[01:28] <twb> soren: hey, you're one of the upstart weenies, right?
[01:29] <twb> As at lucid, what's the right way to start a job when a specific USB device is inserted?
[01:30] <lifeless> udev
[01:32] <twb> And just have it call "start", or emit an event?
[01:34] <lifeless> thats my understanding
[01:34] <lifeless> IMBVW
[01:35] <twb> Which one
[01:36] <twb> Argh
[01:36] <twb> /lib/udev/usb_id /dev/bus/usb/x/y does nothing
[01:37] <twb> WTF, http://paste.debian.net/138403/
[01:38] <twb> Hm, 2011-10-21T12:38:17.139079+11:00 trimserver-lucid usb_id[31119]: unable to access '/dev/bus/usb/001/011'
[01:38] <twb> But I'm root
[01:50] <Kutakizukari> I installed ModSecurity via "apt-get install libapache-mod-security" on Ubuntu 10.04. How do I tell which version of ModSecurity it installed?
[01:51] <twb> dpkg -l libapache-mod-security
[01:53] <Kutakizukari> twb: Thanks
[02:41] <wdaniels> Hi, is this the best place to ask about problems with orchecstra/rsyslog?
[02:41] <wdaniels> Seems that Orchestra configured rsyslog to use SSL certs under /var/lib/orchestra/.ssl but syslog user cannot read there :S
[02:47] <axisys> do I need acpid for a server? I think not but need to verify
[02:48] <qman__> only you can answer that question
[02:48] <qman__> do you require the function it provides?
[04:01] <aviandroid> can anyone recommend a vnc server
[04:01] <TimR_> hey guys how do i check to see if tls is supported by ubuntu-server 10.04.3 lts
[04:01] <aviandroid> the two most common ones that i have seen are tightvnc and vnc4server
[04:19] <lunaphyte_> hi.  i have a 10.10 server i'd like to upgrade to 11.10.  it seems to be aware that natty is available, but do-release-upgrade doesn't work, and complains "warning:root:file 'natty.tar.gz.gpg' missing" - failed to fetch".  i don't have any networking problems - how can i troubleshoot this?
[04:27] <TheEvilPhoenix> lunaphyte_:  are you on a VPS?
[04:27] <TheEvilPhoenix> out of curiosity :P
[04:27] <lunaphyte_> no.  it's a vmware guest, on a host managed by me.
[04:27] <TheEvilPhoenix> that's a VPS then
[04:27] <TheEvilPhoenix> if its a server :P
[04:27] <lunaphyte_> ok
[04:28] <TheEvilPhoenix> i've seen mixed results upgrading VPSes
[04:28] <TheEvilPhoenix> but as to how to diagnose that issue, i'm unsure.
[04:29] <TheEvilPhoenix> but i did want to state that you might run into some weird issues upgrading in place 10.10 -> 11.04 -> 11.10
[04:29] <lunaphyte_> it would seem to me that the host it's trying to retrieve the file from doesn't have it.  i wondered if i might somehow determine another ubuntu host which did have it.
[04:37] <hallyn> zul, the lxc tty fix (lxc-fix-grantpt.patch) is not actually in the new 0.9.6 merge.  (no big deal, and now we can probably just take the fix from upstream since it finally landed there in a slightly different form)
[04:37]  * hallyn out
[05:12] <twb> Bleh.  Do nut's "driver" binaries really need to be in /lib (not /usr/lib) ?
[05:12] <twb> It's not like you need UPS to boot
[06:36] <lunitik> twb: /usr is for application stuff... drivers are not application stuff, so yes
[06:38] <lunitik> twb: drivers are referred to as "modules"... you will note /lib/modules/$(uname -r)/ are where your current drivers are
[06:40] <soren> twb: upstart on its own won't help you with that.
[06:41] <soren> twb: I remember using a package that could run a script when a specific USB device was plugged in, but I'm having trouble finding it now.
[06:42] <lunitik> soren: umount or previously pmount probably
[06:42] <soren> lunitik: No, this was more generic than mounting stuff.
[06:42] <lunitik> soren: how does it get more generic than mounting?
[06:43] <soren> There are other things you can plug into a USB port than disks and memory sticks.
[06:43] <soren> Webcams, serial-USB dongles, modems, etc. Lots of things.
[06:44] <lunitik> soren: usbmount says "automatically mount and unmount USB mass storage"
[06:44] <greppy> hotplug can do interesting things for usb device insertion
[06:44] <soren> lunitik: Yes?
[06:44] <lunitik> soren: depends pmount
[06:45] <soren> greppy: "hotplug"?
[06:45] <lunitik> greppy: pmount uses udev which replaced hotplug like 10 years ago
[06:46] <greppy> lunitik: since I remember when it was new, forgive me :)
[06:46] <soren> lunitik: This is not about mounting things.
[06:47] <lunitik> soren: pmount runs a script... edit the script... dpkg -L usbmount and see if it made a cron job or something
[06:48] <lunitik> soren: I honestly don't recall any others
[06:48] <soren> lunitik: We don't use pmount anymore either, just FYI.
[06:49] <lunitik> soren: udisks seems to do the same based on udisks-glue
[06:50] <lunitik> "simple automount daemon with support for user-defined actions"
[06:50] <lunitik> user-defined actions = scripted stuff
[06:51] <soren> lunitik: So you're proposing to trick udisks into handling devices that aren't disks, but also trick it into not mounting it?
[06:51] <lunitik> soren: udisks does storage in general
[06:51] <soren> Gah.
[06:52] <soren> This. is. not. about. storage!
[06:52] <lunitik> USB devices are storage devices
[06:52] <soren> THEY'RE NOT!
[06:52] <soren> 06:43 < soren> There are other things you can plug into a USB port than disks and memory sticks.
[06:52] <lunitik> Then what are they?
[06:52] <soren> 06:43 < soren> Webcams, serial-USB dongles, modems, etc. Lots of things.
[06:52] <soren> They're devices!
[06:52] <soren> They could be anything.
[06:52] <lunitik> Their function is programmed... program needs storage
[06:53] <soren> http://lmgtfy.com/?q=usb+missile+launcher
[06:53] <soren> Tell me that is a storage device while keeping a straight face.
[06:53] <soren> http://www.vat19.com/dvds/usb-warmer-cooler.cfm  or that.
[06:54] <lunitik> soren: It stores the programming to do whatever it is doing... it also stores files which you can manipulate via devices... including viewing
[06:54] <lunitik> uhh... via modules
[06:54] <soren> The usb missile launcher may have storage for some primitive firmware, but it's not exposed to you. The cup warmer sure as heck don't have firmware. It just draws power from the USB port.
[06:54] <_ruben> *facepalm*
[06:54] <soren> Are you saying your monitor is also a storage device?
[06:54] <soren> And your keyboard?
[06:55] <_ruben> it stores your keys!!
[06:55] <soren> And crumbs and coffee!
[06:55] <_ruben> gotta keep that mind tho, whenever i lost my keys again
[06:55] <soren> On that note..
[06:55] <_ruben> ;)
[06:56] <lunitik> soren: what do you think is transfered over the cable if not data? where is the data coming from to be transmitted?
[06:56] <soren> 06:54 < soren> Are you saying your monitor is also a storage device?
[06:56] <lunitik> so, yes
[06:56] <_ruben> wtf
[06:56] <soren> Ok. Then this conversation is over.
[06:57] <lunitik> Everything is a file... if you don't believe me, explain /dev/usb/
[06:57] <soren> Ok, so where is your monitor mounted?
[06:57] <_ruben> on my desk!
[06:57] <lunitik> soren: /dev/ttyS0 unless they moved it
[06:57] <_ruben> lol
[06:57] <soren> Your monitor is conected over a serial port?
[06:58] <soren> For rizzle?
[06:58] <_ruben> ok .. if this isn't a troll, then i'm santa
[06:58] <lunitik> soren: monitor cables are serial cables
[06:58] <soren> HAHAHAHAHAHAHAHAHAHAAHHA
[06:58] <onre> erm
[06:58] <onre> they are not
[06:58] <soren> Ok, this conversation is really over now.
[06:59] <_ruben> i gotta hand it to ya, lunitik .. you made me smile on this "early" morning :)
[06:59] <onre> but, i actually have a question, too :p  i'm running 10.04 LTS from alestic image on EC2. is there some easy way of getting the instance security group name during bootup? i'm looking to do some customization on the system based on which group it belongs to
[06:59] <onre> or, to be exact, my own image based on alestic image
[06:59] <soren> _ruben: Alright, Santa. For christmas I want my faith in mankind back.
[07:00] <_ruben> soren: be reasonable now ;)
[07:00] <soren> onre: I imagine it's to be found in the meta-data service.
[07:01] <soren> onre: It is.
[07:02] <lunitik> I am dumb, sorry... it is /dev/video0 ... cuz the monitor is plugged into the graphics card  :/
[07:02] <soren> onre: So, query http:://169.254.169.254/latest/meta-data/security-groups
[07:02] <onre> soren, thank you!
[07:03] <soren> lunitik: That's the device node. Where is it *mounted*?
[07:05] <lunitik> soren: actually, it is /dev/agpgart ... I really should stop... technically it is mounted in /sys/modules
[07:06] <lunitik> soren: for me, it is /sys/modules/nvidia
[07:06] <lunitik> soren: there is also various data in /proc
[07:06] <lunitik> soren: it has to be mounted for linux to listen to it
[07:07] <lunitik> soren: these are userspace visibility of kernel space mounts
[07:07] <soren> lunitik: You're funny.
[07:07] <soren> But, alas, I'm busy.
[07:08] <twb> 17:38 <lunitik> twb: drivers are referred to as "modules"... you will note /lib/modules/$(uname -r)/ are where your current drivers are
[07:08] <twb> lunitik: not kernel modules, nut "drivers" are just binaries.
[07:08] <lunitik> twb: those are the definitions for how to mount
[07:08] <twb> soren: re upstart/usb -- never mind, I just used the crappy stock nut sysvinit start scripts
[07:09] <lunitik> twb: the kernel is binary too when you run it
[07:09] <twb> lunitik: plonk.
[07:09] <soren> lunitik: Tell me... What do you do for a living?
[07:10] <lunitik> soren: contract network installations mostly
[07:10] <lunitik> (so, I wire LAN's and WAN's for companies, and they pay me)
[07:10] <soren> Ok.
[07:11] <twb> cable monkey
[07:11] <soren> How do you wire WAN's?
[07:11] <lunitik> twb: *cue Idiocracy* I like money
[07:11] <twb> soren: really long cables :P
[07:11] <soren> That must take a lot of cable.
[07:11] <lunitik> soren: how do you get cable internet?
[07:12] <lunitik> soren: you connect their LAN's to that
[07:12] <twb> soren: I think everything about layer1 is a bit of a muddle to him
[07:12] <twb> Er, s/about/above/
[07:12]  * twb puts down the cider jug
[07:13] <lunitik> twb: nah, other 3 just don't pay bills atm
[07:13] <soren> This just gets better and better.
[07:13] <lunitik> twb: although you have to configure the WAN's etc... so it goes all the way up the stack really
[07:14] <lunitik> soren: it is only better if you are using the OSI modal instead of TCP
[07:15] <soren> Do you have a newsletter I can subscribe to?
[07:15] <twb> lunitik: so you roll out X.25, or what?
[07:16] <soren> twb: Be nice. I'm sure he's moved on to X.26 by now.
[07:16] <twb> Bloody MPLS
[07:17] <lunitik> twb: it isn't a frequent request, usually such companies have people staffed, but sometimes
[07:18] <twb> "As recently as March 2006, the National Airspace Data Interchange Network has used X.25 to interconnect remote airfields with Air Route Traffic Control Centers."
[07:18] <lunitik> twb: usually umm... IP tho
[07:19] <lunitik> ISDN networks use them too
[07:20] <lunitik> Or rather, that probably is an ISDN network...
