[00:05] <owh> It just occurs to me that it might look like I'm asking someone else to do the work. I'm prepared to do the leg-work, but I thought I'd ask about the size of the effort before I volunteered :-)
[00:06]  * racecar56 is sick of trying too get ubuntu working on a hp mediasmart ex475
[00:06] <ajmitch> owh: adding new stuff to an LTS release is generally a not-going-to-happen
[00:06] <ajmitch> unless you could make a really really good case for it
[00:07] <racecar56> blizzkid, i didnt know ubuntu at the time :(
[00:07] <racecar56> blizzkid, i found ubuntu in june/july 2008
[00:08] <blizzkid> racecar56: I've been using ubuntu ever since 4.10
[00:08] <racecar56> blizzkid, cool
[00:08] <racecar56> blizzkid, if i would have known of it i would have used it :(
[00:08] <owh> ajmitch: At this stage I'm coming up ZIP with "a really really good case for it", other than "this should not be the only reason for a user to have to upgrade to 9.04", and that doesn't really cut it does it :)
[00:11] <ajmitch> I'd say 'good luck' if you were trying to convince those in control of stable release updates :)
[00:11] <owh> As-in "Hell will freeze over, but if you manage to succeed I'll applaud"?
[00:11] <ajmitch> pretty much
[00:11]  * ajmitch must go to meeting
[00:11] <owh> Tah
[00:11] <owh> Later
[00:50] <lwizardl> hi
[00:51] <lwizardl> would it be wise to have your store web site and retail store database server on the same machine?
[00:53] <yann2> depends on load I guess
[00:54] <yann2> what point are you worried about?
[00:56] <lwizardl> well I'm working on opening a local game store and the we site will be hosted inhouse using my local isp
[00:56] <lwizardl> but I wasn't sure if I would bottleneck for sales and site on the same machine
[01:00] <blizzkid> lwizardl: I'd never install web and db on one machine myself
[01:01] <blizzkid> unless for testing/developing
[01:04] <lolololo> i need a little advice from someone using command restrictions in .ssh/authorized_keys2
[01:07] <owh> lwizardl: Are you concerned about load, or security?
[01:07] <lwizardl> both
[01:07]  * owh would be concerned about the latter :)
[01:08] <owh> As in, retail store database implies suppliers, credit limits, prices, stock levels, etc. I'd not want that on the same device as the web server.
[01:08] <blizzkid> +1 owh
[01:08] <owh> Unless you're going to do lots of traffic, load is nearly always solvable with money.
[01:09] <blizzkid> owh: crossed cable between frontend and backend on seperate nicks...
[01:09] <lwizardl> ok because I don't think i will be doing sales anywhere near levels of like bestbuy, walmart, etc
[01:09]  * owh observes that that is an extreme simplification, but that's what it boils down :)
[01:10] <owh> money that is.
[01:11] <owh> blizzkid: That is one option. Another is to publish data onto the web server with no live connectivity between them. Another is to virtualize the lot and separate them that way, another is to host the web server with a hosting provider and host the retail database locally. Another is to put them both remote. It's impossible to make recommendations without information.
[01:12]  * owh needs food.
[01:12] <owh> And sleep :)
[01:12] <blizzkid> owh: obviously there's diff approaches ;)
[01:14] <lwizardl> well see my only problem is that my webhost doesn't allow direct access to the mysql db for security reasons
[01:14] <lwizardl> so i figured if it was locally here i could easy access it for the POS terminals
[01:14] <blizzkid> lwizardl: can't you install xen?
[01:14] <blizzkid> and then install 2 virtual servers?
[01:14] <lwizardl> xen?
[01:14] <blizzkid> locally
[01:14] <blizzkid> or kvm
[01:14] <blizzkid> or virtualbox
[01:14] <blizzkid> or vmware
[01:15] <blizzkid> :p
[01:15] <owh> FYI: virtualisation != security
[01:15] <lwizardl> k
[01:15] <blizzkid> lwizardl: owh is right, although it helps
[01:16] <blizzkid> if one virtual machine is compromised, the other is not by default
[01:16] <blizzkid> best option imho is 2 machines separated by a dedicated firewall plus a bunch of other configs
[01:16] <owh> Virtualisation is a means to abstract your problem domain, that is, break the problem into smaller chunks as well as manage hardware resources in a different manner.
[01:17] <owh> Ultimately, the virtual machines are still running on the same actual CPU, so there is scope for transgression, ie. privilege escalation, DOS, whatever.
[01:18] <blizzkid> indeed
[01:19] <owh> lwizardl: Ultimately the question needs to be answered by doing research. Things like: "What will it cost the company if this data is compromised?" "How much will it cost if the server goes down?" "What fall back solutions are they and can I implement them in a cost-effective fashion."
[01:19] <lwizardl> I was thinking a single server for starting out would have been ok, and then add another one later after the business is starting to get stronger known
[01:20] <owh> You need to think of it in terms of: "How much does it cost if this disaster happens?"
[01:20] <lwizardl> k
[01:20] <owh> Sometimes the answer is a $1500 PC, sometimes it isn't.
[01:20] <blizzkid> lwizardl: I started a project for a customer of mine
[01:20] <blizzkid> small business to start with
[01:21] <blizzkid> but I opted for 2 2-node clusters
[01:21] <blizzkid> one for the frontend, and one for the backend
[01:21] <blizzkid> $ 7K in servers
[01:22] <blizzkid> cost of not being online, or stolen data would be much higher
[01:22] <blizzkid> so was it worth it? No doubt
[01:22] <owh> A question in that scenario would be: "Does the business have a full-time IT person with a clue, or not."
[01:22] <owh> The sexier the solution, the more skill is needed to fix stuff.
[01:23] <owh> K.I.S.S.
