[00:44] <saidian> Hello. I'm having trouble mounting a Mac share on Ubuntu server and was wondering if anyone could help out?
[01:15] <kees> has anyone seen broken fonts in cacti after upgrading to 11.04?
[01:16] <jj995> Xlib:  extension "RANDR" missing on display ":0.0". ---- I'm getting this message on the terminal whenever I run an X application, e.g. gedit.  The X application works fine, but I'd like to get rid of this RANR message.  Any ideas?  I tried removing libxrandr2 from Synaptic, but it seems like tons of things depend on it that I want to keep
[02:10] <kexman> hello
[02:11] <kexman> is there any nice 1 liner command for ubuntu server latest, to strip it down to the bare minimum ? like if i selected LAMP or something so that should be gone ... i want a minimal ubuntu install with openssh only but dont want to reinstall
[02:16] <twb> Try aptitude markauto '?installed ?not(?section(metapackages)) ?not(?name(openssh-server))'
[02:16] <twb> But be sure to review the propsed change before you approve it, in case there's extra stuff in there
[03:23] <kexman> what is the difference between what ubuntu minimal installs as "core" and ubuntu-server without selecting any additional packages like lamp, openssh, postgres and such
[03:25] <twb> Zero
[03:25] <twb> Well, possibly the kernel flavour
[03:26] <twb> If you do a ubuntu-server install and keep clicking next, you will get ubuntu-minimal + ubuntu-standard + server kernel flavour + stupid 200MB of thunderbird/firefox/oo.org english language packs
[03:27] <kexman> twb: duhhh
[03:27] <kexman> i dont need all that
[03:27] <twb> So do an expert install
[03:28] <twb> Uncheck "standard packages"
[03:28] <kexman> i just want ubuntu-sver = ubuntu-minima-server :) no X involved no nothing that i dont need :)
[03:28] <kexman> like gentoo minimal :P
[03:28] <twb> Or do it from apt when you're done
[03:28] <twb> kexman: uh, none of what I said included X
[03:28] <kexman> twb: thats good :D
[03:28] <kexman> i didnt say you did :)
[03:28] <twb> ubuntu-standard is stuff like less and w3m
[03:29] <kexman> 200mb thunderbird stuff is what i dont need :)
[03:29] <twb> kexman: yeah, that pisses me off, too
[03:29] <kexman> twb: so standard i do need :) those are really basic
[03:29] <kexman> twb: how to remove it ?
[03:29] <twb> aptitude remove ~i~nlanguage-pack IIRC
[03:30] <twb> Oh, but then you'll need to configure locales by hand, which is something like "locale-gen en_AU.UTF-8; set-locale LANG=en_AU.UTF-8"
[03:30] <kexman> omg :) that hurts :P no click of button for that :P joking :)
[03:30] <kexman> twb: thanks a lot
[03:30] <twb> I think in theory if you are using the CD's preseed file, one of the options disables those
[03:30] <kexman> so other then that ubuntu-standard is nice to have right ? i dont want "cant" use it minimal :)
[03:31] <kexman> twb: hmm ? dunno what your talking about i just clickclickclick without not even selecting openssh right now
[03:31] <kexman> 5,689 kB will be freed :S
[03:31] <twb> kexman: if you boot the installer from the network, you don't get a preseed file by default.
[03:31] <kexman> no 200mb here :)
[03:32] <twb> kexman: so the installer can't "tell" if it's a server or desktop install
[03:32] <kexman> twb: i downloaded a ubuntu-server iso
[03:33] <kexman> btw i need to do set-locale only 1 time and im done ?
[03:33] <kexman> also no set-locale command here ...
[03:33] <twb> kexman: don't bother trying that unless "locale -a" reports only the "C" locale
[03:33] <twb> It's supposed to be automated
[03:34] <kexman> it reports C locale
[03:34] <kexman> twb: i removed the language packs and i generated the en_US.UTF-8 locale
[03:34] <twb> kexman: *only* the C locale?
[03:34] <kexman> now id like my system to use it
[03:34] <kexman> no
[03:34] <twb> Hang on, lemme look up what I do
[03:34] <kexman> en_US.UTF-8 + POSIX
[03:35] <twb> update-locale LANG=${LANG:-en_US.UTF-8}
[03:35] <twb> update, not set
[03:37] <kexman> twb: thanks
[03:37] <kexman> twb: so this will remain no need to do anything more ? its permanent right ?
[03:38] <twb> Yeah, you should be done now
[03:38] <twb> If you are logged in, you might need to log out and back in
[03:38] <twb> You might want to "aptitude purge ~c" too
[03:39] <kexman> twb: whats that for ?
[03:39] <twb> Removes config files of uninstalled packages
[03:45] <keyboardtalk> I am trying to set up userdir with 10.04 and apache2. I am getting a permissions error when I try to access user directories. Any help?
[03:46] <twb> keyboardtalk: root_squash?
[03:48] <keyboardtalk> twb: what is that?