[07:21] <lunitik> Still no closer to soren's answer though... was trying to kill time for other input
[07:22] <lunitik> soren: sorry, maybe check during the day time in America - it is when most seem active
[07:22] <soren> I'll be fine :)
[08:02] <jamespage> Daviey: are you able to mark further duplicates of bug 862129?  I keep getting a timeout
[08:21] <Daviey> jamespage: same issue here, i mentioned it in #launchpad yesterday
[08:21] <jamespage> Daviey: ah - I just did as well
[08:36] <stanman246> err.... we're currenlty using a sbs2003 server and i want to throw the thing away and go for an ubuntu server.
[08:36] <stanman246> but
[08:38] <stanman246> i also want the clients to run ubuntu desktop, to start off i want the users to run ubu desktop first and connect them to the sbs2003 server. I'm using likewise-open, but then the users need to put the domain name before their username. Is there a way to join the ubu desktop to the sbs2003 domain and let users logon with only their username and password?
[08:53] <wfu4422> i need help with MGA drivers and a matrox g450 quad card on 11.04
[08:57] <RoyK> wfu4422: hardly a server issue....
[09:00] <wfu4422> matrox g450 is business class card, not a gaming card. i will use it on my server
[09:01] <wfu4422> i would expect someone in here to know what to do over #ubuntu
[09:03]  * Daviey spies a new upstream version of jonas in oneiric-partner queue.
[09:03] <koolhead17> hello all!!
[09:06] <Daviey> hey koolhead17
[09:10] <jamespage> Daviey: w00t!
[09:14] <soren> wfu4422: What do you need a graphics driver for on a server?
[09:14] <maswan> Anyone know if nfs 4.1 support is there in something newer than lucid? the only reference I have is https://answers.launchpad.net/ubuntu/+source/nfs-utils/+question/142624 which is "closed" without an answer
[09:15] <soren> maswan: nfs 4.1 client support, you mean?
[09:15] <maswan> soren: yeah
[09:16] <maswan> I guess I could bootstrap a VM and chekc it out myself, was just wondering if someone knew and could answer before lunch instead of me doing it after lunch. :)
[09:17] <soren> maswan: It doesn't seem like it. Also, the Kconfig question explicitly says:           Unless you're an NFS developer, say N.
[09:18] <soren> maswan: Did Lucid have it?
[09:19] <jamespage> Daviey: hmm - might need to get that rejected - its still using the sun-jdk package which is not in partner
[09:19] <soren> maswan: AFAICT, it was disabled in Lucid as well.
[09:20] <jamespage> it will still work but you will have to install default-jre-xx first
[09:20] <Daviey> jamespage: I thought the chances of sun/orcacle java still being in distro's was low now?
[09:20] <jamespage> Daviey: its gone in oneiric
[09:20] <jamespage> likely to be removed from earlier releases as well
[09:20] <Daviey> jamespage: Yes, chances of it coming back, i mean
[09:21] <Daviey> jamespage: earlier releases are still OK with the old licence, no?
[09:21] <jamespage> Daviey: yes - I understand its low as well
[09:21] <jamespage> Daviey: well the issue is that we can't security patch it as Oracle are not releasing under the distributor license agreement
[09:21] <Daviey> *sigh*
[09:22] <maswan> soren: I'm more interested in if the mount tools etc are there, custom kernel with just some config options changed I can do but finding nfs utils etc and fixing/backporting them is a bit more strange. Maybe I'll stick with rhel 6.2[-derivatives] for now then.
[09:22] <jamespage> so as soon as we hit something nasty.....
[09:22] <jamespage> we can't fix it
[09:23] <jamespage> Daviey: Oracle want to drive more people to using OpenJDK - as this benefits them in the long run
[09:23] <jamespage> sun-jdk-6 is just an overhead for them now
[09:23] <Daviey> I imagine Oracle will always have a non-free (including cost) jre.
[09:24] <soren> maswan: Do you know what special stuff is needed for that? I would have thought it was all kernel stuff.
[09:24] <jamespage> Daviey: they will - but with Java 7 its 99.9% OpenJDK
[09:25] <Daviey> ah!
[09:25] <jamespage> with Java6 its more divergent
[09:25] <jamespage> so more overhead for them
[09:25] <Daviey> i see
[09:25] <maswan> soren: I will, need to run now though.
[09:25] <soren> maswan: np
[09:26] <jamespage> Daviey: stat for your - Natty - java packages using maven - 53
[09:26] <jamespage> check check in precise - 150!
[09:26] <Daviey> jamespage: It's a timebomb until maven needs to be in main?
[09:26] <jamespage> well I did have an action to prepare maven3 for main inclusion this cycle
[09:26] <jamespage> (i.e. precise)
[09:27] <jamespage> I've only had to migrate one package from maven->ant so far this release to avoid that
[09:27] <jamespage> and maven3 packaging is not complete yet
[09:28] <jamespage> I must have touched another 5 or so that are ripe for ant->maven migration
[09:28] <jamespage> (well that's what I would do if I only packaged in Debian anyway)
[09:34] <jamespage> Daviey: ^^
[09:35] <jamespage> I think that is similar to an issue I've discussed here re dovecot/postfix starting in a racey what
[09:35]  * jamespage goes to look
[09:35] <Daviey> jamespage: Hmm, i thought nova was safe for this.. ie, it did retry?
[09:35] <Daviey> ttx: ^^?
[09:42] <jamespage> Daviey: hmm - not the same as the dovecot/postfix issue
[09:42] <jamespage> well similar
[09:42] <jamespage> maybe
[09:46] <soren> Not entirely.
[09:46] <soren> psotfix talks to dovecot over a UNIX socket, IIRC.
[09:46] <soren> So you can expect them to live on the same host, so adjusting start order actually makes sense.
[09:47] <soren> You can't even expect MySQL and Nova to run on the same box, so the order in which they start isn't terribly interesting.
[09:47] <soren> Nova should just handle it properliy if it's not there.
[09:47] <soren> (whatever that means)
[09:48] <jamespage> soren: agreed
[09:48] <Daviey> jamespage: did you raise a bug for the dupe issue?
[09:48] <jamespage> Daviey: no
[09:50] <jamespage> soren: I was more referring to the fact that quite alot of upstart configurations switched to start on runlevel [2345] in oneiric
[09:50] <jamespage> which pushed back the start of some services
[09:50] <jamespage> revealing this type of issue
[09:51] <flatline> I've just started a VM I built with vmbuilder, but I can't connect to it via ssh, the machine is running, responds to ping and when I try to ssh into it I get connection refused. I didn't include ssh server pkg in vmbuilder line, should I?
[09:52] <Daviey> jamespage: can you 'me too',  bug 879325 please?
[09:52] <jamespage> Daviey: done
[09:52] <Daviey> ta
[09:53] <ttx> Daviey: yes, the bug should probably be returned to nova
[09:54] <ttx> Daviey: maybe it's an old version
[09:55] <ttx> Daviey: will fix
[09:58] <rbasak> bug 879325 affects me too - when I was trying to set more duplicates on the samba bug
[10:11] <Daviey> ttx: thanks
[10:12] <Daviey> rbasak: happy days
[10:16] <Daviey> jamespage / rbasak: I'm going to make another bug the main bug for these ones, we can dupe them together later
[10:16] <jamespage> Daviey: makes sense for the time being
[10:20] <Daviey> jamespage / rbasak : bug 877852
[10:21]  * rbasak makes a note
[10:21] <rbasak> Is anything happening with that bug btw?
[10:21] <Daviey> doesn't look like it. :/
[10:24] <Daviey> jamespage: did you see bug #878877 ?
[10:24] <jamespage> Daviey: yes
[10:24] <jamespage> not seen that before - may be localised
[10:24] <jamespage> Just raising a blueprint and I'll take a deeper look
[10:24] <Daviey> cool
[11:08] <koolhead17> hi all
[11:12] <tyska_> hi guys
[11:12] <tyska_> im having problem with fake raid, can anyone help me?
[11:47] <RoyK> http://i.imgur.com/slOIm.jpg
[11:51] <koolhead17> RoyK: heh
[11:59] <soren> flatline: Sure. No ssh server, nothing's going to respond to ssh.
[12:17] <flatline> soren, oh ok, thought ssh is included in default vmbuilder script, works like a charm now
[12:44] <zul> morning
[12:51] <zul> jamespage: ping did you change the dovecot upstart?
[12:51]  * jamespage tries to remember
[12:51] <jamespage> when? this release or last?
[12:53] <jamespage> I know of two issues with it at the moment
[12:54] <airtonix> so, ec2 (how does it work?) i click "connect to this instance", it provides 1,2,3,4 steps... i follow steps, annnnnnd ssh connection fails becuase : too many authentication failures for user ubuntu.... sigh i wish amazon would give me the right information
[12:55] <patdk-wk> that works fine :)
[12:55] <patdk-wk> you do know that there is no password right?
[12:55] <patdk-wk> you have to use the key you made with amazon
[12:55] <airtonix> yes? i am ...
[12:55] <jamespage> zul: why do you ask?
[12:55] <patdk-wk> if you are it would work
[12:55] <lunaphyte_> TheEvilPhoenix: it appears to have just been an ephemeral issue, presumably with some component on ubuntu's end.  it's working just fine now.
[12:56] <airtonix> patdk-wk: obviously it's not...
[12:56] <patdk-wk> airtonix, oviously it's user error
[12:56] <patdk-wk> and as you haven't supplied any details of how your doing it, can't help
[12:56] <airtonix> patdk-wk: it's hardly a user error when i follow the exact instructions provided by the aws console... there isn't much room for error
[12:57] <patdk-wk> ok, whatever
[12:57] <zoopster> airtonix: that tool assumes you are using root - or it used to
[12:57] <zoopster> airtonix: change it to use ubuntu to login and that will do the trick
[12:58] <patdk-wk> zoopster, what part of " too many authentication failures for user ubuntu" == root
[12:58] <airtonix> zoopster: the instructions provide a tailored ssh command that includes the user
[12:58] <soren> airtonix: Do you have several ssh keys in your ssh-agent?
[12:59]  * patdk-wk always uses -i
[12:59] <airtonix> soren: i have many hosts configured in my ~/.ssh/config and i also created a specific one for this host, i also tried without and just used the specifically tailored ssh command provided by the connect tool
[13:00] <soren> airtonix: I understand, but that's not what I'm asking.
[13:00] <soren> airtonix: Can you provide the output of "ssh-add -l"?
[13:02] <airtonix> soren: no because it contains private data
[13:03] <airtonix> patdk-wk: you tell me how i could interept this incorrectly? http://dpaste.com/638637/
[13:04] <soren> airtonix: a) No, it doesn't. It's just the fingerprint. It's not more private than your public keys. b) Then I can't help you.
[13:04] <airtonix> soren: that paste contains the steps i use... except that i put the .pem file in ~/.ssh and chmod it, and i modify the ssh command to point at that file, i also add port 22 tcp to the inbound rules of the associated security group
[13:05] <zoopster> patdk-wk: thanks for the elegant response - i don't use the aws console, but it used to tell you to use root...obviously they changed it...thanks for pointing that out
[13:05] <patdk-wk> zoopster, I hate it when people reply to people, without reading what they say :)
[13:05] <airtonix> soren: i just ran that command and it does reveal data i do not wish to publicly reveal
[13:06] <soren> airtonix: "ssh-add -l" does?
[13:06] <soren> Then it's broken. Or you are.
[13:06] <zul> jamespage: beccause im thinking of changing the openstack jobs to start on staring rc so it starts after mysql
[13:07] <soren> zul: Does Openstack use MySQL by default?
[13:07] <airtonix> soren: yes, the keys contain comments that i do not wish to reveal to strangers
[13:07] <zul> soren:  not in the packaging
[13:07] <soren> zul: Then don't bother.
