[00:02] <RoyK> baccenfutter: wtf - who did?
[00:04] <baccenfutter> the garrythefish guy
[00:41] <lunaphyte> how can i show the real group and effective group for processes in he output of ps?
[00:57] <thenetduck> hi, i'm setting up postfix on my hardy heron server and I was wondering what would the best option be for a website that sends simple emails for welcome and orders?
[00:57] <thenetduck> there is "internet site" internet site with smarthost"
[00:57] <thenetduck> and some other options
[00:58] <lunaphyte> isn't there already a mail server on your network you can use?
[01:05] <thenetduck> lunaphyte: i'm using a slice from slicehost
[01:05] <thenetduck> I don't know, i'm kind of new to the mail server thing
[01:05] <thenetduck> there is something called sendmail I think
[01:06] <thenetduck> to be honest, I acutally need some realy help getting my mail server set up
[01:06] <lunaphyte> running a mail server isn't trivial.  you really should employ someone who can ensure things go well.
[01:07] <thenetduck> lunaphyte: I have no money, I guess this is more of a learning experience
[01:07] <thenetduck> now if I had a job... haha
[01:07] <lunaphyte> employ doesn't necessarily mean money.
[01:08] <lunaphyte> does slicehost provide a mailserver for their customers to use?  if so i would just use that.
[01:09] <thenetduck> um, I don't know I should ask
[01:09] <lunaphyte> probably a good idea.
[01:10] <thenetduck> I guess the problem is my rails app uses :sendmail for the server
[01:10] <lunaphyte> that's no problem.
[01:11] <lunaphyte> assuming your provider provides a mail server, all you need is a null client, not an mta.
[01:11] <mrchrisadams> lunaphyte: how does that normally work?
[01:11] <lunaphyte> what?
[01:12] <lumpEee> greetings all
[01:12] <mrchrisadams> I have one account with webfaction which has a mailserver at mail.webfaction  that I can send mail through from an app
[01:12] <mrchrisadams> much like thenetduck
[01:12] <mrchrisadams> but I've recently started using a vps
[01:12] <mrchrisadams> which doesn't explicitly have a mail server for this
[01:13] <lumpEee> this is support channel for ubunter server correct?
[01:13] <mrchrisadams> or at least there isn't something like mail.vpscompany.com, like webfaction provided
[01:13] <thenetduck> lumpEee: no it's for elephants who like to eat peanuts
[01:13] <lunaphyte> mrchrisadams: you'd just have to ask, like thenetduck
[01:13] <lumpEee> oh good i fit in there too
[01:13] <thenetduck> haha
[01:15] <lumpEee> and thenetduck, likely i will have more dumb Qs
[01:15] <mrchrisadams> ah, "A null client is a machine that can only send mail. " - http://www.postfix.org/STANDARD_CONFIGURATION_README.html
[01:16] <thenetduck> lumpEee: i'm pretty much the biggest noob on the chanell so ask away ahah
[01:16] <lumpEee> I like ubuntu, have needs best met by a server and am rather a n00b at ubuntu server
[01:16] <lumpEee> i ain't even ready to ask yet
[01:16] <lumpEee> i am reading through docs and figured it be a good idea to lurk here a bit
[01:17] <lumpEee> i love ubuntu though
[01:17] <lumpEee> just not used to any type of server stuff outside of CPanel
[01:19] <thedoble> lumpEee: feel free to lurk, I'm a windows sysadmin who'se slowly converting to the dark side :)
[01:19] <thenetduck> lunaphyte: well they do but it's 10/month for it. I'm not paying that haha
[01:19] <lunaphyte> no surprise there.
[01:20] <lumpEee> thedoble, i still have a ME machine running.. lol
[01:20] <lumpEee> other than that one XP and the other 4 all flavors of ubuntu
[01:22] <thedoble> lumpEee: heh, I am not yet a fan of ubuntu desktop, but I like the server version a lot. I have a macbook now as well so I am slowly weaning myself off microsoft !
[01:23] <lumpEee> i have the studio version, the net book remix version, the desktop (although I mostly use it as storage via the network) and just build a server
[01:23] <mneptok> oh yes. Apple is a *much* more open company than is Microsoft. they treat their users with respect, assuming them to be intelligent and capable.

[01:24] <lumpEee> i am rather fond of the way they did the netbook version
[01:24] <lumpEee> i really like it on this Eee
[01:24] <lumpEee> studio is also very good with a real time kernel
[01:26] <thedoble> lol
[01:26] <lumpEee> but as for the desktop version.  I just set up file sharing and it sits in my lr
[01:28] <lumpEee> my7 migration seem to be more toward ubuntu than "a company that used to be deifned by a band"
[01:28] <lumpEee> heh
[01:29] <lumpEee> sorry i am an apple fan in another way
[01:30] <thedoble> i still struggle to get my head around linux file permissions in samba
[01:30] <lumpEee> thanks for saying that
[01:31] <lumpEee> one of the things i am having problems with myself
[01:31] <lumpEee> corect my n00b arse if wrong
[01:32] <lumpEee> it just don't seem to work right from the GUI
[01:33] <thedoble> i dont know, I don't use the GUI, but I do know that it can be really confusing in the command line too
[01:33] <thedoble> i think its just different
[01:33] <thedoble> and i need to relearn how i think about it
[01:34] <lumpEee> i am not afraid of cli
[01:34] <lumpEee> just reaaaal rusty at it
[01:34] <lumpEee> the last time i played with php it was 3.0
[01:36] <lumpEee> i am just at a point where i really need a server
[01:36] <lumpEee> so it is time to hit the books again so to speak
[01:38] <thedoble> hehe
[01:38] <thedoble> i dont do much work with web servers
[01:40] <lumpEee> more of a database and media server in my case
[01:45] <thedoble> cool
[01:45] <thedoble> yeah i am mostly file/print/domain/email stuff
[01:46] <thedoble> small businesses
[01:46] <lumpEee> i do podcast and internet radio
[01:47] <lumpEee> i do all indie artists
[01:48] <lumpEee> i just need to create a server so that I am not the only one able to enter data to the database and add media to the streaming server
[01:48] <lumpEee> i know it can be done
[01:48] <lumpEee> i believe i am tech competent enough to do it
[01:48] <lumpEee> just not sure exactly how and gotta shake some rust out of me
[01:49] <thedoble> nice one
[01:49] <thedoble> sounds like a good project
[01:49] <thedoble> good luck :)
[01:49] <thedoble> have you read through the ubuntu official documentation ?
[01:49] <thedoble> some of it is brilliant stuff
[01:50] <lumpEee> i am doing that as i go back and forth
[01:50] <lumpEee> i have actually installed server 2x atm
[01:51] <lumpEee> each time i do so, i read more, figure out a bit more and decide the do-over is better
[01:51] <lumpEee> things have improved much since i played much with anything other than a CMS about 10 years ago
[01:53] <thedoble> hehe
[01:53] <lumpEee> and, again, that was all CPanel
[01:53] <lumpEee> i did not actually have to install everything
[01:55] <lumpEee> the docs are good though
[01:55] <lumpEee> i am just going through them with a highligher in oo right now
[02:32] <tonyyarusso> Hi, I was just looking over the server seeds proposals for Lucid.  Two relatively minor questions regarding that:
[02:33] <tonyyarusso> 1)  What's the rationale for dropping vlock?  It's small, but has come in handy for me a few times.  Is there a better replacement these days, or would it be dropped without a replacement in main?
[02:33] <lumpEee> gnight all
[02:36] <tonyyarusso> 2)  Has there been any consideration given to possibly putting an IRCd in main, either for Lucid or later?
[02:39] <twb> tonyyarusso: !anyone :P
[02:44] <ScottK> tonyyarusso: You'll probably have more luck during when mathiaz`or ttx are around.
[02:51] <tonyyarusso> ScottK: all righty
[02:51]  * tonyyarusso idles
[02:53] <marks256> Say if i had a lab of 50 ubuntu computers, could i have an update server so i dont have to download 50 400mb updates (example size). Then i would only have to download 1 400mb set, then distribute it to the lab
[02:53] <twb> marks256: yes.
[02:54] <marks256> twb, how would it work?
[02:54] <twb> It'd be even easier to simply use LTSP, so that you maintain one image and all the lab machines are updated automatically by rebooting
[02:54] <marks256> yeah. LTSP would be nice. but i'm just asking for proof-of-concept
[02:55] <twb> But if you just want a local cache of package updates, I highly recommend debmirror.  If you're very tight on bandwidth, you could try apt-cacher[-ng] or apt-proxy, but both have caused me huge problems in the past.
