[10:50] <munzir> Hi, I downloaded http://cdimage.ubuntu.com/dvd/20060411/dapper-dvd-amd64.iso and it gave me a root prompt after this error: kernel direct mapping tables upto ffff810100000000 @ .... can't access tty; job control turned off.
[10:51] <fabbione> munzir: wrong channel and please try the new image that will be published today
[10:52] <fabbione> or file a bug in malone
[10:53] <munzir> fabbione: ok thx.
[10:53] <munzir> fabbione: why is it wrong channel, isn't that dvd contains the server edition too?
[10:54] <fabbione> because that's a kernel error
[10:54] <fabbione> and it is common to both server and desktop
[10:54] <fabbione> so you could try #ubuntu-kernel
[10:54] <fabbione> or perhaps it's just a bad burn?
[10:55] <fabbione> did you verify the md5sum of the download and the dvd with the one on the site?
[10:57] <munzir> fabbione: I didn't know there is an #ubuntu-kernel. I will check there and I did understand your point. thx
[10:58] <munzir> fabbione: for the server edition there is a default installation and a lamp installation. isn't it confusing what the difference between them?
[10:58] <fabbione> no
[10:58] <fabbione> if you know what LAMP is.. if you don't, you mostlikely don't need it
[10:59] <munzir> fabbione: the default will install what?
[10:59] <fabbione> the installation is exactly the same anyway
[10:59] <fabbione> LAMP adds Apache2 php5 and mysql
[10:59] <munzir> fabbione: so the default doesn't install apache, php5 and mysql, right?
[11:00] <fabbione> right
[11:00] <munzir> fabbione: the confusing point is the L stands for linux so it also means it won't install linux? ;)
[11:00] <fabbione> no comment
[11:02] <munzir> fabbione: ;)
[11:02] <munzir> fabbione: and what does that OEM installation means please?
[11:04] <fabbione> munzir: please these are FAQ. you can ask in #ubuntu
[11:04] <fabbione> and same as lamp, if you don't know what it means you don't need it
[11:07] <munzir> fabbione: ok thx again and sorry for any disturbance
[11:07] <fabbione> np
[12:55] <spike> eeer, is there any more robust alternative to acct/sa/sac ? from time to time I stop receiving reports I scheduled with cron, at times it crashes, but I'm pretty sure the system is fine, ie, no rootkits or anything
[12:55] <spike> only thought I could come up with is problems to deal with broken wtmp because of dead ssh connections... ie, the record isnt properly closed 'cause ssh died so I didnt logout
[12:56] <spike> it's very annoying, and but for that idea I cant really trace down the problem
[01:47] <Kolan> Hi, does anyone have experience in using a MegaRaid 105-4 controller with Ubuntu?
[03:08] <Kolan> Hi, does anyone have experience in using a MegaRaid 105-4 controller with Ubuntu?
[03:11] <neuralis> Kolan: what's the problem?
[03:12] <Kolan> I got it installed with Raid 1
[03:12] <Kolan> it seems like that the server is running quite slow
[03:13] <Kolan> I ran hdparm -tT
[03:13] <Kolan> and the buffered reading is about 1MB/s
[03:13] <Kolan> which isnt good
[03:17] <Kolan> neuralis: any idea?
[03:23] <neuralis> bad performance on a hardware raid device doesn't usually have anything to do with the operating system.
[03:25] <Kolan> neuralis: ok, well this is my first RAID1 system so Im not that experienced in it.
[03:27] <neuralis> Kolan: experience isn't really a factor here -- with hardware raid, the OS doesn't even know you're using raid.
[03:29] <Kolan> neuralis: I get an error message when booting. sda: asking for cache data failed
[03:29] <Kolan> sda: assuming drive cache: write through
[03:29] <Kolan> can this have something to do with it?=
[03:30] <neuralis> sure, but that wouldn't make reads slow.
[03:31] <neuralis> (if you had write-through set as policy on your raid controller, you'd see write performance degradation.)
[03:31] <neuralis> in your case, you can safely ignore the message.