[13:07] <zul> soren: why
[13:07] <zoopster> patdk-wk: oh I read what was written, but it was written...guess you are better than I
[13:07] <jamespage> zul: on my list - I also need to add a pause to workaround unclean subprocess termination on restarts as well
[13:08] <soren> zul: Why would you? If you have 100 nodes in your cloud, at most 1 of them will be on the same machine as the MySQL server. The rest will still be screwed.
[13:08] <zul> soren: k
[13:08] <soren> zul: The order in which they start shouldn't matter.
[13:08] <Daviey> surely the correct fix is to make sure nova doesn't throw up if it can't contact mysql on startup? :)
[13:08] <soren> zul: If it does, *that's* the problem that needs to be fixed. Not the boot ordering.
[13:08] <soren> Daviey: Exactly.
[13:08] <zul> soren: gotcha
[13:09] <soren> airtonix: I don't quite see how "ssh-add -l" can contain comments. How did you make that happen?
[13:09] <soren> Hm... Maybe it's because I'm using gpg-agent.
[13:09] <soren> As my ssh-agent.
[13:09] <airtonix> soren: ssh rsa keys can have comments on the end
[13:10] <soren> airtonix: How many keys are there? Or is that super secret, too?
[13:10] <airtonix> soren: 6
[13:10] <soren> That's probably your problem, then.
[13:10] <soren> It tries each of them in order.
[13:10] <airtonix> soren: but why would it? i specify the -i and point at the pem key
[13:11] <soren> So it has more than the acceptable... er... how many was it? 3 failures?
[13:11] <Daviey> airtonix: I'm not sure we can help further without, ssh -vvv -i ~/foo/bar.pem
[13:11] <airtonix> but why would it bother trying the keys in my ~/.ssh unless ssh on ubuntu doesn't accept pem keys
[13:12] <soren> airtonix: MAybe it tries those first. *shrug*
[13:12] <soren> Try:
[13:12] <freakynl> Hi, I'm having extremely slow raid-6 performance, it's 8 disk 2tb sata raid-6 (mdadm). It's still syncing (at a whopping 70k, yes k, /s) whilst I'm copying data on it (over iscsi from a windows box, at a whopping ~15MB/s, or well, that's what it says but at 15MB/s it should be done copying the 800GB in after the 16 hours it's been running now)
[13:12] <soren> SSH_AUTH_SOCK= ssh -i whatever.pem ubuntu@whereever
[13:12] <freakynl> any ideas? when it still was on raid-10 it was fine... iostat output is shocking, disks don't write more than 2.5MB/s a piece
[13:13] <airtonix> soren: that worked. thanks
[13:13] <airtonix> soren: seems a bit retarded that if you specify a very specific key to use to authenticate with that it should then assume you didn't actually mean to use a specific key and try all these other keys
[13:14] <airtonix> i mean if i wanted it to try other keys i simply would have left out the -i parameter
[13:14] <patdk-wk> freakynl, what cpu?
[13:15] <freakynl> core i3, but the raid6 process doesn't consume more than 2-5% cpu and the average of the process is even below that
[13:15] <freakynl> i3 530 to be exact
[13:17] <patdk-wk> how are the drives connected?
[13:18] <freakynl> all 8 connected to a LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03)
[13:18] <patdk-wk> so an lsi9201?
[13:19] <patdk-wk> dunno, unless it's just random seeking like nuts
[13:19] <patdk-wk> cause of the raid build + copy your doing
[13:19] <freakynl> that seems right yea
[13:19] <patdk-wk> iscsi forces sync
[13:20] <freakynl> it was *MUCH MUCH* faster on raid-10. Even when it just started up (no iscsi, nothing else using the set) it didn't pass 70MB/s on sync which I find rather low. Thought it was the checksum calculation, but cpu was hardly loaded
[13:20] <patdk-wk> did you adjust the max sync speed?
[13:20] <patdk-wk> I think default is like 200MB/sec, so that might be a max of 100MB/sec on raid10, not sure how it computes
[13:21] <freakynl> max is at the default of 200MB, with 70k/s I have plenty of room :)
[13:21] <patdk-wk> but yes, raid10 is always going screem faster than raid6
[13:21] <patdk-wk> yes, but each sync write from iscsi causes several writes to happen
[13:21] <patdk-wk> that could really be killing your performance
[13:22] <patdk-wk> I'm not sure what block size your iscsi is using
[13:22] <patdk-wk> or what stripe size you used for the raid6
[13:22] <patdk-wk> but that all matters heavily on performance
[13:22] <freakynl> patdk-wk: yea I know, but even then... it should atleast do something along the lines of 100MB/s... not when it's heavily random, but there's just 3 processes pulling on it, 2 moves and the sync itself (although at 70kbyte/s I'd hardly count that)
[13:23] <freakynl> stripe size is 64k, not sure on iscsi blocksize, ntfs blocksize 64k
[13:23] <patdk-wk> hopyfully you aligned the raid6 and ntfs blocks on the same locations
[13:24] <patdk-wk> and I wouldn't call 2 moves plus a sync, sequentual
[13:24] <patdk-wk> at best that is 3 random locations
[13:24] <freakynl> 2008 aligns automatically (usually by using X MiB offsets)
[13:24] <patdk-wk> heh? what does the 2008 lsi card have to do with this?
[13:24] <freakynl> compared to our 40vm's that where on the raid-10 before, that's pretty sequential :)
[13:25] <freakynl> patdk-wk: ? 2008 server (windows, the iscsi initiator)
[13:25] <patdk-wk> if you talk about windows, say windows
[13:25] <freakynl> patdk-wk: i don't like the word :)
[13:25] <freakynl> or well... actually I don't like it's vendor
[13:26] <freakynl> sdh              24.50    90.00  104.00  145.50   514.00   936.00    11.62     7.07   26.97   59.90    3.44   1.30  32.50
[13:26] <freakynl> and average from iostat -x (after 5 outputs) kinda... shitty :)
[13:27] <freakynl> 24.5 rrqm/s, 90 wrqm/s, 514kB/s read, 936kB/s write
[13:28] <_ruben> syncing + copying = killing performance
[13:28] <freakynl> 32.5% util, but never quite understood how to interpret that, kinda hard to measure when a disk is at max capacity and according to the manual this value is based on cpu
[13:28] <patdk-wk> that is how much cpu time is spend waiting on the disk
[13:28] <freakynl> _ruben: sync alone didn't go above 70MB/s
[13:29] <freakynl> :/ will see what it does at completion then... currently just 350k mins away
[13:29] <patdk-wk> do you have bitmap table on or off?
[13:30] <freakynl> I didn't set it, not sure what the default is... mdadm --detail has no mention of bitmap
[13:45] <scalability-junk> mh I have a small problem. I want to resize my home partition, but can't because /etc/lvm/cache/.cache.tmp has not enough space left any work around?
[13:47] <_ruben> so your root partition is completely full?
[13:48] <scalability-junk> yeah sort of mis copied some stuff ;)
[13:48] <scalability-junk> *misscpoied
[13:48] <scalability-junk> *misscopied :D
[13:50] <freakynl> scalability-junk: free some space on /, even if it's just temporary?
[13:50] <patdk-wk> empty /tmp :)
[13:50] <scalability-junk> freakynl: yeah got it
[13:51] <patdk-wk> purge old /var/log/
[13:51] <scalability-junk> already found a file to delete thanks
[13:51] <patdk-wk> :)
[13:52]  * scalability-junk and again copy 200 GB of data ;)
[13:56] <jose__> Hi, Im running 10.04 LT and my computer keeps rebooting once in a while, and I do not know why. Where do I start looking to find out whats going on?
[13:58] <patdk-wk> no idea
[13:58] <patdk-wk> probably log files :)
[13:58] <jose__> patdk-wk, any ideas which ones?
[14:00] <patdk-wk> the ones on the server that is rebooting
[14:02] <Duvrazh> Hi all, seriously screwed up here. 11.10 server, was running a find exec mv on my raid array to get rid of ._* files made by iMacs, now whenever I try to connect through ssh (or before I disconnected to run apt-get or find or anything) I get /bin/bash: No such file or directory.   ------- no backups to restore through webmin, as if they'd work. Suggestions?
[14:04] <Duvrazh> Additional info: I still have a cyberduck sftp session open
[14:04] <pmatulis> cyberduck?
[14:04] <Duvrazh> Yeah, Cyberduck.
[14:04]  * pmatulis shivers
[14:04] <Duvrazh> Only thing that will connect to the server right now.
[14:05] <orudie> when using man, to scroll down the text I press Enter, how about to go up ?
[14:05] <Duvrazh> I'm in the /bin directory and it says all the files types are Kind Unknown
[14:05] <pmatulis> orudie: k
[14:05] <patdk-wk> orudie, I normally use the arrow keys or pageup/down
[14:06] <hallyn> Daviey, on bug 876768, how do you come to the conclusion that it's being undertaken in debian? it'd be nice if it was, but i see no evidence that anyone but me is, nor that the packaging request there is going anywhere
[14:07] <hallyn> maybe what i shoudl ask is, how can we kick that along in debian :)  (being more optimistic)
[14:08] <hallyn> zul, libvirt merge is missing the lxc-fix-grantpt.patch
[14:08] <Duvrazh> are the files in the bin directory supposed to have extensions?
[14:08] <zul> hallyn: yeah i saw last night its on my todo list for today
[14:08] <hallyn> Daviey, all right, i guess i'll start by testing my package under debian and then bug you
[14:08] <hallyn> zul, great, thanks
[14:12] <Duvrazh>  when I login via ssh it says /bin/bash: No such file or directory Connection to duvrazh-ms closed., any ideas on what happened there? before I closed the ssh session (retarded) apt-get and find also had similar output
[14:13] <pmatulis> Duvrazh: sounds like the /bin/bash binary is gone
[14:13] <Duvrazh> I can see the file
[14:14] <Duvrazh> there are no file extensions in /bin though, is that normal?
[14:14] <Duvrazh> Bash Size: 928.6 KB Owner 0 Group 0 Permissions rwxr-xr-x (755) Kind Unknown
[14:15] <patdk-wk> what are file extentions?
[14:15] <patdk-wk> you can probably see it, cause soemthing is using it
[14:16] <patdk-wk> it's been deleted, but not purged yet, cause it's still in use
[14:16] <patdk-wk> once everything stops using it, it will be gone
[14:16] <Duvrazh> well
[14:16] <Duvrazh> I still have 99% of a /bin folder
[14:16] <Psi-Jack> Is it still possible to dit-upgrade from Ubuntu 8.10 somehow to 10.04?
[14:16] <Psi-Jack> dist-upgrade that is.
[14:16] <Duvrazh> any recommendations to undelete the goddamn thing?
[14:16] <patdk-wk> psi-jack, nope
[14:16] <patdk-wk> Duvrazh, nope
[14:16] <Psi-Jack> Not at all, eh?
[14:16] <patdk-wk> ext* doesn't support undelete
[14:17] <Duvrazh> can someone email me a /bin folder then?
[14:17] <patdk-wk> psi-jack, come on, you have to upgrade to 9.04 -> 9.10 -> 10.04
[14:17] <patdk-wk> but those have been gone for awhile
[14:17] <Psi-Jack> Yeah. heh
[14:17] <Psi-Jack> I got an 8.10 server I'm trying to get up with PHP 5.3 at least until we can decommission that server.
[14:17] <patdk-wk> Psi-Jack, they might exist on the archive
[14:18] <Duvrazh> i'll just load it in a vm
[14:18] <Psi-Jack> 8.10 does exist in the archive. old-releases.ubuntu.org
[14:18] <Psi-Jack> err, .com whatever it is. ;)
[14:18] <patdk-wk> well do that :)
[14:18] <patdk-wk> personally, I have found just changing sources.list to 10.04, apt-get update, apt-get dist-upgrade
[14:19] <patdk-wk> seems to work, well enough
[14:19] <Psi-Jack> Hmmm
[14:19] <patdk-wk> defently not supported method
[14:19] <Psi-Jack> I'll try it. I'm doing it on VM's now to test.
