[00:54] <vz07mk> Havng trouble with dell poweredge 1400sc install of ubuntu 810
[00:54] <vz07mk> giving incomptible chip error
[00:55] <vz07mk> dell server is an older one with scsi aic 7899 controller
[00:55] <vz07mk> wondering if anyone has senn this b4
[00:56] <vz07mk> error is somethng like 8139c+ is not a compatible chip
[00:56] <vz07mk> then says to try the 8139t00 driver instead
[00:57] <altf2o> NIS + NFS appear to be working just fine, i CAN login as an NIS 'exported' user, however i get: Your home directory /home/<user> doesn't appear to exist, do you want to use / (root) as your home dir? (You likely won't have access to much) , or similar. I haven't found a tutorial that mentions how to make sure the users are using their home dir on a central server? And NOT on the local computer.
[00:58] <vz07mk> alert shows something like /dev/disk/by-uuid/b728fabt...does not exist
[00:58] <vz07mk> dropping to shell
[01:00] <vz07mk> anyone out there?
[01:00] <vz07mk> Is this a server help chat
[01:04] <Bizzeh> hi, i am just about to reinstall my home server, and looking at replacing my hacked together system built on win2k3 (running apache rather than iis).. does ubuntu server force that POS shoshin (sp?) patch on me or does it allow me to run a vanilla php ?
[01:07] <rdw200169> Bizzeh, you mean suhosin?
[01:08] <Bizzeh> thats the one
[01:10] <rdw200169> Bizzeh, yeah, ubuntu's got it, this should help you though: http://ambitonline.com/nextrelease/archives/113-How-to-Ubuntu-PHP-Remove-Suhosin.html
[01:11] <Bizzeh> why isnt it an optional extra?
[01:11] <rdw200169> who knows...
[01:12] <Bizzeh> seems a bit of a dictatorship there.. "your going to use this patch because we say so"
[01:12] <hads> hah
[01:13] <rdw200169> i think it would be a little two difficult to offer both options
[01:13] <rdw200169> i don't know though; all i *do* know is that it can't be an 'option' because of how significant the php patch is
[01:14] <Bizzeh> what? php5.2.x-vanilla and php5.2.x-suhosin as 2 conflicting packages isnt doable?
[01:16] <Bizzeh> does it add features, or significantly change features where that if i was to transplant the app written in php to a non suhosin server, it would break?
[02:11] <altf2o> hmm, update to my question. I have a user on my server (scubasteve) and if i'm on my client and create an empty: /home/scubasteve , then manually: sudo mount <server_ip>:/home/scubasteve /home/scubasteve , i can log in without question! So it appears NIS + NFS is working correctly, it's simply a matter of how to tell the clients to automount /home/<user> from the server?
[02:12]  * altf2o looking @ autofs , but not finding many /decently coherent/ tutorials explaining the process. I'm gonna keep going in the mean time.
[03:37] <altf2o> alright incase anyone is interested, i got it simply by adding: /home /etc/auto.home -rw , in /etc/auto.master , then in /etc/auto.home itself: * <serverIP>:/home/&  , restarted autofs\nis, works exactly as expected! :). This is actually a really neat way of doing things. Now to try and secure\lock it down as much as possible.
[04:31] <Aug> Hi2All
[04:31] <Aug> Iam use Ubuntu Server 8.04. it work all ok. i changed motherboard from ASUS P5B-VM to ASUS P5K-VM... Problem: it is long loading on load modprobe modules... very long on: [    6.221367] ACPI: Power Button (CM) [PWRB] [  185.761627] loop: module loaded    How it tune?
[04:32] <Aug> 3 minutes it freeze on load
[06:40] <PC_Nerd> Hi,  is there any program/daemon etc  that allows me to monitor CPU, memory and network usage of selected || all proccesses running on the machine?
[06:40] <PC_Nerd> ultimately im looking at creating a small web based panel (CGI script) that will display graphs of recent activity, logs of network usage etc.
[06:42] <jtaji> PC_Nerd: nagios, munin
[06:43] <jtaji> they are web based with graphs
[06:43] <PC_Nerd> Thanks, Ill look into them
[09:06] <kraut> moin
[09:21] <quizme> on ubuntu should prefix=/usr instead of prefix=/usr/local  ?
[09:23] <andol> quizme: In which situations?
[09:23] <jpds> quizme: I think all packages install their stuff to /usr.
[09:23] <jpds> Hey andol.
[09:23] <quizme> like for example
[09:23] <andol> jpds: good morning
[09:24] <quizme> i just installed ruby, rubygems, openssl, from source
[09:24] <quizme> and it seems they are going in /usr/local/bin
[09:24] <hads> Why?
[09:24] <quizme> instead of /usr/bin
[09:24] <andol> quizme: If you install yourself from source I'd put them in /usr/local, or somewhere completly else.
[09:24] <_ruben> to avoid conflicts with packaged stuff, i'd put such stuff in /usr/local
[09:24] <quizme> oh
[09:24] <quizme> so it's good that i kept it in /usr/local ?