[02:55] <marks256> twb, aah ok. repos dont mind keeping dumps of them though?
[02:55] <twb> In particular, one of them (apt-cacher?) is vulnerable to injection attacks by any client, which -- although being caught by checksumming -- might be overridden by an ignorant user.
[02:56] <twb> marks256: debmirror magically maintains a partial copy of the Ubuntu archive.
[02:57] <twb> e.g. you can tell it "keep a copy of hardy and hardy-security for i386 and amd64, but don't mirror the "games" section or the "universe" category."
[02:57] <marks256> twb, neat!
[02:58] <twb> IIRC a single-arch, single-release mirror is on the order of 10 to 30GB.  The updates each week are negligible (for stable releases) or significant (for ubuntu+1, testing and sid)
[02:59] <twb> The other thing you could do, which I haven't tried yet, is setting up apt-bittorrent.
[02:59] <marks256> oh my that's not bad
[02:59] <marks256> only 30gb?
[03:00] <twb> 31GB for --arch i386 --nosource --dist hardy,hardy-updates,hardy-security,hardy-backports --section main,restricted,universe,multiverse
[03:01] <twb> Unfortunately I don't keep a record of downloads per week
[03:02] <marks256> wow
[03:02] <marks256> i though it'd be more along the lines of a few hundred gb!
[03:03] <twb> Where that kind of size comes in is if you want to track all the arches that Debian supports
[03:04] <twb> Which according to type-handling, is: cpus [alpha amd64 arm armeb armel avr32 hppa i386 ia64 lpia m32r m68k mips mipsel powerpc ppc64 s390 s390x sh3 sh3eb sh4 sh4eb sparc] × systems [darwin freebsd hurd kfreebsd knetbsd kopensolaris linux netbsd openbsd solaris uclibc uclinux]
[03:09] <tonyyarusso> marks256: I've been keeping local mirrors for a while.  For fine-grained control, ie for just one release, one architecture, I've been very happy with apt-mirror.  I recently decided to do a full mirror, and I'm pulling that down with the ubumirror utility.
[03:09] <tonyyarusso> (Which, btw, I put in my PPA if you'd rather apt-get than bzr branch)
[03:09] <marks256> twb, aah ok ok
[03:10] <marks256> tonyyarusso, ok cool. this sounds quite useful
[03:10] <tonyyarusso> marks256: It's about 250GB for all supported releases, btw.
[03:10] <tonyyarusso> i386 and amd64.
[03:11] <marks256> tonyyarusso, :o oh my
[03:16] <wolfrein> hi everybody
[03:16] <wolfrein> i have a problem with one of my ipsec tunnels
[03:17] <wolfrein> it doesnt come up when i do a racconctl fs isakmp followed by fs ipsec
[03:17] <wolfrein> the preshared key is not issued
[03:17] <wolfrein> when i check status, it shows all 0000000 in the preshared key position
[03:17] <wolfrein> please advise how i can rectify this
[03:21] <twb> tonyyarusso: have you used both apt-mirror and debmirror?  Would you care to compare them briefly for me?
[03:21] <tonyyarusso> twb: I haven't tried debmirror, no.
[03:25] <marks256> tonyyarusso, twb, thanks for the hlep guys
[03:27] <wolfrein> please advise about the ipsec tunnel
[03:56] <twb> Why doesn't ntpd like to bind to my IPv6 address? http://hpaste.org/fastcgi/hpaste.fcgi/view?id=13480#a13480
[04:04] <twb> Also, how can I just tell my 8.04 system to not use IPv6 at all?
[04:04] <Clusty_> hey
[04:04] <Clusty_> are there generally any issues when it comes to blooth dongles?
[04:04] <Clusty_> or most chips are really supported
[04:15] <twb> Answer: change /etc/modprobe.d/aliases:alias net-pf-10 ipv6 from "ipv6" to "off".
[05:00] <MBCR> !ops
[05:01] <MBCR> !staff
[05:01] <ScottK> MBCR: What are you doing?
[05:01] <MBCR> calling the ops duh
[05:01] <MBCR> !ops
[05:02] <ScottK> Apparently not.  Why?
[05:08] <Ninjix> hello all
[05:11] <Sorell> hi Ninjix
[05:15] <Ninjix> greetings Sorell
[05:18] <Ninjix> I'm looking to discuss Intel ICH10R "fakeraid" with anyone that has any production experience with them
[05:20] <twb> !anyone
[05:25] <Ninjix> twb: Hi there
[05:35] <jmarsden> Ninjix: https://help.ubuntu.com/community/FakeRaidHowto
[05:36] <jmarsden> Basically, it doesn't make sense to use fakeraid on a server, software RAID is more flexible and performs as well.
[05:37] <Ninjix> jmarsden: Yes, I have always believed that but so many servers come with fakeraid now
[05:37] <Ninjix> jmarsden: and the tech has been around a few years now
[05:38] <jmarsden> Preconfigured Ubuntu servers that use fakeraid?  Interesting.  Just because the motherboard chip can do it doesn't make it a good choice for a Linux/Ubuntu server.
[05:40] <Ninjix> True
[05:41] <jmarsden> What would you hope to gain from using fakeraid -- how would it benefit your users or your sysadmins or your profit margin, compared to using software RAID ?
[05:41] <Ninjix> I having good success with Adaptec 2258100-R cards but would rather not have to shell out extra $400
[05:42] <Ninjix> I trying to move as much of our data center to commodity hardware as possible
[05:43] <jmarsden> I don't think you'lld get hardware RAID performance from fakeraid; I think youu'd fine software RAID performance about equal to fakeraid... so then the chocie is between software RAID or hardware RAID... right?
[05:43] <Ninjix> Disk IO is a tough obstacle, though
[05:43] <jmarsden> Interesting... a lot of people are moving the other way -- a few big servers and virtualization, rather than many cheap servers.
[05:43] <Ninjix> ah.. don't get me wrong. I'm moving down that path as well
[05:45] <jmarsden> At the same time?  So... a few really big fast but cheap commodity servers?  :)  Where can I find some of those? :)
[05:45] <Ninjix> this year we are moving off VMware and onto Ubuntu KVM virtualization
[05:48] <Ninjix> after spending tons of money on Dell and HP "enterprise" class hardware for years now, I sat back and ask my team how much of those enterprise features we were really using
[05:49] <Ninjix> the answer turned out to be very little
[05:50] <Ninjix> this made us rethink where we spend our money, especially with the economic condition of the last year
[05:51] <jmarsden> Sure... but you've had hard drives fail and may even have hot-swapped them when that happened... that's a different level of "enterprise-class" than the "lights out management" and similar things, I'd think.
[05:51] <Ninjix> agreed
[05:52] <freeflying> ha all
[05:52] <freeflying> I can ping other machine from an instance in UEC, but can't access to the instance from other machine besides from cloud clontroller
[05:53] <Ninjix> and now we're returned to rolling our own white-box servers that have redundancy where we use it the most like hot swap bays
[05:53] <jmarsden> Ninjix: So maybe you really want to move towards commodity server hardware, but still hardware RAID cards and hot swap disk chassis... yup.  Makes sense.
[05:55] <jmarsden> Maybe you can build two such boxes, one software RAID and one hardware RAID, and benchmark for your workload and then decide if the $400 is worth it for your situation?
[05:56] <Ninjix> so far I've had mix results in our tests
[05:57] <Ninjix> the fakeraid can hang with the hardware RAID on 0,1 and 10
[05:59] <Ninjix> moving up to RAID 5 there's where the parity hurts
[05:59] <jmarsden> I guess that makes sense.  Did you compare fakeraid with pure software (mdadm) RAID too?
[06:00] <Ninjix> same hardware with Windows drivers produces better results by some extra % points
[06:01] <jmarsden> But for the cost of Windows Server you can buy a hardware RAID card :)
[06:01] <Ninjix> I haven't tested the ICH10R against Linux RAID yet
[06:01] <Ninjix> Good point. Love FOSS cost savings. :)
[06:03] <Ninjix> freeflying: I wish I could help but I still waiting to get the chance to play with UEC
[06:03] <jmarsden> Ninjix: I'd think hard about doing RAID5 fakeraid in production... if things go bad and you end up needing to rescue a broken array, you're more likely to find people and tools for messing with the guts of an mdadm setup than with a (Linux) fakeraid-based array, I suspect.
[06:05] <jmarsden> So unless fakeraid has significant reliability or performance benefits (which I don't think it does), I'd use mdadm software RAID rather than fakeraid.  But that's just me.