[14:24] <Psi-Jack> At least if I confirm it's possible now, I can schedule downtime for it. ;)
[14:26] <Psi-Jack> The upgrade website does say 8.10 to 9.04 is supported, but obviously not.
[14:31] <zul> hallyn: fixed
[14:31] <hallyn> zul, rockin'.
[14:32] <zul> hallyn: we should add lxc to the qrt test for libvirt
[14:32] <hallyn> zul, yeah we should.
[14:33] <hallyn> though that particular bug is weird enough i'm not sure it's worth an autotest
[14:33] <hallyn> (since it's gonna be fixed upstream in next version)
[14:33] <hallyn> (and weird)
[14:33] <hallyn> zul, maybe we should discuss testcases in orlando
[14:33] <zul> hallyn: sounds like a good idea
[14:33] <hallyn> spend an hour getting started
[14:49] <hallyn> zul, oh hey, so libvirt 0.9.6 should support block devices for lxc now?
[14:50] <RoyK> any idea how many files per directory that may be a practical limit with ext4? beore dir_index, anything > 10k or so was hell - now I have 46k files in a single directory (apple time machine store) and I can't notice any problems...
[14:50] <zul> hallyn: should :)
[14:50] <hallyn> nifty
[14:54] <Psi-Jack> Heh sheash.
[14:55] <Psi-Jack> update from jaunty to lucid fails too, because the update-core system tries only to update to lucid, not karmic. Sigh.
[14:57] <RoyK> seems bug 879020 is fixed by upgrading to netatalk 2.2.1, replacing 2.2-beta, which is the one in onieric - any idea what I can do to help get this fixed?
[15:09] <air_> RoyK: yeah, something wrong with the beta.
[15:09] <air_> I don't understand why they allow beta versions to go through to releases?
[15:10] <RoyK> that's _very_ common :)
[15:10] <air_> :)
[16:07] <scalability-junk> what is the best way to backup a virtual machine running on lvm?
[16:09] <RoAkSoAx> zul Daviey should I register a cobbler blueprint?
[16:09] <zul> Daviey: sure
[16:21] <zul> smoser: bug #870405 still relevant?
[16:22] <smoser> zul, yes.
[16:23] <cr3> hi folks, is there a puppet backports ppa for lucid? googling returns stuff from mathiaz back in 2010
[16:25] <RoyK> cr3: there's one in backports, just comment out backports in /etc/apt/sources.list
[16:26] <smoser> zul, commented in bug
[16:26] <cr3> RoyK: that was easy, thanks!
[16:26] <RoyK> :)
[16:27] <koolhead17> nijaba: https://twitter.com/#!/nijaba/status/127419370895978496
[16:30] <scalability-junk> is there any kvm way of backing up whole vms? or is the easiest way to make a lvm snapshot?
[16:30] <nijaba> koolhead17: indeed
[16:30] <koolhead17> nijaba: awesome!! news!! :)
[16:30] <zul> smoser: cool ill add it then
[16:31]  * koolhead17 wishes LTS server to have same luck soon :P
[16:31] <nijaba> koolhead17: server LTS have always benefited of 5y maintenance
[16:31] <RoyK> scalability-junk: I'd rather back them up as if they were physical machines
[16:32] <koolhead17> nijaba: i been told 3 years!! :(
[16:32] <RoyK> makes restores of single files a wee bit easier :P
[16:32] <RoyK> koolhead17: desktop LTS is 3 years, server 5 years
[16:32] <scalability-junk> RoyK: but something like just restore a full operating system in a few minutes would be great too
[16:33] <RoyK> scalability-junk: but then, you'll need to shut down the guest to make a proper backup
[16:33] <patdk-wk> heh? backups?
[16:33] <patdk-wk> what are those?
[16:33] <RoyK> for cowards, I know
[16:33] <koolhead17> RoyK: nijaba seems like i confused it with 3 yrs LTS support to desktops then :D
[16:34] <patdk-wk> just takes so much time, make backup, put backup someplace, test that backup worked, ...
[16:34] <RoyK> patdk-wk: much more fun trying to restore data from a defective disk :þ
[16:34] <scalability-junk> RoyK: so the best way would be using lvm snapshots and then restore them using the ubuntu rescue stuff?
[16:34] <patdk-wk> royk, never had that issue :) I always found it easy, as long is it spins :)
[16:35] <RoyK> scalability-junk: I just backup VMs as if they were machines
[16:35] <patdk-wk> lvm snapshot is not consistant
[16:35] <patdk-wk> unless the stuff on the lvm, is off, or synced at the moment you make the snapshot
[16:35] <RoyK> scalability-junk: and LVM snapshots suck rather hard - they add very much I/O to the running system
[16:35] <patdk-wk> ya, snapshots drop performance from 100% to 20%
[16:36] <RoyK> scalability-junk: also, with ext4 or other filesystems with delayed commit, they may not even be consistent
[16:36] <patdk-wk> lvm snapshots of zfs? :)
[16:36] <scalability-junk> I was more looking for like making snapshot back it up to a second server and wait for the next backup in 24 hours or so
[16:36]  * RoyK slaps patdk-wk 
[16:37] <RoyK> scalability-junk: can't you just install bacula-fd on those VMs and treat them as machines?
[16:37] <koolhead17> RoyK: hahaha. :P
[16:37] <patdk-wk> making a lvm snapshot will work for a backup of the machine, but it WILL have some corruption
[16:37] <patdk-wk> if that bothers you or not, is up to you :)
[16:38] <scalability-junk> I just thought there would be something similar to cloning and backing up the clone or so
[16:39] <RoyK> scalability-junk: then the guest will have to be notified to close all file handles etc, which won't work too well
[16:39] <RoyK> scalability-junk: keep it simple, backup the VM from its OS
[16:39] <patdk-wk> ya, would need something like windows volume shadow copy
[16:40] <RoyK> patdk-wk: Bacula uses that
[16:40] <scalability-junk> ok just thought it would be great to have some restore in a few minutes without installing a whole new system and put the data back in stuff
[16:40] <patdk-wk> royk, I have veeam installed, it uses that to backup windows vm's on esx
[16:40] <patdk-wk> it triggers that, then makes a vmclone->backup
[16:41] <patdk-wk> not sure what method it uses for linux
[16:41]  * RoyK sticks to good old server backups with bacula for linux
[16:41] <RoyK> works well and is very fast
[16:42] <patdk-wk> well, lucky my vm machines never change, EVER, except for updates
[16:42] <patdk-wk> no local i/o ever on them
[16:42] <RoyK> then what do they do?
[16:42] <patdk-wk> webservers mainly
[16:42] <patdk-wk> mailserver do have some local i/o, for the spool
[16:42] <RoyK> ic
[16:43] <patdk-wk> logs are all syslog to a logging machine
[16:43] <patdk-wk> there is just no point to local i/o usage, when it's all nfs mapped
[16:43] <RoyK> apache logs to syslog?
[16:43] <patdk-wk> then it's easy to do a nfs backup
[16:43] <patdk-wk> royk, thought about it, but no
[16:43] <patdk-wk> apache -> perl script -> mysql
[16:43] <RoyK> sounds like a rather bad idea to me....
[16:43] <patdk-wk> my own perl script
[16:43] <RoyK> ok
[16:44] <patdk-wk> perl script buffers and reconnects, and fixs a few apache annoyances :)
[16:44] <RoyK> :)
[16:45] <patdk-wk> so the only real things that need local backups, is the database machines
[16:45] <RoyK> http://i.imgur.com/tZwKg.jpg
[16:46] <patdk-wk> is that liquid or solid ounces to bl?
[16:46] <patdk-wk> lb?
[16:46] <RoyK> nfi
[16:47] <dkn> anyone working with  a usb 3.0 card with good support?
[16:48] <patdk-wk> hmm, 16oz weight (by mass, not liquid) is .99lb
[16:48] <patdk-wk> 16 fl oz could be any amount of lb, depending on fluid
[17:05] <pmatulis> man, you guys really need to move to the metric system...
[17:08] <SM0K3SCR33N> Lovin' it
[17:13] <pmatulis> SM0K3SCR33N: what?
[17:15] <dkn> the metric system?
[17:15] <dkn> 1000 mm in a meter,
[17:15] <dkn> how many inches in a yard?
[17:17] <Ursinha>  /11
[17:17] <Ursinha> argh
[17:31] <Daviey> hallyn: bug 876768, the activity on the debian report suggested that between you and the ther people on the report - it was making movement.
[17:31] <Daviey> Did i miss grok that?
[17:32] <Daviey> RoAkSoAx: Please do, i was planning to - but more than happy for you to.
[17:32] <Daviey> :)
[17:32] <Daviey> RoAkSoAx: it would be useful f it was somewhat drafted before the session IMO.
[17:33] <hallyn> Daviey, the upstream netcf developers helped out.  but no debian folks involved.  that's where i read it differently from you.
[17:33] <hallyn> but i'm building a debian schroot right now to try and get it to build for debian (it doesn't out of the box, for some reason)
[17:33] <hallyn> then i'll see if i can make any progress in a week or 3.
[17:33] <hallyn> if not, we can hopefully package for ubuntu anyway
[17:34] <Daviey> hallyn: Yeah, it would be better if it was in debian first.. but yes, no blocker.
[17:34] <hallyn> Daviey, i just fjeer getting into the same situation with vde at end of this cycle as we did with spice at end of o
[17:35] <Daviey> hallyn: Based on your judgement of the others involved on the debian report, would that leave you maintaining it?
[17:35] <hallyn> i'd like our ducks in a row :)
[17:35] <hallyn> perhaps
[17:35] <hallyn> i'm fine with that
[17:35] <Daviey> hallyn: good point.
[17:35] <Daviey> lets get /something/ in Ubuntu, then push for Debian inclusion. Sound good?
[17:36] <hallyn> Daviey, that sounds good
[17:37] <hallyn> for now i'll just keep tryin gto figure out why my pkg doesn't build in sid with pbuilder, but does on oneiric :)
[17:37] <hallyn> i'm sure i did something stupid
[17:37] <RoAkSoAx> Daviey: yeah i'll do that next week
[17:37] <RoAkSoAx> Daviey: just wanted to get the BP's filed first
[17:39] <hallyn> zul, Daviey, smoser - are there any other virt-related blueprints you think we need?  (I have general libvirt and kvm and lxc ones (servercloud-p-%s if you want to look), and two that robbiew carried over from o
[17:40]  * smoser has to do blueprints.
[17:40] <smoser> i will take a look at those, hallyn
[17:40] <zul> hallyn: im good with them
[17:40] <hallyn> smoser, thanks
[17:40] <zul> i added xen though
[17:40] <hallyn> zul, cool
[17:40] <hallyn> zul, uh, cool to you're good, not to xen :)
[17:40] <hallyn> j/k
[17:40] <zul> hallyn: xen jokes just bounce off me
[17:41] <hallyn> and stick to me.  ouch
[17:41] <Daviey> hallyn: last time i used pbuilder with sid, it had an issue with the archive key..
[17:41] <Daviey> (known issue)
[17:41] <Daviey> I had to hack around it.
[17:41] <hallyn> i don't think that was my problem
[17:42] <hallyn> i need to specify --with-driver=debian to configure there, not in oneiric.  actually that sounds like configure is not detecting /etc/*release correctly i guess
[17:42]  * Daviey plans to go through the blueprints next week
[17:42] <hallyn> but i'm sure there's more
[17:42] <hallyn> that i did wrong, i mean
[17:44] <ersi> So! I figure this is a freakin' good time to contribute. Anyone need a hand? Or have a hand to point towards something? Or to hold? :)
[17:45]  * ersi glares in Daviey's general direction
[17:48] <Daviey> ersi: Hey!
[17:48] <Daviey> ersi: What sort of stuff do you want to touch?
[17:49] <ersi> I'm not sure actually. But I was interested in fiddling a little with packaging/FTB's possibly - but right in the end of the last cycle seemed like a bad time :)
[17:49] <Daviey> ersi: actually, that was a GOOD time :)
[17:50] <Daviey> (to fix ftbfs)
[17:50] <ersi> Yeah.. but there was a lot of ARM stuff..