[09:25] <andol> quizme: Actually, putting them in /usr/local works pretty well with existings paths, library inclussions, etc
[09:25] <_ruben> building from source is highly discouraged though
[09:25] <hads> If you really have to install something by hand then, yeah, use local
[09:26] <quizme> it's just that, it seems like there is a dependency tree, where the packages i install in /usr/local look for dependents in /usr/local as well, so i have to double install all the dependents in /usr/local even if they were previously installed in /usr (thanks to apt-get)
[09:26] <quizme> ok
[09:26] <_ruben> if you need newer versions/specific features .. rebuild an existing package with a newer source tarball .. or create a new package from scratch
[09:26] <quizme> so i guess i'm doing the right thing
[09:26] <_ruben> building from source is bound to cuase a dependency hell later on
[09:26] <quizme> just by accident
[09:26] <quizme> yeah
[09:26] <quizme> i'm experiencing it
[09:26] <quizme> it's just that
[09:26] <quizme> i want to use the latest version of ruby
[09:26] <hads> The easy way to fix it is to not install from source :)
[09:27] <_ruben> the properly buid a .deb for the latest version of ruby
[09:27] <quizme> and i'm on 8.04 and it installs only ruby 1.8.6
[09:27] <_ruben> if you start building from source directly, you might as well install slackware or gentoo
[09:27] <quizme> i don't know how to get the latest .deb
[09:27] <hads> Slack isn't a source distro :)
[09:28] <quizme> how do i get the latest .deb file ?
[09:28] <_ruben> http://www.debian.org/doc/maint-guide/ch-update.en.html
[09:28] <_ruben> either find one on the net, backport the package from intrepid/jaunty, or roll your own .deb as described in the url i pasted
[09:29] <quizme> i'm using 8.04
[09:29] <quizme> should i be using 8.10 if i want to be current like that ?
[09:30] <_ruben> depends on your definition of "current"
[09:30] <hads> Does ruby change that much?
[09:30] <_ruben> 9.04a3 probably has even newer packages
[09:30] <quizme> mainly i want to use Ruby 1.8.7
[09:30] <quizme> hads: not really, but i'm OC like that
[09:30] <hads> Oh, then just don't be.
[09:31] <quizme> hads: they recommend using 1.8.7 on the ruby on rails main page
[09:31] <quizme> that's the main reason
[09:31] <barduck> Hey...can anyone recommend good bittorrent client that can run as daemon without a gui and and has built-in web interfance for remote controlling ?
[09:36] <lukehasnoname> rtorrent might?
[09:37] <barduck> it doesn't seem to have a web interface though
[09:37] <hads> transmission+clutch
[09:40] <barduck> hads: I'll check, thanks
[09:56] <quizme> got disconnected :(
[10:01] <quizme> how do u install from a deb file if apt-get is not giving you what you want?
[10:02] <quizme> you just need to download a .deb file?
[10:23] <quizme> who is smarter? the guys who built RoR or the people who use RoR ?
[10:30] <domas> quizme: neither?
[10:31] <quizme> domas: you are correct, you win $10,000
[10:32] <domas> transfer them to paypal account donate@wikimedia.org
[10:33] <quizme> hehe
[10:33] <quizme> ok
[10:33] <quizme> ummm later
[10:33] <_ruben> if only we could properly integrate mediawiki into our sharepoint environment .. damn windows network :p
[10:46]  * [gnubie] waves
[10:46] <[gnubie]> for production setup, do you use LVM on the database partition? if yes, does it have a big impact with its disk i/o or tolerable? please share your ideas. thanks in advance.. ;)
[10:52] <gate_keeper_> any recommends for ssl jabber server and good tutorial too?
[10:52] <quizme> [gnubie] i heard that an xfs freeze stops database activity for around 10-15 seconds.
[10:52] <domas> [gnubie]: LVM snapshot is 5x slowdown
[10:52] <domas> quizme: no need to do xfs freeze
[10:52] <andol> gate_keeper_: I like jabberd2. Its main documentation is pretty good.
[10:52] <domas> at least with new kernels
[10:53] <domas> [gnubie]: I run special LVM-based replication slaves I do snapshots on, not on main machines
[10:53] <[gnubie]> gate_keeper_: look for the ejabberd
[10:53] <domas> [gnubie]: and... if you go for high performance/high concurrency environment, LVM can add 20% overhead in normal operation
[10:54] <[gnubie]> quizme and domas: does it mean that i will not use LVM for database storage?
[10:54] <domas> um, maybe 10%, not 20%, I'd have to rerun my benchmarks
[10:54] <domas> depends on your load
[10:54] <quizme> [gnubie] i don't know about this stuff.  ask domas
[10:54] <domas> is your machine doing 10000 queries a second?