[06:05] <Ninjix> yes. That seems to be the general consensus. I know guys that won't use hardware raid for that reason.
[06:06] <Ninjix> No one want to get burned by proprietary drivers or even hardware
[06:06] <jmarsden> Yes... although real hardware raid on a controller from a decent vendor is slightly different; the vendor can often help you rescue things, and since the RAID is hardware they don't need to know Linux to do so... only their own hardware.
[06:07] <Ninjix> thankfully I have only come close to that with Dell PERC controllers
[06:09] <jmarsden> We've seen some Windows servers where older Adaptec SCSI RAID5 broke badly, and their techs can do what looks near-miracles to get things going again... I think twice in the last 4 years I have seen that.
[06:10] <Ninjix> yes, Adaptec hasn't ever let me down. Still have a old 2940 SCSI card around here somewhere. :)
[06:10] <jmarsden> That's pretty old now, I remember those :)
[06:11] <Ninjix> how about yourself? what kind of disk IO solution are you running your Ubuntu servers on?
[06:15] <jmarsden> Software RAID1; we're doing smaller lower cost servers mostly where the cost of a hardware RAID card doesn't seem to be justified.  For a couple of clients that did need more serious IO we did SAS drives and an LSI MegaRAID card, I don't remember the exact model.
[06:22] <Ninjix> 15k SAS drives?
[06:27] <fallous_> I was always a symbios fan myself.  875-based cards were the bomb
[06:27] <jmarsden> Yes, Seagate.  I'm really more of a network/sysadmin, I only get involved with low level hardware details when I have no choice :)
[06:33] <Ninjix> nice chatting with you this evening
[06:34] <jmarsden> Likewise.  Sometimes it is nice *not* to be helping out teenage newcomers with minimal Linux background :)
[06:47] <billybigrigger> hey im 24
[06:47] <billybigrigger> thank you very much
[06:47] <billybigrigger> :P
[06:48] <jmarsden> :)  Hi billybigrigger
[06:48] <billybigrigger> jmarsden, howdy
[07:34] <tonyyarusso> ttx: Hey, I was wondering if you could comment on either of these things re: server seed proposals for Lucid:
[07:35] <tonyyarusso> 1)  What's the rationale for dropping vlock?  It's small, but has come in handy for me a few times.  Is there a better replacement these days, or would it be dropped without a replacement in main?
[07:35] <tonyyarusso> 2)  Has there been any consideration given to possibly putting an IRCd in main, either for Lucid or later?
[07:36] <ttx> 1) I don't know, feel free to ask the question in the discussion section of the spec
[07:36] <ttx> 2) I guess yes
[08:13] <Maleko> how do you view parition details, aside from using fdisk -l
[08:15] <jmarsden> Maleko: YOu could use cfdisk if you prefer :)  What is wrong with using fdisk -l ?
[08:16] <Maleko> it does not give in-depth details such as partition filesystem, total size in gb, free space etc
[08:17] <jmarsden> How could it... it is just reading the partition table, which does not contain that information.
[08:18] <jmarsden> If you know what fs is on a given partition you can use fs-specific tools to look at that fs...
[08:20] <Maleko> hmm. what about ext* ?
[08:22] <jmarsden> tune2fs -l   would work on ext2/ext3 filesystems, lots of info there, for example...
[09:27] <_ruben> looks kinda "ironic" to me: process `sysctl' is using deprecated sysctl (syscall) net.ipv6.neigh.default.retrans_time; Use net.ipv6.neigh.default.retrans_time_ms instead.
[10:55]  * soren grabs some lunch
[11:53] <alvin> Good to see there's another Server survey.
[11:56] <ivoks> when is the meeting?
[12:00] <Aison> hello
[12:00] <Aison> I installed ubuntu server on two machines
[12:01] <Aison> on one machine, /etc/fstab root is monted at /dev/mapper/mediaserv-root
[12:01] <nijaba> ivoks: meeting is at 2PM UTC, 2h from now
[12:01] <Aison> on the other machine, root is defined by an UUID
[12:01] <Aison> why this difference?
[12:01] <ivoks> right in the middle of working hours :/
[12:01] <ivoks> Aison: it's newer version of ubuntnu
[12:03] <Aison> but I installed both from the same ubuntu cd?
[12:03] <ivoks> which one?
[12:03] <ivoks> ubuntu ...?
[12:04] <ivoks> 8.04?
[12:04] <ivoks> 9.10?
[12:04] <Aison> 9.10 amd64
[12:04] <ivoks> so, on one you installed LVM, and on the other you didn't right?
[12:05] <Aison> hmm, I just noticed that also
[12:05] <Aison> ;)
[12:05] <ivoks> hehe
[12:10] <Aison> is there a good howto for ubuntu 9.10 to build my own custom kernel? because I've got 12 DVB-s Cards in this machine, I need to change some constant in the kernel sources
[12:10] <Aison> by default, the kernel support just 8 DVB Cards at the same time
[12:12] <Aison> the wiki here https://help.ubuntu.com/community/Kernel/Compile  is not working somehow
[12:22] <ivoks> there's git
[12:23] <ivoks> https://help.ubuntu.com/community/Kernel/Compile#Build%20the%20kernel%20%28when%20source%20is%20from%20git%20repository,%20or%20from%20apt-get%20source%29
[12:23] <ivoks> this should work
[12:34] <Aison> thx
[12:34] <Aison> :)))
[12:37] <zul> morning
[12:46] <acalvo> how can I grow a LVM partition, if it is mounted as /?
[12:47] <alvin> acalvo: The same way as you grow other logical volumes. It works.
[12:47] <acalvo> but, shouldn't it be unmounted first?
[12:48] <alvin> no, you can resize online
[12:48] <alvin> first, use lvresize, and then resize2fs
[12:48] <acalvo> even with ext4?
[12:48] <alvin> yes, even with ext4
[12:49] <acalvo> thanks alvin, I'll try it
[13:08] <acalvo> is it worth to have vmware tools installed on a production server?
[13:15] <zul> morning
[13:29] <nijaba> morning zul
[13:30] <nijaba> acalvo: if your server is running on ESX or a VMware product, yes, it makes a lot of sense
[13:30] <acalvo> thanks nijaba
[13:30] <acalvo> last question then: are the vmware packages available in any repository?
[13:32] <soren> open-vm-tools.
[13:32] <soren> I don't know if it works. Haven't used it in....
[13:32] <soren> err...
[13:32] <soren> a very long time.
[13:32]  * soren desperately needs coffee
[13:41] <acalvo> soren: quick question: where you from?
[13:44] <soren> acalvo: Denmark
[13:44] <acalvo> ok :)
[13:46] <zul> aka vikingland
[13:47] <soren> The very same.
[13:57]  * zul smacks net-snmp around some
[14:00] <soren> jdstrand: http://libvirt.org/git/?p=libvirt.git;a=commitdiff;h=32f021f2664290cffe34723c52435ac4a62fb365 fixes the eventtest thing.
[14:00] <soren> jdstrand: If you'd rather leave that to me, that's perfectly cool as well.
[14:01] <jdstrand> soren: ok, thanks
[14:01] <soren> ..but there it is.
[14:02] <jdstrand> soren: I ran into some other issues (hal), but am working through them. I'll apply that commit too
[14:25] <cyphermox> I'm trying to setup a full kickstart for a ubuntu server using the kickstart compatibility config format. What would be the way to skip the creation of the initial user?
[14:36] <cipix> hi all. I'm having an issue on a ubuntu server with grub2 installation in that when server boots it gives me the error "You need to load the kernel first" and I tried grub-install and update-grub and I don't know what else to check to see what's wrong and to fix it. can you help me please?
[14:46] <RoyK> u hm... http://www.grist.org/article/2009-11-18-oil-enough-energy-to-melt-glaciers
[14:52] <mathiaz> ttx: byobu isn't brought into main only by Recommends - http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.lucid/rdepends/byobu/byobu
[14:52] <mathiaz> ttx: it's a dependency from a bunch of packages
[14:53] <ttx> mathiaz: ok, good.
[14:54] <mathiaz> ttx: what I'm grepping for is packages that  only have Reverse Recommends and that they're brought in by one of the server seed (ultimately)
[14:54] <mathiaz> ttx: the goal was to have a (rather) short list to review by hand
[14:54] <mathiaz> ttx: and make proposal from it
[15:00] <xfrogman5> Are ssl features installed and enabled with a standard Ubuntu LAMP installation?