[03:49] <twb> keyboardtalk: not that, then
[03:51] <jmarsden> keyboardtalk: check that www-data or whatever your apache2 runs as has read permissions to the ~/public_html/ directories concerned?  And that /home/whoever is not 0700, for instance?
[03:51] <twb> jmarsden: oh, mea culpa; I thought apache2 ran as root
[03:51] <twb> I guess only the master proc does
[03:51] <jmarsden> twb: It drops root ptivs as soon as it binds to port 80, basically.
[03:51] <jmarsden> *privs
[03:51] <twb> Righto
[03:56] <keyboardtalk> www-data is the owner of public_html
[03:57] <keyboardtalk> I gave everything read permissions
[03:59] <twb> You may also need to dance with <directory> options in the apache configuration
[03:59] <twb> #httpd can help with that
[04:00] <jmarsden> keyboardtalk: are perms on /home/whoever sufficient that www-data can traverse it to find /home/whoever/public_html ?
[04:02] <twb> su - www-data -c 'find /home/fred/public_html -ls' >/dev/null
[04:02] <twb> See what errors (if any) it gives
[04:02] <keyboardtalk> the parent folder needs read access as well?
[04:02] <jmarsden> keyboardtalk: needs x, for directory search, I think.
[04:03] <twb> Sounds right
[04:03] <twb> posix permissions hurt my brain tho
[04:03] <jmarsden> So /home/whoever probably needs to be 711 not 700.
[04:04] <keyboardtalk> does that compromise the security of anything inside the home directory?
[04:04] <jmarsden> It means all users can traverse that directory... if that is a compromise in your situation, I can't say :)
[04:05] <keyboardtalk> well it's working, thank you very much
[04:05] <jmarsden> You're welcome.
[04:16] <kexman> twb: how do i remove config files of uninstalled packages ? where would does hide ? :)
[04:21] <jmarsden> apt-get purge PACKAGENAME
[04:22] <twb> 12:38 <twb> You might want to "aptitude purge ~c" too
[06:15] <jazz2> hi, how can I get back to a 2.6.28 kernel on my 10.04-server? since I upgraded from 9.04 the wlan connection is dead slow with the 2.6.32 kernel (I even made a clean install to check if the upgrade was faulty, but unfortunately there is no 2.6.28 kernel available anymore, as it was when I upgraded) or do I have to go back to 9.04?
[06:19] <twb> jazz2: that's extremely non-trivial
[06:19] <jmarsden> jazz2: Maybe you should be asking "how can I fix my wlan connection" instead?
[06:19] <twb> !xyproblem
[06:19] <twb> Grr, this bot doesn't know *anything* :-/
 [xyproblem] People often falsely diagnose problems because they are looking too closely at a problem: they have got stuck at a particular point (Y) doing something (X) and so ask about step Y, not realising that there is an easier or better way to do X in the first place.  See http://mywiki.wooledge.org/XyProblem or http://homepages.tesco.net/J.deBoynePollard/FGA/put-down-the-chocolate-covered-banana.html
[06:20] <jmarsden> twb: You are just too used to dpkg :)
[06:20] <twb> Actually greybot in this case
[06:22] <jmarsden> jazz2: If you only have one machine to deal with, it would seem quicker and easier to replace the wlan NIC with one that is well supported, than to try using old non-standard kernels etc etc.
[06:22] <twb> Even if you had 100 machines
[06:23] <twb> I'm not convinced the kernel is the problem, tho
[06:25] <jazz2> twb, jmarsden, I would like to have a wlan fix, but it seems there are some ppl having similar problems (from what I could google) so I just thought I could go back; In the upgraded version I could change from between 6.28/31/32 and the problem wasn't there in 2.6.28
[06:26] <jazz2> so I thought it would be esiest for the shortterm (until it gets a fix) to use the old kernel
[06:26] <jazz2> easiest*
[06:27] <jmarsden> I really doubt it.  Easiest is probably a $10 wireless NIC that you know is supported well in 10.04.2 :)
[06:27] <jazz2> it's a zydas which worked great up until now
[07:11] <BuenGenio> hello
[07:11] <BuenGenio> Apache2 on my 11.04 server doesn't restart properly
[07:11] <BuenGenio> it just hangs saying ... waiting ...........................
[07:11] <BuenGenio> but never restarts
[07:11] <BuenGenio> additionally apache2 can only be killed with -S KILL
[09:52] <xperia> hello to all. i have heavy big problems for some strange reason with my DNS Server. if i try several times as a example to load a website with wget it hangs allways at the resolving of the domain name. need help with this one. problem is with other ISP it works all fine when the sites are loaded just with my ISP it makes problems
[09:53] <xperia> when i sniff the requests i get also this errors here
[09:53] <xperia> 00:04:41.281 0.132 426 393 GET 302 text/html (NS_ERROR_REDIRECT_LOOP) http://www.mypicx.com/images/logo.jpg
[09:53] <xperia> what could be wrong. have lot of time outs and a heavy lag with the DNS
[09:56] <jazz2> xperia, sometimes there is an issue, that the dns tries to resolve using ipv6 first and only resolves ipv4 when that times out
[09:58] <xperia> hmmm i looked the last days in the bind log and it is full of strange messages like
[09:58] <xperia> Jun 20 10:56:48 HauptServer named[7854]: error (network unreachable) resolving 'ns1.nzz.ch/A/IN': 2001:500:2d::d#53
[09:59] <xperia> Jun 20 10:56:53 HauptServer named[7854]: error (unexpected RCODE REFUSED) resolving 'www.zischweb.ch/AAAA/IN': 168.144.1.155#53
[09:59] <xperia> What should i do ?
