[00:09] <hggdh> Daviey: you there?
[00:37] <jeeves_Moss> I'm having some issues with my script.  I can't figure out how to get my script to dload the files it greps out and skips the files that already have been dloaded.  curl http://boards.4chan.org/b/ | grep -o 'http://[^"]*' | grep "http://images.4chan.org/b/src/"
[01:09] <jr> someone can help me to configure a network wireless in ubuntu server
[01:45] <hey_joe> im trying to setup ldap on 10.04, or 9.10, but the dpkg-reconfigure slapd no longer asks for the password for the admin account to be created... am I missing something or what? all howtos say this, and yet it doesnt exist.
[01:45] <hey_joe> whats up with this?
[01:52] <wad> Hi folks. I just installed Ubuntu 10.04 server edition on a new box. It has no network yet (don't ask). The one thing that's blocking me right now is installing mdadm on this box. I know it's on the install CD somewhere, but uncommenting the cdrom source in /etc/apt/sources.list doesn't help; it tried to install, but says it can't find files. Ideas?
[01:54]  * wad looks around hopefully
[01:54] <pmatulis> wad: how did you try to install it?
[01:55] <wad> apt-get update
[01:55] <wad> apt-get upgrade
[01:55] <wad> apt-get install mdadm
[01:55] <wad> ^^
[01:55] <pmatulis> wad: well, what happened?
[01:55] <wad> It said it would install mdadm, and I confirmed with Y,
[01:55] <wad> and then it hit the CD for a moment, and came back with errors about not finding files.
[01:56] <pmatulis> wad: maybe pastebin your sources.list file
[01:56] <wad> I can go and get the exact errors, they are still on-screen (two floors away; no network; don't ask)
[01:56] <wad> It's a brand new install.
[01:56] <wad> As in, this is the very first thing I tried doing after the install.
[01:57] <wad> I need to pull all the files off of a disk from a raid, so that I can configure this box.
[01:57] <wad> That's why I need mdadm. :)
[01:58] <wad> The only line uncommented is the first one, specifying the cdrom as the source.
[01:58] <wad> cd /media/apt #this gets me into the CD-ROM, which has files and stuff in it.
[02:00]  * wad goes downstairs to try some more things with it
[07:48] <_chris_> hej all
[07:48] <_chris_> when logging in to my server with putty, i always get some kind of small stats summary
[07:48] <_chris_> and there it tells me "6 packages can be updated, 6 updates are security updates"
[07:49] <_chris_> when i run update though, he doesnt update anything
[08:19] <RoAkSoAx> ttx: new debdiff attached for etckeeper :)
[08:19] <ttx> RoAkSoAx: arh :)
[08:19] <ttx> Will sponsor during my sponsoring time, in two hours :)
[08:21] <RoAkSoAx> ttx: no prob!! I just stood up from bed to fix that so that I can sleep :)
[08:22] <ttx> ok, you can sleep now :)
[08:22] <twb> I got a nasty surprise when I found the new LTS defaults to silly bzr for etckeeper
[08:22] <RoAkSoAx> ttx: indeed. nighty!!
[08:23] <twb> At least git is in main
[08:24] <ttx> twb: well, bzr is in main too
[08:25] <twb> I meant "at least I can still manually make etckeeper use git, without leaving main"
[08:25] <ttx> ah, ok :)
[08:57] <ben99`> Hello world!
[09:15] <huats> morning
[10:26] <Gorlist> Hi! running 8.04 with plesk 9.5.2 - last week we've been having allot of denied dns query cache errors cropping up in the logs pointing to various domains. I suspect we are under DNS spoofing attack, searching on google now but can anyone here give any advice on blocking or prevent the problem? fail2ban?
[10:28] <twb> Ideally you would first understand the attack.
[10:30] <Gorlist> where could I find more information out? ive been skimming the logs and the only ref. i can find is in the syslog
[10:30] <twb> Is it *your* DNS cache that is being poisoned?  Or are you saying that random people in the internet are having their caches poisoned for your records?
[10:30] <Gorlist> one sec, let me paste the error
[10:30] <Gorlist> client 77.255.9.81 query (cache) 'mail.kssogo.com.tw/MX/IN' denied: 1 Time(s)
[10:31] <Gorlist> and as a result Im suspect we are outputting packets to other DNS servers
[10:31] <twb> What is that message from?  bind?
[10:31] <Gorlist> yes I believe so. Appears in syslog
[10:31] <Gorlist> to be honest im slightly confused by it, I can only seem to trace it back to the bind server
[10:32] <twb> Well, some things you can do are 1) turn on more logging in bind; 2) talk to the bind people / read their docs
[10:32] <twb> mail.kssogo.com.tw doesn't resolve, here.
[10:33] <twb> 77.255.9.81 appears to be some Polish host.
[10:33] <Gorlist> yes, normally about 5 attempts from each ip to different URLS/DNS
[10:33] <Gorlist> and about 10 to 15 attempts during the day
[10:34] <Gorlist> okay will check into bind more and see what I can see. I was wondering if I could change the bind ports as well
[10:34] <twb> Are either 77.255.9.81 or mail.kssogo.com.tw your customers?
[10:35] <Gorlist> no, nothing todo with me
[10:35] <Gorlist> let me put a bit in pastebin
[10:35] <Gorlist> http://pastebin.com/n7bPUnRj
[10:35] <Gorlist> thats over period of 24 hours..
[10:35] <twb> It sounds to me like you're (ab)using bind both outward-facing, serving your DNS records, and inward-facing, serving other people's records to your LAN.
[10:36] <Gorlist> hmm slightly worrying.
[10:36] <twb> And bind is (correctly) telling random strangers on the internet to go to hell, when they ask your bind to resolve something for them.
[10:37] <Gorlist> right,
[10:37] <Gorlist> thats given me something to look for atleast. Thank you
[10:37] <twb> Well, you could just host your DNS records elsewhere, and close that port on the internet-facing interface.
[10:37] <Gorlist> how do you mean host elsewhere?