[10:54] <domas> 10000 write transactions a second, that is
[10:54] <domas> :)
[10:54] <gate_keeper_> ok, thx guys
[10:55] <[gnubie]> quizme: i'm still about to setup a database server
[10:55] <domas> LVM gives you quite some freedom
[10:55] <quizme> [gnubie] i run on ec2, do you use that
[10:55] <domas> but snapshots cost
[10:55] <[gnubie]> domas: sms targetting (hopefully) a million sms a day
[10:55] <domas> [gnubie]: it all depends on your application :)
[10:55] <domas> anyway, if you can tolerate 5x lower i/o performance, you can use LVM snapshots
[10:56] <domas> if you can tolerate 5% lower i/o performance in general cases, LVM is all good
[10:57] <[gnubie]> domas: if let's say, i will not use lvm.. i will have raid-1.. if in case the hard disks are almost used up, how do you extend or migrate to a bigger drives without or limited downtime?
[10:58] <domas> you grow your RAID array :)
[10:58] <domas> quite a few RAID controllers support online growing
[10:58] <[gnubie]> domas: oh, there will be a downtime for adding 2 hard drives and removing the old ones afterwards
[10:59] <[gnubie]> domas: you mean, just add another 2 hd drives
[10:59] <[gnubie]> ?
[10:59] <domas> yes
[10:59] <domas> hehe, some or our db boxes: http://commons.wikimedia.org/wiki/File:Wmf_sdtpa_servers_2009-01-20_42.jpg ;-)
[11:00] <[gnubie]> domas: does it mean that there will be 4 HDDs configured as RAID-1?
[11:00] <domas> raid 10
[11:00] <domas> or raid 0+1
[11:00] <domas> depends how you call it
[11:00] <domas> :)
[11:02] <[gnubie]> domas: currently, i will be planning to set the 2 HDDs as RAID-1.. does it mean that later if i add 2 additional HDDs, i can easily add these 2 HDDs to the RAID 1 setup?
[11:02] <domas> depends on your controller
[11:02] <domas> I wonder how kernel would detect a volume size increase
[11:02] <domas> might need a reboot :)
[11:02] <domas> of course, you can use LVM
[11:03] <domas> hehehe, we do manage servers in a bit different way
[11:03] <domas> any hardware change = immediate reinstall
[11:03] <domas> with full data wipe
[11:03] <[gnubie]> domas: btw, do you prefer hardware raid 1 over software raid 1?
[11:03] <domas> yes, I prefer hardware raid
[11:04] <domas> software raid for databases is shooting into one's foot
[11:04] <[gnubie]> domas: i wonder red hat prefers software raid and lvm for their setup.. any idea?
[11:04] <domas> do they?
[11:04] <domas> LVM makes managing easier
[11:04] <[gnubie]> domas: not sure.. but to think that they are actively promoting software raid and lvm even on their docs
[11:05] <domas> [gnubie]: software raid is probably better than no raid
[11:05] <domas> could be a reason for promotion
[11:05] <domas> hardware raid is much much much better
[11:05] <domas> there's one very simple reason
[11:05] <[gnubie]> domas: i agree..  ;)
[11:06] <domas> write-behind caching with battery
[11:06] <domas> you can't beat that with software ;-D
[11:06] <[gnubie]> domas: what raid controller can you recommend if the server is just a clone amd64 based server?
[11:06] <domas> *shrug*, any expensive one ;-)
[11:07] <[gnubie]> domas: what are those?  ;)
[11:07] <domas> adaptec,lsi,etc :)
[11:07] <[gnubie]> domas: 3ware?
[11:07] <domas> didn't have good experiences with 3ware :)
[11:08] <domas> oh, and had some bad :)
[11:09] <[gnubie]> i see..
[11:09] <domas> didn't have too many 3ware based boxes
[11:10] <domas> but random corruption isn't something I like to keep in-house :-)))
[11:10] <[gnubie]> going back, i wonder if adaptec or lsi raid controllers supports online extending of raid 1 arrays
[11:10] <[gnubie]> too bad..
[11:11] <domas> ghm, ones I have I guess do
[11:11] <[gnubie]> domas: what do you have in there?
[11:12] <domas> we got a batch of Sun X4240s shipped with 'sunraid' raid controller, which is Sun branded Intel-manufactured stuff with Adaptec software ;-D
[11:13] <[gnubie]> domas: how sweet.. ;)
[11:13] <[gnubie]> domas: if i can just go back from the past..  ;)
[11:14] <[gnubie]> domas: maybe i will try to check their websites and see..
[11:15] <[gnubie]> domas: another question.. in the early 200x, i used ReiserFS for Maildir partitions and XFS for database.. nowadays, what can you recommend?
[11:16] <domas> hhahahaha, I remember very nice reiserfs corruption on some maildir hosts ;-)
[11:16] <domas> I prefer XFS
[11:19] <[gnubie]> domas: i heard that xfs has been discontinued
[11:19] <domas> by whom?
[11:19] <[gnubie]> domas: is that true?
[11:19] <domas> join #xfs and you'll find most lively filesystem development channel ;-)
[11:20] <[gnubie]> ok
[11:20] <domas> anyway, XFS is the only linux filesystem that supports concurrent DIO ;-)
[11:21] <domas> if you have hardware raid, make sure you turn off write barriers though :)
[11:22] <[gnubie]> domas: write barriers? where can i turn it off?
[11:22] <domas> 'nobarrier' mount option
[11:23] <domas> you may want to use it if you have direct disk access, and have write-behind caching at _disk_ level
[11:23] <[gnubie]> domas: ok.. thanks for the advice..