[01:23] <lwizardl> its just me in the business so i'm still new to this stuff
[01:23] <blizzkid> owh: they don't, but I do the support for them, and am (frankly) being quite well paid to do so ;)
[01:23] <lwizardl> my first webserver was a xbox console running gentoo
[01:23] <owh> lwizardl: Well, it's a start :)
[01:23] <owh> At least you've come to u-s :)
[01:24] <lwizardl> to this day i still love that distro but ubuntu is on all my computers now
[01:24] <owh> One day you'll understand that optimized compilation is not better than apt-get :)
[01:24] <lwizardl> all 9 of them including the 3 that will be my POS
[01:24] <blizzkid> lol owh
[01:25] <lwizardl> i use apt-get all the tim from bash i've always hated GUI configs
[01:25] <owh> The biggest thing that most small businesses do is over estimate their income and under estimate their expenses. Unless you've got a pot of gold, spend less.
[01:25] <owh> That doesn't mean, "be a cheap skate" btw.
[01:26] <lwizardl> yeah i've been buying used fixtures and such to save money for later use on bills etc
[01:26] <owh> If today your web traffic is zilch, then spend $9.95 per month on a virtual server in the cloud. Get your systems working, make money, expand. Rinse and Repeat.
[01:26] <blizzkid> I don't completely agree with you owh better safe than sorry, even if that comes at a higher cost
[01:27] <owh> blizzkid: The road is littered with failed start-up companies. I'm still here after 10 years <grin>
[01:28] <blizzkid> owh: true, but what I meant to say it's all a matter of roi/tco and all that "crap" ;)
[01:28] <blizzkid> If you know your data getting compromised will cost you at least $ 100K
[01:28] <blizzkid> what's an investment of $ 10K?
[01:28] <owh> That's only true if you have $10K :)
[01:29] <blizzkid> obviously :)
[01:29] <blizzkid> and in my example (the 2 clusters) it wasn't a startup company, but a new project for a long-existing company, so that's slightly different
[01:30] <owh> Yeah, different kettle of fish.
[01:30] <lwizardl> and whats a good hardware firewall for server use?
[01:33] <owh> lwizardl: That's the same question really :)
[01:34] <lwizardl> I was looking for like brands. only kind of hardware firewalls i have used so far are Linksys routers and most have been flashed with OpenWRT or DD-WRT
[01:36] <blizzkid> lwizardl: for startup I'd go with an old pc and install some software fw
[01:36] <lwizardl> blizzkid, so basically IPcop
[01:37] <blizzkid> lwizardl: I don't know IPcop as such, but I'm sure there's some good software fw's out there
[01:39] <lwizardl> i just know of 3 software firewalls all i think are linux based, m0n0wall, shorewall, and ipcop
[01:39] <blizzkid> I guess monowall is good
[01:42] <owh> lwizardl: Have you read the u-s guide? http://doc.ubuntu.com/ubuntu/serverguide/C/index.html
[01:43] <lwizardl> I think so but I will look at it again
[01:44] <owh> It all gets much more personal if you need to make it work for a business :)
[01:44] <owh> Gotta run.
[01:46] <lwizardl> well he left before I could thank him for his help
[01:46] <blizzkid> lol
[01:46] <lwizardl> but thanks owh
[01:46] <lwizardl> and blizzkid thanks for you help also. this is the reason why i use ubuntu all the time now
[01:47] <blizzkid> same here :)
[01:47] <blizzkid> I've done lots of messing up in the past :)
[01:47] <blizzkid> the key is experimenting
[01:47] <blizzkid> One advice I give you though: for production servers stick to LTS versions
[01:52] <lwizardl> so 8.10 over 9.04
[01:52] <blizzkid> yeah
[01:53] <lwizardl> k i always just go to the newest version
[01:53] <blizzkid> for production servers that's not the best option imho
[01:53] <ScottK> That's generally worked for me too.
[01:53] <ScottK> I have a test server that I verify everything with first.
[01:53] <lwizardl> but all my servers are desktops being used as servers
[01:54] <blizzkid> ScottK: I've messed up a software raid on a production server upgrading to jaunty... Not something you want :) Took me about 6 hours to fix that
[01:55] <ScottK> blizzkid: I did that once, but I'd taken a 4 hour maintenance window for the box, just in case.
[01:55] <blizzkid> lwizardl: that's not servers imho :p
[01:55] <lwizardl> yeah i know
[01:55] <lwizardl> servers are rack mounted cases
[01:55] <blizzkid> ScottK: in this case it was one node of a 2-node cluster, so no real damage was done, but still...
[01:55] <blizzkid> lwizardl: not necesarily
[01:56] <blizzkid> tower servers are perfectly fine too
[01:56] <ScottK> It depends.  I've been in situation where newer functionality was actually useful.
[01:56] <blizzkid> but desktop os != server os
[01:56] <lwizardl> yeah but generally you see the 1U/2U type servers used
[01:56] <blizzkid> ScottK: obviously if you need that newer functionality, it's reasonable to upgrade
[01:57] <ScottK> Actually my oldest system is Dapper and it's a desktop (and used as such).
[01:57] <yann2> enjoying firefox 1.5?
[01:57] <blizzkid> nice, I didn't think those would still be in the wild :)
[01:58] <ScottK> yann2: It works well enough when I use it.  It still gets security support.
[01:58] <ScottK> I tend to use Konqueror more on and more on that particular machine though.
[01:58] <yann2> firefox3 still significantly faster on heavier webaps :)
[01:59] <ScottK> Yes.  FF3 is the first FF I've liked more than FF 1.0.
[01:59] <yann2> Its still OpenOffice 1 isnt it?
[01:59] <yann2> sounds painful not to be able to open ODF
[01:59] <twb> I thought OO was up to 2.4 or 2.6 or something
[01:59] <ScottK> No.  It's 2
[01:59] <ScottK> 2.4 I think.
[01:59] <yann2> right
[01:59] <ScottK> Works fine with ODF.
[02:00] <yann2> sounds surprising  I thought it was 1.x or 2.0
[03:02] <Noah0504> Does anyone have any experience with LVM and or software RAIDs?
[03:03] <benspaulding> I am trying to configure daemontools on Jaunty server and it needs to edit /etc/inittab in order to work.
[03:03] <benspaulding> But my Jaunty install doesn’t have an /etc/inittab.