[10:02] <airtonix> anyone installed bind9.7 from haukes ppa on maverick?
[10:02] <airtonix> https://launchpad.net/~hauke/+archive/bind9
[10:03] <xperia> i have bind9 running on natty but dont know if i have the haukes ppa
[10:03] <airtonix> i need the dnssec extensions to get samba4 running properly
[10:03] <jazz2> xperia, do you have/use ipv6? can that be the issue?
[10:04] <xperia> well i guess i dont have ipv6 any suggestion how to check and do it on my ubuntu server ? looks like problem is related to this ipv6 change then that has been done a few days ago
[10:05] <jazz2> maybe this is helpful: http://www.ubuntugeek.com/how-to-disable-ipv6-in-ubuntu.html
[10:13] <xperia> jazz2 found this one too as it is more related to bind =>http://ubuntu-tutorials.com/2009/03/21/configure-bind-9-for-ipv4-or-ipv6-only/
[10:13] <xperia> thanks a lot for your help. need now to look if this was the problem
[10:14] <jazz2> xperia, good luck :)
[10:18] <speakman> Still having serious fatal MCE's :( http://pastebin.com/KdUHbpdC Any suggestions are welcome!
[10:29] <DigitalFlux_> Hi Guys
[10:29] <DigitalFlux_> So i've installed cobbler under Ubuntu
[10:29] <DigitalFlux_> and while i try to network-boot natty, i get the error that it can't find the preconfig file, it's searching in http://127.0.0.1/cblr/svc/op/ks/profile/ubuntu-natty-server-x86_64
[10:30] <DigitalFlux_> Any hints why is that ?
[11:31] <fulc> hello, i would like to ask for some help with regards to Ubuntu Enterprise Cloud
[12:15] <kexman> morning :)
[12:23] <_johnny> anyone remember some months/a year ago where a security update was pushed fixing some openssh bug/vulnerability? anyone got a link for which it was, or which ubuntu version it was? got a friend on 10.04 who might be a victim
[13:24] <Daviey> kirkland: Around?
[13:40] <zul> Daviey: how did you fix mod-perl?
[13:42] <Daviey> zul: build failure?
[13:42] <zul> yeah
[13:42] <Daviey> zul: it was failing due to a test failure
[13:43] <zul> Daviey: so how did you fix it? :)
[13:43] <Daviey> the test suite only worked with HTTP 1.0 not HTTP 1.1
[13:43] <Daviey> cherry picked patches from upstream
[13:44] <Daviey> and like a good cookie, fixed it in Debian :)
[13:44] <Daviey> http://bugs.debian.org/cgi-bin/bugreport.cgi?msg=10;filename=230-test-failures-fix.patch;att=1;bug=628296
[13:48] <Daviey> zul: why do you ask?
[13:48] <zul> im just curious
[13:53] <Kathusyas> whats the diffrent of ubuntu / ubuntu server is it same distro?
[13:55] <Daviey> uvirtbot: Maybe one day, you'll give us good news?
[14:14] <hallyn> Daviey: hey, how is status on spice packages?
[14:17] <Guest81013> hey guys, i need to change the grub loader on my ubuntu to text so that it doesnt load the gui, i cant remember where the grub files are kept.  can anyone help me by telling me where the files are kept?
[14:18] <Daviey> hallyn: Hello sir!  Good to have you back :)
[14:18] <Daviey> hallyn: was this something you asked me to sponsor?
[14:19] <hallyn> yup :)
[14:19] <pwnguin> anyone selfhosting openID providers?
[14:19] <Guest81013> hey guys, i need to change the grub loader on my ubuntu to text so that it doesnt load the gui, i cant remember where the grub files are kept.  can anyone help me by telling me where the files are kept?
[14:19] <hallyn> the packages in ppa:serge-hallyn/spice2 for universe
[14:19] <hallyn> Guest81013: edit /etc/default/grub i think
[14:21] <Guest81013> hallyn: that is correct. thankyou
[14:26] <Guest81013> hallyn:  i changed it to text but its still loaded the gui.  any ideas exactly what i should change it to in order to stop the gui loading by default?
[14:28] <hallyn> you cleared out 'GRUB_CMDLINE_LINUX_DEFAULT' and then ran update-grub?
[14:28] <hallyn> Guest81013: ^
[14:29] <Guest81013> ahhh yes
[14:29] <Guest81013> update
[14:29] <Guest81013> lol thats what i missed
[14:45] <w00> hi
[14:46] <w00> I have in /etc/apt/apt.conf.d/20auto-upgrades set: APT::Periodic::Update-Package-Lists "1"; but i would like to know how can i check that it updated the package list?