[17:54] <hallyn> speaking of ftb, kvm-pxe (etherboot) fails
[17:54] <hallyn> at least under sbuild
[17:54] <hallyn> i wonder if this cycle we should make kvm-pxe based on ipxe
[18:09] <adam_g> Daviey: cobbler-enlist... this gonna fall under a cobbler next steps spec or somewhere else?
[18:11] <Daviey> adam_g: I think it needs it's own BP tbh.
[18:12] <Daviey> I planned to raise one today regarding that, but if you want to drive it - feel free.
[18:12] <zul> adam_g: yeah Daviey wants a nice relaxing uds with a pina colada by his side at the pool
[18:13] <adam_g> Daviey: sure
[18:13] <koolhead17> lynxman: hellos
[18:13] <Daviey> zul: How did you guess?
[18:14] <adam_g> zul: servercloud-p-waterpark
[18:14] <Daviey> BTW, i trust everyone attending UDS server sessions will be in the early morning swims?
[18:14] <RoAkSoAx> lol
[18:14] <RoAkSoAx> Daviey: how so?
[18:14] <lynxman> koolhead17: hey there
[18:14] <Daviey> RoAkSoAx: getting out of bed, and going for a swim before the day stars? :)
[18:14] <Daviey> starts*
[18:14] <RoAkSoAx> Daviey: lol
[18:15] <zul> adam_g: nah servercloud-p-holylandadventure
[18:15] <koolhead17> hello RoAkSoAx sir :)
[18:15] <RoAkSoAx> koolhead17: howdy
[18:15] <koolhead17> lynxman: how are you?
[18:15] <RoAkSoAx> Daviey: we'll have to do some kind of swimming competition
[18:15] <koolhead17> RoAkSoAx: am good. thanks
[18:15] <RoAkSoAx> Daviey: hope it's not cold though
[18:15] <koolhead17> *swimming
[18:15] <zul> RoAkSoAx: bastard.. (not cold)
[18:15] <koolhead17> hey Daviey zul
[18:15] <koolhead17> haha
[18:16] <lynxman> koolhead17: pretty good, enjoying the cold ;)
[18:16] <Daviey> RoAkSoAx: In Dallas, there was ice on the side of the pool.
[18:16] <koolhead17> lynxman: its 23 degrees for me :P
[18:18] <aviandroid> guys what is the best way to remote connect to ubuntu server 11.10 now
[18:18] <aviandroid> vnc rdp
[18:18] <aviandroid> other than ssh
[18:19] <utlemming> koolhead17: C or F?
[18:19] <hggdh> hallyn: do you want access to the lab (re bug 868753)?
[18:19] <koolhead17> utlemming: C :)
[18:19] <hallyn> hggdh, i do have access to the lab, i guess i'll take a machine and try to reproduce.  though if you have a host in the lab still installed that you can still reproduce there, that would rock
[18:20] <RoAkSoAx> zul: ehehehe its cold today.. but for you is pfrobably nice and warm
[18:20] <RoAkSoAx> Daviey: hehehe but it was a warm water pool... wasn't it?
[18:20] <hggdh> hallyn: yes, I do :-) -- aldebaran, or wazn (this is the QA lab, not the server's)
[18:20] <zul> RoAkSoAx: low of 8C tonight here :(
[18:21] <Daviey> RoAkSoAx: can't remember TBH
[18:21] <RoAkSoAx> zul: hehe here it will be ~15C which is cold lol
[18:21] <RoAkSoAx> Daviey: i'll take my swimming suit this time
[18:22] <Daviey> RoAkSoAx: either way. :)
[18:22] <zul> RoAkSoAx: can you make sure Daviey takes his as well
[18:22]  * zul shudders
[18:22] <RoAkSoAx> lol
[18:22] <RoAkSoAx> Daviey: you want me to stop on my wat there and get you one? lol
[18:23] <hallyn> if they're cheap get me one too.
[18:23] <hallyn> (last time i got one at cvs accross the street)
[18:23] <RoAkSoAx> lol
[18:23] <RoAkSoAx> that's where I was planning to stop on my wau there
[18:23] <RoAkSoAx> way*
[18:23] <hallyn> as i recall i spent 30 mins trying to decide between medium and large
[18:24] <hallyn> and now i can't recall which i got
[18:24] <Daviey> heh.
[18:24] <hallyn> (i left it in the room)
[18:24] <hallyn> hggdh, and this is an oneiric laptop accessing natty server, or vice versa?
[18:25] <RoAkSoAx> lol
[18:25] <hggdh> hallyn: an Oneiric laptop accessing Natty
[18:25] <hallyn> ok
[18:34] <Daviey> rbasak: still around?
[18:35] <rbasak> Daviey: yeah but on phone brb
[18:36] <Daviey> rbasak: no hurry.
[19:02] <Demosthenes> so, on lucid, using external USB storage i'm creating luks encrypted filesystems, and they keep going offline. the drive itself stops showing in /dev/disk/by-id. any ideas?
[19:04] <rbasak> Daviey: pong
[19:06] <zul> for those who are daring enough and like to live on the razor wire precise has new libvirt, nova, swift, glance, and novaclient, keystone is a bit lagging because it needs new build dependencies in the archive
[19:07]  * RoAkSoAx will be back in an hour
[19:09] <Daviey> rbasak: It was about the xmlrpc-c branch, the changes confused me until i read the bug comments.
[19:09] <Daviey> rbasak: Are you certain it covers lucid->precise upgrades aswell as oneiric->precise ?
[19:10] <rbasak> Daviey: No, I haven't checked that. I can't see any reason that it wouldn't unless there are further breaks/replaces that I need to add. I can compare the file manifests from the lucid package to check for that too
[19:10] <Daviey> zul: can i deploy it into production?
[19:10] <zul> Daviey: yes in bizzaro world
[19:10] <Daviey> heh
[19:11] <Daviey> rbasak: I might ask cjwatson to take a look over it aswell.
[19:11] <rbasak> Daviey: yeah that's a good idea I wondered about that when adding you as a reviewer
[19:12] <rbasak> Daviey: I'll check the lucid->precise path tomorrow. Wish I'd written some scripts rather than doing ad-hoc comm/uniq/sort/awk stuff by hand now :)
[19:12]  * rbasak wonders if there's some tool for doing this that he's missing
[19:13] <rbasak> In theory the breaks/replaces for this kind of case might be worked out automatically
[19:13] <Daviey> I do not believe there are helpers for this.
[19:16] <rbasak> erm...s/tomorrow/monday/
[19:17]  * rbasak has plans for tomorrow!
[19:17] <Daviey> sounds interesting!
[19:18] <rbasak> might be a bit windy :-/
[19:19] <koolhead17> soren: ping
[19:19] <soren> wazzup?
[19:21] <koolhead17> soren: i wanted your suggestion. i had that 46 revision of dashboard which had code without keystone and it was working fine for cactus
[19:21] <cjwatson> rbasak: the reason they aren't is that there is more than one possible resolution for a file being in two packages
[19:21] <cjwatson> rbasak: but http://conflictchecker.ubuntu.com/ at least has a report for cases that haven't been resolved at all
[19:22] <cjwatson> (yes, its output is giant, at present nobody gardens it)
[19:22] <koolhead17> but since diablo is on keystone and dashboard what shall we do with catctus documentation?
[19:22]  * koolhead17 is confused
[19:23] <soren> koolhead17: Sorry, I don't understand the question.
[19:23] <rbasak> cjwatson: yeah sure, dpkg still needs the fields to work out what to do. I meant a tool that a maintainer could run to say "I've just rearranged the files built from this single source package; tell me what breaks/replaces I need"
[19:24] <rbasak> cjwatson: is that possible? Although maybe a bit too rare a case to be worth doing?
[19:24] <koolhead17> soren: in the cactus/natty documentation we pointed dashboard at launchpad for download, revision 46. Now since there is nomore dashboard there the cactus documentation becomes wrong
[19:24] <koolhead17> :(
[19:25] <cjwatson> rbasak: that would be possible, although the tool would have to have the previous binaries available to it
[19:25] <cjwatson> debdiff can tell you about moves, though it doesn't produce sample Breaks/Replaces fields
[19:26] <koolhead17> i am not able to find same code on github which will not have keystone integrated in it, atleast to maintain the cactus documentation. it seems now i will have to remove that dashboard part form cactus doc
[19:26] <rbasak> cjwatson: yeah, I generated the file I attached to the bug by building the oneiric and precise binaries and poking them
[19:26] <cjwatson> rearranging files in a single source package is not uncommon
[19:26] <soren> koolhead17: Oh.
[19:27] <cjwatson> realistically I suspect it would make little difference unless lintian told you about it, and lintian's interface doesn't give it access to the previous binaries
[19:27] <soren> koolhead17: There should be a mirror on Launchpad still.
[19:27] <koolhead17> soren: nopes :( it used to be earlier
[19:27] <cjwatson> the usual problem is that people forget they need to set the fields at all, not what they should be set to (which is a problem when it's the first time you've done it, but not an ongoing one)
[19:27] <koolhead17> it was your branch only i think :P
[19:28] <soren> koolhead17: https://code.launchpad.net/horizon/diablo
[19:29] <koolhead17> soren: yes
[19:30] <soren> koolhead17: If anyone really cares about Cactus anymore, I guess they could update the docs or something. Dashboard wasn't even an official openstack project back then.
[19:31] <koolhead17> soren: exactly that is what am thinking :)
[19:31] <rbasak> thanks cjwatson, I'll check the lucid->precise path on monday
[19:32] <Daviey> hallyn: Are you considering a merge from lxc in sid?
[19:32] <koolhead17> although 65th revision on openstack-dashboard(launchpad) dated 20/6/11 had latest code minus keystone. i found that
[19:32] <hallyn> Daviey, we're going to do one, yes
[19:33] <soren> koolhead17: I'm not really involved in the docs, and certainly not the docs for Cactus. I also have very little to do with Horizon. I think you've got the wrong guy :)
[19:33] <Daviey> hallyn: is that on your plate?
[19:33] <hallyn> if noone else does it before i get to it, yes
[19:33] <hallyn> i figure zul or stgraber might get to it first.  dont' really care
[19:33] <koolhead17> soren: np.  I will prefer putting/showing love to keystone :)
[19:33] <koolhead17> hehe
[19:34] <hallyn> i'm looking at annoying libvirt network code right now
[19:34] <zul> im missing something
[19:34] <Daviey> hallyn: ahhhh.. don't look too closely.
[19:34] <Daviey> zul: we knew that :)
[19:34] <zul> hallyn: what should i be caring about supposedly?
[19:34] <hallyn> hggdh is making me
[19:34] <zul> hallyn: oh nm
[19:35] <zul> hallyn: no i dont care about it
[19:35] <soren> hallyn: Perhaps i can be of help?
[19:35] <hallyn> zul, i was just sayin gi don't really care if you do it first :)
[19:35] <hallyn> soren, with libvirt?
[19:35] <soren> hallyn: Yeah.
[19:35]  * hggdh sits down to read
[19:35] <zul> hallyn: right
[19:35] <hallyn> soren, that'd be great - there is some buffer overrun or 64/32bit int or something going on,
[19:36]  * hallyn tries to find the bug#
[19:36] <zul> hallyn: im going to play around with the libvirt-lxc block device support this weekend (hoepfuly)
[19:36] <hallyn> bug 868753
[19:36] <hallyn> zul, great, maybe we can get away from using nbd then for openstack-lxc
[19:36] <zul> hallyn: thats the plan
[19:38] <RoyK> it's friday and it's not early and I'm not trying to fix bugs :þ
[19:39] <soren> hallyn: Ah, that sort of thing. Not today, then :)
[19:39] <hallyn> soren, thx anyway :)
[19:43] <zul> have a good weekend
[19:46] <hallyn> is there a good macro or envvar to use in debian/rules to set --srcdir= to?