[10:37] <twb> I mean zonedit or dyndns or similar.
[10:38] <Gorlist> I agree though closing or adjust ports is the quickest temp solution until I can figure out it further
[10:38] <Gorlist> okay
[10:38] <twb> If you aren't hosting your own records, there's no reason that port should be open in the first place
[10:39] <Gorlist> im hosting our own
[10:39] <Gorlist> 17 domains
[10:39] <twb> Then you should know bind better ;-P
[10:39] <Gorlist>  :) I should know lots of things better, but im human after all :)
[10:39] <Gorlist> and I don't pretend to be an expert!
[10:44] <drurew> Hey people...I need a bit of help running a network mapper on a remote interface...anyone know where to start on "forwarding a local script to a remote interface"? More specificly Im trying to map our network, and to get the best results I would like to sniff the gateway...the gateway however is a junos (juniper). I have ssh but no method to run (or install) the libs or script.
[10:46] <Jeeves_> drurew: Create a mirrorport on your switch
[10:47] <drurew> thanks Jeeves_
[10:47] <Jeeves_> the router will probably not show you traffic other than traffic directed to the router itself
[10:47] <Jeeves_> To the Routing Engine, that is
[10:48] <drurew> cant I just do it via ssh ?
[10:48] <drurew> is it possible ?
[10:52] <twb> drurew: I don't think that makes sense
[10:57] <Jeeves_> drurew: If it works, it will probably break your router because it's too busy tcpdumping :)
[10:57] <drurew> yeah Im seeing that...theoreticaly I dont need to sniff the GW...I could also just sniff a central utility server
[11:02] <twb> tcpdump is sufficient to capture a pcap, which you can push out over a different interface to a host with nonvolatile storage, and then run tshark on it to analyse it.
[11:09]  * ttx curses quilt
[11:12] <twb> ttx: you prefer gbp or something?
[11:13] <ttx> twb: no... but that doesn't prevent me from cursing :)
[11:13] <twb> Do you want help, or are you just cursing?
[11:13] <ttx> this source format just doesn't behave the way I want it to. So I'm just cursing
[11:17] <Daviey> ttx: Love the quilt!
[11:20] <ttx> Daviey: if only it didn't persist to think I was creating ANOTHER patch...
[11:20]  * ttx lunches
[11:22] <Daviey> ttx: If you are usign DEP-3, and you make native changes to the upstream source - dpkg-source will create a new patch for you. :)
[11:43] <RoyK> anyone that knows if there's any good software for mapping a network (layer 2) graphically?
[11:44] <rew> RoyK: lanmap2
[11:45] <twb> graphviz
[11:46] <twb> plus whatever you use to generate the non-graphical data; arp or nmap or whatever
[12:24] <xperia2> hello to all. i am going in some weeks for hollidays home and i am planning to deploy there some internet caffe based on ubuntu. my question is what is best hardware solution for clients at the moment that works best with ubuntu ?
[12:25] <xperia2> also exist clients that can multiple monitors and keayboards/mouses be attached at one such client ? say one client server three or four Users ?
[12:27] <twb> I believe normal Xorg with the evdev HID module will Just Work with respect to multiple concurrent desktops
[12:28] <twb> I haven't tried it myself.
[12:29] <xperia2> twb: thank you a lot. will just lookup for this "evdev HID module" !
[12:31] <xperia2> twb: this evdev are more ubuntu software related. i am looking at the moment more for hardware clients to buy
[12:31] <Gorlist> twb, right I think I figured the problem was plesk allows by default external allow-query-cache,  so ive prevented that and also changed the logging for fail2ban :) touch wood and fingers crossed! thanks for the help
[12:32] <twb> Gorlist: no worries.
[12:41] <Daviey> hggdh: Hey, are you around yet?
[12:57] <zul> morning
[13:08] <ttx> Daviey, kirkland, smoser, SpamapS, hallyn, zul: We are missing a few candidates for the next round of papercuts: if you find some low-hanging fruit in your usual bug lists that could be considered a usability issue, please mark as also affecting server-papercuts
[13:09] <ttx> This game is open to anyone listening ^
[13:09] <ttx> (A3 nominations deadline is eod today)
[13:09] <Daviey> ttx: already doing it :)
[13:10] <ttx> Daviey: thx !
[13:10] <Daviey> ttx: feel free to Invalidate them if you don't like 'em! :)
[13:10] <kirkland> ttx: okay
[13:10] <Daviey> morning kirkland o/
[13:10]  * kirkland high fives Daviey 
[13:11] <Daviey> rockin'
[13:16] <tydeas> Why do i get this http://pastebin.com/GDnqMQ7j
[13:16] <tydeas> ???
[13:16] <tydeas> when trying to apt-get upgrade
[13:21] <SkyX> tydeas: try this sudo pkill apt
[13:21] <SkyX> Now you should be able to update as you would normally.
[13:23] <tydeas> I checked there was a apt-get running on the server at the same time
[13:25] <sommer> morning all
[13:25] <Daviey> hey sommer!  How is it going?
[13:26] <Daviey> zul: Would you be able to sponsor bug 590111, it seems you are probably the best person to review it! :)
[13:26] <zul> Daviey: sure
[13:26] <Daviey> rockin'
[13:30] <FC34> hi there
[13:31] <FC34> does any people played with cluster filesystem, san and ubuntu ?
[13:32] <TeTeT> kirkland: can you remind me where the cloud on a stick image lives?
[13:35] <Daviey> FC34: people have, yes.. If you ask what you want to know about them, you will probably get a better answer :)
[13:39] <FC34> Daviey
[13:39] <FC34> tahnk you
[13:39] <FC34> hum ...