[03:32] <neuralis> the kernel is asking the drive for caching parameters, but the request is getting intercepted by the raid card that doesn't know what to do with it, causing the problem report you're seeing.
[03:33] <allee> Kolan, neuralis: a collegue had trouble with megaraid and a Intel SRCS 16?? card.  AFAIR 2.4.30 as okay, 2.4.31 and all 2.6 kernels he tired where horrible slow.
[03:34] <allee> he found no solution and eventually gave up frustrated
[03:34] <neuralis> allee: that's possible, but surprising.
[03:35] <Nite_Hawk> heya guys, are any of you using ubuntu-server in a production environment?
[03:35] <neuralis> allee: although the transition you mentioned is the new megaraid code branch, if i remember correctly -- so the problem can be fixed by manually compiling the kernel with the old megaraid driver.
[03:35] <neuralis> Nite_Hawk: yes.
[03:35] <allee> neuralis: yeah, seems counterintuitive but happens (I've a problem that raids disk are not detected at all.  But I've to investigate more on this ...)
[03:37] <allee> neuralis: mhmm, I'll ask him.  but with all the 2.6 API changes I doubt that old megaraid will compile with 2.6
[03:37] <Nite_Hawk> neuralis: I'm an old debian guy, but at my new place of employment almost everyone are RHEL/SLES fans.  I may have a chance to introduce an ubuntu/debian solution if I play my cards right.
[03:38] <neuralis> allee: i *believe* the old megaraid module was kept in-tree, and still works. you'd have to check, though.
[03:38] <neuralis> Nite_Hawk: sounds good.
[03:38] <neuralis> Nite_Hawk: let us know if you have any questions, certainly.
[03:40] <Nite_Hawk> neuralis: Cool.  One question, do you know if there are any problems doing TSM backups?
[03:40] <neuralis> Nite_Hawk: i'm not a tivoli user; wouldn't know, sorry.
[03:41] <allee> Nite_Hawk: I use TSM client on dapper.  Works fine afaics, but I had not to recover much yet on dapper
[03:42] <allee> Nite_Hawk: you need sun java with TSM.  dsmj does not the free one
[03:42] <Nite_Hawk> oh, was ubuntu-server hit with the installer plaintext password bug that was present in breezy?
[03:42] <allee> dsmj does not like that is
[03:43] <Nite_Hawk> allee: ok, that's probably fine.  You can build a sun jvm easily enough using java-package right?
[03:44] <allee> neuralis: thx.  I'll forward the info to him.  2.6.15 has at least CONFIG_MEGARAID_LEGACY=m
[03:44] <neuralis> allee: yeah, i thought that was still there.
[03:44] <allee> Nite_Hawk: yes.  I've run in some problem but at least for dsmj they are not critical ;)
[03:46] <allee> Nite_Hawk: something about mimetype registrationa and java policy stuff failed because install script tries to update them below /etc/
[03:46] <Nite_Hawk> allee: hrm...
[03:46] <allee> Nite_Hawk: normal user can't and of course those changes would not find their way into the deb
[03:46] <allee> if run by root
[03:47] <allee> Nite_Hawk: try.  maybe they fixed it in the last ~ 3 months ;)
[03:48] <Nite_Hawk> allee: yeah, I may try to install ubuntu-server on one of my test machines.
[03:49] <allee> Nite_Hawk: ah, one addition.  that was on i386.  TSM on amd64 is still on TODO.
[03:50] <allee> Nite_Hawk: up to now all closed source stuff refused to work on amd64 because they still expect some i386 apps. I keep my fingers crossed TSM is different.  Pah!  Dreamer! :)
[03:51] <Nite_Hawk> hrm... It won't affect my machines, but possibly some others around here.
[03:51] <Nite_Hawk> Good to know though, thanks! :)
[03:54] <allee> Nite_Hawk: yeah, I hope that amd64-has-no-32bit-libs will not eventually kill kubuntu usage on my cluster.  Unfortunately some (stupid?) apps need more that 4 GB of mmaped address space :(
[04:00] <Nite_Hawk> allee: Yeah, I think it's going to be a major uphill battle around here.