[15:24] <brianthelion> Hi all! Just trying to get some eyes on my krb5/ldap/nfsv4 bug -- https://bugs.launchpad.net/ubuntu/+bug/794112
[15:25] <brianthelion> feedback greatly appreciated
[15:32] <Daviey> hallyn: Hmm.. I reviewed it and seem to remember i had some questions.
[15:32] <Daviey> gah
[15:32] <Daviey> hallyn: i should have made a note
[15:33] <hallyn> Daviey: going to plumber's?
[15:35] <Daviey> hallyn: unlikely
[15:35] <hallyn> :(
[15:38] <Daviey> hallyn: Think it'll be useful?
[15:39] <hallyn> well i've not seen a schedule.  But i'd think so, yes.  Good people are there usually.
[15:39] <hallyn> (i'm gonna be hiding at a min-conf on the 8th)
[15:42] <xperia> hello to all. i have just connected my ubuntu server to my adsl modem and i am able to download content from the internet but if i try to access the webserver nothing happens.
[15:42] <xperia> ssh connection to the server from outside is possible. what could be the problem ?
[15:45] <xperia> okay got it running
[15:46] <xperia> it was the reversy proxy that maked problems
[16:08] <zul> RoAkSoAx: can you push the mini.iso stuff for cobbler upstream
[16:08] <RoAkSoAx> zul: yes will do!
[16:10] <zul> RoAkSoAx: if i get some spare time ill add arm detection support this week sometime i hope
[16:11] <RoAkSoAx> zul: cool... are there any arm ISO's yet?
[16:12] <negronjl> good morning all
[16:12] <zul> RoAkSoAx: no im suspecting it will behave a bit differently
[16:12] <RoAkSoAx> zul: ok
[16:12] <RoAkSoAx> negronjl: morning
[16:12] <negronjl> hi RoAkSoAx
[16:15] <zul> RoAkSoAx: as in slit my wrists differently
[16:16] <Daviey> RoAkSoAx: arm doesn't yet use ISO's.
[16:16] <Daviey> That will hopefully come this cycle.
[16:16] <RoAkSoAx> zul: hehehehe yeah I cna imagine looking at all the documentation :)
[16:16] <RoAkSoAx> Daviey: cool!
[16:16] <Daviey> Currently it's a dd image with OEM mode ubiquity on first boot.
[16:35] <xperia> hello to all. i am trying to turn my ubuntu server into a wirelsee access Point based on this howto here
[16:35] <xperia> https://help.ubuntu.com/community/WifiDocs/WirelessAccessPoint
[16:35] <xperia> the thing that i dont understand is the interface br0
[16:36] <xperia> have installed the bridge utils like written but dont see still the br0 interface
[16:40] <atotclic> hello
[16:42] <airtonix> xperia: without reading that page I think you need to manually add the br0 interface to your /etc/networks or something ?
[16:43] <airtonix> xperia: does your wifi device successfully enter master mode ?
[16:44] <xperia> airtonix: i am not that versatile with wifi. the server has wlan0 interface but if it is possible to run as master mode i cant say
[16:45] <xperia> any sugestion how to find it out ? have just running lspci and found the wlan card
[16:45] <xperia> 03:00.0 Network controller: Broadcom Corporation BCM43225 802.11b/g/n (rev 01)
[16:48] <smoser> cjwatson, so, for libpng, using pkg-config is fine, but it looks to me that libx11 and libjpeg62 do not provide pkgconfig files, so it doesn't work fo rthem.
[16:49] <smoser> you suggest making pkgconfi files and bugs for those packages or... something else?
[16:49] <airtonix> xperia: lspci -nn
[16:50] <airtonix> xperia: then search the manufacturer code and device id on some websites like hotspot ap or something
[16:50] <airtonix> tbh it's much easier to buy a 35$ tplink 1043nd and flash it with openWRT
[16:51] <airtonix> then you can opt install all kinds of linux stuff on it
[17:04]  * zul lunch
[17:09] <cjwatson> smoser: x11.pc exists
[17:10] <cjwatson> smoser: for libjpeg, hmm.  a Debian-specific patch for this is to add /lib/$(DEB_HOST_MULTIARCH) and /usr/lib/$(DEB_HOST_MULTIARCH) to your search path (you'll need to 'export DEB_HOST_MULTIARCH := $(shell dpkg-architecture -qDEB_HOST_MULTIARCH)' somewhere, perhaps in debian/rules, or fetch the output of 'dpkg-architecture -qDEB_HOST_MULTIARCH' in config_unix.py)
[17:11] <cjwatson> that would handle the cases you've dealt with using pkg-config too, although pkg-config is generally better
[17:18] <xperia> airtonix one small question my eth0 interface has this data here at the moment
[17:18] <xperia> inet Adresse:192.168.1.33  Bcast:192.168.1.255  Maske:255.255.255.0
[17:18] <xperia> My Question is what i need to write for the br0 interface as ip adresses ?