[13:40] <FC34> in fact i'm search information about opensource cluster FS
[13:40] <FC34> and i'm think to use ubuntu to to that
[13:40] <FC34> i know GFS as cluster FS
[13:40] <FC34> perhaps it exist others FS
[13:41] <FC34> and On OPenGFS web sites, news are from 2004
[13:41] <FC34> it looks old
[13:41] <cef> not a server q.. but who's a good person/channel to take ubuntu legal stuff to? I've found someone using the ubuntu logo for something not related to ubuntu.. (quite odd actually)
[13:41] <Daviey> I've never heard of OPenGFS :)
[13:41] <Daviey> cef: trademarks@ubuntu.com
[13:41] <FC34> hummm or GFS if you prefer
[13:41] <cef> Daviey: ta
[13:42] <cef> http://cgi.ebay.com.au/NEW-MACH-MINI-BENCH-METAL-LATHE-VARIABLE-SPEED-TOOLS-/230492316545 - top pic on that fwiw
[13:42] <cef> (not the first object pic)
[13:43] <Daviey> cef: agreed, odd!
[13:43] <jpds> That is quite random.
[13:51] <cef> yeah.. oh well.. sent off to trademarks@ anyway
[13:52] <cef> (this is what you find when you think about buying a lathe on ebay...)
[13:53] <cef> for a sec I had visions of it being computer controlled AND having ubuntu driver support. ;)
[14:07] <zul> Daviey: can you look at #598182, im not qualified
[14:08] <Daviey> zul: wilco
[14:10] <Daviey> zul: i commented on it an hr ago
[14:10] <zul> Daviey: thanks
[14:14] <pnunn> What are my chances of recovering a raid array with a manual resync?
[14:14] <pnunn> Booted from a rescue disk.
[14:15] <Daviey> pnunn: linux raid / mdad?
[14:15] <pnunn> yep
[14:15] <Daviey> pnunn: And what do you mean by manual resync?
[14:16] <pnunn> Daviey: Booted of a ubuntu disk, installed mdadm and then configured mdadm.conf and started the array.  /proc/mdstat shows a resync happening now.
[14:18] <Daviey> pnunn: that *looks* promising then
[14:18] <Daviey> pnunn: The disks keep meta data about what position they have in the riad
[14:18] <Daviey> raid*
[14:19] <Daviey> so you shoudl be good!
[14:19] <Daviey> (hoping you have backups)
[14:19] <pnunn> Daviey: Thanks, wasn't sure. The thing was failing to boot the local filesystem after a power snaffu (UPS that died).  Sadly, not my system, a clients, so no backup in store...
[14:26] <Daviey> pnunn: you shoudl be able to mount the raid staright away
[14:26] <Daviey> pnunn: mount it, and check the files are there :)
[14:26] <pnunn> Daviey: OK, I'll give it a go and see what's there.
[14:27] <Daviey> pnunn: fingers crossed!
[14:28] <pnunn> Daviey: except it looks like its lvm.. .bugger.. have to work that out now.
[14:30] <Jeeves_Moss> how can I modify this script to d-load ONLY the output from the script"  curl http://boards.4chan.org/b/ | grep -o 'http://[^"]*' | grep "http://images.4chan.org/b/src/"
[14:36] <hggdh> Daviey: good morning/afternoon
[14:36] <Daviey> hggdh: \o
[14:37] <hggdh> Daviey: thank you again. Every time I try to package Chris' fixes I screw up somewhere...
[14:37] <zul> ttx: adding a upstart job would be considered for a papercut?
[14:37] <Daviey> hggdh: the lucid package is quite exciting :)
[14:38] <engemec> Hi everybody!
[14:42] <ttx> zul: it could be, yes... for easy ones
[14:45] <pnunn> Daviey: Looking good... seems I can see the data.. I'll let the sync finish, then try a reboot.
[14:45] <pnunn> Atleast I know I can get the data back if that doesn't work.
[14:54] <Snadder> How come I get lots of "Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock'" when I put "event_scheduler = 1" into /etc/mysql/my.cnf under [mysqld] ?
[14:55] <Snadder> Happening when I restart mysql
[14:58] <ccheney> appears we now have a working kernel
[15:03] <ttx> sommer: saw https://bugs.launchpad.net/ubuntu/+source/amavisd-new/+bug/165184/comments/33 ?
[15:03] <Snadder> * Stopping MySQL database server mysqld ERROR 1053 (08S01) at line 1: Server shutdown in progress ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2)
[15:04] <sommer> ttx: ya, saw that this morning... I'll ping mdke about updating doc.u.c thanks
[15:04] <ttx> sommer: cool, thanks
[15:31] <docmur____> does anyone have preseed working with partitioning disks
[15:31] <docmur____> when ever I run the install I get stuck at partitioning disks screen
[15:31] <Jeeves_Moss> docmur____, what error are you getting?
[15:32] <docmur____> nothing, just that I always get hit with the screen that wants me to select how to partition
[15:32] <Jeeves_Moss> docmur____, and you've partitioned the disks already?  you set a swap, you've set a root, etc?
[15:33] <docmur____> no, I'm trying to have preseed partition the disk, yes the disk has been partitioned before but that shouldn't matter, it should just wippe it and set it up via the preseed config
[15:34] <Jeeves_Moss> you may have to nuke the disk
[15:35] <docmur____> that doesn't make any sense, what difference would it make if the disk is blank or not.  The preseed shouldn't care because partitioning is partitioning regardless of what's currently on the disk
[15:36] <Jeeves_Moss> docmur____, this is true, however, there is obviously an issue (otherwise you wouldn't be here).  you need to figure out WHY.  Will it let you edit the partitions, etc?
[15:36] <docmur____> if I do it manually yes
[15:37] <Jeeves_Moss> sorry, I'm not sure then
[15:48] <Italian_Plumber> my networking has spontanteously stopped working; I've rebooted and still no dice.  I have not looked a logs and I'm not near the computer right now.  My first thought was hardward dorked up, and to try shutting down the machine for a few minutes and re-starting.  The only thing I did that may have caused this was screw on the case sides which had been just sitting in place
[15:48] <Italian_Plumber> the rest of my internal network is operating properly.