[19:46] <hallyn> besides `pwd`
[19:53] <soren> hallyn: $(CURDIR)?
[19:53] <soren> (I'm not entirely sure what you're asking, tbh)
[19:53] <hallyn> some TOP_SRCDIR var
[19:54] <soren> hallyn: Ok, then $(CURDIR).
[19:54] <hallyn> but nm, the problem is deiff anyway and i need to punt on that for now
[19:54] <hallyn> ok thx
[19:55] <hallyn> sigh, i don't mind one of those fail days on occasion, but i don't like them to be fridays
[20:28] <Duvrazh> Hey, when I try to ssh my server, it says /bin/bash: No such file or directory, Connection closed.            Yet I can SFTP in and see that /Bin/bash is there. What gives?
[20:29] <Duvrazh> Server is still running all active programs and everything but I can't send ANY commands to it
[20:30] <RoyK> what ubuntu version?
[20:30] <RoyK> never seen that error
[20:30] <Duvrazh> latest
[20:31] <RoyK> latest what? LTS or 11.10?
[20:31] <Duvrazh> 11.10
[20:31] <Duvrazh> sorry
[20:31] <RoyK> dunno, really. for servers, I always stick to LTS
[20:31] <Duvrazh> I'm afraid to reboot it..
[20:32] <Duvrazh> How about this
[20:32] <RoyK> why do you run 11.10 on a server?
[20:32] <Duvrazh> If I reinstall 11.10, will the partman be able to pick up my software raid without deleting everything?
[20:32] <Duvrazh> Because I knew know reason not to
[20:32] <Duvrazh> *no
[20:32] <RoyK> the software raid will be readable whatever version you use
[20:32] <RoyK> Duvrazh: LTS versions are far better tested
[20:33] <RoyK> try 10.04
[20:33] <Duvrazh> So I should be able to install it on a new harddrive and just swap os drive no problems?
[20:33] <RoyK> yes
[20:33] <Duvrazh> That's good to know
[20:33] <RoyK> 10.04 works well, so does even 8.04
[20:33] <RoyK> 8.04 may even be better for some uses :P
[20:34] <Duvrazh> I've always just been a fan of being as up-to-date as possible, I only recently switched from windows to mac&ubuntu (and still testing distros for eee)
[20:34] <Duvrazh> my uses are Twonky, Dropbox, Transmission. :P Do you recommend 10 or 8 for that?
[20:35] <RoyK> 10.04
[20:35] <RoyK> it's not 10, it's 10.04 as in april 2010
[20:35] <Duvrazh> I was typing lazy
[20:36] <RoyK> the LTS releases don't have all the latest versions, but for most use, you won't need that anyway
[20:36] <RoyK> but then, LTS is _stable_
[20:36] <Duvrazh> I'll take your recommendation...
[20:37] <air_> hopefully no beta versions in LTS then :P
[20:37] <RoyK> I have some 30-odd servers at work, most on LTS
[20:37] <Duvrazh> I just use it as a VERY powerful media streamer
[20:37] <Duvrazh> large storage, automatic transmission, upnp, headless and out of the way
[20:37] <RoyK> air_: there may be beta software on LTS as well, but it'll be patched along the way to fix issues
[20:38] <air_> RoyK: just what I thought.
[20:38] <RoyK> air_: if Microsoft were to use the versioning regime used in open source, Vista would be flagged as alpha
[20:38] <air_> and windows me? :D
[20:38] <air_> probably just a nightly
[20:39] <RoyK> :)
[20:40] <air_> RoyK: yeah, I get that sometimes beta is "good enough" and it may have been tested well enough. Still, it feels like if stable release of the package has been released, it would make sense to pull it into the distribution. But of course, it's extra work, and in cases where not so many use the package (netatalk), it might work with beta.
[20:40] <Duvrazh> 10.04 is latest LTS?
[20:41] <genii-around> Yes, until 12.04 arrives
[20:41] <RoyK> air_: that's all to do with whoever wrote it - one guy may call his version 2.0.0.0.0.0 stable, another may release a 0.0.4rc9 and it'll be better
[20:42] <air_> of course.
[20:42] <RoyK> air_: I used to work with asterisk and a package called spandsp, for faxing, the spandsp still is in 0.0.something and works very well - asterisk, on the other hand, is "stable", but is buggy as hell
[20:43] <air_> RoyK: yes, but I referred to versions of the same package.
[20:43] <air_> RoyK: 2.2-beta-4 vs 2.2.0 stable.
[20:43] <RoyK> seems spandsp is at  0.0.6_pre12 atm
[20:44] <air_> If I maintained that package, I'd pull the stable on in before making a dist release (if time allows of course)
[20:44] <RoyK> air_: LTS releases tend to use more stable builds, or at least I hope so :P
[20:45] <RoyK> don't blame me - I'm not packaging things, I just use it :P
[20:45] <air_> yeah, I wont blame you. And I also hope not to see too unstable things in LTS. :)
[20:45] <RoyK> Filesystem            Size  Used Avail Use% Mounted on
[20:45] <RoyK> /dev/md0              5.4T  1.2T  4.2T  22% /raid
[20:45] <RoyK> that's a nice home storage :D
[20:46] <air_> RoyK: you beat me on that.
[20:46] <air_>  /dev/md0              3.6T  2.5T 1011G  71% /Volumes/RAID
[20:46] <air_> 3*2TB in raid-5.
[20:46] <RoyK> waiting for some 3-in-2 thing to extend it
[20:47] <Duvrazh> I'm right up there with you Royk, same size
[20:47] <Duvrazh> 4*2TB raid 5
[20:47] <RoyK> going for 7 drives in RAID-6 soon
[20:47] <air_> I actually have an unused 2TB drive somwhere, hmm.
[20:48] <RoyK> luckily linux can convert to raid-6 :)
[20:48] <air_> RoyK: considering you only store 1.2TB in there, it wouldn't be a big deal to redo it all :P
[20:48] <RoyK> air_: it's filling up rapidly....
[20:49] <air_> ah, ok.
[20:49] <RoyK> and then, a backblaze account for $5 a month and it's all backed up :)
[20:50] <RoyK> http://blog.backblaze.com/2009/09/01/petabytes-on-a-budget-how-to-build-cheap-cloud-storage/
[20:55] <air_> $ lsb_release -d; apt-show-versions | grep beta
[20:55] <air_> Description:	Ubuntu 10.04.3 LTS
[20:55] <air_> liblzma1/lucid uptodate 4.999.9beta+20091116-1
[20:55] <air_> xz-utils/lucid uptodate 4.999.9beta+20091116-1
[20:55] <air_> I just had to. :P
[20:57] <hallyn> hggdh, well this is messed up.  in gdb, 'len' is stil 'optimized out' after it is set with cdr_u_int(&xdr, &len), but it suddenly is defined (to the large value) when it is checked.
[20:58] <RoyK> air_: jag vet...
[20:59] <air_> hmm, var du svensk? eller norsk?
[20:59] <RoyK> norsk
[21:00] <RoyK> men svarte på svensk, siden finner stort sett forstår sånt :P
[21:00] <air_> åtminstone finlandssvenska finnar.
[21:00] <air_> :D
[21:00] <RoyK> :)
[21:01] <RoyK> det er fredag, skål!
[21:02] <air_> skål!
[21:04] <RoyK> air_: do you work with linux as well as abusing it at home?
[21:05] <air_> RoyK: I've had a few servers for internal purposes on my last work. Just switched to a new place, and just got the first server installed there as well. Just to handle some version control, project tracking, etc.
[21:06] <air_> RoyK: so, I do use it, but not in a large scale.
[21:06]  * RoyK manages some 60-odd unix servers, most of them on ubuntu
[21:07] <air_> My current employer doesn't have any hw-servers at all. just some hosted things that are in someone elses control.
[21:07] <air_> or, we do have one mac-mini server :D
[21:07] <RoyK> give or take another solaris 8 machine, some hpux, some VAXes, som windoze boxes from the last millenium :P
[21:08] <air_> :)
[21:09] <air_> At my previous employer, my team had a few vsphere servers in a cluster, and some ~10 win 2k8 servers and maybe 5-10 linuxes.
[21:09] <air_> some old suse servers :/
[21:09] <air_> running on hw. everything new was ubuntu lts running on vsphere.
[21:10] <RoyK> we had this little issue, an instrument computer (connected to an ICP-MS) had a fucking hardcoded IP in the 192.168.0.x network, which we use for our internal network
[21:11] <RoyK> so to get that to work with our network, we either had to let it steal an address from one of our servers, or do som NAT hack
[21:11] <RoyK> I did the latter
[21:11] <air_> :P
[21:12] <RoyK> I tried to talk to the engineer setting up the system and asked, well, can we please set another IP on this?
[21:12] <RoyK> no - well - we can, but that'll require a new firmware, which will cost NOK 100k or so
[21:12] <RoyK> idiots
[21:12] <air_> omg.
[21:12] <air_> seriously.
[21:12] <RoyK> indeed
[21:13] <air_> on my previous employer, we ran this one software that used to be hell to install and configure.
[21:13] <air_> of course, we needed to switch hostname on one server, and it was nowhere to change in the software.
[21:13] <RoyK> the system boots on a floppy (can you beleive it?), so I did a binary search for the IP and found it
[21:13] <air_> I contacted the vendor, and they gave me some bs about the hostname being "compiled in" at install time :D
[21:14] <RoyK> changed it, and found the system booted up with the new IP and then changed it to its old one :P
[21:14] <air_> I grepped for the hostname, found it in a bunch of files, and replaced it. :)
[21:14] <air_> RoyK: so, the ip was hardcoded in several places?
[21:14] <RoyK> so the system had hardcoded its IP not only once, but at least twice
[21:15] <air_> haha, that should reach dailywtf.
[21:15] <koolhead17> Daviey: :P
[21:15] <RoyK> so I gave up, found an old pix router and set it up to do some port forwarding :P
[21:17] <RoyK> air_: the worst thing was the engineer said "We have never heard of such a problem..."
[21:17] <RoyK> as in "I Don't know shit about IPv4"
[21:17] <RoyK> nor computing :P
[21:17] <air_> the world is full of idiots.
[21:18] <RoyK> he was very happy when I did a copy of that floppy disk
[21:18] <el_seano> so I thought I would be clever and save myself a process by using `tail -f /var/log/dmesg` to see what's happening on my machine, but it would appear that copy of dmesg is either from an earlier boot, or the dmesg binary doesn't write to that part of the disk until some other specified point.  Does anyone know where I can throw my tail -f to watch the current dmesg log?
[21:18] <air_> and people rarely want to accept that they are wrong.
[21:18] <air_> RoyK: floppy disk, what year was this?
[21:18] <RoyK> air_: this guy didn't know too much
[21:19] <RoyK> air_: 2011
[21:19] <air_> :O
[21:20] <RoyK> air_: it's an ICP-MS http://en.wikipedia.org/wiki/ICP-MS - a rather expensive set of hardware that analyzes things
[21:21] <RoyK> boots from a floppy and runs a small RTOS
[21:23] <air_> sounds dangerous
[21:23] <air_> I wouldn't want anything expensive relying on a floppy drive to get going.
[21:24] <RoyK> heh - we had an issue a year ago
[21:25] <RoyK> the floppy disk died - the sectors were almost invisible
[21:25] <RoyK> it took us a week to get a new floppy
[21:25] <RoyK> I dd'd it to a file so I guess we're safer now :P
[21:26] <air_> :)
[21:30] <genii-around> RoyK: This gadget looks possibly useful http://hxc2001.free.fr/floppy_drive_emulator/index.html#SDCARDFloppyemulator
[22:20] <dabukalam> this is what ubuntu server startup looks like on a chip: http://pastebin.com/pxr2228W
[22:21] <dabukalam> The network configuration takes way too long. I don't need network for the time being. Is there a way I can get it to not look for network?