[17:18] <xperia>     #address 10.1.1.1
[17:18] <xperia>     #network 10.1.1.0
[17:18] <xperia>     #netmask 255.255.255.0
[17:18] <xperia>     #broadcast 10.1.1.255
[17:18] <airtonix> for next time, use pastebin
[17:19] <xperia> okay
[17:19] <airtonix> or dpaste or any other other derivatives
[17:19] <airtonix> from memory i think your bridge is going to be anything that isn't on that eth0 network
[17:20] <airtonix> it's essentially a different subnet of it's own
[17:20] <xperia> ahh then i have to use the wan IP adress then ?
[17:20] <airtonix> well i'm not sure about that
[17:22] <airtonix> are you looking to create an access point that more than one wifi client can connect to or are you just wanting to do a point to point connection ?
[17:27] <xperia> airtonix yes this is my idea
[17:33] <airtonix> access point or tunnel?
[17:34] <airtonix> because you can just use ufw's NAT features to do the latter
[17:34] <airtonix> (which is just a simplified interace to IPtables NAT features)
[17:35] <airtonix> interface*
[17:35] <airtonix> time to sleep
[17:39] <xperia> okay airtonix have nice sleep
[18:13] <hallyn> mdeslaur: hey, do you know what is going on with bug 795427 ?
[18:16] <mdeslaur> hallyn: I have no idea what that's about
[18:16] <mdeslaur> hallyn: it's booting from the disk instead of from the cdrom? I dunno
[18:16] <hallyn> ok :)   me neither, wanted to ask you before askng for clarification
[18:52] <nonotza> could someone help me out with this odd behavior in ubuntu? when I start up I get two start up messages like this: http://pastebin.com/uGX2J7cY
[19:03] <nonotza> ** when I login ...
[19:03] <SpamapS> nonotza: thats been fixed in a recent update actually
[19:04] <nonotza> oh ok, I just updated actually
[19:04] <nonotza> let's see if it worked
[19:04] <nonotza> do you know how I can change that text?
[19:04] <nonotza> what do you know ... its fixed now hehe
[19:05] <SpamapS> nonotza: read 'man motd' and 'man motd.tail'
[19:05] <nonotza> thank you
[19:05] <SpamapS> nonotza: ^5 for testing our updates. :-D
[19:05] <nonotza> hehe
[19:05] <AndroidLoverInSF> i got postfix mail delievered to my maildir new folder alright. but when i type mail. it says i have no new mail. so what wrong and how to fix? ideas?
[19:29] <hggdh> Daviey: there?
[19:37] <SpamapS> Andre_Gondim: mail checks your system mail queue, Maildir is something different.
[19:38] <SpamapS> Andre_Gondim: you might want to look at 'mutt' or 'sup-mail' if you want to see your mail via the console.
[20:02] <orudie> how can I archive a directory with preserving the entire files and subdirectory structure ?
[20:04] <pmatulis> orudie: use rsync or cp
[20:04] <pmatulis> 'rsync -a --delete' or 'cp -a'
[20:08] <RoyK> pmatulis: cp -a won't remove old stuff
[20:09] <RoyK> orudie: rsync -avP will do it verbosely and allow large files to be resumed (-P == --progress --partial)
[20:09] <pmatulis> RoyK: should
[20:09] <RoyK> should what?
[20:09] <pmatulis> RoyK: should remove old stuff
[20:09] <RoyK> cp -a doesn't remove anything
[20:10] <orudie> so I should use rsync and not scp ?
[20:10] <RoyK> I'd use rsync
[20:10] <RoyK> rsync uses rdiff for large binary files, so if a small set is changed in a large binary file, mostly the differences will be transferred again
[20:11] <orudie> RoyK, I need to migrate the content of /var/www/ which contains a number of websites, about 5 GB total from one host to another. can you give an example of rsync command line to be used ?
[20:11] <RoyK> rsync -avP /var/www /new/path
[20:11] <RoyK> or somehost:/new/path
[20:12] <RoyK> it'll tunnel over ssh unless you ask it otherwise
[20:12] <orudie> and to specify ssh port ?
[20:12] <RoyK> -p
[20:12] <orudie> k
[20:12] <patdk-wk> -e 'ssh -p 222'
[20:12] <RoyK> erm
[20:12] <RoyK> no
[20:12] <RoyK> yes, patdk-wk's right
[20:12] <patdk-wk> rsync supports ssh port directly now?
[20:12] <RoyK> no, my wrong
[20:13] <RoyK> --port=PORT
[20:13] <RoyK> from the manual
[20:13] <RoyK> so, yes, it does, but not with -p
[20:14] <RoyK> again, I may be wrong, seems --port is more for the rsyncd works
[20:14] <RoyK> rather do as patdk-wk said
[20:14] <patdk-wk> ya, that is for rsync protocol only
[20:14] <patdk-wk> easier to use the uri syntax for it, rsync://user@host:port/
[20:14] <RoyK> or ssh:// .. ?
[20:15] <patdk-wk> doesn't say that is supported
[20:15] <RoyK> rsyncd is rather outdated imho
[20:15] <orudie> so this would be good way ? rsync -aevP 'ssh -p 222' /var/www/ user@host:/new/path ?