[16:07] <ccheney> Daviey, should 588861 really be confirmed against euca, probably is 'invalid' for it
[16:07] <ccheney> Daviey, unless it is just a tracking task
[16:07] <ccheney> in as much as there is nothing to fix in euca anyway :)
[16:08] <Daviey> ccheney: Yeah.. mainly for tracking, and not knowing what the issue was.
[16:08] <Daviey> but ack marked invalid :)
[16:08] <Daviey> (we aint gonna forget this one :P)
[16:08] <Daviey> ccheney: Is there any news, with the latest upstream patch + kernel?
[16:09] <ccheney> Daviey, yep, fixed will be in ppa and uploaded to the archive the day after freeze
[16:13] <Daviey> ccheney: Awesome!  All that upset over a single line patch eh? :)
[16:13] <ccheney> heh yea
[16:14]  * ccheney is going to try to get OOo in shape for desktop today while my wife is either doing better or in the eye of the storm (hurricane metaphor) ;-)
[16:14] <smoser> hallyn, do you have thoughts on bug 598649
[16:14] <ccheney> i hope its better or i'll be in for a long ride :-\
[16:14] <smoser> i hope for "better" too, ccheney
[16:15] <ccheney> smoser, thanks :)
[16:15] <ccheney> i think it might be a combination of the two with her thinking she can handle it, we'll see
[16:15] <ccheney> and we have numerous other people around to watch her closely
[16:17] <SpamapS> smoser: hey so I discovered libcloud on Friday..
[16:18] <smoser> libcloud is lowest common denominator
[16:18] <smoser> ie, i dont even think that you can pass user data to a launched instance
[16:19] <smoser> which would make it insufficient for our needs
[16:20] <hallyn> smoser: not really (any thoughts).  i'll look at changelog...
[16:20] <smoser> hallyn, well, i'm fuzzy here
[16:20] <smoser> but i *think* that upstream does not build its bios blobs
[16:20] <smoser> and we do
[16:20] <hallyn> smoser: me too, i've never used a non-linux-kernel for -kernel
[16:21] <Daviey> hggdh: Any news on that latest lucid build?
[16:21] <smoser> but upstream works.
[16:21] <hallyn> hm
[16:21] <smoser> so its a matter of figuring out why one does and one doesnt
[16:21] <smoser> we *have* to build our blobs
[16:21] <hggdh> Daviey: I was busy with a phone call, but I am updating the rig with 30.3.5, and will satrt test soon
[16:22] <Daviey> hggdh: Can you update as soon as you can report success/failure; then we can put this bug to bed!
[16:22] <Daviey> Thanks hggdh!
[16:22] <hggdh> Daviey: most certainly. I am eager to get this done ;-)
[16:22] <Daviey> hggdh: I'm eager to forget about the bug :)
[16:23] <hggdh> LOL
[16:24] <smoser> hggdh, so uec-testing-scripts are down to kernel oopses now ?
[16:27] <SpamapS> smoser: I think though, long term, that our time would be better served by improving libcloud, given it works for multiple cloud providers.
[16:27] <smoser> well, its generally a fundamental design issue
[16:27] <inveratulo> Good Morning All -- I am having problems installing ubuntu-server on a Xen host using 'virt-install'.  It has worked with other redhat-based distros AND ubuntu-9.04.  Is there any chance that 10.04 has changed the way this works?  I now get the following error: http://dpaste.com/212322/
[16:27] <smoser> their target is the lowest common denominator
[16:28] <smoser> you would have to change their unified api to allow for cloud-specific api points
[16:28] <smoser> which would somewhat defeat the purpose of a unified api
[16:28] <hallyn> smoser: 590bf491a49670843ee902c47f7ab1de5e9acd06 perhaps
[16:28] <Dennis2> Hello
[16:29] <Dennis2> Using Eucalyptus and having a very odd behavior.
[16:29] <Dennis2> I can't seem to get more than 10 Public IP addresses without it saying out of resources
[16:31] <SpamapS> smoser: the author was there at devopsdays and seemed fairly open to ideas.. I only spoke with him for about 30 seconds though.
[16:32] <smoser> SpamapS, i'm not suggesting that it can't be done
[16:32] <smoser> just that its at least an annoyance.
[16:32] <SpamapS> smoser: ultimately the LCD is very low right now.. list, start, stop .. but I'd hope there will be more as the providers mature.
[16:33] <smoser> additionally, its just a library, which doesn't solve the configuration file format (i dont think) that we sort of wanted to address.
[16:33] <smoser> unless it has a 'parse-config' or something method.
[16:36] <SpamapS> well I'm not sure libcloud is the one ring to bind them all, but it seems euca+ec2 is about 70% of the market now, and will be about 30% of the market in 18 months...
[16:45] <smoser> SpamapS, what hole did you grab "70%" out of ?
[16:45] <smoser> just wondering
[16:48] <smoser> dice roll ? coin flip ?
[16:50] <zul> ouch :)
[16:50] <SpamapS> smoser: I saw it in some presentation by Rightscale.
[16:51] <SpamapS> euca is 0.0003% of that, so I just rounded it down. ;)
[16:52] <SpamapS> smoser: maybe you have something more accurate than the rightscale guys who pimp for amazon quite a bit? ;)
[16:52] <smoser> well i can roll dice
[16:52] <smoser> i think its just about an impossible thing to guess
[16:52] <smoser> for *current*
[16:53] <smoser> and then to forecast that for 18 months
[16:53] <smoser> :)
[16:53] <smoser> take a random number, and then decide if that random number will be higher or lower when you take it again in 18 months
[16:54] <SpamapS> smoser: thats just a guess based on how many people are flocking to rackspace and other providers because they're cheaper and have better IO
[16:56] <hallyn> smoser: test pkg is building, i'll test when i get back from bouncing in 1-2 hrs.  <bouncy bouncy>
[16:56] <Dennis2> Is there any community for ubuntu-cloud? I"m looking through the forums and such and seems to be pretty small
[16:57] <ttx> SpamapS: any chance you can get bug 583933 in order by eod ?