[11:40] <andol> domas: next time I have to do disk optimization I'll have to remember to ask for your advice :)
[11:40] <domas> andol: just read my blog
[11:40] <domas> most of my 'wisdom' is there ;-)
[11:41] <andol> domas: nice, url?
[11:41] <domas> dammit.lt, or google for 'domas'
[11:41] <domas> ;-)
[11:41] <andol> thanks
[11:43]  * andol takes a closer look at the article on ZFS and MySQL...
[11:43] <domas> hahaha, caused some internal noise with this :)
[13:26] <palt> I have a ubuntu server which corrupts big files. I have some dvd isos that doesn't get the right md5sum. Small files say 200MB however works just fine.
[13:26] <palt> I have 2 discs in a software raid 1.
[13:27] <palt> I have tried several dvd's and both wget them and scp from another box, but they still gets corrupt
[14:38] <geekboxjockey> good morning
[14:39] <geekboxjockey> I was wondering if anyone here who runs Xen could answer a quick question for me,
[14:39] <ScottK> !ask > geekboxjockey
[14:41] <geekboxjockey> I am using disk images for Xen, not LVM and was wondering what others may recommend for VM backups with this setup. (I am fine with full machine image backups, not sure about snapshots)
[15:08] <jdstrand> didrocks: re around?> I am now :)
[15:10] <jtaji> geekboxjockey: I do a mysqldump then rdiff-backup the whole system from my home machine
[15:11] <jtaji> daily
[15:14] <geekboxjockey> jtaji: so rdiff-backup is incremental and works with a .img right?
[15:15] <geekboxjockey> so if i have 20-50  gig images, and run rdiff-backup for say 3-4 machines, would that cripple operations for the duration of the backup?
[15:19] <jtaji> geekboxjockey: it's incremental in that it stores diffs to go back in time
[15:21] <jtaji> it doesn't deal with the image, it runs on the running filesystem
[15:22] <jtaji> geekboxjockey: example: http://www.linode.com/wiki/index.php/AutomatedRdiffBackup
[15:24] <geekboxjockey> so if i need to restore, i just drop files in a fresh image?
[15:24] <geekboxjockey> thank you for the link :)
[15:26] <jtaji> geekboxjockey: I guess if you have access to dom0 you can just backup the images... I have never done it from that end
[15:28] <jtaji> I think Linode is working on implementing that type of backup solution for their customers
[15:33] <geekboxjockey> the only issue with image backup is that the images are the full size of the disk, regardless of the actual usage size
[15:33] <geekboxjockey> ive seen some documentation for LVM but was advised to stay away from LVM
[15:41] <padhu> Dear Xperts
[16:05] <robertj> I installed 8.04.1 Server x64 on a R200 and it all works fine but I don't have video if I leave it unplugged and plug it back in a while later
[16:05] <robertj> (like a month)
[16:07] <Faust-C> anyone tried out samba4 yet
[16:07] <palt> Isn't it in alpha version still?
[16:08] <Faust-C> yeah but ive read of ppl actually using it (idk how much truth is to it being used in production)
[16:09] <palt> Cool, well I guess samba alpha is quite stable, but I haven't tried it yet :/
[16:09] <Faust-C> theyve had success w/ it being a DC for windows systems
[16:10] <Faust-C> like a 2003 DC not 2000 mixed mode
[16:37] <isaac> uhm, I am converting LDAP schema files to LDIF as explained in https://help.ubuntu.com/8.10/serverguide/C/openldap-server.html
[16:37] <isaac> and it's splitting the lines in a weird way
[16:37] <isaac> is that normal?
[16:38] <isaac> example: http://paste.ubuntu.com/109874/
[16:42] <sommer> isaac: yes, it should be fine
[16:42] <isaac> sommer: ok, thanks :)
[16:42] <sommer> isaac: I believe as long as there's a space at the start of a new line it's valid LDIF
[16:43] <isaac> great :)
[16:45] <Blah78> anyone knows how to make fb work with ubuntu-server in a vmware? Can't find /dev/fb0
[16:50] <phreestyle-work> hello all. I'm having a very strange problem with postgresql: everytime I reboot the server, all data in the tables is gone. All the sequences are still intact but none of the data is...any thoughts?
[16:50] <incorrect> how much disk space do i need to build my own repository?
[16:54] <andol> incorrect: A complete mirror, or just a repo for a few packages you've built yourself?