[20:15] <patdk-wk> I use rsyncd for some things
[20:15] <RoyK> well, I guess -e 'sssh -p someport' should work, then
[20:15] <RoyK> orudie: looks good
[20:16] <patdk-wk> hmm, is it legal to have the e floating in the middle like that?
[20:16] <RoyK> sssh == Schutzstaffel-shell :P
[20:16] <RoyK> sorry, one s too much
[20:16] <RoyK> patdk-wk: rsync -avP -e 'ssh -p someport' usr@host:/somepath
[20:16] <patdk-wk> that is what I do, it's *safer*
[20:17] <RoyK> -e is just another argument and 'ssh... is just the value of that
[20:17] <patdk-wk> I've had programs take vP as the aurg to -e i nthe first example
[20:17] <patdk-wk> probably cause they don't use getopt or whatever?
[20:17] <RoyK> the -vP parts will give you quite a lot of output and not very much needed
[20:18] <RoyK> getopt() just takes whatever's given and parses it
[20:18] <RoyK> rsync -a -v -P -e ... == -avPe
[20:18] <patdk-wk> ya, but does rsync -aevP = rsync -a -e vP
[20:18] <RoyK> patdk-wk: no
[20:19] <RoyK> patdk-wk: -e takes an argument, the command, 'ssh' or 'ssh -p someport' etc
[20:21] <RoyK> patdk-wk: so try "ssh -e 'ssh -p 2020' -avP /my/dir user@somehost:/their/dir
[20:22] <RoyK> the -e 'blah' can be anywhere in the string, just before /my/dir
[20:23] <patdk-wk> for the record: rsync -aevP 'ssh -p 222' = rsync -ae 'vP' 'ssh -p 222'
[20:23] <patdk-wk> rsync -avPe 'ssh -p 222', works as expected
[20:23] <pythonirc101> I've a 200GB hard drive on which I want to install ubuntu and then i've 5 other drives on which i want to run RAID 5. Is this easy to configure during installation?
[20:24] <patdk-wk> sure
[20:24] <RoyK> pythonirc101: just use manual partitioning
[20:25] <RoyK> pythonirc101: setup the root drive with whatever's needed there, create partitions on the data drives for 'physical drive for used with RAID' and after doing that, configure the RAID
[20:25] <sw0rdfish> hey guys add-apt-repository needs to be installed? its giving me 'command not found'
[20:25] <patdk-wk> install it?
[20:25] <RoyK> apt-get install python-software-properties
[20:25] <RoyK> erm, that was perhaps wrong
[20:25] <pythonirc101> RoyK: Lemme try
[20:25] <patdk-wk> ya that is it, heh
[20:26] <patdk-wk> my search wouldn't bring it up
[20:26] <RoyK> anyway, adding something to /etc/apt/sources.list.d should work well
[20:27] <RoyK> I'm on Maverick on this box (or VM)
[20:27] <RoyK> tried to upgrade it to natty the other day, but grub messed up :P
[20:28] <pythonirc101> RoyK: I guess its easier to just install the OS on the 200GB harddrive and then setup the RAID later?
[20:28] <pythonirc101> RoyK: I dont see any raid options in the installer?
[20:29] <orudie> well i tried rsync as a test , didn't work.. not sure what I did wrong there http://paste.ubuntu.com/629990/
[20:29] <RoyK> pythonirc101: with the server installation, the options are there
[20:29] <pythonirc101> RoyK: In the partitioning phase?
[20:29] <RoyK> pythonirc101: just choose 'manual partitioning', and do as I said above
[20:29] <RoyK> pythonirc101: you need to create a 'physical partition for raid' on each of the data drives
[20:30] <pythonirc101> I did goto manual partitioning, but it asks me for mount points and such...no mention of raid anywhere
[20:30] <RoyK> after that, exiting the partitioner, it's simple
[20:30] <pythonirc101> ok, lemme try again
[20:30] <RoyK> pythonirc101: the standard partition type is ext4, press enter there and choose RAID partition
[20:30] <RoyK> or filesystem type, even
[20:31] <orudie> anyone? what did I do wrong there ?
[20:32] <RoyK> orudie: 21:23 < patdk-wk> for the record: rsync -aevP 'ssh -p 222' = rsync -ae 'vP' 'ssh -p 222'
[20:32] <RoyK> so move the e after the P
[20:32] <RoyK> since e takes an argument, P doesn't
[20:33] <pythonirc101> with 5 1TB drives, should i go with RAID 5 / 6/ 10?
[20:33] <pythonirc101> I need reliability in case one drive fails, and fast IO
[20:33] <cloakable> 5 probably
[20:33] <RoyK> pythonirc101: depends on the needs
[20:33] <patdk-wk> pythonirc101, depends what you call fast io
[20:33] <RoyK> pythonirc101: RAID-5 is rahter slow for random access
[20:33] <RoyK> pythonirc101: RAID-6 even slower
[20:33] <patdk-wk> but only random writes
[20:34] <pythonirc101> what about raid 10?