[04:00] <neuralis> Nite_Hawk: what are the reasons that rhel/sles are preferred at the moment? support?
[04:01] <Nite_Hawk> allee: We run CentOS right now on the premise that Redhat offers support contracts for basically the same thing, and even though we don't buy the support, it must be good if they "offer" it.
[04:01] <Nite_Hawk> On some of the larger systems we actually run SLES or RHEL for the support contract (and presumed better compatability with things like TSM).
[04:02] <Nite_Hawk> A couple of the admins are pretty rabidly anti-debian.
[04:03] <allee> :)
[04:06] <Nite_Hawk> allee: I'd rather be running suse than redhat, but a number of people just left that is going to cause a lot of confusion/power vacuum, so I'm trying to decide what I can get away with. ;)
[04:06] <allee> neuralis: FWIW here, we need oracle server due to 3rd party software.  So sles for this.
[04:07] <neuralis> allee: we've got ubuntu certified for DB2; i'm hoping oracle isn't too far off.
[04:07] <Nite_Hawk> neuralis: actually, oracle server support would be a big consideration for us on some (though not all) of our systems too.
[04:08] <neuralis> Nite_Hawk: canonical should have support options on par with redhat for dapper.
[04:08] <allee> Nite_Hawk: fwiw: there were not big stress tests but at least oracle client runs on amd64
[04:08] <neuralis> Nite_Hawk: that won't solve pacify the rabid anti-debian crowd, of course. what's their problem with debian?
[04:10] <neuralis> that's "solve or pacify".
[04:13] <h3sp4wn> If it is just to run oracle what is wrong with solaris x86 ? If you need support sun's is second to none (anything is better than rpm)
[04:28] <allee> Nite_Hawk: I've created some i386 debs for TSM ba client.  Much better than alien deb, nevertheless there are still quite somethings that could be enhanced, aka Works for me quality
[04:46] <Nite_Hawk> neuralis: they haven't been willing to go into it, and I haven't pushed them very hard yet.
[04:47] <neuralis> Nite_Hawk: reluctance and/or inertia doesn't equate with being rabidly anti-debian. which is it? (the ofrmer is obviously a lot easier to deal with.)
[04:48] <Nite_Hawk> neuralis: Well, it's "There's no way we will run debian on these machines and I'm not going to talk about it"
[04:48] <Nite_Hawk> or more "We will run debian over my dead body"
[04:48] <neuralis> Nite_Hawk: i see. buy a gun? :)
[04:50] <Nite_Hawk> neuralis: Still, I may be able to sneak ubuntu-server in if I can present some advantages and they keep refusing to say why they don't like it.
[04:51] <Nite_Hawk> I'm trying to guess what their arguments will be before hand (TSM compat, support, etc) so that I can be ready for them. ;)
[04:51] <Nite_Hawk> And also so that I don't present it too early (IE, oracle should be supported soon, but isn't quite yet)
[04:52] <Nite_Hawk> I'd rather wait 6 months if it means I have a much stronger argument.
[04:53] <neuralis> Nite_Hawk: i don't know if there are any official negotiations with oracle, actually.
[04:54] <Nite_Hawk> neuralis: Hrm... ok.  btw, do you have any kind of rough idea what the costs of a support contract will be like through canonical?  SLES seems to be quite a bit more expensive than RHEL.
[04:56] <Nite_Hawk> gotta run to a meeting, bbl
[04:57] <neuralis> Nite_Hawk: http://www.ubuntu.com/support/supportoptions/paidsupport
[05:46] <nicola_> hallo, I'm planning to install ubuntu 5.10 on a HP Proliant DL380 G4, I've tried to find some certification for the hardware but I could'nt has anybody installed ubuntu on such hardware? Some problems? NIC or other?
[05:47] <fabbione> we do have DL380 at the Ubuntu Datacenter and we do run hmm.. Ubuntu on them.. i guess ;)
[05:50] <nicola_> Ok, that makes me very happy :)
[06:41] <Nite_Hawk> neuralis: any advantages to Canonical paid support vs other distros?