[22:22] <rcsheets> dabukalam: what's in /etc/network/interfaces ?
[22:22] <dabukalam> it's a fresh install
[22:22] <dabukalam> so i'd imagine nothing
[22:23] <dabukalam> or just loopback
[22:23] <rcsheets> well it's almost certainly not nothing
[22:23] <dabukalam> or whatever
[22:23] <rcsheets> egrep '^auto' /etc/network/interfaces
[22:23] <dabukalam> yeah, i'm just booting in one sec
[22:23] <rcsheets> oh right, we have to wait 10 minutes for the system to boot. ;)
[22:23] <dabukalam> exactly
[22:24] <dabukalam> I don't remember natty doing this
[22:24] <rcsheets> is this oneiric?
[22:24] <dabukalam> yeah
[22:24] <rcsheets> it has quite a bit of new stuff in terms of network autoconfiguration.
[22:24] <rcsheets> i'm not sure what the installer leaves you with if it doesn't detect anything. all my boxes have manually configured networking.
[22:25] <dabukalam> I surmised as much
[22:25] <rcsheets> seems like there's a bunch more ipv6 support, etc.
[22:25] <rcsheets> so you might be waiting for a lot of that kind of stuff to time out
[22:25] <rcsheets> but once you get the system up, show me the output of that egrep
[22:26] <dabukalam> well i'm not configuring anything networking wise until I get that boot time down, but it will eventually have a network connection
[22:26] <dabukalam> if this install boots in less than 25 secs
[22:26] <dabukalam> no interfaces file
[22:26] <dabukalam> at all
[22:26] <rcsheets> hmm
[22:26] <dabukalam> has that changed?
[22:26] <dabukalam> there are if-down etc files
[22:26] <rcsheets> no... not that i'm aware. did you do some kind of custom install?
[22:26] <dabukalam> nope
[22:26] <rcsheets> hmm
[22:27] <dabukalam> well it's an omap image for arm
[22:27] <dabukalam> but it's ubuntu server
[22:27] <dabukalam> i'll just make one
[22:27] <dabukalam> one sec
[22:27] <rcsheets> what /etc/init/network* files do you have?
[22:28] <rcsheets> if you want to make one, just do "auto lo\niface lo inet loopback\n"
[22:29] <dabukalam> aiight
[22:29] <dabukalam> rebooting
[22:29] <dabukalam> let's see
[22:29] <rcsheets> but that might not do what you want, if this ARM image has some wonky network setup.
[22:29] <rcsheets> i haven't played with any ARM stuff
[22:29] <dabukalam> later on it'll have a static ip
[22:30] <dabukalam> aiight it worked, 24 secs. but i didn't start it properly and missed the start button
[22:30] <dabukalam> one more time
[22:30] <rcsheets> well as long as the interface exists, you can go ahead and give it a static ip and time it again.
[22:30] <rcsheets> i'm curious what it was sitting there doing with no interfaces file. i'd think 'ifup -a' would just exit in that case.
[22:31] <dabukalam> yeah it's weird
[22:31] <dabukalam> anyways 31 secs
[22:31] <dabukalam> do you think removing a bunch of unnecessary apps will cut down the boot time?
[22:31] <dabukalam> can i install minimal
[22:31] <dabukalam> and get it to remove everything?
[22:31] <rcsheets> i'd install bootchart if you can
[22:31] <rcsheets> then you can see what's taking the most time
[22:32] <dabukalam> blegh gotta conf networking
[22:32] <dabukalam> k one sec
[22:32] <rcsheets> have you seen bootchart?
[22:32] <dabukalam> no, sounds cool though
[22:32] <rcsheets> gives you a chart like http://www.bootchart.org/images/bootchart.png
[22:33] <dabukalam> ifup -a will gimme a dhcp config?
[22:33] <rcsheets> 'ifup -a' will configure every interface in /etc/network/interfaces that has an 'auto' line and that is not already up.
[22:33] <dabukalam> so i need to add auto eth0
[22:33] <dabukalam> then do ifup?
[22:33] <rcsheets> you'd need
[22:33] <rcsheets> auto eth0
[22:34] <rcsheets> iface eth0 inet dhcp
[22:34] <rcsheets> ... at a minimum, for dhcp
[22:34] <dabukalam> yeah
[22:34] <dabukalam> i've done that before
[22:34] <dabukalam> what i meant was
[22:34] <dabukalam> is there a utility that does that for me?
[22:34] <rcsheets> a utility that adds those two lines to /etc/network/interfaces? that would usually be the installer.
[22:34] <dabukalam> i've done it already, but just out of curiosity
[22:34] <dabukalam> yeah yeah
[22:35] <rcsheets> i don't know of another utility that does that
[22:35] <dabukalam> but the installer probably uses some package to do that
[22:35] <rcsheets> there could be something
[22:35] <rcsheets> oh yeah it's probably some udeb that the installer environment has
[22:35] <rcsheets> i don't know that much about the installer, sorry
[22:35] <dabukalam> s'fine ^^
[22:39] <dabukalam> rcsheets, haha bootchart wants to install 22MB of shit
[22:39] <dabukalam> a bit counter-productive don't you think?
[22:39] <dabukalam> I'll just purge it when I finish I guess...
[22:39] <rcsheets> 22MB? wow.
[22:39] <dabukalam> dependencies bro
[22:40] <rcsheets> yeah i guess i'm used to installing it on systems that have a few more packages. when i install it it only brings along one dependency ;)
[22:40] <dabukalam> maybe on a desktop
[22:40] <dabukalam> that's why
[22:40] <rcsheets> though a lot of the deps are probably just so it can generate the graphics
[22:40] <dabukalam> wait this is command line logging right?
[22:40] <dabukalam> what graphics?
[22:40] <dabukalam> brah this is server
[22:40] <dabukalam> -.-
[22:40] <rcsheets> yes, so are all the systems i've ever used bootchart on.
[22:41] <rcsheets> your server can't have .png files on it?
[22:41] <dabukalam> it's gui-less
[22:41] <dabukalam> it can
[22:41] <rcsheets> okay... and ssh-less?
[22:41] <dabukalam> haha
[22:41] <dabukalam> another thing to install
[22:41] <rcsheets> surely you have some access to its file system
[22:41] <dabukalam> no
[22:41] <dabukalam> i'll just ssh in and remove that as well
[22:41] <rcsheets> wait, you'll ssh in?
[22:41] <dabukalam> >.>
[22:42] <dabukalam> i mean i'll download and install openssh-server and then purge that when i'm done
[22:42] <dabukalam> i want to cut out all the fat
[22:42] <rcsheets> wel you don't need the server
[22:42] <dabukalam> yeah. I don't.
[22:42] <rcsheets> how did you edit /etc/network/interfaces?
[22:43] <dabukalam> on a monitor
[22:43] <rcsheets> connected to the console of this system?
[22:43] <dabukalam> yeah
[22:43] <rcsheets> ok and it's going to have some kind of networking set up
[22:43] <dabukalam> yeah
[22:43] <dabukalam> eventually
[22:44] <rcsheets> but you don't plan on having any way to get files on/off the box?
[22:44] <rcsheets> that might be useful.
[22:44] <dabukalam> rcsheets, this is a research project ;) and this is a demo machine
[22:45] <rcsheets> i'm not sure how either of those facts negates the utility of remote access to its file system
[22:45] <rcsheets> anyway, with bootchart you end up with a png file or something
[22:45] <rcsheets> you copy that to a system where you can view it
[22:45] <rcsheets> and then you view it
[22:45] <dabukalam> it's installed
[22:45] <dabukalam> so all i do is reboot?
[22:46] <rcsheets> yeah i believe so
[22:46] <rcsheets> it should Just Work[TM]
[22:46] <Jasonn> hi, when I try to SSH into my server, I get this problem: ssh_exchange_identification: Connection closed by remote host
[22:51] <rcsheets> Jasonn: try ssh -v
[22:51] <rcsheets> Jasonn: it might shed some light
[22:53] <Jasonn> stefan@stefan:~$ ssh -v ytalk.us
[22:53] <Jasonn> rcsheets: OpenSSH_5.8p1 Debian-7ubuntu1, OpenSSL 1.0.0e 6 Sep 2011 | debug1: Reading configuration data /etc/ssh/ssh_config | debug1: Applying options for * | debug1: Connecting to ytalk.us [91.121.113.42] port 22. | debug1: Connection established. | debug1: identity file /home/stefan/.ssh/id_rsa type -1 | debug1: identity file /home/stefan/.ssh/id_rsa-cert type -1 | debug1: identity file /home/stefan/.ssh/id_dsa type -1 | debug1: identity file /hom
[22:53] <Jasonn> e/stefan/.ssh/id_dsa-cert type -1 | debug1: identity file /home/stefan/.ssh/id_ecdsa type -1 | debug1: identity file /home/stefan/.ssh/id_ecdsa-cert type -1 | ssh_exchange_identification: Connection closed by remote host
[22:55] <rcsheets> Jasonn: pastebins are awesome.
[22:55] <Jasonn> rcsheets: sorry :3
[22:55] <Jasonn> You want me to pastebin now?
[22:55] <rcsheets> Jasonn: pasting that much stuff into the channel is less awesome
[22:55] <rcsheets> yes, that would make it much more legible, and it wouldn't scroll away while we talk about it
[22:56] <Jasonn> ok
[22:56] <Jasonn> one sec
[22:57] <Jasonn> rcsheets: http://paste2.org/p/1731529
[22:58] <rcsheets> cool. that's exactly the same output i get.
[22:58] <Jasonn> heh
[22:58] <Jasonn> yeah, I know
[22:58] <rcsheets> O_o
[22:58] <Jasonn> i tried to connect at school too
[22:59] <rcsheets> oh, i was wondering how you got into my box
[22:59] <Jasonn> lol
[22:59] <Jasonn> [18:59:17] rcsheets is rcsheets!rcsheets@kumquat.picosecond.org
[22:59] <Jasonn> well, thats not an obvious security flaw ;)
[22:59] <Jasonn> anywayyy
[22:59] <rcsheets> um
[22:59] <Jasonn> how shall I fix thais?
[22:59] <rcsheets> if your hostname has to be a secret, you're doing something wrong.
[22:59] <Jasonn> :3
[23:00] <Jasonn> ok anyway
[23:00] <rcsheets> anyway, is the host otherwise responsive? does it have a non-empty /etc/hosts.allow?
[23:00] <Jasonn> it does not have a non-empty hosts.allow file
[23:00] <Jasonn> but
[23:00] <Jasonn> it was working fine last night
[23:00] <rcsheets> w00t double negative
[23:00] <Jasonn> and I tried to connect this morning
[23:00] <Jasonn> :3
[23:00] <rcsheets> i see
[23:01] <rcsheets> is anything else running on it that you can check? like a web server?
[23:01] <Jasonn> and telnet works
[23:01] <Jasonn> and yeah, I have a web server, and IRC server running
[23:01] <Jasonn> both work
[23:01] <rcsheets> you're running telnetd and you're telling me *I'm* insecure?
[23:01] <Jasonn> I can get into the box, if thats what im askin
[23:01] <rcsheets> ok, well if you can get into the box then have you looked at the logs?
[23:01] <rcsheets> specifically auth.log?
[23:01] <Jasonn> I am running telnetd, usually killed, which I start with a web manager when I need it ;)
[23:01] <Jasonn> where is taht auth located?
[23:01] <rcsheets> in /var/log
[23:03] <Jasonn> I am looking now
[23:03] <flatline> Hi, I'm trying to have an install over ssh, I have my ssh ready, partitioned my hdd, I'm looking for someone help me through the way
[23:04] <Jasonn> flatline: how do you intend to install over ssh? doesnt the OS have to be installed to ssh into it?
[23:04] <Jasonn> or am I wrong?
[23:04] <flatline> Jasonn, booted with ubuntu-server 11.04 cd had that installation bug related with the NEC drives
[23:05] <flatline> so I set up a ssh connection and left work, came home, will try to do the installation over ssh
[23:05] <rcsheets> oh, interesting.