[16:57] <ttx> SpamapS: otherwise I'll have to defer it to A3
[16:57] <SpamapS> ttx: working on it right now actually.
[16:57] <ttx> SpamapS: cool, thanks !
[16:57] <SpamapS> ttx: nearly done, I'll submit it for sponsorship in a couple of hours.
[16:57] <SpamapS> zul: will you have time to sponsor that one today?
[16:57] <zul> SpamapS: ill have a look
[16:57] <ttx> zul is lunching
[16:58] <ttx> zul is back from lunch
[16:58] <SpamapS> zul: ok I'll ping you again when its ready.
[17:13] <hallyn> smoser: yes, i created a pkg with just cherry-picking 590bf491a49670843ee902c47f7ab1de5e9acd06, and it seems to work now
[17:14] <hallyn> i'm leaving for real now, back in awhile
[17:14] <smoser> hallyn, rock on.
[17:23] <smoser> just so every one is clear, i would like to clarify that i was not trying to be as big a prick as i appeared above.
[17:23]  * smoser apologizes to SpamapS 
[17:28] <SpamapS> smoser: no worries  :)
[17:28]  * SpamapS had already begun to forget about it.
[17:36] <bogeyd6> smoser, i think we all have those days
[17:50] <Doonz> Hey guys im looking at the easiest way to mount a directory from one ubuntu server into another. what im trying to do is i have a storage server and would like my webserver to be able to use some of the space there any suggestion on how to do it
[18:01] <SpamapS> Doonz: what if the storage server goes down?
[18:02] <SpamapS> Doonz: I mean, does it need to be available 99.99% of the time?
[18:02] <Doonz> no
[18:02] <SpamapS> Doonz: ok, NFS is probably the simplest way to do this
[18:02] <Doonz> and the server wont go down unless the place blows up
[18:03] <SpamapS> Doonz: you have to make sure the user ids match on both servers
[18:03] <SpamapS> Doonz: hahahahahahaha
[18:03] <Doonz> yes
[18:03] <SpamapS> Doonz: servers go down all the time. :)
[18:03] <Doonz> brb baby need diaper changes
[18:14] <Doonz> how can i check to see what users exist on my system
[18:15] <pmatulis> Doonz: baby ok?
[18:20] <Doonz> yep
[18:21] <Italian_Plumber> for those who may or may not have been interested in my post aobut my server losing connectivity:  It appears to have been a bad cable.
[18:21] <Doonz> damn cables
[18:21] <Doonz> anyone have a link with the easiest how-to for nfs setup
[18:22] <pmatulis> Doonz: server guide
[18:27] <zul> SpamapS: how did the bof go?
[18:27] <SpamapS> zul: pretty well I think.
[18:28] <zul> SpamapS: did people actually showed up?
[18:28] <SpamapS> zul: we didn't really bother with a UEC demo.. we just chatted with a few users.
[18:28] <SpamapS> zul: not a lot.. the 3 of us + Monty Taylor and then a couple of users (though one was Wikia)
[18:28] <zul> ah
[18:28] <zul> wikia?
[18:29] <SpamapS> Yeah, the guys who run wikipedia
[18:29] <zul> oooh....yeah...
[18:29] <zul> knowing that wikipedia runs server makes me nervous...hmmm...i broke wikipedia today
[18:30] <SpamapS> and some university library software integrator who has 200+ installations of 8.04 in libraries and is looking at migrating them to UEC
[18:30] <SpamapS> zul: hahahaha
[18:31] <SpamapS> question.. the memcached package previously ran as 'nobody' .. but we're moving it to run as 'memcache' .. however, the file that defines that is /etc/memcached.conf ... so when people upgrade, it will stil say '-u nobody' ... can we force replacing that file if it hasn't changed?
[18:32] <SpamapS> zul: actually their main point was that they don't use *any* of our packages.
[18:32] <mathiaz> SpamapS: it == the user/
[18:32] <mathiaz> SpamapS: ?
[18:32] <SpamapS> zul: they install, and move on
[18:32] <SpamapS> mathiaz: it == /etc/memcached.conf
[18:32] <mathiaz> SpamapS: if the configuration file hasn't changed dpkg will override automatically
[18:33] <zul> SpamapS: oh thank god
[18:33] <SpamapS> mathiaz: ahh, in this case the override didn't work for me
[18:33] <mathiaz> SpamapS: is /etc/memcached.conf a conffile?
[18:33]  * SpamapS will try on a base maverick
[18:34] <SpamapS> mathiaz: yeah its used to configure the startup for memcached
[18:35] <ccheney> rebooting irc machine, bbs
[18:37] <adurity> I'm looking to do some network I/O benchmarks.  What tools do you recommend?
[18:38] <zul> adurity: bonnie
[18:38] <adurity> isn't that for disk I/O?
[19:10] <RoyK> anyone here tried to use zfs-fuse?
[19:10] <RoyK> seems rather buggy
[19:30] <ccheney> taking late lunch, bbia 1hr
[19:33] <prodcutnews1> im using ubuntu 9.04, always a update screen for 9.10 appears onto screen, i never updated ... incase i do update, whatz next...does it asks me for 10.4 update or not ?  is it possible jumps from 9.04 to 10.4 update .. what i need to do
[19:36] <preseed> ok I can not get preseed to partition, I've tried in both ubuntu and now debian so I'm faily sure it's how I'm doing it.  Here is the preseed config code to partition, does it look okay. http://pastebin.com/F6uBBBTt
[19:37] <RoyK> preseed: you can jump from LTS to the next LTS or from a non-LTS to the next non-LTS, so 9.04 will upgrade to 9.10, then you can do-release-upgrade again to turn it into 10.04
[19:37] <RoyK> prodcutnews1: that was for you, sorry
[19:38] <prodcutnews1> thanks RoyK
[19:39] <prodcutnews1> i understood , we need to follow the sequence
[19:39] <RoyK> yep
[19:39] <RoyK> it won't take too long unless you have a truckload of packages installed
[19:40] <hallyn> smoser: trying to figure out whether it's fixed or not.  When I run your kvm command, i now get:
[19:40] <hallyn>                                   error: hd0 cannot get C/H/S values.