[03:04] <benspaulding> Has that been moved? Is there somewhere else I need to look? Or am I just doing my daemontools setup wrong?
[03:09] <ScottK> I don't think Ubuntu has had that since Dapper.
[03:10] <benspaulding> Oh. Alright.
[03:10] <benspaulding> Any ideas how I can get daemontools set up then?
[03:11] <ScottK> What was your editing /etc/inittab intended to accomplish?
[03:16] <benspaulding> It adds a line ``SV:123456:respawn:/usr/local/bin/svscan-start`` that scans to be sure certain processes are running.
[03:19] <twb> upstart replaces inittab with /etc/event.d.  upstart-sysv-init has partial compatibility with inittab.
[03:19] <twb> Sorry, upstart-sysv-compat
[03:22] <Mal3ko> hey folks, how do i disable users from making ssh tunnel?
[03:23] <benspaulding> Cool. Thank you.
[03:29] <Noah0504> Bleh.  I'm trying to figure out if I should use RAID5 or LVM and and back up to NAS.
[03:30] <twb> Where "partial" basically means "bugger all"
[03:30] <twb> Noah0504: as opposed to what?
[03:32] <Noah0504> Well, basically I have a desktop system that is going to aslo act as a file server.  I plan on picking up two more TB drives this weekend to make a total of four.  That would make a 3TB RAID5 system... but it seems a little confusing to set up as a software RAID.  I was ust thinking about not doing RAID and using LVM to have different partition on different drives.  Maybe one or two drives for /srv and so on.
[03:33] <Noah0504> Or maybe just / on multiple HDDs with LVM and backup as neccesary.
[03:33] <twb> Noah0504: then a failure of any one disk will lose ALL the data on that disk
[03:33] <twb> But using a server as a desktop also is a Dumb Thing.
[03:34] <Noah0504> Ha.  I know.  It's a lightly used desktop.
[03:34] <Noah0504> Not a main machine.
[03:34] <Noah0504> But, it's the occasional file server, so I thought different partitions on different disks would speed things up a bit, and still add some protection.
[03:38] <twb> You're hurting my brain with your pseudologic
[03:41] <Noah0504> Well, thanks for not offering your logic then...
[03:48] <twb> As I said, 12:37 <twb> Noah0504: then a failure of any one disk will lose ALL the data on that disk
[03:49] <Noah0504> Would a BIOS RAID be any good?
[03:57] <twb> Noah0504: only if it's real raid and not fakeraid.
[03:57] <twb> Noah0504: did you pay at least $100 for your RAID card?
[03:58] <twb> (Oh, and hardware raid is usually more of a hassle to set up and maintain.  I would just put up with the slower performance of md RAID for a non-critical system.)
[03:59] <Noah0504> for md RAID, I would basically create idental partitions on n drives and go from there?
[04:00] <twb> The ubuntu-server installer will help you set up md raid.
[04:01] <twb> Broadly, you will create partitions of equal size on each disk, then create arrays of those partitions (one for /boot and one for LVM), then  create an LVM PV on the latter array, then create volumes and filesystems on the PV.
[04:11] <Yingying_Zhao> Does Ubuntu 9.04 have IA64 build?
[04:11] <ScottK> Yes.
[04:12] <twb> ScottK: an official one, or an unofficial one?
[04:12] <Yingying_Zhao> official one
[04:12] <ScottK> twb: Community supported, not by Canonical
[04:12] <twb> Yingying_Zhao: note that 64-bit Intel systems are usually AMD64 architecture -- IA64 is for Itanium.
[04:12] <twb> ScottK: righto
[04:13] <Yingying_Zhao> i know. so only community support for IA64 distro, right?
[04:13] <ScottK> Yes.
[04:13] <twb> ScottK: I suppose it doesn't matter much unless you were intending to buy a support contract from Canonical?
[04:14] <ScottK> twb: It does matter somewhat in the focus during development.  Problems in community supported archs are by definition not show stoppers for release.
[04:14] <twb> Do community supported architectures benefit from canonical infrastructure, such as buildds and launchpad?
[04:14] <ScottK> Yes
[04:14] <twb> ScottK: ah, good point
[04:14] <twb> Like m68k on Debian
[04:14] <ScottK> And it's not like Canonical people don't care/work on problems with them, it's just not their priority
[04:15] <ScottK> twb: No m68k isn't in Debian at all anymore.
[04:15] <twb> Hmm, really?
[04:15] <ScottK> It was dropped right after Lenny released.
[04:15] <twb> dnsmasq is available for m68k in sid...
[04:15] <twb> Where "dnsmasq" is just the first package name I could think of
[04:16] <twb> Oh well, it doesn't matter.
[04:16] <ScottK> Some of the tools still report stuff.
[04:16] <sbeattie> and indeed, I believe there may be installability issues around the kernel on ia64; Yingying_Zhao, you probably want ot ask around on #ubuntu-ports.
[04:16] <ScottK> Jaunty was a good release for ports in general.
[04:16] <ScottK> I don't recall specifically about IA64
[04:17] <twb> Does Ubuntu have anything like emdebian (an embedded / arm SIG)?
[04:17] <ajmitch> ubuntu-mobile?
[04:17]  * twb reads
[04:18] <twb> Apparently Ubuntu Mobile is for atoms (i386).
[04:18] <ScottK> We have an armel architecture, but nothing as extreme as embeddian.
[04:18] <ScottK> twb: Actually we have a separate lpia architecture for LPIA.
[04:18] <Yingying_Zhao> sbeattie: Thanks for your suggestion~
[04:18] <ScottK> LPIA/Atom
[04:18] <twb> ScottK: I thought that was deprecated
[04:18] <ajmitch> I'm fairly sure there's ongoing work on ARM
[04:18] <ScottK> Not yet.
[04:19] <ajmitch> wasn't jaunty officially released for ARM?
[04:19] <ScottK> It was released, but as a developer preview.