[23:05] <Jasonn> flatline: wont work, you will lose connection  many times
[23:05] <Jasonn> why not just move your monitor over to the box and connect to it?
[23:06] <Jasonn> rcsheets: nothing there that I see shouldnt be there
[23:06] <rcsheets> Jasonn: context = bootchart?
[23:06] <Jasonn> :3
[23:06] <Jasonn> rcsheets: Can I just give you telnet access to the box?
[23:07] <flatline> Jasonn, that box is at work, I'm home, cd drive is not working properly, so my other option is getting myself to work at 2:06AM
[23:07] <rcsheets> oh whoops i got you confused with someone else
[23:07] <flatline> so I'm trying to make sure I really have to go to work to do that before leaving the warmth of my pyjamas
[23:07] <rcsheets> Jasonn: you could, but i'm a stranger on the internet.
[23:07] <flatline> also, I've did this with a live CD before
[23:07] <Jasonn> rcsheets: so can you do it for me :D
[23:08] <Jasonn> ?
[23:08] <rcsheets> Can I? Sure.
[23:08] <Jasonn> heh
[23:08] <Jasonn> well, I mean
[23:08] <Jasonn> now you got me doubting myself ;)
[23:08] <rcsheets> My rate for sysadmin tasks is $100/hour. I accept PayPal sent to rcsheets@acm.org.
[23:08] <rcsheets> You'll need to pay for the first two hours in advance, of course.
[23:08] <flatline> all you need is chroot and a bit of luck I suppose, and damn binutils which is not included in ubuntu-server installer environment :(
[23:11] <flatline> rcsheets, how much for not the task but help?
[23:12] <rcsheets> flatline: help in the channel is free.
[23:12] <Jasonn> rcsheets: honestly, how do I fix this :x
[23:12] <rcsheets> flatline: however i have no idea how to install ubuntu over ssh, so someone else will have to help with that.
[23:13] <foenix> can anyone give me advice about stopping a raid device?
[23:13] <rcsheets> Jasonn: so sshd isn't logging anything?
[23:13] <Jasonn> flatline: I am telling you, it cant be done
[23:13] <Jasonn> rcsheets: oh, that, no
[23:13] <flatline> rcsheets, :( I'm pretty new in ubuntu but have had my experience in fedora (since 8) and arch linux lately, if there was a through guide I would be more than happy to follow
[23:13] <foenix> I've checked fuser and no one's using it
[23:13] <rcsheets> flatline: *shrug* i've never tried such a thing.
[23:13] <flatline> rcsheets, also accepting possible guide whereabouts
[23:13] <flatline> ok :( thanks
[23:13] <rcsheets> sorry
[23:14] <rcsheets> maybe try the forums?
[23:14] <rcsheets> Jasonn: try restarting ssh?
[23:14] <Jasonn> rcsheets: done, same problem :x
[23:14] <Jasonn> I am not that stupid :p
[23:14] <rcsheets> Jasonn: turn up the sshd debug level? i think there's an option for that in sshd_config
[23:14] <Jasonn> yeah?
[23:14]  * Jasonn goes
[23:15] <foenix> I'm getting the following when I check my block holders: https://gist.github.com/1305236
[23:15] <foenix> I can't stop the disk RAID using mdadm or umount
[23:15] <foenix> umount freezes and mdadm returns "mdadm: fail to stop array /dev/md0: Device or resource busy
[23:17] <Jasonn> rcsheets: fixed :D
[23:17] <Jasonn> twas the damned hosts.deny
[23:17] <Jasonn> was preset to *
[23:17] <rcsheets> ah
[23:17] <rcsheets> that doesn't explain it working yesterday but not today.
[23:17] <Jasonn> well
[23:17] <Jasonn> its fixed
[23:17] <Jasonn> now
[23:18] <Jasonn> and thats all I need toknow
[23:18] <Jasonn> \o/
[23:18] <Jasonn> rcsheets: and my server is running debian, btw
[23:18] <foenix> any advice on raid arrays or should I consult another channel?
[23:19] <Jasonn> !volunteers | foenix
[23:19] <Jasonn> :3
[23:19] <Jasonn> just meant the last paty foenix
[23:19] <rcsheets> Jasonn: interesting choice of channel
[23:19] <Jasonn> rcsheets: ubuntu ~ debian
[23:19] <rcsheets> i like debian though
[23:19] <Jasonn> yeah
[23:19] <Jasonn> its a nice OS
[23:20] <Jasonn> ubuntu is overfilled with crap now
[23:20] <rcsheets> well, that's the cost of popularity
[23:20] <Jasonn> and I wouldnt trust it with a server I need to be up for a long time
[23:20] <rcsheets> foenix: i'm not that familiar with md, sorry. is a file system mounted from dm-1?
[23:21] <Jasonn> rcsheets: I still use the desktop though :D
[23:21] <foenix> i don't really know
[23:21] <rcsheets> foenix: ...or a swap space
[23:21] <foenix> no swap
[23:21] <Jasonn> even though I detest unity
[23:21] <rcsheets> foenix: mount | grep dm-1
[23:21] <foenix> it's a 10 disk array
[23:21] <Jasonn> its nice overall
[23:21] <foenix> rcsheets: no return on that command
[23:21] <foenix> I don't care if I wipe the array
[23:21] <rcsheets> well what *is* mounted?
[23:22] <foenix> rcsheets: I updated https://gist.github.com/1305236 to show what is mounted
[23:23]  * rcsheets clicks
[23:23] <foenix> rcsheets: the bottom entry has disappeared
[23:23] <rcsheets> foenix: can you umount /mnt/data and then retry your mdadm command?
[23:24] <rcsheets> oh
[23:24] <foenix> I did the umount and it said it was never mounted
[23:24] <rcsheets> that's ... pretty weird
[23:24] <foenix> i think there's something seriously wrong with mapper
[23:24] <rcsheets> this is why i prefer hardware raid :)
[23:24] <foenix> I've been having a ton of issue with this server
[23:24] <rcsheets> sorry i'm not sure what to tell you.
[23:24] <foenix> anyway to setup hardware raid with this server? Any way to force the kernel to unmount the fs/raid?
[23:25] <foenix> hw raid over ssh seems kinda impossible :\
[23:27] <foenix> rcsheets: this all started when the RAID was catastrophically degraded, I lost 9 out of 10 disks
[23:27] <foenix> I have no clue why
[23:27] <rcsheets> that is a lot of disks.
[23:27] <foenix> my intution is telling me that ubuntu is trying to mount one of the slave disks rather than the md
[23:28] <foenix> which is likely considering the clowns I work with (and who have server access)
[23:28] <foenix> mdadm -D /dev/md0 is hanging
[23:31] <rcsheets> i am sadly quite clueless for mdadm related things
[23:31] <rcsheets> and no one else appears to be around
[23:34] <foenix> yeah, and I'm the most capable tech in my non-profit org
[23:34] <foenix> capable = can use the command line
[23:38] <rcsheets> foenix: what is your relatively tech-light nonprofit doing running a server/
[23:38] <rcsheets> ?
[23:38] <elz89> On a fresh install of 11.10, when I run sudo apt-get update, the last line says the following "W: Failed to fetch gzip:/var/lib/apt/lists/partial/gb.archive.ubuntu.com_ubuntu_dists_oneiric_main_binary-i386_Packages  Hash Sum mismatch
[23:40] <rcsheets> elz89: also using gb.archive.ubuntu.com on oneiric, i am not seeing that.
[23:40] <foenix> rcsheets: our non-profit builds websites for non-profits
[23:40] <foenix> I have a lot of django apps I've been running with vagrant and want to try using the server to create virtualization
[23:41] <rcsheets> foenix: perhaps it would be worth outsourcing the infrastructure so you can focus on the web stuff?
[23:41] <foenix> rcsheets: we got the server donated and the hosting will be free at the library
[23:41] <foenix> believe me, I want to focus on the web stuff
[23:41] <rcsheets> yeah but when it breaks your customer nonprofits are not going to be benefitting much
[23:41] <RoyK> bug 202009
[23:42] <rcsheets> just a thought
[23:43] <foenix> my servers don't usually break. I'm fairly certain a newbie to the team fucked something up when she was trying to figure out how to set up RAID+LVM. I had set it up, but she wanted to learn how. She probably forgot to stop something so I've spent the last 6 hours fixing it
[23:43] <rcsheets> elz89: oh i'm on 64-bit though, so i'm not getting the same stuff you are anyway.
[23:43] <foenix> fuck it, I'm going to call the co-lo centre and have them do a manual reset
[23:43] <rcsheets> foenix: your library is a colo? O_o
[23:44] <elz89> rcsheets: youve lost me?
[23:44] <elz89> rcsheets: how do you know i'm talking about 32?
[23:44] <rcsheets> elz89: well now i'm confused, because you said "i386_Packages" but now i see on my 64-bit box i'm also updating i386_Packages
[23:45] <foenix> rcsheets: the city library has co-lo services for us for free because we were the first free Internet provider in my city
[23:45] <elz89> rcsheets:  ;-)
[23:45] <foenix> we actually still have free dial-up access :\
[23:45] <rcsheets> foenix: interesting setup.
[23:45] <foenix> yeah
[23:45] <rcsheets> foenix: good luck with the raid :(
[23:45] <foenix> when it works
[23:45] <foenix> ty
[23:45] <foenix> you know of any other freenode channel I could ask
[23:45] <foenix> ?
[23:46] <rcsheets> i dunno, you could ask alis
[23:46] <RoyK> ffs - bug 202009 was reported some three years ago and is still not fixed
[23:47] <rcsheets> RoyK: feel free to work on it. :)
[23:48] <rcsheets> RoyK: are you experiencing the problem on something more recent than lucid?
[23:48] <RoyK> see the bug - it was there on hardy as well
[23:49] <rcsheets> RoyK: right, sorry, what i meant to ask was whether you're experiencing the problem on something more recent than whatever was the most recent release mentioned in the bug.
[23:49] <RoyK> and no, I don't want to upgrade my servers to something non-LTS
[23:49] <jiriki> amen
[23:49] <jiriki> ubuntu LTS has been a great repo overall
[23:49] <jiriki> just upgrading from LTS to another one was... a disaster
[23:50] <RoyK> rcsheets: if canonical or whatnot can't fix critical bugs, who can?
[23:50] <rcsheets> RoyK: grub is open source. anyone can fix its bugs.
[23:51] <jiriki> btw are there any good feeds (blog or whatever) on ubuntu server stuff, some tags on delicious work ok but havent' tuned to anything else
[23:51] <jiriki> apart from the mandatory mailing lists
[23:51] <RoyK> rcsheets: may I please ask you to be nice or would you like me to tell you to fuck off?
[23:51] <rcsheets> RoyK: I didn't intend to be anything other than nice. If you'd like to tell me to fuck off, I'll be happy to /ignore you.
[23:52] <RoyK> rcsheets: if a bug was posted some three years ago and still isn't fixed, what would that do to the ubuntu name?
[23:53] <rcsheets> RoyK: Then don't use Ubuntu? What do you want from me?
[23:53] <RoyK> I don't know who you are....
[23:53] <rcsheets> RoyK: then why be so hostile?
[23:54] <rcsheets> RoyK: do you think i'm canonical? should i personally go fix this bug for you?
[23:54] <RoyK> just replying in the manner of your chat
[23:54] <rcsheets> Best of luck to you, RoyK.
[23:55] <RoyK> idiot
[23:55] <qman__> agreed
[23:57] <jiriki> haters gonna hate
[23:57] <RoyK> and no, that's not a grub issue, it's an ubuntu issue
[23:57] <qman__> yeah, it's clearly an issue with the upgrade process
[23:57] <RoyK> packaging, updating configs
[23:58] <RoyK> qman__: all those servers have been installed with lucid
[23:58] <RoyK> no upgrading