[19:40] <hallyn>                                   grub rescue>
[19:41] <hallyn> smoser: the grub-mkimage command you gave can't be quite right, it complains no -O was specified
[19:41] <smoser> not fixed
[19:41] <smoser> oh. wait.
[19:41] <smoser> so, i should have mentioned this.
[19:41] <smoser> to get it working for me, i used:
[19:41] <smoser> make install INSTALL_BLOBS=1
[19:42] <prodcutnews1> incase of update from 9.04 to 9.10 was successful, and i had any prob , how can i get back myself 9.10 to 9.04 , i insist update is successful, but there probs like sound card, screen reolution, drivers issues or any . how can i downgrade the version which favoured my machine hardware .
[19:42] <smoser> off bc3a9ccc5ddea4c0c713ef6fb3c11d9a88cec169 of git://git.kernel.org/pub/scm/virt/kvm/qemu-kvm.git
[19:43] <RoyK> prodcutnews1: backup the system before updating, I don't think there is such a thing as a rollback - that may come with btrfs and clones, but when that happens, I don't think even the gods know
[19:43] <hallyn> smoser: confused-  what are you saying about bc3a9ccc5ddea4c0c713ef6fb3c11d9a88cec169 ?
[19:43] <smoser> i built from that and it works for me.
[19:44] <smoser> i built like:
[19:44] <smoser> ./configure --target-list="" --prefix=$PWD/../kvm-testbin --interp-prefix=/etc/qemu-binfmt/%M --disable-blobs --disable-strip --audio-drv-list="pa,alsa,sdl,oss"
[19:44] <smoser> then did
[19:44] <smoser> make install
[19:44] <smoser> make install INSTALL_BLOBS=1
[19:44] <smoser> so, i'm fairly sure i'm using the blobs from pc-bios/ at that tip
[19:44] <prodcutnews1> why i ask this, i did earlier update from 9.04 -> 9.10 , it would be strange ghost story for few whoever listens, a sound was coming out of my laptop which i could see outside to my ears, i installed and 9.04, it was ok.
[19:45] <smoser> hallyn, that make sense ?
[19:45] <smoser> and if you 'git-log pc-bios/multiboot.bin' you'll notice it was last touched in November.
[19:46] <hallyn> smoser: i see, that's the last commit in your tree :)  got it
[19:46] <RoyK> prodcutnews1: doesn't sound like a server issue, though
[19:47] <smoser> hm.. i just now noticed that my multiboot.bin that i'm using , and the one in pc-bios are not bit for bit.
[19:47] <smoser> so let me check something
[19:49]  * smoser has to get his build/source directories off of encrypted home.
[19:51] <hallyn> but the "error: hd0 cannot get C/H/S values." is *different*...
[19:52] <hallyn> all right guess i have to try from kvm git to make sure...  <sigh>
[19:56] <smoser> hallyn, i have seen that error.
[19:56] <smoser> when i tried to mix and match bios files
[19:57] <smoser> ie, tried pointing the lucid kvm at the bios from the git tip or vice versa
[19:57] <TeTeT> kirkland: what's the recommended approach to change the networking mode of UEC? eucalyptus.local.conf states it should never be edited directly but with a tool
[19:57] <smoser> i've verified nwo, that my explanation of what i did above is correct.
[19:58] <TeTeT> kirkland: the question popped up on the answer tracker, feel free to reply directly to https://answers.launchpad.net/ubuntu/+source/eucalyptus/+question/116040
[19:58] <smoser> hallyn, please ping me if you need help, i'm sorry that i was confusing above.
[19:59] <hallyn> smoser: in the command:
[19:59] <hallyn> grub-mkimage --output=grub-loader.img --config=mycfg.cfg \ --prefix=/boot/grub biosdisk ext2 part_msdos
[20:00] <hallyn> did you use -O multiboot or something?
[20:00] <smoser> no.
[20:00] <hallyn> smoser: that command just will not work forme
[20:00] <hallyn> (kvm-git is compiling right now...)
[20:00] <hey_joe> anyone know why ldap is broken in ubuntu 10.04?
[20:00] <hallyn> (actually no, i'm still doing apt-get build-dep :)
[20:01] <smoser> hallyn, i just verified that it works for me here on lucid on my laptop
[20:02] <hallyn> hm, it was on maverick it wasn't working for me
[20:08] <hey_joe> what are you guys talking about anyways?
[20:10] <smoser> hallyn, sorry, that is confusing. i was using lucid grub
[20:11] <RoyK> grub2
[20:11] <RoyK> smoser: if you installed lucid from scratch, it'll have grub2. if you upgraded, it'll probably have grub1
[20:12] <smoser> RoyK, well, grub-mkimage is grub2 only as far as I know. but mine is from grub2.
[20:12] <smoser> hallyn, i've just verified that : echo "set prefix=(hd0)/boot/grub" > mycfg.cfg ; grub-mkimage --output=grub-loader.img --config=mycfg.cfg --prefix=/boot/grub biosdisk ext2 part_msdos
[20:12] <hallyn> smoser: yeah /usr/sbin/grub-mkimage now has a new -O option, apparently mandatory, with no indication of what is the equiv of the old default :)
[20:12] <smoser> works on alpha1 maverick
[20:13] <hallyn> can you give me `dpkg-query -s /usr/sbin/grub-mkimage` ?
[20:13] <smoser> yeah, i'm going to upgraading that really quick
[20:15] <hallyn> on lucid it comes for me from grub-pc, in maverick from grub-common
[20:16] <smoser> grub-pc 1.98-1ubuntu6
[20:16] <smoser> is what i had on my maverick
[20:16] <smoser> i will upgrade it and figure something out for you
[20:20] <rberger> With Ubuntu 10.4 UEC server running Eucalyptus  http://169.254.169.254/latest/meta-data/hostname returns an IP address not a FQDN  How do I get UEC on Eucalyptus to set the FQDN of the hostname so that hostname -f returns a FQDN?