[16:54] <incorrect> complete mirror
[16:54] <incorrect> i am going to have a server farm of ~500 systems,  thought it would be easier for me to mirror
[16:55] <andol> incorrect: http://www.ubuntu.com/getubuntu/mirror/2
[16:56] <incorrect> that is nothing,  i had 4TB reserved for it
[17:22] <crackintosh> can I run a cron job as www-data? do I just change the user in /etc/crontab or do I have change www-data's cron
[17:28] <incorrect> crackintosh, sure, give it a shell then su - www-data
[17:30] <Deeps> crackintosh: you can do it eitherway. using /etc/crontab or /etc/cron.d/ is easier to maintain later, than having separate per-user crontabs though, especially for services
[17:31] <crackintosh> Deeps: I have some php software that requires me to run a cron script: *  *    * * *   www-data    cd /var/www/sugarcrm; php -f cron.php > /dev/null 2>&1
[17:31] <crackintosh> I have never seen a cron with all asterisks
[17:32] <Deeps> means it runs every minute
[17:32] <Deeps> of every hour
[17:32] <crackintosh> perhaps I should change it to: 1-59  *    * * *   www-data    cd /var/www/sugarcrm; php -f cron.php > /dev/null 2>&1
[17:32] <Deeps> of every day of the month
[17:32] <incorrect> you can set */1 just to make it easier to read
[17:32] <Deeps> or just *
[17:32] <crackintosh> the script keeps failing.
[17:32] <Deeps> failure isnt due to the asterisks
[17:32] <Deeps> redirect output to a file
[17:33] <Deeps> stdout and stderr into a /tmp file
[17:33] <Deeps> *  *    * * *   www-data    cd /var/www/sugarcrm; php -f cron.php > /tmp/sugarcron 2>&1
[17:33] <Deeps> and see why it's failing
[17:33] <Deeps> your asterisks are not the problem
[17:34] <incorrect> run it manually
[17:35] <incorrect> when you are happy set up the cronjob
[17:35] <incorrect> also are you sure your script runs in less than 1 minute?
[17:36] <incorrect> you aren't running hundreds of them ?
[17:36] <Deeps> if that does occur, file a bug with sugarcrm as they should account for that in their cron.php
[17:37] <incorrect> i once saw this guy who had an rsync job on cron, one day it look longer than 5 minutes,, that was fairly amusing
[17:37] <incorrect> food
[17:37] <incorrect> bye
[18:40] <crackintosh> Apparently cron was setup correctly, something within the application was not
[19:07] <Deeps> crackintosh: i believe i said that ;)
[19:37] <DawnLight> hello. i'm trying to set up exim4 as an MTA for an "internet site". i'm not sure why, though blindly following the simple instructions, exim doesn't ask for smtp authentication.  how do i change this and why is this default behaviour?
[19:42] <Deeps> !serverguide | DawnLight
[19:42] <Deeps> covers how to enable smtp auth
[19:42] <Deeps> including over ssl/tls
[19:42] <Deeps> as to why it's the default, good question, seems a bit silly really, there may be some info in /usr/share/doc/exim*
[19:44] <Deeps> then again, it may not relay mail for domains: * by default
[19:44] <Deeps> which would explain why
[19:49] <DawnLight> Deeps: i'm sorry. i've followed https://help.ubuntu.com/8.04/serverguide/C/exim4.html only. i don't see anything else and in this document i don't see anything about requiring passwords, which is what exim isn't currently doing
[19:50] <Deeps> SMTP Authentication
[19:50] <Deeps> This section covers configuring Exim4 to use SMTP-AUTH with TLS and SASL.
[19:50] <Deeps> mind, you may need to change that url to reflect the version of ubuntu you're using, heh
[19:50] <Deeps> ie if you're using intrepid, change the 8.04 to 8.10
[19:52] <DawnLight> heh - my bad. the gnome thingo was keeping my password for that account even though i've deleted that account from evolution... it does ask for a password. smart exim! love exim!
[19:52] <DawnLight> thanks Deeps
[19:52] <Deeps> no problem
[19:52] <DawnLight> :)
[19:52] <DawnLight> ciao
[20:41] <melter> is there an alpha 3 server release of jaunty?
[20:59] <martin___> hi all
[20:59] <martin___> hope someone can help me
[20:59] <martin___> i try to make a linux software raid
[20:59] <martin___> raid 5 to be exact
[21:00] <ivoks> ok
[21:00] <martin___> something strange happens
[21:00] <martin___> when i create the raid 5 with my three disks all seems fine but when i reboot the 3rd drive is kicked out "failed"
[21:01] <ivoks> did you check dmesg?
[21:01] <martin___> it is not the disk (tried 2 brand new ones and moved between slots
[21:01] <knoppix> did you tryed remounting redecting somehow ?
[21:01] <martin___> what is interesting is that 2 of the disks are on the motherboard; standard sata II
[21:01] <martin___> the other disk is on a sil 3124 RAID PCI controller
[21:02] <martin___> that is the disk that gets kicked out at reboot
[21:02] <martin___> I see with lsmod also that dmraid is active (although i don't want and use it)
[21:03] <martin___> could that somehow conflict with mdraid?
[21:03] <martin___> ivoks: yes but i could only see it being recognized as failed
[21:03] <martin___> knoppix: Yes added, repartitioned and tried again several times
[21:04] <martin___> it is a long proces rebuilding; almost 8 hours per try
[21:04] <ivoks> martin___: right, purge dmraid
[21:04] <ivoks> you don't have to wait for it to rebuild
[21:04] <martin___> ivoks: I wish I would know how.