[20:34] <sw0rdfish> i'm just trying to add the repository for nginx
[20:34] <RoyK> pythonirc101: 5 drives in RAID-1+0 isn't very easy :P
[20:34] <patdk-wk> raid10 is fast for random writes
[20:34] <sw0rdfish> add-apt-repository ...etc gives me command not foudn
[20:34] <sw0rdfish> found*
[20:34] <RoyK> sw0rdfish: and reads...
[20:34] <sw0rdfish> lol
[20:35] <sw0rdfish> ohhh
[20:35] <pythonirc101> can raid 10, 5 , 6 all survive one disk failure? more?
[20:35] <patdk-wk> random reads would be faster with raid5/6 than raid10
[20:35] <RoyK> pythonirc101: if you don't have a bad-ass database or a lot of video streams to serve, just use RAID-5
[20:35] <sw0rdfish> I can add it to /etc/apt/sources.list :D
[20:35] <patdk-wk> raid10 depends, can handle 1 or more failures
[20:35] <RoyK> pythonirc101: RAID-6 can survive two drive failuers, RAID1+0 can survive one drive failure on each mirror
[20:35] <pythonirc101> ok, raid 5 seems to be the middle ground then?
[20:36] <RoyK> pythonirc101: for five drives, just use RAID-5 - it'll probably suffice
[20:36] <pythonirc101> one drive failure? + good read/write?
[20:36] <pythonirc101> thanks
[20:36] <patdk-wk> if a disk fails in raid5 or raid6, though, your speed will ONLY be 20% what it used to be
[20:36] <RoyK> pythonirc101: anyway - you can change it to RAID-6 later by adding a drive and doing some mdadm magick
[20:36] <patdk-wk> till the new drive is rebuilt
[20:36] <pythonirc101> so i need raid 5 -- and i make all my 5 drives active?
[20:37] <patdk-wk> that will give you 4*drive space available
[20:37] <RoyK> working with ZFS with rather large volumes (some 350TB total) Linux RAID is a heaven when it comes to flexibility
[20:37] <pythonirc101> when one fails , i replace it?
[20:37] <pythonirc101> RoyK: How do you even build the 350TB storage? one machine? or distributed zfs?
[20:38] <RoyK> pythonirc101: think of it as this - in the ZFS world, we talk about RAIDz1 or RAIDz2 or even RAIDz3. RAIDz1 has one parity level, so it can survive a dead drive (per VDEV). Same applies to RAID-5. RAIDz2 is like RAID-6. RAIDz3 is like - well - RAID-6 with another parity drive
[20:38] <RoyK> pythonirc101: two 100TB boxes and a few smaller ones
[20:39] <RoyK> that's net storage, after RAIDz2 overhead :P
[20:39] <pythonirc101> RoyK: how many active drives for raid 5 should i use?
[20:39] <RoyK> 5-6 max
[20:39] <RoyK> for RAID-6, a bit more, perhaps up to 12
[20:39] <pythonirc101> RoyK: So if i put 5 active, and one crashes, I'll have to find out which one crashed, and replace it manually?
[20:39] <RoyK> yep
[20:40] <RoyK> and if you replace the wrong drive, tough luck :P
[20:40] <pythonirc101> sounds good...checking hard drives should be easy... :)
[20:40] <patdk-wk> why it's nice to check serialnumbers
[20:40] <pythonirc101> well i can always check the drives one by one using my external sata plugs?
[20:40] <pythonirc101> patdk-wk: ah, the serial numbers tell which one failed then?
[20:40] <patdk-wk> if you use smartctl on the drive to read that info
[20:41] <^rcaskey> 350TB is a hefty number
[20:41] <patdk-wk> or use smartctl to see what drives are left working :)
[20:41] <patdk-wk> mdadm will tell you what device failed
[20:41] <patdk-wk> smartctl on that device will give serial
[20:42] <pythonirc101> nice thanks
[20:42] <pythonirc101> RoyK: Who builds 100tB boxes? how much do they cost?
[20:43] <RoyK> ^rcaskey: 100+100+60+50+14+12 - well, not quite 350, but still ok
[20:43] <^rcaskey> RoyK, why broke out like that?
[20:43]  * RoyK kicks uvirtbot 
[20:43] <RoyK> ^rcaskey: different use pattern
[20:43] <RoyK> the three 100+100+14 are backup servers for bacula
[20:43] <^rcaskey> would 96 be close enough?
[20:44] <RoyK> 60+50 are scientific work data + backup
[20:44] <RoyK> the last 12TB box is a file server
[20:44] <RoyK> the latter with all striped mirrors and some nifty SSDs for caching
[20:44] <pythonirc101> RoyK: Who builds the 100TB boxes?
[20:44] <RoyK> 100TiB or 110TB for the big ones
[20:45] <pythonirc101> how much do they cost?