[15:11] <nijaba> xfrogman5: yes they are installed but not enabled
[15:12] <xfrogman5> Can you point me to a detailed howto for enabling ssl within apache2
[15:12] <nijaba> xfrogman5: https://help.ubuntu.com/9.10/serverguide/C/httpd.html
[15:13] <xfrogman5> thx
[15:13] <nijaba> xfrogman5: see ubuntu.com/server/doc for a summary of resources available about ubuntu server edition
[15:22] <heath|work> can some pastie there source.list for 7.10, I'm getting nothing but 404's
[15:23] <heath|work> or did support end and I not know it
[15:25] <mjeanson> heath: http://old-releases.ubuntu.com
[15:29] <soren> heath|work: Support ended more than 6 months ago. You should upgrade ASAP.
[15:29] <heath|work> we are moving servers, I just needed to install git for a test thanks soren
[15:31]  * MenZa makes mental note to upgrade his 8.10 box the moment 10.04 is out
[15:31] <soren> MenZa: 8.10->10.04 is not a supported upgrade path, fwiw.
[15:31] <MenZa> aye, I know
[15:31] <soren> It may Just Work, but we put no effort into testing anything at all.
[15:32] <MenZa> I'm scared to do it
[15:32] <MenZa> I might just do a full backup and install 10.04
[15:32] <soren> Bah. It's only bits and bytes.
[15:32] <soren> :)
[15:32] <MenZa> I regret not installing Hardy back when I set my box up
[15:32]  * soren is still running hardy on most of his production systems.
[15:32] <MenZa> I ought to on my server
[15:32] <MenZa> I never considered using it for anything else than irssi, but I'm now running multiple websites and stuff on it
[15:34] <soren> I suppose my primary work laptop is sort of a production system (at least in the sense that I'd be severely impaired if it broke), but it's running Lucid.
[15:36] <MenZa> soren: So, you mean Jaunty with a new toolchain? and /etc/apt/sources.list?
[15:36] <MenZa> er
[15:36] <MenZa> s/Jaunty/Karmic
[15:36] <MenZa> I can't keep up with releases anymore
[15:36]  * MenZa has lived to see Breezy, Dapper, Edgy, Feisty, Gutsy, Hardy, Intrepid, Jaunty, Karmic and Lucid
[15:36] <MenZa> Christ, that is a lot of releases.
[15:36]  * MenZa feels old.
[15:40] <soren> MenZa: No, Lucid is waaay past just being an updated toolchain.
[15:40] <MenZa> soren: I suppose - Alpha 1 is out soon
[15:41] <smoser> mathiaz, or anyone, have you done command line compltion before (ie, added bash completion hooks for a package) ?
[15:48]  * soren goes to pick up his daughter at day care
[15:48] <TeTeT_> how to exchange the self signed certificate for UEC with a CA signed?
[15:50] <Rascal999> have setup up rsa/dsa so i don't need password for ssh but i get this error pop up in auth.log Error attempting to add filename encryption key to user session keyring; rc = [1]
[15:56] <MTecknology> If my server is running on xen and I don't want xen to control the kernel; what kernel can I use that already has xen support built in?
[15:56] <MTecknology> I tried -generic and -virt but neither seem to have xen support
[16:04] <smoser> MTecknology, it depends on what version of xen you have. ... newer xen can boot either of the above kernels (i think) in pv_ops
[16:05] <smoser> additionally there is the xen hvm, which i htink you mean by your statement of "don't want xen to control the kernel"
[16:11] <MTecknology> smoser: I'm running on a Linode if that helps
[16:12] <smoser> doesnt help me, sorry.
[16:18] <MTecknology> smoser: I'm getting this error   ERROR Invalid kernel: elf_xen_note_check: ERROR: Will only load images built for the generic loader
[16:18] <MTecknology> or Linux images
[16:18] <MTecknology> xc_dom_parse_image returned -1
[16:19] <smoser> MTecknology, i'm sorry, i can't be much more help. sorry to get your hopes up.  maybe someone else here can help.
[16:21] <MTecknology> ok, thanks
[16:33]  * ttx hates C
[16:37] <Rascal999> ttx, why do you hate C?
[16:38] <ttx> Rascal999: struggling with some memory management, I forgot how painful it can be.
[16:38] <Rascal999> ttx, oh, not been there yet
[16:39] <Rascal999> based on what I've learnt and done in C so far, I'm liking it
[16:39] <Rascal999> but I haven't really touched on memory management, multi-threading or the rest
[16:41] <kirkland> ttx: where is your eucalyptus SRU branch?
[16:41] <kirkland> ttx: last i recall, it was a private branch owned by you
[16:41] <kirkland> ttx: could you push it somewhere ubuntu-core-dev owned?
[16:42] <ttx> kirkland: done already
[16:42] <kirkland> ttx: sweet, link me :-)
[16:42] <ttx> kirkland: http://bazaar.launchpad.net/~ubuntu-core-dev/eucalyptus/ubuntu-karmic/changes
[16:42] <ttx> committed as rev725..726
[16:42] <kirkland> ttx: thanks; i'll hack the new CLEAN=1 operations today
[16:43] <kirkland> ttx: fixing the non-determinism seems, um, a bit harder
[16:43] <ttx> kirkland: I didn't observe any non-determinism in my testing
[16:44] <ttx> kirkland: doesn't mean there isn't any, by definition
[16:44] <kirkland> ttx: https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/490382/comments/3
[16:44] <ttx> i.e. stop eucalyptus-cc CLEAN=1 always PASSed
[16:44] <kirkland> ttx: you wrote "Looks quite non-deterministic from here."
[16:44] <ttx> arrh
[16:45] <kirkland> ttx: typo?
[16:45] <ttx> yep
[16:45] <ttx> commenting
[16:45] <kirkland> ttx: ah, would you add another comment to that effect?
[16:45] <ttx> i progress
[16:45] <kirkland> ttx: i was happy you had seen something similar to me :-)
[16:46] <ttx> kirkland: https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/490382/comments/5
[16:46] <ttx> sorry about that
[16:46] <kirkland> ttx: no problem
[16:46] <kirkland> ttx: so i'll add the new CLEAN=1 support, test that a bit, and push to proposed?
[16:46] <ttx> yes, sounds good
[16:47] <ttx> If you still observe some non-determinism, let me know which commands you tested
[16:49] <ttx> I couldn't make it fail with "stop eucalyptus-cc CLEAN=1"
[16:49] <ttx> but that's rather counter-intuitive
[16:53]  * soren goes to dinner... bbl
[16:57] <kirkland> ttx: question about https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus/+bug/490382/comments/3
[16:57] <kirkland> ttx: your comment about the env not being passed from eucalyptus to eucalyptus-cc upstart script ...
[16:57] <kirkland> ttx:  is that supposition, or confirmed operation?
[16:58] <kirkland> ttx: ie, do i need to chase keybuk down and ask him about that ?
[16:58] <ttx> kirkland: supposition.
[16:59] <kirkland> ttx: okay, i'll confirm with scott
[16:59] <ttx> kirkland: I see no reason why it would be propagated to upstart tasks that get triggered by another event
[17:00] <ttx> kirkland: but sometimes I'm surprised by upstart :)
[17:03] <xfrogman5> Looking for detailed howto for converting an IIS6 ssl cert to be used on a Ubuntu Apache2 Openssl system.
[18:24] <Aison> is there some ldap howto (samba, pam, radius, etc...)   for 9.10?
[18:24] <Aison> I tried https://help.ubuntu.com/community/OpenLDAPServer
[18:24] <Aison> but eg. on 9.10 /etc/ldap/ldaps.conf is a directory
[18:25] <Aison> err slapd.conf
[18:42] <ahasenack> Aison: look for the server guide
[18:43] <Aison> this? https://help.ubuntu.com/9.10/serverguide/C/index.html
[18:43] <Aison> :D
[18:44] <ahasenack> I was looking for it, let me see
[18:45] <ahasenack> ah, help.ubuntu.com
[18:45] <ahasenack> Aison: yes, that's the one
[18:45] <ahasenack> Aison: https://help.ubuntu.com/9.10/serverguide/C/openldap-server.html
[18:51] <Aison> with sudo dpkg-reconfigure slapd I can't set the default suffix :(  I also can't set the manager password
[18:55] <Aison> brb, reboot, back soon ;)
[19:07] <tuxcrafter> hi all
[19:07] <tuxcrafter> yesterday i updated my server to ubuntu 9.10
[19:07] <tuxcrafter> and i had a mediawiki install
[19:07] <tuxcrafter> and now it is gone
[19:08] <tuxcrafter> it is asking me to setup the wiki first
[19:08] <tuxcrafter> i am so afraid i am lossing all the data
[19:08] <Aison> re
[19:12] <Aison> back to my question ;)  with sudo dpkg-reconfigure slapd I can't set the directory or ldap manager password :(
[19:12] <tuxcrafter> i do got daily mysql dumps of the media wiki
[19:12] <ahasenack> Aison: in karmic, there is no such password anymore in the default tree
[19:13] <ahasenack> Aison: using -H ldapi:// -Y EXTERNAL in your ldap* command-line tools is enough to become the rootdn
[19:13] <adurity> I need to update the ESM firmware on a Dell PowerEdge server.  Does anyone have experience doing this under Ubuntu?