[04:19] <ajmitch> ok
[04:19]  * ajmitch only read the subject line of that mail 
[04:19] <Noah0504> twb: So, if I make three partitions on each drive (boot, /, and a swap... will those be combined and synced?
[04:19] <twb> Noah0504: only if you make RAID arrays of the partitions.
[04:20] <twb> Noah0504: normally you would either put swap inside LVM inside the RAID5 (slow, but you can allocate more/less swap space as needed), or a separate swap partition on each disk (much better performance).
[04:20] <twb> I don't worry about swap too much because by the time my systems are swapping, they are already screwed.  Hooray for the 2.6 kernel....
[04:21] <ajmitch> or stuff just sits in swap for months on end
[04:24] <twb> IMO the main reason for a modern system to have swap is 1) suspend-to-disk; and 2) buffering of disk blocks.
[04:27] <Noah0504> Eh, my Internet cut out.  twb, is there anyway to just create the RAID array and have Ubuntu take care of the partitions?
[04:28] <twb> Noah0504: not really, no.
[04:28] <twb> Noah0504: setting up RAID still sucks on d-i (Debian and Ubuntu)
[04:29] <Noah0504> Well, it's no big deal, I'm just trying to get an idea of everything.  I still need to get the rest of the drives to complete a RAID 5 setup which will happen on Sat.
[04:29] <twb> There really ought to be a guided partitioning choice like "Erase all my disks and use LVM and RAID", which would use RAID1 for a two-disk system and RAID5 for more disks.
[04:30]  * twb goes to pester #debian-boot about it
[04:30] <Noah0504> That's what I was hoping would have been done by now!
[04:30] <Noah0504> :)
[04:31] <ajmitch> twb: it's on the list of things to do, I believe
[04:31] <Noah0504> It's not roo much trouble to make the partitions, I'm just working on the research right now.
[04:31] <twb> ajmitch: yeah, I know
[04:31]  * ajmitch knows how annoying it was to go through & setup this desktop with RAID & LVM
[04:31] <ajmitch> even though I've done it several times before, it was just tedious
[04:33] <Noah0504> So, if I make a 10GB swap on a each of the drives in a 5 disk RAID 5 setup, is my swap partition going to be 40GB?
[04:34] <twb> Noah0504: if you make swap partitions outside the array, then you get size × number of disks.
[04:35] <Noah0504> If they're inside the array?
[04:35] <twb> Noah0504: if you make swap inside the array, then only size × 1
[04:35] <twb> Noah0504: how big are your disks?  1TiB?
[04:35] <Noah0504> Yes.
[04:35] <twb> Yeah, so I would make (say) a 1GiB or 8GiB swap partition on each disk *outside* the array.
[04:36] <twb> Then tell /etc/fstab to mount them all with the same priority, so they are striped
[04:36] <twb> This stuff is much easier in d-i than it is on an already-running system, I think.
[04:36] <Noah0504> The system will be reloaded.  So, it will all be done during install.
[04:37] <Noah0504> So, I should make a /boot, LVM (basically /) and a swap partition on each drive?  Then make the array on the LVM or / partition?
[04:38] <twb> Let me find the notes I wrote the other day for someone else
[04:39] <Noah0504> Thanks a lot.
[04:39]  * ajmitch would do LVM on top of RAID rather than RAID on LVM
[04:39] <twb> ajmitch: yes
[04:44] <philsf> isn't there a TUI that makes it all easy?
[04:45]  * philsf thinks it's evms
[04:45] <ajmitch> I only care about it at install time, usually
[04:46] <philsf> sure
[04:46] <philsf> except when something fails
[04:47] <ajmitch> then I'd rather use mdadm & lvm tools myself than relying on evms
[04:47] <philsf> pacakges.u.c tells evms was dropped after hardy
[04:47] <ajmitch> evms had issues
[04:47] <twb> I thought evms was the opposite of lvm
[04:47] <ajmitch> evms was sort of all-encompassing
[04:47] <twb> i.e. you picked one or the other
[04:47] <ajmitch> I recall it & lvm disagreeing on the size of some logical volumes once
[04:47] <ajmitch> which worried me a little
[04:48] <ajmitch> I'm fairly sure you could have them both installed, it's been a little while since I had it installed though
[04:48] <philsf> I recall discovering about it when accessing my volumes wasn't possible with lvm tools
[04:48] <philsf> because the devices were moved to /dev/evms/*
[04:49] <philsf> took me some hair to find that out
[04:49] <philsf> because dapper installed and enabled it by default
[04:50] <twb> Noah0504: http://paste.lisp.org/display/80184
[04:50] <Noah0504> Thanks, twb.
[04:51] <twb> Noah0504: note that there I advocate swap inside LVM, but you can easily have separate swap partitions outside of LVM, i.e. sd[abc]3 4GiB swap
[04:51] <twb> With 1TiB disks, permanently allocating several GiB to swap doesn't matter too much
[04:51] <Noah0504> So, sticking them outside of the array would be fine?
[04:51] <ajmitch> yes
[04:51] <Noah0504> Faster even?
[04:52] <ajmitch> unsure if it'd be noticeable
[04:52] <Noah0504> Ha.  Well, like twb mentioned, you shouldn't need to be too dependent on it.
[04:53] <twb> It'll be faster because swap will strip across the disks -- RAID0 instead of RAID5
[04:53] <ajmitch> yes, I just setup 2 swap partitions rather than using RAID 1 on here
[04:53] <twb> (Note that you swap is RAID0 automatically -- do NOT actually create a RAID0 array.)
[04:54] <Noah0504> So, the /boot is actually being mirrored to each other, while LVM will actually stripe?
[04:55] <Noah0504> (With parity, of course.)
[04:56] <twb> Noah0504: /boot is RAID1 (mirror) because it has to be, for GRUB
[04:56] <twb> So it's a mirror of three disks
[04:56] <Noah0504> Which will cause it to still boot upon fail...
[04:56] <Noah0504> I get it.
[04:57] <ajmitch> which is sort of important & all
[04:57] <twb> The main LVM PV partition is a RAID5 array, which basically means it does the right thing, and you can lose up to one disk safely.