[20:20] <smoser> hallyn, i can pretty much verify that grub2 on maverick is borked
[20:20] <smoser> :)
[20:21] <hallyn> )$*%*($&%*(&
[20:22] <smoser> hallyn, i can verify that this command returns success on grub2 from current maverick (1.98+20100614-2ubuntu3) : grub-mkimage -O i386-pc --output=grub-loader.img --config=mycfg.cfg --prefix=/boot/grub biosdisk ext2 part_msdos
[20:23] <smoser> but that it doesn't boot properly
[20:23] <smoser> when i try -kernel with the result i get:
[20:23] <smoser> error: symbol not found: `grub_list_pop'.
[20:23] <jeiworth_> hi, just encountered a package linux-virtual that apparently "is geared for running linux as virtual guest", but additional information is a bit sparse... anyone know what advantages this kernel has in comparison to the standard server kernel, i am running various vm's under kvm/qemu?
[20:25] <hey_joe> anyone have experience with openldap in 10x?
[20:26] <mreynolds> I'm using the Ubuntu 10.04 images with EC2 and I'm running into occasional frozen image behavior (can't SSH, can't get console).  Is anyone else seeing this?  Is there someone I can bug about this? :)
[20:26] <smoser> hallyn, stragely, 'grub_list_pop' appears no where in the maverick source at the moment
[20:27] <smoser> (other than references to removing it in the changelog)
[20:27] <hey_joe> jeiworth_: its a kernel built for virtualization... specifics I don't have, but it will exclude stuff, and include other stuff obviously...
[20:27] <hggdh> smoser: I am running a 2,000-instance run now
[20:28] <hggdh> smoser, Daviey: close, a cigar, but no lighter. Almost there
[20:28] <hggdh> in 1,000 instances, 4 metadata issues
[20:28] <smoser> and those metadata issues are public id issues ?
[20:28] <jeiworth_> hey_joe: thanks, yeah, i think so too but i can't seem to find any specific details googling... hmm, well maybe the ubuntu docs yield something
[20:28] <smoser> and, on the ohter errors, are those kernel oops ?
[20:29] <jeiworth_> anywho, trying it on a dev machine right now :D
[20:29] <hey_joe> Jeeves_: i couldn't find anything either...
[20:29] <hey_joe> are you running a private cloud mreynolds or on the actual amazon cloud?
[20:29] <mreynolds> hey_joe: Actual amazon cloud
[20:30] <hey_joe> do those services that freeze up come back?
[20:30] <mreynolds> I'm trying the east images just in case.  I've seen this before, but it was intermittent.
[20:30] <hey_joe> are do you have to reboot the vm?
[20:30] <hey_joe> prolly just a connection issue
[20:30] <mreynolds> hey_joe: Rebooting the VM doesn't seem to help.
[20:30] <hey_joe> yeh
[20:30] <hey_joe> i haven't played with it much..
[20:30] <mreynolds> hey_joe: I have an EBS image up and it can't connect from inside the cloud either
[20:30] <hey_joe> but I would watch your traffic and see if it drops
[20:31] <hey_joe> ah
[20:31] <mreynolds> hey_joe: I'm inclined to think it's not connectivity since I can't connect from an existing, up connection.  It's using the same security group, so I know I can ssh in and all that.
[20:32] <hey_joe> mreynolds: im sorry... i haven't played with it at all.. was going to setup a private cloud, but went with seperate physical installs, and ldap authentication..
[20:32] <hey_joe> does EC2 support authentication from a centralized place for all the virtual servers?
[20:32] <mreynolds> hey_joe: NP, thanks for the comment.  My company is internally screwed up enough that the best we can do is get Amazon servers :)
[20:33] <hey_joe> like if I add a user... can that account be used on all virtual servers?
[20:33] <smoser> hey_joe, it supports it in that you can set up a centralized authentication service and configure your isntances to use it
[20:37] <ocatacoo> can someone help me with named.conf
[20:38] <ocatacoo> I am just looking for a working example
[20:42] <mreynolds> ocatacoo: I think there are tons if you google for them.  This is probably something people won't respond to here (I'm new here, so I"m guessing this isn't the forum for that question)
[20:45] <mreynolds> Is there a preferred way of upgrading images on EC2?  Right now I'm just running an instance image, dist-upgrade'ing, and then rebooting, then saving the image.
[20:54] <hey_joe> smoser... so i could setup a cloud, node controller on one machine with the auth service, then add all my machines to the node?
[20:54] <smoser> hey_joe, yes.
[20:54] <hey_joe> sweet
[20:54] <smoser> wait.
[20:54] <smoser> wait.
[20:54] <hey_joe> what auth service it it? does it use ldap or something?
[20:55] <smoser> what do you mean "setup a cloud"
[20:55] <smoser> you asked about EC2
[20:55] <smoser> which *is* a cloud
[20:55] <hey_joe> private cloud
[20:56] <smoser> UEC (based on eucalyptus) has no support that I know of to integrate with an existing authentication service.
[20:56] <hey_joe> k
[20:56] <smoser> however, you could set up an LDAP server, and then configure your AMIs such that, when launched, they would authenticate against that ldap server.
[20:56] <hey_joe> thats what i wanted to know :-)
[20:56] <smoser> hallyn, i'm going to ask in #grub about mkimage if you want to join
[20:56] <hey_joe> smoser... that is true, however openldap seems to be broken in Ubuntu 10.04 and 9.10
[20:57] <rberger> Why does UEC on Eucalyptus return an IP address and not a FQDN for the hostname meta-data call http://169.254.169.254/latest/meta-data/hostname?
[20:57] <smoser> i dont know much about that. are there bugs open addressing its broken-ness ?
[20:57] <hey_joe> smoser, unfortunately
[20:57] <rberger> Is there a way to get the FQDN out of the Eucalyptus CLC?
[20:58] <smoser> rberger, i do not belive that there is.