[19:13] <ahasenack> Aison: there is a regexp in the config mapping that form of local authentication to the directory administrator user
[19:13] <Aison> k
[19:13] <ahasenack> Aison: try "ldapwhoami -H ldapi:/// -Y EXTERNAL"
[19:14] <ahasenack> Aison: it will tell you that you are cn=localroot,cn=config IIRC, which is the rootdn
[19:14] <Aison> ok, yes
[19:14] <Aison> eg. there's a sample ldapsearch call to view the tree
[19:14] <Aison> ldapsearch -xLLL -b cn=config -D cn=admin,cn=config -W olcDatabase={1}hdb
[19:15] <Aison> there a password is requested....
[19:22] <ahasenack> because you added -W
[19:23] <ahasenack> is that in the doc?
[19:23] <Aison> yes
[19:23] <ahasenack> ugh, then it's wrong
[19:23] <Aison> it's copy&past from https://help.ubuntu.com/9.10/serverguide/C/openldap-server.html
[19:23] <ahasenack> replace all authentication options with "-H ldapi:/// -Y EXTERNAL"
[19:23] <ahasenack> and the commands should work locally, when run on the same server
[19:23] <ahasenack> so, drop -x -D -W -w
[19:23] <Aison> the doc also tells me that "sudo dpkg-reconfigure slapd" let me configure the default suffix
[19:24] <Aison> but I cant
[19:24] <ahasenack> note it's a bare bones tree, so you will eventually want to add something to it, like users and so, so at some point you will be using -D, -W, etc
[19:24] <ahasenack> yeah, looks like it's outdated then
[19:24] <Aison> well, it also tells me, that the suffix is taken from the FQDN
[19:25] <Aison> but I don't remember that i've ever set the FQDN at installation
[19:25] <ahasenack> there is no suffix yet
[19:26] <ahasenack> it's from the older distros documentation
[19:26] <ahasenack> in 9.10 it's really bare bones, just a bare cn=config database
[19:26] <ahasenack> if you want a script to populate it, you can check out launchpad.net/openldap-dit
[19:26] <ahasenack> although it highlighted a bug in openldap, let me commit a workaround
[19:28] <ahasenack> hmm
[19:30] <Aison> on the old distribution I used mysql for users and such, but now I changed to ubuntu (also changed my RAID device, raised raid capacity to 8TB, etc....)
[19:31] <Aison> and I would like to use ldap ;) so far, everything works nice
[19:31] <Aison> bind setup was easy, etc..
[19:55] <mathiaz> Aison: the ldap section of the ubuntu server guide in 9.10 hasn't been updated to match the new slapd configuration
[19:55] <mathiaz> Aison: most of the ldif files are good, but the command won't work, and you'd have to create a default database/tree as well.
[19:56] <ahasenack> mathiaz: how does its updating work? Is it a wiki, does a diff need to be pasted in some LP ticket, or just a discussion in the mailing list?
[19:57] <mathiaz> ahasenack: the server guide is actually a bzr branch and a package
[19:57] <ahasenack> hmm
[19:57] <mathiaz> ahasenack: so the update uses the standard SRU process
[19:57] <ahasenack> mathiaz: what's the branch?
[19:57] <mathiaz> ahasenack: https://wiki.ubuntu.com/ServerTeam/KnowledgeBase#Documentor%20resources
[19:57] <mathiaz> sommer: lp:ubuntu-docs
[19:57] <mathiaz> sommer: ?
[19:57] <mathiaz> ahasenack: ^^
[19:58] <Aison> mathiaz, too bad :)  i'm new to ldap, so it's not that easy for me ;)
[19:58] <ivoks> it's quite easy once you understand it
[19:58] <ivoks> ....once...
[19:58] <mathiaz> ahasenack: I meant to write a blog post about using slapd in Karmic
[19:58] <ahasenack> cof cof
[19:58] <mathiaz> ivoks: a couple of days ago you mentioned that we switch to MariaDB - why?
[19:59] <ivoks> mathiaz: they listen
[19:59] <ivoks> mathiaz: we had a bug in mysql for ages
[19:59] <ivoks> reported a year ago
[19:59] <ivoks> until uds and meeting with mneptok, nothing moved
[19:59] <RoAk> ivoks, should corosync be merged from debian/testing or should we just work on the packages provided by madkiss?
[19:59] <ivoks> then in two days all was resolved :)
[20:00] <ivoks> RoAk: i haven't looked at merges yet; i plan to do that this night
[20:00] <ivoks> mathiaz: they have newer stuff
[20:00] <mathiaz> ivoks: well - there are plenty of bugs - I'd rather not base such a decision because 1 bug was fixed promptly
[20:00] <ahasenack> mariadb is a mysql fork or what?
[20:01] <mathiaz> ahasenack: yes
[20:01] <RoAk> ivoks, ok I was planing on grab the packages of Madkiss and recompile them for ubuntu. He has also made available packages for karmic in his repo
[20:01] <ivoks> mathiaz: of course, but as i said, they have newer featuers
[20:01] <ivoks> features
[20:02] <ivoks> once 5.1 is released, i'll do some testing and if everything goes as advertised, i'll replace mysql with mariadb on my servers
[20:02] <mathiaz> ivoks: right - the maria db engine, a few other patches?
[20:02] <mathiaz> ivoks: I think that's perfectly reasonalbe
[20:02] <ivoks> mathiaz: it's mysql 5.6 + patches
[20:02] <mathiaz> ivoks: 5.6? IIRC they're merging from 5.1
[20:02] <ivoks> so mariadb 5.1 is at the same level as mysql 5.6
[20:03] <mathiaz> ivoks: my understanding was that MariaDB 5.1 is 5.1 + patches
[20:03] <ivoks> true
[20:03] <ivoks> but those patches come from 5.6
[20:04] <ivoks> not all, of course
[20:04] <ivoks> anyway, once it is released, i'll test it
[20:04] <mathiaz> ivoks: I think we should make it as easy as possible for people to experiment with mariadb
[20:04] <ivoks> and, if you are interested, i could provide some analysis
[20:04] <ahasenack> mathiaz: I branched that url, but I don't see a file related to openldap in it
[20:05] <mathiaz> sommer: ^^ - could you ahasenack with the server guide? I haven't looked at it lately
[20:05] <mathiaz> ivoks: I'm always interested in such analysis
[20:05] <mathiaz> ivoks: the main issue for now is that haven't released anything
[20:05] <ahasenack> hmm, found it
[20:05] <ahasenack> network-auth.xml
[20:05] <ivoks> mathiaz: correct
[20:05] <mathiaz> ivoks: which is a bit annoying for the LTS
[20:06] <ivoks> of course
[20:06] <mathiaz> ivoks: there was a discussion about MariaDB during the last UDS - were in there?
[20:06] <ivoks> yes it was
[20:06] <ivoks> i was
[20:06] <mathiaz> ivoks: ok - I think that what mark suggested in the end is the best option for LTS
[20:07] <ivoks> imho, having mariadb at least in universe would be very smart thing to do
[20:07]  * mathiaz agrees
[20:07] <ahasenack> the branch is already more up-to-date than the published page
[20:07] <ivoks> who knows what could happen to mysql and having a fail over would be awesome
[20:07] <ivoks> in universe, but very well maintained
[20:08] <mathiaz> ahasenack: bug 463684
[20:14] <enquora> I have an Asus EEE box with 8.04lts on it that I want to upgrade to 9.04 server. do-release-upgrade complains that no upgrade is available and booting from external CD fails when installer can't locate driver for external drive. any suggestions?