[06:42] <neuralis> Nite_Hawk: you'd want to ask jbailey@canonical.com, likely.
[06:42] <Nite_Hawk> thanks!
[06:43] <neuralis> sure.
[06:43] <intelnux> does ubuntu require anything other then telling bios not to worry for headless operation? or are there some tweeks that need to be made to the system?
[06:52] <neuralis> intelnux: no tweaks, it'll work fine.
[06:53] <intelnux> thanks
[10:54] <nawty> gday matey's :)
[10:58] <Nite_Hawk> ahoy
[10:59] <nawty> :)
[10:59] <nawty> Nite_Hawk: any ideas why ubuntu-server isn't on the lists.ubuntu.com front page?
[11:01] <Nite_Hawk> nawty: huh... no idea. :)
[11:03] <fabbione> a bug in mailman
[11:04] <fabbione> it's not exported for some reasons to the web interface
[11:04] <fabbione> but it's there
[11:07] <spike> do you know of any alternative to acct to get info about commands ran/resource usage per user?
[11:07] <spike> acct isnt being very stable, and I've read of quite a few cases where it misbehaves
[11:08] <spike> I know of other couple accounting patches, wondering if there's any alternative to them
[11:09] <spike> actually what I want to do is monitoring, not billing, so I'm fine with not so high precision
[11:09] <spike> but I cant really think of anything else but accounting patches/sw to extract info like those mentioned
[11:10] <Nite_Hawk> spike: there's http://oss.sgi.com/projects/csa/, but I'm not sure it'd be very useful to you over acct.
[11:10] <nawty> spike: no ideas here. I'm an unfortunate single use system person.
[11:11] <nawty> fabbione: logged it in launchpad ? ;)
[11:11] <Nite_Hawk> spike: I actually am the primary developer on the accounting system for a super computer institute.
[11:11] <nawty> Nite_Hawk: too many toys then ;)
[11:13] <Nite_Hawk> nawty: tell me about it.  Each supercomputer vendor uses their own queueing systems and job schedulars.  IBM's AIX systems don't even provide ascii logs.
[11:13] <spike> Nite_Hawk: I know csa, it was one of the two of the "couple accounting patches" I mentioned above
[11:13] <Nite_Hawk> spike: yeah, I figured as much.
[11:13] <spike> Nite_Hawk: have you actually used it in production?
[11:14] <spike> I played with it but never had a chance to see it even used in production...
[11:14] <Nite_Hawk> spike: Nope.  I don't think it will actually do anything for me over acct.
[11:14] <nawty> Nite_Hawk: don't mention that evil world there :P
[11:14] <nawty> s/world/word/
[11:14] <nawty> s/there/here/
[11:14] <nawty> damn, it's too late.
[11:15] <spike> Nite_Hawk: sounds like you've never had a proble with acct...
[11:15] <Nite_Hawk> spike: Well, it's not so much that I've never had a problem with it, as I've got about 20 fires that are 10 times as big to worry about until I finish rewriting our accounting system. ;)
[11:18] <Nite_Hawk> Right now I'm a lot more concerned with getting accurate data from our job queues.
[11:19] <Nite_Hawk> btw, when I said AIX systems, I specifically meant load leveler, not AIX itself. ;)
[11:22] <nawty> Nite_Hawk: aah, :)
[11:46] <Plug> Hi all.  The Dapper Flight 5 CD detects my Areca SATA-RAID controller and installs to it, but seems the installed kernel doesn't support the controller and can't find the disk?  Anyone got around this?
[11:50] <allee> Plug: check it device name during install and on boot are the same (been there ;). If they are, add the missing driver(s) to /etc/mkinitramfs/modules and dpkg-reconfigure linux-image-<vers>
[11:51] <allee> Plug: and last but not least open a lauchpad bug (I assume pkg hw-detect)
[11:56] <Plug> I will have to see if I can use the installer CD to get a shell on the installed FS :)