[20:58] <hey_joe> thnks guys
[20:58] <smoser> rberger, https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/475354
[20:58] <rberger> smoser not having the hostname set to a fqdn seems to break things like Chef and puppet and some apt-get package installs
[20:59] <smoser> rberger, that bug shows how we've addressed that in the ubuntu images.
[20:59] <rberger> smoser it seems to imply that only the hostname part is addressed not the FQDN
[20:59] <smoser> i'm not sure i follow
[21:00] <smoser> read comment 10.
[21:00] <smoser> if you want a FQDN in the hostname field, then you have to enable dns in eucalyptus
[21:01] <rberger> smoser, the last entry says " set hostname to a hostname-like value (ip-W-X-Y-Z)  if instance metadata includes a ip address rather than a hostname"
[21:01] <hallyn> smoser: JINKEYS!  well, the multiboot.bin installed by my modified package (which doesn't work) is exactly the same (per sha1sum) as the one compiled by kvm-git (which does work)
[21:02] <hallyn> so that commit does need to be there, and i'm now just onto a different bug
[21:04] <rberger> smoser right now on my Unbutu 10.4 with latest patches hostname returns euca-172-19-1-2, but hostname -f returns "hostname: Name or service not known"
[21:06] <smoser> rberger, so, i think that to address that you'll have to run eucalyptus dns service.
[21:07] <rberger> smoser, I am running the Eucalyptus DNS service as per their wiki. They don't implement PTR yet.  Could someone give me a hint where hostname gets set when booting as a UEC eucalyptus instance? I can at least hack around things and get on with my work if I knew that.
[21:07] <smoser> rberger, it is set by cloud-init (assuming you're using a Ubuntu lucid guest)
[21:08] <smoser> i think /usr/bin/cloud-init iirc
[21:09] <rberger> smoser I'll take more of a look into cloud-init. I was poking around there and couldn't figure out if that was actually doing it or not.. Thanks
[21:12] <smoser> rberger, thats the rigth place to look. it will only do it on first boot of an instance
[21:25] <hallyn> smoser: if i want you to test a particular .deb, am i better off pushing to my ppa, or posting jsut the deb to launchpad?
[21:27] <smoser> probably push to your ppa, or just push somewhere http
[21:27] <smoser> (people.canonical.com is used for that sort of thing, or some personal web space you have)
[21:27] <hallyn> it's just htat pushing to my ppa makes lauunchpad recompile it
[21:27] <smoser> right.
[21:27] <smoser> so i'm good with taking a binary from you
[21:28] <smoser> assuming you've inserted all sorts of keyloggers and rootkits of course
[21:28] <hallyn> hm, how do i put it on people.canonical?
[21:28] <hallyn> i do have  a source tree pushed to bzr so you could rebuild if you have a really fast machine
[21:28] <hallyn> well shall i just email you?
[21:28] <hallyn> (and have my rootkit email me back your keys)
[21:29] <smoser> hallyn, here.
[21:30] <smoser> hm..
[21:30] <smoser> well.
[21:30] <smoser> you should add some ssh keys to https://launchpad.net/~hallyn/
[21:30] <smoser> then i could let you put them somewhere by putting your keys somewhere.
[21:32] <hallyn> huh, thought my keys were there
[21:32] <hallyn> emailed
[21:48] <smoser> hallyn, so yeah, i get the same error you're getting now.
[21:48] <smoser> maybe you need to install the updated qemu-kvm package
[21:49] <hallyn> ?
[21:49] <smoser> (ie, a tie between kvm binary and bios)
[21:49] <hallyn> well on my system i installed them all,
[21:49] <smoser> ah.
[21:49] <smoser> ok
[21:49] <smoser> i saw errors like you are seeing there when I was trying to mismatch the two
[21:49] <smoser> (ie, bios from lucid and kvm from git)
[21:49] <smoser> have you seen this work ?
[21:50] <smoser> ie, were you able to recreate my success ?
[21:50] <hallyn> yes
[21:50] <smoser> good.
[21:50] <smoser> so i'm not completely insane
[21:50] <hallyn> hm, maybe i didn't do sbuild -A, maybe i *did* miss out on something obscure...
[21:51] <smoser> i dont think thats it.
[21:52] <smoser> zul, i believe we never came up with "package-of-the-day" for kvm, is that right?
[21:52] <smoser> kirkland,
[21:52] <kirkland> ?
[21:52]  * hallyn goes to look some more at the kvm changelog
[21:52] <kirkland> smoser: sup?
[21:55] <smoser> above
[21:56] <smoser> do we have a package-of-the-day for qemu-kvm ?
[22:27] <MTecknology> Is it possible to only update packages that were released 1wk ago or more?
[22:38] <zul> smoser: disabled right now...itll be back for alpha3
[23:35] <hallyn> smoser: now it just occurred to me, you and i are comparing to the kernel.org kvm version, not the sf.net qemu , which is what we're currently based on
[23:54] <SpamapS> mathiaz: how's hadoop?
[23:54] <mathiaz> SpamapS: are you refering to the summit?
[23:56] <SpamapS> mathiaz: yes. :) is that today?
[23:56] <mathiaz> SpamapS: nope - tomorrow
[23:57] <SpamapS> mathiaz: ah...
[23:57] <SpamapS> mathiaz: I am playing with bzr-build .. how do you handle situations where people have added all generated files like 'configure' to their .bzrignore .. can the recipe call autoconf scripts?
[23:58] <mathiaz> SpamapS: bzr builder or bzr builddeb?
[23:58] <hallyn> smoser: hm, well, just confirmed that it still works with upstream qemu too :)
[23:58] <hallyn> $*(&%(*$&
[23:58] <SpamapS> mathiaz: builder
[23:59] <mathiaz> SpamapS: IIRC bzr builder --help prints a good message
[23:59] <mathiaz> SpamapS: IIRC there is a way to specify commands to run after the merge has been done
[23:59] <mathiaz> SpamapS: bzr help builde
[23:59] <mathiaz> SpamapS: bzr help builder