[20:14] <enquora> Sorry, want to upgrade to 9.10 server
[20:15] <ivoks> first of all
[20:15] <ivoks> you can't do 8.04 -> 9.04
[20:15] <ivoks> you have to do 8.04 -> 8.10 -> 9.04 -> 9.10
[20:15] <enquora> I know that
[20:15] <ivoks> or wait for 8.04 -> 10.04
[20:16] <ivoks> now, by default, only upgrades from LTS to LTS are enabled in LTS releases
[20:16] <enquora> I assumed that when I saw 'no upgrade available'
[20:16] <enquora> Is there a way to over-ride that?
[20:16] <ivoks> if you want to update to non-LTS, you have to edit /etc/update-manager/release-upgrades
[20:16] <ivoks> Prompt=normal
[20:18] <enquora> ivoks: I see the setting. Thks.
[20:19] <ivoks> np
[20:24] <Aison> uhm, what is recommended as ldap web frontend
[20:25] <ivoks> as in management?
[20:26] <Aison> something like phpldapadmin
[20:26] <Aison> I just tried this, but no chance ;)
[20:26] <ivoks> my preference is apache directory studio
[20:27] <ivoks> it's not web based
[20:27] <Aison> well, if it's not webbased, that's ok als
[20:27] <Aison> also
[20:27] <ivoks> then go for it
[20:27] <ivoks> it rocks.
[20:28] <ivoks> http://directory.apache.org/studio/
[20:28] <mneptok> ivoks / mathiaz: FYI, i'll be talking to Norbert Tretkowski (Debian MySQL packager) about MariaDB packages for Debian. so hopefully Lucid can inherit universe packages directly from Debian.
[20:29]  * mneptok and nobse have a call scheduled for this weekend
[20:29] <Aison> thx
[20:29] <Aison> but first I need somehow to setup my ldap
[20:29] <Aison> that crap is not working ;)
[20:29] <mathiaz> mneptok: great - that's the best option IMO
[20:29] <ivoks> learn it
[20:29] <Aison> the docs of 9.10 aren't really up to date
[20:29] <Aison> :(
[20:30] <mathiaz> Aison: true - any help in updating them is welcome
[20:30] <ahasenack> Aison: would you like to try the openldap-dit script?
[20:30] <Aison> ahasenack, well, if that helps, why not ;)
[20:30] <mneptok> mathiaz: also, i pointed bytee (Colin Charles) in your direction as regards AppArmor profiling for MariaDB. if he has questions, he might poke you.
[20:30] <ivoks> 'what is dit?'
[20:30] <ivoks> :D
[20:30] <Aison> ahasenack, since my ldap is empty anyway, I can't damage much
[20:30] <ahasenack> Aison: it's supposed to be run right after installing the karmic slapd package
[20:31] <Aison> ok, so I reset my small changes, sec...
[20:31] <ivoks> hm, i could give that a try :)
[20:31] <ivoks> and probably provide patches for email services :D
[20:31] <ahasenack> Aison: people.canonical.com/~andreas/openldap-dit-0.20.tar.gz
[20:32] <Aison> ok, and now?
[20:32] <ahasenack> Aison: after opening the tarball, run "sudo make install" and then as root "/usr/share/slapd/openldap-dit-setup.sh" and answer the two questions
[20:32] <ahasenack> (I never tried "sudo /usr/share/slapd/openldap-dit-setup.sh" now that I think about it, but should work)
[20:32] <ivoks> mneptok: there's a new patch for croatian collation in mysql 6.0
[20:33] <ivoks> mneptok: it should be merged
[20:34] <ivoks> mneptok: i'll file a bug or reopen the old one
[20:35] <Aison> ahasenack, ok, finished
[20:35] <Aison> no error so far ;)
[20:35] <mneptok> ivoks: perfect
[20:35] <ahasenack> Aison: I need to leave for about 30min, but will be back
[20:35] <Aison> ok, cu
[20:35] <ahasenack> Aison: now just use that admin dn as the root one
[20:35] <mneptok> ivoks: bump the old bug. if there's no reaction within 24h, i'll start screaming.
[20:36] <Aison> ok
[20:36] <ahasenack> Aison: it installed a basic tree with people and group locations
[20:36] <ahasenack> bbl
[20:36] <ivoks> mneptok: :)
[20:37] <kirkland> mathiaz: ping
[20:37] <kirkland> mathiaz: re: uec-testing
[20:37] <mneptok> ivoks: some people in the US have learned that solving problems in the Balkans before they grow too large is a good idea.  ;)
[20:37]  * mneptok giggles
[20:37] <ivoks> hahaha
[20:37] <kirkland> mathiaz: " * Eucalyptus to make their stress test suite publicly available: TODO"
[20:37] <MTecknology> You guys have any experience with uptrack?
[20:38] <kirkland> mathiaz: can we mark that DONE, now that we have the tests from nurmi?  or is that not "public" enough?
[20:38]  * mathiaz o^14 kirkland 
[20:38] <MTecknology> !info python-yaml
[20:38]  * kirkland tries to figure that one out ...
[20:38] <mathiaz> kirkland: hm - for the scope of uec-testing I think so
[20:38] <mathiaz> kirkland: making them public means putting a license on it
[20:38] <mathiaz> kirkland: so that they could be included in the packages
[20:39] <kirkland> mathiaz:
[20:39] <kirkland>  * Obtain the Eucalyptus test suite: DONE
[20:39] <kirkland>  * Eucalyptus to make their stress test suite publicly available: TODO
[20:39] <mathiaz> kirkland: seems like a good plan
[20:40] <kirkland> mathiaz: okay, can we take a quick look at https://blueprints.edge.launchpad.net/ubuntu/+spec/server-lucid-uec-testing together?
[20:40] <ivoks> mneptok: such a small area and so much problems :)
[20:40] <mathiaz> kirkland: sure
[20:40] <kirkland> mathiaz: i think you went over the work-items and split a few up
[20:40] <mathiaz> kirkland: yes
[20:40] <kirkland> mathiaz: mdz asked me to go over these and carve up any that are more than a day or two worth of effort
[20:41] <mathiaz> kirkland: right - jos pinged me earlier as well
[20:41] <mathiaz> kirkland: 10:47 <jib> i think the main thing missing is the plan for the eucalyptus test integration. could you please amend that?
[20:41] <kirkland> mathiaz: i'm looking at the tests from nurmi right now
[20:41] <mathiaz> kirkland: right - these are stress testing
[20:41] <kirkland> mathiaz: right ... so these would run on one machine (through testbox), and point at an existing cloud setup
[20:42] <kirkland> mathiaz: getting the cloud setup itself is outside of the scope of checkbox, AFAICT
[20:42] <MTecknology> !search jps
[20:42] <mathiaz> kirkland: yes
[20:42] <mathiaz> kirkland: and I wouldn't integrate with checkbox yet
[20:42] <mathiaz> kirkland: that was my main update to the WI
[20:42] <kirkland> mathiaz: cool
[20:43] <mathiaz> kirkland: checkbox doesn't have the concept of running tests spanning multiple systems
[20:43] <kirkland> mathiaz: right
[20:43] <kirkland> mathiaz: okay so looking at the work items for those that might take >2 days ...
[20:43] <kirkland>  * Automate the installation of UEC for the different topologies on the assigned hardware: TODO
[20:43] <kirkland> mathiaz: ^ could be complex
[20:43] <mathiaz> kirkland: well - I was thinking about preseeding
[20:44] <kirkland> mathiaz: but that really just depends on how many and how complex the topologies might be
[20:44] <mathiaz> kirkland: and may be throw some puppet in there
[20:44] <mathiaz> kirkland: right - I have to loook into that
[20:44] <kirkland> mathiaz: sure ...  if we're talking about 3-4 reasonable topologies, i think that's 2 days worth of work
[20:44] <mathiaz> kirkland: but maximum 6 physical machines?
[20:44] <mathiaz> kirkland: oh right
[20:44] <mathiaz> kirkland: I see what you mean
[20:45] <mathiaz> kirkland: let me look up the different topologies
[20:45] <kirkland>  * Automate the installation of UEC for the 3-4 different topologies on the assigned hardware: TODO
[20:45] <kirkland> mathiaz: i scoped it like that ^
[20:45] <mathiaz> kirkland: well - I'd split in one WI per topologie
[20:46] <kirkland> mathiaz: right!
[20:46] <kirkland> mathiaz: do you want to agree on a few topologies right now?
[20:46] <kirkland> mathiaz: i think that's one of the TODO items :-)
[20:46] <kirkland> mathiaz: i'm not sure who needs to sign off on that .... though
[20:46] <mathiaz> kirkland: yes - ttx already defined them actually
[20:46] <mathiaz> kirkland: mdz told me so
[20:46] <kirkland> mathiaz: oh?  where?