[04:57] <twb> If you have *five* disks, you might want to consider using *two* parity disks, which will give you more safety in exchange for less storage space.
[04:58] <Noah0504> Great.  I think I have it all figured out now.  Doesn't seem like it should be too hard to get figured out when I actually go to do it.  I just need to pick up the extra disks on Sat.
[04:58] <twb> RAID5 with two parity disks is sometimes (incorrectly) referred to as RAID6.
[04:58] <Noah0504> I'll be going with 4 or 5 disks.
[04:58] <Noah0504> And then a couple eventually for drive failures.
[04:59] <Noah0504> Which hopefully won't happen soon.
[04:59] <Noah0504> :)
[04:59] <twb> Oh, and just to be clear: RAID5 doesn't mean 5 disks.  RAID5 works with 3 or more disks.
[04:59] <Noah0504> That I do know.  I'm not really a stranger to RAIDing... just software RAID with md.
[04:59] <Noah0504> Ha.
[05:00] <Noah0504> I screwed myself though.  I'm locked in on getting this Seagate drives when the WD Greens are $90 now!
[05:00] <Noah0504> Oh well, these Seagates have 5 year warranties.  Which is why I ususally go with them.
[05:12] <twb> Man, $90 for 1TiB
[05:12] <twb> Remember MFM?
[05:15] <Noah0504> I've been paying $99 for these Seagate drives... still not bad.
[05:15] <Noah0504> RAM and HDD prices have plumeted.
[05:15] <ajmitch> twb: if only it were 99 AUD
[05:15] <Noah0504> I'm going to set up a test RAID of 0 on these drives.
[05:16] <Noah0504> I just have two... take it for a spin and make sure I have a grasp of it.
[05:16] <twb> ajmitch: Seagate 3.5" 1TB 1000GB 7200rpm SATA or SATA II Serial ATA100 HDD Hard  AU$144
[05:16] <ajmitch> that's pretty decent
[05:16] <twb> AU$139 for hitachi, meh
[05:17] <Noah0504> Not worth it.
[05:17] <Noah0504> Ha.
[05:17] <twb> AU$197 for 1.5TiB seagate
[05:17]  * ajmitch is still looking at over NZ$230 for 1TB drives on a local site
[05:17] <Noah0504> Ouch.
[05:18] <twb> ajmitch: yeah, well.  I'm looking at the listings for "warehouse with an asshole and a desk in front of it" retailers
[05:18] <ajmitch> aha, I'm looking at a reliable site with free delivery in NZ
[05:18] <ajmitch> so I'd expect a little more
[05:18] <Noah0504> I'm just glad to have Newegg.com here.
[05:19] <twb> Oh, of course warehouse-asshole does not do delivery at all.
[05:20] <twb> I mainly use his listings to find out what the lowest possible price is, so that I know what the markup of other retailers is :-)
[05:20] <twb> http://twb.ath.cx/Preferences/.bin/msy
[05:22] <Noah0504> ext2 or 3 for boot?
[05:22] <twb> Use ext3 unless you know better
[05:23] <Noah0504> Gotcha.
[05:36] <mib_a6s4uz3b> HI, I have a Ubuntu Server 9.04 consisting of DHCP, NTP, DNS, OpenLDAP and SAMBA. I want to join my Ubuntu 9.04 Desktop. How can I do so?
[05:38] <mib_a6s4uz3b> I have configured this as per the server documentation.
[05:39] <mib_a6s4uz3b> But instructions for joining a Ubuntu Client is not clear to me.
[05:47] <mib_a6s4uz3b> Hey anyone around
[05:51] <Rafael> how can i see/check the size of the boot area
[05:59] <mib_a6s4uz3b> Hi, anybody around
[06:02] <dhruba> Hi, anybody around?
[06:04] <dhruba> I have configured a PC with NTP, DHCP, DNS, OpenLDAP and SAMBA as per the documentation of Ubuntu 9.04 Server. But I am unable to join my Ubuntu 9.04 Desktop with it.
[06:06] <pteague> do i simply need to do a `/etc/init.d/networking restart` after changing an ip address or do i need to do something else?
[06:08] <dhruba> No restarting the network through the command is sufficient
[07:49] <Kamping_Kaiser> just over 100 5.04 cds. knew i had 'some' left, didnt know it was enough to start a company
[09:04] <[HU]gnanet> Hi, anyone got an idea how to downgrade rdiff-backup nicely if my backup server and the server i do my backups to are different versions of ubuntu?
[09:04] <[HU]gnanet> and therefore rdiff-backup is also different?
[09:09] <[HU]gnanet> ok i read on forums its a rdiff-backup problem not the version diff (8.10 ubuntu and 1.1.16 rdiff-backup)
[09:14] <andol> ttx: Believe I might have goten the hang of merging nagios3. Just have a few questions...
[09:26] <ttx> andol: shoot
[09:34] <andol> ttx: Seems like the only Ubuntu changes which I might want to keep is the Maintainer field. Is that something we still want to keep, even if the rest of the package just is a copy of the debian version?
[09:35] <ttx> andol: no. You should keep it only if there are Ubuntu changes
[09:35] <ttx> andol: so in this case it should be a sync request.
[09:35] <ttx> and I can confirm it is.
[09:36]  * andol starts reading about syncs...
[09:39] <andol> ttx: So, what I do now is submit a sync request in LP, and subscribe ubuntu-main-sponsors?
[09:40] <ttx> andol: yes. Manually or using requestsync -s
[09:45] <Blackhold> hello
[09:45] <Blackhold> could someone help me please?
[09:46] <ttx> !ask | Blackhold
[09:46] <Blackhold> ubottu: gutsy repos don't appear
[09:46] <Blackhold> I have the system not updated about a 9 months
[09:47] <Bergcube> Ubuntu 9.04 with zabbix-server-mysql and zabbix-frontend-php installed. Upon first connect from the webbrowser I get the error "Timezone for PHP is not set. Please set "date.timezone" option in php.ini." I have set the timezone in /etc/php5/apache2/php.ini but that didn't help.  What should I do to fix this?