[21:04] <ivoks> apt-get purge dmraid
[21:04] <martin___> I entered in grub nodmraid
[21:05] <martin___> ivoks: I removed that dmraid package but the modules are still loaded by ubuntu server
[21:05] <knoppix> when power gets late they are ignored by linx
[21:05] <knoppix> i see this problem commonly
[21:05] <knoppix> but im not sure if its some as your pblem
[21:06] <ivoks> martin___: add /etc/modprobe.d/my-custom-config
[21:06] <ivoks> martin___: and put in it 'blacklist dmraid'
[21:06] <ivoks> martin___: run depmod
[21:06] <ivoks> martin___: and update-initramfs -u
[21:07] <knoppix> anyway what im talking about is not linux spesific
[21:07] <knoppix> its just about bioss detect time..
[21:07] <martin___> ok
[21:07] <martin___> going to write that down
[21:08] <martin___> what should i put in my-custom-config file?
[21:08] <ivoks> blacklist dmraid
[21:09] <martin___> btw: the modules i talk about are called dm_mirror dm_log and other two i forgot
[21:09] <ivoks> you need those
[21:09] <martin___> so I should enter lines like:
[21:09] <martin___> blacklist dm_mirror
[21:09] <ivoks> silly me
[21:09] <martin___> ?
[21:10] <ivoks> dmraid module doesn't exist at all
[21:10] <ivoks> martin___: you need dm_* modules
[21:10] <ivoks> martin___: that's software raid
[21:10] <martin___> can you use that wildcard or specify them one by one?
[21:11] <ivoks> martin___: don't blacklist them
[21:11] <ivoks> martin___: i was wrong, dmraid module doesn't exist
[21:11] <kaje> I changed the permissions of /etc/sudoers to 640 in an attempt to edit it and now I can't sudo to anything because it complains the permissions are wrong. I can't even sudo chmod 440 to fix it... How do I fix this?
[21:11] <ivoks> kaje: add init=/bin/bash argument in grub
[21:12] <ivoks> fix it and reboot
[21:12] <ivoks> martin___: everything is fine with your modules
[21:12] <kaje> ugh, there is no way to fix it without rebooting? this is a server I'd rather not take down...
[21:12] <maxb> Or, acquire root by another means, e.g., ssh, su
[21:12] <martin___> ivoks: but aren't those modules i mention like dm_mirror for fakeraid (so raid on a software raid card) instead for linux software raid?
[21:12] <ivoks> martin___: put output of dmesg on paster.ubuntu.com
[21:12] <ivoks> martin___: no, that's linux software raid
[21:13] <ivoks> dm is from device mapper
[21:13] <martin___> ah ok
[21:13] <martin___> then i am going to reboot my newly created raid 5 with 4 disks now
[21:13] <ivoks> ok
[21:14] <martin___> before i do, any other interesting logs i should safe after creating the raid before rebooting?
[21:15] <ivoks> no, just reboot
[21:16] <martin___> going down :)
[21:19] <martin___> while booting another complete different question: during boot i get a message like about something: CPU does not support KVM extensions
[21:19] <martin___> any idea what that is?
[21:19] <martin___> i have an atom processor in it
[21:19] <ivoks> atom doesn't support kvm
[21:19] <ivoks> so, that message is normal :)
[21:19] <martin___> so no problem, just a message, thx
[21:20] <ivoks> kvm is for virtualisation
[21:20] <martin___> 64bit ubuntu didn't work also while the specs says 64bit capable
[21:20] <martin___> ok it has rebooted
[21:20] <martin___> going to pastebin
[21:21] <ivoks> martin___: which atom is that?
[21:22] <ivoks> not every atom supports 64bit
[21:22] <martin___> it is a different one as in my eeepc
[21:22] <martin___> becuase it shows 4 CPUs in system monitor
[21:23] <ivoks> that's doesn't mean anything
[21:23] <ivoks> 64bit atoms are atom 230 and atom 330
[21:23] <martin___> model name	: Intel(R) Atom(TM) CPU  330   @ 1.60GHz
[21:23] <martin___> so a 330
[21:24] <martin___> but after grub loaded initrd with the setup ubuntu server crashed
[21:24] <martin___> so I had to install 32bit
[21:25] <ivoks> did you try live cd?
[21:25] <martin___> nice device but I run into some issues with it and ubuntu :( (no 64bit, raid issue). It is a BBS2 from tranquilPC
[21:26] <ivoks> raid issues have nothing to do with processor
[21:26] <martin___> Yes, I booted from USB stick (only option, no CDROM avaialbe)
[21:26] <martin___> btw paster.ubuntu.com gives address not found
[21:26] <ivoks> paste.ubuntu.com
[21:27] <martin___> that's better :)
[21:29] <martin___> ok it's pasted
[21:29] <martin___> under Martin
[21:30] <martin___> For completeness I also now pasting the dmesg from before the reboot where you can see the creation of the raid array is complete
[21:30] <ivoks> could you paste the link?
[21:31] <ivoks> martin___: link?
[21:31] <martin___> aft reboot: http://paste.ubuntu.com/110019/
[21:32] <martin___> ivoks: before reboot: http://paste.ubuntu.com/110020/
[21:35] <ivoks> raid5: raid level 5 set md0 active with 4 out of 4 devices, algorithm 2
[21:35] <ivoks> ?