[20:46] <mathiaz> kirkland: I'm looking for the spec right now
[20:47] <mathiaz> kirkland: the intstaller improvmeent
[20:47] <kirkland> mathiaz: url?
[20:47] <kirkland> mathiaz:  * Define a comprehensive, finite list of topologies to be tested: TODO
[20:47] <kirkland> mathiaz: you can mark that DONE, then ;-)
[20:47] <mathiaz> kirkland: :) - that's the one I'm searching for
[20:48] <mathiaz> kirkland: https://wiki.ubuntu.com/FoundationsTeam/UECInstallerEnhancement
[20:48] <mathiaz> kirkland: https://blueprints.launchpad.net/ubuntu/+spec/foundations-lucid-uec-installer-enhancement
[20:48] <kirkland> mathiaz: cool, i'm subscribed now, wasn't before
[20:49] <mathiaz> kirkland: since this is the work to be done in the installer - it makes sense to prepare a test plan for these topolobies
[20:49] <kirkland> mathiaz: ack
[20:50] <mathiaz> kirkland: there is also https://blueprints.launchpad.net/ubuntu/+spec/server-lucid-euca-remote-autoregister
[20:51] <kirkland> mathiaz: that looks like 5 different topologies
[20:51] <kirkland> mathiaz: are those the 5?
[20:51] <mathiaz> kirkland: where do you see 5 topologies?
[20:51] <kirkland> mathiaz: https://blueprints.edge.launchpad.net/ubuntu/+spec/server-lucid-euca-remote-autoregister/
[20:52] <kirkland> mathiaz: grep "("
[20:52] <mathiaz> kirkland: same as https://wiki.ubuntu.com/EucalyptusRemoteAutoRegistration - Test/Demo plan section?
[20:52] <kirkland> mathiaz: hmm, 3 of them are the same, never mind
[20:53] <kirkland> mathiaz: okay, well, there's 5 there too
[20:53] <mathiaz> kirkland: yes
[20:54] <mathiaz> kirkland: so that 5 WI
[20:54] <Aison> ivoks, this apache directory editor is really nice :D
[20:54] <Aison> thx
[20:55] <kirkland> mathiaz: the separate networks ... is that something you can automate?
[20:56] <mathiaz> kirkland: you mean - (CC2+SC2 will be registered manually)?
[20:56] <mathiaz> kirkland: I think what needs to be tested here is manual entry of IP - ie not using avahi
[20:56] <kirkland> mathiaz: http://pastebin.ubuntu.com/333382/
[20:56] <kirkland> mathiaz: how does that look?
[20:56] <kirkland> mathiaz: i split that 1 item up like that
[20:57] <mathiaz> kirkland: looks good to me
[20:57] <kirkland> mathiaz: refresh https://blueprints.edge.launchpad.net/ubuntu/+spec/server-lucid-uec-testing
[20:58] <kirkland> mathiaz:  * Enable Eucalyptus upstream test suite: TODO
[20:58] <kirkland> mathiaz: that looks like the hardest / most ambiguous remaining item
[20:58] <kirkland> mathiaz: otherwise, I think it looks okay now
[20:59] <mathiaz> kirkland: hm - we're not testing "three adjacent networks" and "two remote networks (CC2+SC2 will be registered manually) "?
[20:59] <mathiaz> kirkland: last two test cases from https://wiki.ubuntu.com/EucalyptusRemoteAutoRegistration?
[20:59] <mathiaz> kirkland: upstream test suite - right - I don't know how hard it is
[20:59] <kane_> kirkland, mathiaz if that's any more than 1-2 days of work, we'll want to split that out. and i'm trusting at least one of you's seen the code & test suite in question?
[20:59] <kirkland> mathiaz: that's what i was asking you before ... i don't think that's going to be easy to automate the testing of
[20:59] <mathiaz> kirkland: I just remember dan mentioning they had a test suite
[21:00] <kirkland> kane_: both of us have the tarball, i'm looking at it now
[21:00] <kirkland> kane_: neither of us have run it though
[21:00] <mathiaz> kirkland: I've actually run it
[21:00] <mathiaz> kirkland: it requires some small changes
[21:00]  * kirkland bites his tongue
[21:00] <mathiaz> kirkland: but overall it works
[21:01] <ivoks> Aison: great
[21:01] <kirkland> mathiaz: kane_: okay, in that case, I think everything remaining in that spec is less than 1-2 days work
[21:01] <Aison> ivoks, what's the difference between New Entry and New Context Entry?
[21:01] <mathiaz> kane_: what did you mean exactly with 10:47 <jib> i think the main thing missing is the plan for the eucalyptus test integration. could you please amend that?
[21:02] <ivoks> Aison: you are using it on 9.10? 32 or 64 bit?
[21:02] <kirkland> mathiaz: kane_: Drafting -> Review now
[21:02]  * ahasenack is back
[21:02] <mathiaz> kane_: one of the WI in the blueprint is to actually write up the plan for integration testing
[21:02] <Aison> well, the ldap server runs on 64bit 9.10
[21:03] <mathiaz> kane_: and then another WI is to automate that plan
[21:03] <kane_> mathiaz: it was specifically " * Enable Eucalyptus upstream test suite: TODO"
[21:03] <mathiaz> kane_: ah ok.
[21:03] <ivoks> Aison: i'm asking about apache dir.?
[21:03] <mathiaz> kane_: I'd defer that to kirkland  - it's a packaging WI
[21:03] <kane_> which could be anything from 'add line to shell script' to 'massive refactoring required to work'
[21:04] <mathiaz> kane_: right - may be we should first investigate the feasility of it
[21:04] <kirkland> mathiaz: kane_: I don't think it's going to be possible to enable the tests that I've seen in the package build, as they have to be run against a real live cloud on your network
[21:04] <kane_> mathiaz: that'd be great, or at least document the uncertainty there
[21:04] <Aison> ivoks, gentoo ;) 64bit
[21:04] <mathiaz> kane_: hm - well - if it's possible to do it
[21:04] <kirkland> mathiaz: kane_: that's different than, say, the perl binary testing 64-bit floating arithmetic against itself
[21:05] <kane_> understood
[21:05] <mathiaz> kirkland: right - I have no clue about what dan was refering to as their test suite
[21:05] <mathiaz> kirkland: I thought it was a kind of API unit testing suite
[21:05] <kane_> so, let's figure out what we're dealing with and what it would take for us to " * Enable Eucalyptus upstream test suite: TODO"
[21:05] <kirkland> mathiaz: kane_: in my opinion, that item is either not a packaging one, or should be dropped altogether
[21:05] <mathiaz> kirkland: I don't think he was refering to the stress testing scripts I send you
[21:05] <kirkland> hmm, okay
[21:06] <kirkland> then we need to ask them that tomorrow, kane_
[21:06] <kane_> well, we have a very strong wish to enable upstream tests in our processes, so let's figure out what's needed
[21:06] <mathiaz> kirkland: I thought there was a test suite similar to the mysql/openldap test suite
[21:06] <kane_> kirkland: ok, then let's add that to list if it isn't already on mdz's mail
[21:06] <mathiaz> kirkland: may be we should first talk to upstream about what kind of test suite they have (1 WI)
[21:06] <mathiaz> kirkland: and then enable the one that makes sense in the build (1 WI)
[21:07] <kirkland> mathiaz: agreed; i'm working with nurmi right now
[21:07] <kirkland> mathiaz: i'll ask him about it shortly
[21:07] <kirkland> mathiaz: right now, we're reviewing eucalyptus/debian/patches/*
[21:07] <kane_> mathiaz, kirkland: if you guys can demistify that bullet point for me, i'm happy to have it in review
[21:07] <kirkland> kane_: k
[21:07] <kane_> if that takes a call with Euca, then so be it. earlier is better though.
[21:08] <mathiaz> kirkland: yeah - it may well turned out that I misunderstood what dan was saying
[21:08] <mathiaz> kirkland: kane_: this WI is based on one sentence that dan said when we were in Austin talking about the stress test scripts
[21:09] <mathiaz> kirkland: "we also have a test suite we run against the internal API/components..." something like that - that's what I remember
[21:11] <MTecknology> I want to run Apache on my website but I want it to only run as the httpd user (www-data). It seems that it's default is to run as root...
[21:11] <kane_> mathiaz, kirkland: updated the spec to reflect the above point, just for record keeping purposes
[21:12] <MTecknology> Am I right or wrong on that? If I'm right; how do I change that without breaking anything?