[09:47] <twb> Would that be because Gutsy is not a LTS, and it is no longer supported?
[09:47] <twb> Bergcube: maybe you also need to restart apache or php?
[09:48] <Bergcube> twb~  Well, I've even rebooted.  Sorry for not including that....  :-S
[09:48] <twb> Bergcube: dunno, then.  Be patient and hope someone else knows
[09:48] <Bergcube> twb~  Thx.
[09:48] <Blackhold> twb: yes
[09:48] <twb> Blackhold: you should schedule an upgrade to 8.04 ASAP
[09:49] <Bergcube> Hm, I guess there must be a PHP channel here too...
[09:49] <Blackhold> wich is the codename?
[09:49] <twb> Blackhold: hardy
[09:49] <Blackhold> twb: then, where appears gutsy I should put hardy?
[09:49] <twb> Blackhold: I think you can change the sources.list to point elsewhere and still get gutsy, but obviously it won't be supported after it is EOLd
[09:49] <Blackhold> and then # apt-get safe-upgrade ?
[09:49] <twb> Blackhold: you should read the release notes
[09:49] <Blackhold> sorry apt-get update && apt-get safe-upgrade
[09:49] <twb> !gutsy->hardy
[09:49] <twb> Grr.
[09:50] <twb> !dist-upgrade
[09:52] <andol> ttx: Since you seem to be around, does bug #376397 do the trick?
[09:56] <ttx> andol: looking...
[09:58] <ttx> andol: adding the Debian changelog since latest Ubuntu helps in seeing the delta was merged
[09:58] <ttx> andol: see bug 372699 for example
[10:02] <ttx> andol: basically you have to say that the only Ubuntu delta was that prerm apacheconf purge and show in the Debian changelog that it was adopted in Debian
[10:03] <ttx> It will help the sponsors job.
[10:03] <andol> ttx: Got it, changing the description now...
[10:08] <andol> ttx: better?
[10:16] <ttx> andol: looks good
[10:16] <andol> ttx: thanks
[10:17] <ttx> now subscribe the relevant sponsor team so that they ACK your conclusion
[10:17] <ttx> I can't do that myself for main (yet)
[10:18] <andol> done
[10:20] <ttx> andol: cool, now, on to nagios-plugins for level 2 :)
[10:20] <andol> ttx: Will do, just have to start doing some laundry first...
[10:21] <ttx> andol: bah, who needs clothes
[10:22] <andol> ttx: Well, I'm going to ride the bus tomorrow. Not sure if that is kosher to do without any clothes on :-)
[10:38] <sergevn> hi i want to install IBM DB2 client for php5, is there any package or tutorial ?
[10:49] <safl> Hey people, I'm used to having the output of cron send to roots mailbox, this is disabled by design in ubuntu-server, how can i "get it back"? I've tried installing mailutils, but i guess theres more to it than that... the $MAILTO var is still empty...
[11:10] <andol> ttx: Since I assume you already know the "correct answer" of nagios-plugins... I take it that there should be a merge preserving 99_check_ntp_segfaults.dpatch as well as the structure of nagios-plugins-extra?
[11:11] <ttx> andol: I seem to remember we must also preserve usage of libradius1-dev as build-dep
[11:13] <andol> ttx: You'r right, I missed that.
[11:14] <andol> ttx: Anyway, seems like the merge-o-matic took care of all those things. So what I should do now is write up a merge request, explaining the deltas we keep and fix debdiffs to upload?
[11:16] <ttx> you should also write a changelog entry
[11:16] <ttx> regenerate the source package to include that changelog entry
[11:16] <ttx> generate debdiffs from that
[11:17] <ttx> + forward the relevant fixes to Debian
[11:17] <ttx> (if any)
[12:33] <ttx> andol: + also check the existing bugs list to see if any would get fixed by the new version
[12:40] <andol> ttx: Don't see anything in the 1.4.12-5 changlog entry which matches current LP bug reports on nagios-plugins.
[12:41] <ttx> andol: me neither
[12:53] <reya276> How can I restrict users connecting through OpenSSH to only read,write and view to their home directories?
[12:58] <andol> zul: See you did the last merge of nagios-plugins? How did you handle/note changes in the ./po/ folder?
[12:58] <zul> andol: no i just woke up
[13:00] <andol> zul: Well, the second last I guess now when I take a closer look, but ttx refers the question to you. good morning by the way.
[13:00] <zul> ttx: ok thanks..
[13:01] <reya276> How can I restrict users connecting through OpenSSH to only read,write and view to their home directories?
[13:01] <ttx> zul: you're welcome ;)
[13:02] <reya276> I tried changing the user shell to /bin/bash -s but that does not work
[13:03] <zul> andol: keep them and note the changes in the changelog
[13:03] <andol> zul: Specific po changes, or just refer to them as general changes to po/?
[13:03] <zul> just general
[13:04] <andol> zul: That was the answer I was hoping for, thanks :)
[13:16] <\sh> moins guys...
[13:28] <andol> zul, ttx: Any particular reason why the nagios-plugins package doesn't have any XSBC-Original-Maintainer in debian/control?
[13:29] <zul> nope just add it
[13:56] <andol> ttx: Before I subscribe sponsors to bug #376459, is there anything in particular wrong with it?
[13:56] <ttx> andol: looking
[14:01] <ttx> andol: looks ok, <nitpicking>You don't need to add a changelog entry about the XSBC-Original-Maintainer change because it is an implied change</nitpicking>
[14:02] <ttx> andol:  I'd say do not bother redoing the patches unless your sponsor objects to it
[14:02] <andol> Ok, thanks
[14:09] <henriquelm> Hello There
[14:10] <henriquelm> Where are located the header files and libraries of the X window system un Ubuntu Server 8.04?
[14:19] <W8TAH> ive a snort machien that i just moved to a new physical location -- stupid me forgot to mark the cables - -is there a way that i can have the box tell me which nick is which ?