[21:35] <ivoks> looks ok to me
[21:35] <martin___> yeah: just did a cat /proc/mdstat
[21:36] <martin___> this is strange, very strange
[21:36] <martin___> itried for days
[21:36] <martin___> it seems using 4 instead of 3 disks helped
[21:36] <ivoks> 'before reboot' paste was... what?
[21:36] <ivoks> these are different configurations
[21:36] <martin___> the dmesg just after creating the array with 4 disks instead of 3
[21:37] <ivoks> you mean before?
[21:37] <ivoks> couse, i can see here two raids
[21:37] <ivoks> one mirror and one raid5
[21:38] <ivoks> i'm talking about http://paste.ubuntu.com/110020/
[21:38] <martin___> that's right: that's another try i did
[21:38] <ivoks> there's mirror of sdb1 and sdc1
[21:39] <martin___> after trying to build raid 5 with 3 disks several times (that failed) I made a RAID0 on the sil3124 controller disks
[21:39] <ivoks> and raid 5 of 3 disks, but only two are found; sde1 and sdf1
[21:40] <martin___> i did that to expiriment (out of sensefull options)
[21:40] <martin___> another thing i did was assigned as partition type linux raid to all disks instead of linux
[21:41] <martin___> so compareds to the failing config with 3 disk (where one was marked failed) it either helped i changed partition types or using 4 instead of 3 disks in the array
[21:41] <martin___> What do you think?
[21:42] <martin___> oh yes, i also made a chunk size of 128KB instead of 64K now
[21:43] <ivoks> err...
[21:43] <ivoks> noting failed in these logs
[21:43] <martin___> yes, as i said
[21:43] <martin___> it did in the situation with 3 disks but i don't have the log
[21:43] <ivoks> oh
[21:44] <martin___> going to create ext3 on the array now
[21:44] <ivoks> it's hard to say why it failed, if there aren't any logs
[21:44] <martin___> that is mkfs.ext3 /dev/md0 right?
[21:45] <martin___> ivoks: i understand, till now i am just happy it seems to work now after reboot. But I am very sorry that i asked for help of you while it works now. I assumed too much my last fiddling didn't help after having fiddled for days now :)
[21:46] <martin___> ivoks: but your help is very much appreciated anyway
[21:46] <ivoks> anyway, a friend of mine does have 64bit ubuntu on atom 330
[21:47] <martin___> that is strange huh? I really got a dump message. Well it is not that big of an issue cause it is just a home server
[21:47] <martin___> i think performance wise it won't make a real big difference
[21:48] <ivoks> probably, yes
[21:49] <martin___> is mkfs.ext3 /dev/md0 the right command?
[21:51] <mathiaz> kirkland: are you looking into the iscsi SRU for hardy? https://bugs.launchpad.net/ubuntu/hardy/+source/open-iscsi/+bug/236640
[21:51] <kirkland> mathiaz: not particularly
[21:51] <ivoks> martin___: yes
[21:51] <ivoks> mathiaz kirkland ;) hi
[21:52] <kirkland> ivoks: howdy
[21:52] <mathiaz> ivoks: bonjour!
[21:52] <ivoks> mathiaz: ca va?
[21:52] <martin___> ivoks: thx again m8
[21:52] <mathiaz> kirkland: well - you're assigned to the bug
[21:52] <mathiaz> ivoks: oui - et toi?
[21:52] <ivoks> mathiaz: bien, merci
[21:52] <mathiaz> kirkland: and slangasek assigned another iscsi SRU for hardy to me.
[21:53] <mathiaz> kirkland: so it seems one of us should look into doing one SRU for hardy
[21:53] <kirkland> mathiaz: i won't stand in your way, if you're volunteering ;-)
[22:09] <binspace> How do you do a non-interactive install of postfix? I already tried export DEBIAN_FRONTEND=noninteractive.
[22:10] <lamont> binspace: mostly it's with a crowbar
[22:11] <lamont> used as a bludgeoning instrument
[22:11] <binspace> lamont: Aah, so expect?
[22:11] <lamont> it _can_ be preseeded, but it's not pretty
[22:11] <lamont> binspace: nah - you have to tell it answers
[22:11] <lamont> so you _could_ use expect, but preseeding is easier
[22:11] <lamont> it just kinda doesn't make it trivial
[22:11] <binspace> lamont: What do you mean by preseeding?
[22:12] <lamont> binspace: providing the debconf answers to debconf in a file is generally done through the process called 'preseeding'
[22:12] <lamont> wherein you hand debian-installer a preseedfile
[22:12] <lamont> learning how to do that is on my todo list. :=(
[22:13] <binspace> cool. we're already doing this for mysql-server, but what are the debconf options for postfix?
[22:13] <lamont> that's the part that postfix is not exactly friendly about
[22:13] <binspace> bummer.
[22:14] <lamont> I had my reasons for making it unfriendly, but it's still unfriendly in that way
[22:14] <binspace> So how can I automate the install?