[21:12] <ScottK> kane_: It might be nice if you sent a mail to the ubuntu-server mailing list introducing yourself.
[21:12] <kane_> MTecknology: the parent process runs as root, and then drops privs afaik
[21:12] <kane_> ScottK: good point, thanks
[21:12] <kirkland> kane_: ack, cheers
[21:13] <MTecknology> kane_: /var/www and /var/www/* are owned by root:root
[21:16] <MTecknology> kane_: I guess a fun way to test that would be to make a simple php script that writes to a file that's 700 root:root and 700 www-data:www-data
[21:17] <smoser> jjohansen, so the lucid ec2 kernel is in archive...
[21:17] <ivoks> urgh... hate windows
[21:17] <smoser> but linux-image-ec2 seems to still depend on linux-image-2.6.31-302-ec2,
[21:17] <smoser> so i wont pick up the kernels in nightly builds.
[21:17] <smoser> hello mr erichammond
[21:19] <mathiaz> kirkland: anything else on the /server-lucid-uec-testing spec?
[21:20] <kirkland> mathiaz: i don't think so
[21:20] <kirkland> mathiaz:  i marked it for review
[21:20] <kirkland> mathiaz: i need to get back with you when i understand the different test suites dan and eucalyptus have
[21:21] <mathiaz> kirkland: ok
[21:22] <kirkland> soren: around?
[21:22] <soren> smoser: I've figured out why /dev might be nuked by vmbuilder cleaning up.
[21:22] <kirkland> soren: question about this commit: http://bazaar.launchpad.net/~ubuntu-core-dev/eucalyptus/ubuntu/revision/543#debian/patches/04-axis2c-1.6.0-rampart-1.3.0.patch
[21:22] <soren> kirkland: Oui, oui.
[21:22] <kane_> MTecknology: ps aux|grep apache will show you the children run with www-data privs. apache default installation comes with nothing more than an index.html that says 'it works'
[21:23] <kirkland> soren: i'm curious why a new aclocal.m4 file was added to that patch
[21:23] <MTecknology> kane_: I just tested with php too :P
[21:23] <MTecknology> kane_: thanks
[21:23] <kane_> np
[21:23] <soren> kirkland: I don't see the string "aclocal" anywhere on that page?
[21:24] <soren> kirkland: Ah, sorry. My bad.
[21:24] <kirkland> soren: see debian/patches/05-axis-alternative-repository.patch
[21:24] <smoser> soren, thats good.
[21:24] <soren> kirkland: For PKG_CHECK_MODULES, I think.
[21:24] <smoser> if your box is hosed, you can get it back with udevadm trigger
[21:25] <soren> kirkland: ..but why it's not in /04-axis2c-1.6.0-rampart-1.3.0.patch instead... I don't know.
[21:25] <kirkland> soren: hrm, yeah, i think it might make more sense there
[21:25] <soren> smoser: It's not. I was just looking at the code and realised where it might go wrong.
[21:26]  * soren is about to call it a day
[21:30]  * soren does so
[21:30] <MTecknology> Would there be any issue in changing www-data shell to bash; or is there a chance that could hurt performance?
[21:30] <MTecknology> I only ask because I intend to only modify web data as that user from now on
[21:32] <MTecknology> I have a feeling that'll fix a lot of issues I've been having with Apache; but if it's at the cost of performance - I can get used to dash or w/e it uses -- or perhaps I should ask #httpd
[21:34] <jjohansen> smoser: https://edge.launchpad.net/ubuntu/+source/linux-ec2
[21:34] <jjohansen> smoser: but there is still some meta packaging work to be done so that updates can happen
[21:36] <smoser> jjohansen, hm...
[21:37] <jjohansen> yeah its not ideal, but I thought it best to let you guys know as soon as it hit the archive instead of waiting until the meta packaging was updated
[21:38] <smoser> ah. ok. so it will happen.
[21:38] <smoser> and the karmic dailies will just pick it up and publish it when it appears there.
[21:39] <smoser> jjohansen, we're ramdisk free riht now with the -31
[21:39] <jjohansen> really? nice
[21:39] <jjohansen> Then we should be able to be with -32 too
[21:39] <smoser> yeah, so hope that the -32 still boot :)
[21:40] <smoser> i tested boot to mount root on kvm, uec and ec2 with the 20091201 images.
[21:40] <jjohansen> hehe it better, I did smoke testing of both i386 x86_64 before handing it to andy
[21:40] <jjohansen> and then I did it again with his ppa
[21:40] <smoser> you're the ubuntu-server kernel person, right?
[21:40] <jjohansen> yep
[21:40] <smoser> so this would fall on your lap eventually...
[21:41] <smoser> if we're thinking we want to be ramdisk free as much as possible for virtual systems.
[21:41] <jjohansen> hrmm, I suppose
[21:41] <smoser> i dont know what extra hard ware support would need to be turned to 'y' in the kernel to support microsoft and/or vmware vms without ramdisk
[21:42] <smoser> virtualbox obviously another one. i've not tested that.
[21:42] <jjohansen> ugh, me neither
[21:42] <smoser> i do think its reasonable (given the fairly small amount of hardware) that -virtual would support just about everything
[21:42] <jjohansen> yeah it should
[21:43] <jjohansen> well within reason
[21:43] <jjohansen> there is some strange stuff out there
[21:46] <sbeattie> smoser|jjohansen: if you've got kernels that need testing in virtualbox or vmware esx, I can help out.
[21:48] <jjohansen> sbeattie: thanks, basically we need to step through and test booting sans initramfs
[21:48] <jjohansen> hitting as many platforms as possible
[21:48] <smoser> sbeattie, well... yeah. http://uec-images.ubuntu.com/lucid/current/unpacked/
[21:49] <smoser> you can just grab the lucid-uec-{i386|amd64}-vmlinuz-virtual kernel there and try it.
[21:49] <smoser> you should be able to moutn root filesystems on an existing image.
[21:50] <smoser> the images there are'nt terribly friendly today to environments other than UEC or ec2, but we're working on that.
[22:09] <Aison> re
[22:09] <Aison> why do I have to add organizationalUnit for posixGroup entry?
[22:09] <Aison> else it's not accepted
[22:24] <kirkland> mathiaz: okay, i confirmed with nurmi that these tests are not all of the tests they are planning to give us
[22:25] <kirkland> mathiaz: he said we'd talk more about the test suites in tomorrow's meeting with Eucalyptus
[22:25] <mathiaz> kirkland: ok
[22:25] <mathiaz> kirkland: so far we only have the stress tests right?
[22:26] <kirkland> mathiaz: right
[22:26] <kirkland> mathiaz: they have a functional suite too
[22:26] <mathiaz> kirkland: that's probably what dan was refering to
[22:26] <mathiaz> kirkland: and I though about enabling it in the build (if possible)
[22:28] <kirkland> mathiaz: if possible -> sure
[22:28] <sbeattie> jjohansen|smoser: one issue with virtualbox is that it can (and in newer versions defaults to) use a virtual e1000 nic.
[22:29] <jjohansen> hrmm vmware has that as an option too
[22:30] <jjohansen> though it defaults to its lance or vmxnet
[22:33] <MTecknology> What's the 'correct' way to set my hostname? I tried to do   123.123.123.123  server.domain.com     server   in /etc/hosts but that didn't seem to change anything
[22:33] <baccenfutter> MTecknology: also edit /etc/hostname
[22:33] <MTecknology> oh. thanks
[22:34] <jmarsden|work> MTecknology: hostname -F
[22:34] <baccenfutter> or that
[22:34] <baccenfutter> which does the same
[22:34] <MTecknology> thanks
[22:35] <MTecknology> actually /etc/hostname doesn't exist..
[22:35] <sbeattie> jjohansen: vbox also supports pcnet interfaces
[22:37] <MTecknology> there we go - thanks :D
[22:41] <jjohansen> sbeattie: thanks, that is good to know
[22:44] <Aison> argh, i'm getting crazy ;) ldap nss is working now, but so far no chance for samba
[23:01] <Mike_lifeguard> Apache's error logs show lots of IPs trying to access stuff that doesn't exist (/var/www/mysql, /var/www/pma, /var/www/mysqladmin, /var/www/phpadmin etc) - I guess those are looking for something to try to break into? In any case, is there a way to get rid of them? Should I not worry about it?
[23:37] <baccenfutter> anybody happen to have a nice pic of system respond times
[23:38] <baccenfutter> like a visualization or anything fancy you could put in a presentation