[14:19] <andol> ttx: Well, now it is time for me to enjoy some of that outdoors. Thanks for the help today. I've learned lots.
[14:20] <ttx> andol: you're welcome
[14:22] <henriquelm> Can someone help me out with the x11vnc installation? http://rafb.net/p/49hZC730.html
[14:22] <soren> henriquelm: Try in #ubuntu.
[14:36] <angelleye> I'm a PHP developer but my local test enviorment has always been Windows.  I'd like to get an actual linux dev enviornment up and running.  I just installed ubuntu 9 and lamp.  Now I'm sitting here with some questions.  Can anybody help me out real quick?
[14:38] <ScottK> !ask | angelleye
[15:48] <Hamzifer> kirkland: here?
[15:50] <gourgi> hey guys, i have a problem with vlans in 9.04. same configuration is just fine under 8.10
[15:50] <Hamzifer> kirkland: oh, heh, i've just seen bug #375309, will post on there instead
[15:50] <gourgi> when i restart networking i get this http://pastebin.com/m17c7325b
[15:50] <gourgi> here is my interfaces http://pastebin.com/f77236772
[15:52] <gourgi> also i noticed that in 9.04 lsmod |grep 8021 give two modules:  8021q and garp, while in 8.10 only 8021q. is this relevant?
[15:52] <kirkland> Hamzifer: yes, thanks.
[15:53] <gourgi> anyone else seen this vlans problem with 9.04 ?
[15:55] <gourgi> btw i have no nfs mounts in me fstab
[15:55] <moonpup> hi, can someone tell me when setting up samba if adding the root user (smbpasswd -a root) is still a required step for functioning properly?
[15:56] <moonpup> for some reason i can't get OS X to mount home directories although other shares mount ok
[16:20] <reid_> hmm.. x was just working on my server..  now when I startx, it says "Saw signal 11. Server aborting"
[16:23] <ScottK> If it involves X, this isn't the channel for it.
[16:23] <reid_> fair enough
[16:36] <big_ham> anyone around to help with NIC issue in Intrepid?
[16:41] <ivoks> RoAk: there?
[16:41] <RoAk> ivoks, heya master, how's it going
[16:41] <ivoks> great
[16:41] <ivoks> i've talked with fabbione
[16:41] <RoAk> ivoks, what he said?
[16:42] <ivoks> there's an intention to merge rhcs and pacemaker
[16:42] <ivoks> basically, heartbeat is a dead project
[16:42] <ivoks> http://www.gossamer-threads.com/lists/linuxha/pacemaker/51414
[16:42] <RoAk> ivoks, so basically, they want to do: pacemaker + openais = rhcs
[16:42] <ivoks> no, merge them
[16:42] <ivoks> parts of rhcs + parts of pacemaker = new cluster stack
[16:43] <RoAk> ivoks, I see then.. so what they wan't to do, is to create a new cluster stack and then rhcs and pacemaker will disappear as a single project?
[16:45] <ivoks> correct
[16:45] <ivoks> but this will not happen in next year
[16:45] <ivoks> there's no date when they will achieve this
[16:45] <ivoks> that's the goal and they started with first steps
[16:45] <ivoks> so, what we should do is talk with debian-ha-maintainers
[16:45] <RoAk> ivoks, yeah.. it will take time for them to create it, however in the link I can see that they'll keep the heartbeat subsystem
[16:45] <ivoks> parts of it
[16:45] <RoAk> ivoks, i've been on that, and debian-ha-maintainers are only contacts of drbd8 and csync2
[16:45] <ivoks> ok
[16:46] <RoAk> so it's a dead end, unless they start working on other package.. anyways.. we'll have to talk to them and see if they have something thought about HA
[16:46] <RoAk> otherwise, what do you think we should do? at least for karmic?
[16:48] <ivoks> i'll work on rhcs
[16:48] <ivoks> i think we should have both rhcs and pacemaker in main
[16:48] <RoAk> if we do so, we should also have heartbeat and openais
[16:48] <ivoks> openais is in main already
[16:48] <RoAk> because, what i've been seen lately is that many people is now trying to use pacemaker/heartbeat
[16:48] <ivoks> rhcs depends on it
[16:50] <ivoks> that's true
[16:50] <RoAk> so, pacemaker won't do any good if it works by itself, since it is intended to work with heartbeat or openais
[16:50] <ivoks> RoAk: i think we should stick with openais
[16:50] <ivoks> we should also join linux-cluster and linux-ha channels
[16:51] <RoAkSoAx> ivoks, an what should be do about heartbeat (in my experience, much more people use heartbeat rather than openais)
[16:52] <ivoks> RoAkSoAx: make it usable and leave it in universe?
[16:55] <RoAkSoAx> ivoks, ok then... so these new project is gonna have a little bit of everything, heartbeat, pacemaker, rhcs
[16:55] <ivoks> it seems so
[16:55] <ivoks> but we will know more once we join linux-ha and linux-cluster
[16:55] <RoAkSoAx> ivoks, i do think it is a good idea to take the best of every stack and make one single stack, which should work better
[16:56] <ivoks> that's their plan :)
[16:57] <RoAkSoAx> k :)
[16:57] <RoAkSoAx> ivoks, btw.. have you take a look to my latest merge?? I'm not sure if what I did is the right thing to do... I added quilt as patch system for the Ubuntu change...
[16:58] <ivoks> i haven't yet
[16:59] <RoAkSoAx> ivoks, k, when you have a little time please take a look and give me some feedback about it
[16:59] <ivoks> i will today
[16:59] <ivoks> in couple of minutes
[16:59] <RoAkSoAx> ivoks, now.. how's the process to work with FTBFS
[16:59] <ivoks> RoAkSoAx: bbl
[17:16]  * ScottK wonders if which mysql to use for Karmic needs a spec?
[17:16] <ivoks> :)
[17:17] <ivoks> let's move to MariaDB
[17:17] <ivoks> that way we can still call it a LAMP stack
[17:17] <ScottK> Heh
[17:19] <mathiaz> ivoks: did you get a change to test drbd 8.3 with 2.6.30 in karmic?