[22:15] <lamont> maybe bugs.debian.org/436294 has ideas?  dunno
[22:16] <binspace> ok. For now, I'm thinking apt-get install postfix <<< '\n\n\n'
[22:18] <binspace> lamont: Thanks. I'll give it a shot.
[22:39] <Baptiste_Manson> hi there
[22:40] <Baptiste_Manson> i'm trying to find a good dedicated hoster in USA
[22:40] <Baptiste_Manson> with ubuntu server
[22:40] <Baptiste_Manson> any tips ?
[22:47] <Baptiste_Manson> To sum up I'm trying to find a hoster in USA, with dedicated servers under ubuntu servers for app. 10 servers
[22:50] <_Cid> Baptiste_Manson:  no idea here - I thought several of the providers out there allowed you to determine your own OS?
[22:50] <_Cid> I run a "hobby" server myself (of a dynamic IP now, eeew)   you could do the same to :P
[22:51] <Baptiste_Manson> _Cid: not really, the US server market seems mainly centered around centos and redhat
[22:52] <_Cid> Baptiste_Manson:  I am sure you researched, but top google hit: http://www.m5hosting.com/ubuntu-dedicated-server.php :)
[22:53] <_Cid> oh, dyndns offers ubuntu as an option - those guys rock:  http://www.dyndns.com/services/springserver/
[22:53] <Baptiste_Manson> _Cid: right, it was somehow to see if one of the american hoster was far better than others from your experience
[22:54] <_Cid> Baptiste_Manson:  never used dyndns for hosting, but using a few different other services of theirs, I like them .... shooting in the dark, I might go with them ;-)
[22:54] <Baptiste_Manson> _Cid: thx
[23:02] <ScottK> Baptiste_Manson: I use http://softlayer.com/ and have been happy with them.
[23:03] <Baptiste_Manson> ScottK: thx
[23:07] <kaje> Does anyone know how to make module-assist pass a -y option to apt-get so I don't have to accept the packages it wants to install every time?
[23:08] <Baptiste_Manson> Thx for the testimonials about hosters providing dedicated servers under ubuntu server in USA (I got 3 names now : m5hosting, softlayer and dyndns)
[23:10] <kaje> cari.net also does, but they have $75 setup fee if you choose ubuntu (which I think is dumb)
[23:10] <kaje> I use cari.net for a server and they have been very professional and reliable...
[23:11] <maxb> ubuntu, as opposed to?
[23:11] <kaje> as opposed to fedora, suse, and one other I can't remember
[23:11] <maxb> how peculiar
[23:11] <maxb> What's to stop you just reimaging as soon as you're set up?
[23:11] <kaje> yeah, I agree... it is probably just that they don't have a lot of experience with it...
[23:11] <erichammond> Baptiste_Manson: You might consider Amazon EC2.  I maintain Ubuntu images for EC2 listed on http://alestic.com and there is a support group at http://ec2ubuntu-group.notlong.com
[23:12] <kaje> well, nothing I guess... except how would you reinstall ubuntu through an SSH connection?
[23:13] <kaje> hmm, maxb... that's an interesting idea...
[23:13] <kaje> do you know of someone that has been able to remotely image a system and blogged about it?
[23:14] <maxb> The systems guys where I work have done ~10 remote slackware->debian overwrites.
[23:15] <maxb> I think via the magic of "borrowing" the swap partition to be a minimal temporary OS
[23:17] <kaje> that's a pretty cool hack =)
[23:21] <Baptiste_Manson> erichammond: amazon ec2 can be a good technical solution though, sadly it is really hard to say to a client that their data will be hosted by amazon
[23:22] <binspace> ec2?
[23:22] <Baptiste_Manson> binspace: Elastic Compute, this is cloud computing as far as I know
[23:22] <binspace> sorry. Yes, I recommend ec2 :-)
[23:22] <binspace> or I've used it and have been happy
[23:23] <binspace> lamont: Do you know which config file the postfix system mail name is stored in?
[23:24] <binspace> Actually, I'm trying to set the postfix debconf, and run into the System mail name prompt
[23:24] <lamont> binspace: /etc/mailname, per policy
[23:24] <binspace> And I'm using the template from bugs.debian.org/436294
[23:25] <binspace> thanks
[23:35] <martin___> HI all, anyone who can help me with a RAID issue
[23:36] <martin___> I create a linux software RAID5 with 4 1TB disks, chunk size 128K and ext3 filesystem on it
[23:36] <martin___> Now the issue is I have really reasonalbe read speeds but unacceptable slow write speeds: Writing is only done at 2,8MB/s
[23:42] <martin___> none who can tell me why I have such slow write speeds on my RAID5?
[23:57] <thehook_> Hello :) I am making some simple scripts to setup and configure a mailserver, and was wondering if it's possible to "supress" the big blue screen asking for input during a scripted apt-get installation of mysql-server for example? And maybe have them as variables in the script or something..
[23:59] <thehook_> I have also made such script for installing a nginx web server with php support and mysql. And would really appreciate some feedback :) You can find it on http://insecure.no/2009/01/easy-nginx-php-and-mysql-installer-script/