it should work in theory
[21:03] <SpamapS> smoser: you'd need an instance if you want it to run for *each* and not just for the *first*
[21:07] <smoser> instance ?
[21:07] <smoser> SpamapS,
[21:07] <SpamapS> smoser: as in, instance $JOB
[21:07] <SpamapS> I think its UPSTART_JOB btw
[21:07] <SpamapS> I can't remember
[21:08] <smoser> man 5 init says JOB
[21:08] <SpamapS> yeah ok man 5 init is right
[22:07] <jcastro> zul: do you have a defaco overall blueprint that covers openstack for 12.04?
[22:07] <jcastro> defacto even
[22:07] <zul> jcastro: yep https://blueprints.launchpad.net/ubuntu/+spec/servercloud-p-openstack
[22:09] <jcastro> zul: I'm writing up your tweet into a post.
[22:09] <jcastro> zul: assuming from this blueprint, you're doing weekly snapshots?
[22:09] <zul> jcastro: cool i was going to blog about it this weekend
[22:09] <jcastro> blog it anyway
[22:09] <zul> yeah weekly snapshots
[22:09] <jcastro> in 12.04 or seperate PPA?
[22:13] <jcastro> zul: ^
[22:14] <zul> 12.04
[22:58] <pythonirc1012> can anyone tell me what is this "none" in df-- http://paste.pocoo.org/show/521914/ ?
[22:59] <pythonirc1012> also, if i want to build a binary copy of the machine's hard drive to boot it remotely, is this possible to do? if so how?
[23:10] <koolhead17> Daviey: around :P
[23:16] <SpamapS> hmmm
[23:16] <SpamapS> 27713 root      20   0  135m 5732 1320 S   91  0.1 394:29.88 powernapd
[23:27] <Daviey> koolhead17: no.
[23:29] <koolhead17> Daviey: ok. :D