[17:20] <ivoks> chance? not yet
[17:20] <ivoks> i could do it during UDS
[17:24] <ivoks> mathiaz: why do you ask? any problems with it?
[17:24] <mathiaz> ivoks: nope - just checking
[17:24] <ivoks> ok
[17:24] <mathiaz> ivoks: I know we run into problems in previous releases
[17:24] <ivoks> 8.3.1 was the last drbd
[17:24] <mathiaz> ivoks: with issues between userspace and kernel
[17:24] <ivoks> it was tested with 2.6.29
[17:24] <ivoks> i know...
[17:24] <mathiaz> ivoks: yeah - since karmic is on 2.6.30 - thinks may be broken now
[17:25] <ivoks> i doubt...
[17:25] <ivoks> Wed, 29 Apr 2009 14:06:04 +0000
[17:25] <ivoks> that's the last commit in drbd 8.3 :)
[17:25] <mathiaz> ivoks: and there is the ubuntu specific patch about CN_IDX_DRBD
[17:25] <mathiaz> ivoks: is this something that may have changed in 2.6.30?
[17:25] <RoAkSoAx> ivoks, btw.. DRBD8 has not reached the archives yet, right?
[17:26] <mathiaz> RoAkSoAx: the merge for karmic - no
[17:26] <ivoks> mathiaz: ignore that patch
[17:26] <mathiaz> RoAkSoAx: I'm looking at it now
[17:26] <ivoks> mathiaz: it's just a number under which module is registred
[17:26] <mathiaz> ivoks: ? it's still in the merge though
[17:26] <RoAkSoAx> mathiaz, :)
[17:26] <ivoks> mathiaz: so, if you merge it, it won't work in karmic (i need to send in a patch for kernel)
[17:26] <mathiaz> ivoks: ok - so it's something specific to ubuntu and the integration of the kernel module into lum
[17:27] <ivoks> correct
[17:27] <ivoks> even the upstream knows that
[17:27] <mathiaz> ivoks: hm - I didn't understand - is the CN_IDX_DRBD
[17:27] <mathiaz> ivoks: hm - I didn't understand - is the CN_IDX_DRBD patch still needed in karmic?
[17:28] <ivoks> it is, but that change doesn't influence the drbd it self
[17:28] <mathiaz> ivoks: ok
[17:28] <ivoks> CN_IDX_DRBD is a number for linux's connector
[17:28] <big_ham> guys ... I had a VMServer running perfectly on Intrepid with a static WAN IP ... I went to put it behind a router and change to a static LAN IP and now both NIC's show down
[17:29] <big_ham> I changed everything back to the way it was and still everything shows down
[17:29] <big_ham> literally all I did was to change the network file
[17:30] <big_ham> I am baffled
[17:47] <ivoks> mathiaz: you can sign it with you own key, you know that? :)
[17:48] <mathiaz> ivoks: I always sign with my own key - why do you ask?
[17:48] <ivoks> never mind :)
[18:11] <andol> If I want to merge packages related to the server team. Is there any reason why I might want to focus on packages in main or on packages in universe?
[18:12] <mathiaz> andol: well - focusing on universe will help get you experience for becoming a MOTU
[18:13] <andol> mathiaz: That's a good point, thanks.
[18:14] <mathiaz> andol: bug 376459 - could you please send the ntp patch to Debian?
[18:16] <andol> mathiaz: Yes, but before I do that I want to make sure the error also applies to a Debian system. Or is it obvious enough just to assume it does?
[18:17] <mathiaz> andol: I think it applies to debian too - I haven't checked - but the patch is taken from upstream svn
[18:17] <mathiaz> andol: so there is a good change it applies in Debian too
[18:18] <andol> Ok, then I'll submit it to the debian bts.
[19:00] <dayo> anyone know how to manually config a connection to an ADSL modem?
[19:37]  * gourgi is desperate , needs help with vlans in 9.04 !
[19:38] <gourgi> anyone willing to help me with a vlan problem in 9.04 ?
[19:39] <gourgi> here i describe the problem http://ubuntuforums.org/showthread.php?t=1159283
[19:42] <wizardslovak> anyone has played with postfix here?
[19:51] <ivoks> nope
[19:51] <ivoks> what's that? :)
[19:58] <Vog-work> wizardslovak: I do regularly but postfix questions are best made in #postfix
[19:59] <Vog-work> nevermind see you there now
[20:07] <wizardslovak> ok so when i type "telnet localhost 25" i am getting "connection refused"
[20:07] <wizardslovak> what should i do?
[20:08] <RoAkSoAx> mathiaz, I added a new debdiff for openvpn :)
[20:11] <ivoks> good night
[20:26] <wizardslovak> "apt-get remove "software"" will remove program , and if i want to remove file and config what should i erase?
[20:27] <gourgi> wizardslovak apt-get purge
[20:28] <wizardslovak> ""apt-get purge software"??
[20:28] <gourgi> wizardslovak correct
[20:29] <gourgi> wizardslovak apt-get --help
[20:29] <wizardslovak> thx
[20:32] <ScottK> Or man apt-get
[21:11] <TimReichhart> I need some help on getting my network to come back up and when I did the lshw -C network  it came up with network card disabled
[21:12] <TimReichhart> and I am using 3com 3c905b cyclone
[21:39] <wizardslovak> ok when i type "netcat mail.domain.com 25"
[21:39] <wizardslovak> it doesnt show anything
[22:31] <n0ah> i kinda need the default php.ini that comes with ubuntu server 9.04
[22:32] <n0ah> messed it up bad
[22:51] <wizardslovak> how can i change server name?
[23:02] <MTecknology> What do you guys think of this? http://pastebin.ca/1423111 - I'm going to build my own sexxy server. I don't think I'm missing any parts...
[23:22]  * KaptenRodSkagg_ is away: Jag är upptagen
[23:42] <Wiseguy> hey guys, im lookin through my smb.conf file tryin to find where the default location for samba shares is located but cant seem to find it, does anyone know?