[20:45] <^rcaskey> RoyK, i'm way out of my league, I'v only got 2TB in mine and I never got around to my SSDs due to ram + nvram
[20:45] <RoyK> this was back in, what, november, perhaps october, and they cost us about NOK 130k a piece
[20:46] <^rcaskey> If you are willing to do your JBOD in software Dell R510s are cheap
[20:46] <RoyK> I learned one lesson, though - don't mix SuperMicro SAS expanders and WD desktop drivers
[20:46] <RoyK> s/drivers/drives
[20:46] <pythonirc101> I'm using seagate + supermicro sata ports
[20:47] <RoyK> seagate and hitachi works well
[20:47] <patdk-wk> royk, what wd drives did you attempt to use?
[20:47] <RoyK> but there's something fishy about the timing with WD desktop drives
[20:47] <RoyK> blacks
[20:48] <RoyK> dunno if this is an openindiana/solaris issue, but sometimes we see a whole backplane/sas expander go bad, starting to produce massive i/o errors
[20:48] <RoyK> but only one
[20:48] <RoyK> with four backplanes/sas expanders per server, it's quite easy to pinpoint the black sheep
[20:49] <patdk-wk> I have a bunch of the yellow currently
[20:49] <RoyK> yellow?
[20:49] <patdk-wk> next level up from black
[20:49] <RoyK> is that a cheap green thing?
[20:49] <patdk-wk> re3/4
[20:50] <RoyK> oh, RE
[20:50] <RoyK> ok
[20:50] <RoyK> RE is ok, but bloody expensive compared to blacks
[20:51] <RoyK> thing is, Seagates and Hitachis work flawlessly, but adding a bunch of WD green/black drives to a chassis, the whole thing starts acting erronously
[20:51] <patdk-wk> I have nothing but issues with green drives
[20:51] <RoyK> they work well directly attached
[20:52] <RoyK> but through an LSI 3810, they are chaos
[20:52] <RoyK> so are WD Blacks
[20:52] <RoyK> with LSI 9201 it works better, though, but not quite
[20:52] <^rcaskey> my how I hate hardware
[20:53] <RoyK> yeah, better off without it - use Semaphore!
[20:54] <^rcaskey> My file server rocks the two-disk raid 1
[20:55] <^rcaskey> I'd bet in the average week a single piece of data doesn't get read off a disk
[20:55] <^rcaskey> Well excluding whatever smart does
[20:55] <hggdh> 1
[20:55] <^rcaskey> one
[20:55]  * RoyK is setting up a new 2-drive RAID-5 for a private server soon
[20:56] <^rcaskey> ummm howzat work?
[20:57] <^rcaskey> mirrored parity data on reserved portions of both disks :P?
[20:57] <RoyK> it's just that simple
[20:57] <RoyK> you need 1+n data drives and one parity drive
[20:57] <RoyK> n >= 0
[20:57] <patdk-wk> it's like raid1, but slower :)
[20:58] <patdk-wk> cause of cpu overhead
[20:58] <RoyK> but more flexible :)
[20:58] <RoyK> when I want to add another, I expand the RAID-5, adding another drive
[20:58] <RoyK> whops, lots of more capacity
[20:58] <RoyK> and then another, and another
[20:58] <RoyK> and if I grow past 4-5 drives, I change it to RAID-6 and keep on growing
[21:06] <pythonirc101> RoyK: Somehow grub wont install on my raid/other drive...still trying...
[21:06] <RoyK> pythonirc101: grub won't install on a RAID-5
[21:07] <RoyK> pythonirc101: use a single drive or a mirror for the root, and data drives for the rest
[21:07] <pythonirc101> the problem is , its showing me /dev/mapper for installing grub
[21:07] <pythonirc101> I dont know where to install it right now..I've a raid 5 and another separate drive named /dev/sd.... -- dont remember
[21:08] <RoyK> pythonirc101: then just reinstall, use manual partitioning, create a small /boot of 1-2GB, a swap partition about half your RAM size (or twice, if you expect it to overcommit a lot), then the root for the rest of the drivee (or mirror)
[21:08] <RoyK> then install and create the RAID-5 volume manually
[21:08] <pythonirc101> here is the problem...I've two drives 200GB and 5x1TB...the 200GB is also a raid, and  grub wont install there either...
[21:08] <RoyK> it should work fine with the installer, but there may be issues
[21:09] <pythonirc101> 40GB x 5 = 200GB
[21:09] <RoyK> pythonirc101: did you add a mirror for the root?
[21:09] <pythonirc101> nope.
[21:09] <RoyK> then use the two 200GB drives for the root alone
[21:09] <pythonirc101> I was hoping to install things on the 200GB hard drive and keep the 5x1TB for data/storage
[21:09] <RoyK> just mirrors
[21:11] <RoyK> pythonirc101: in the installer, create a fresh partition on each of the two 200GB drives, on each, create a 2GB partition for /root (ext2), a 2GB partition (or so) for swap, and the rest for the root. Mark all those as physical RAID partitions
[21:11] <RoyK> when done, go back and create ext2/swap/ext4 on those partitions, and leave the data out of it
[21:11] <RoyK> then, when installed, create a new RAID with mdadm
[21:11] <pythonirc101> I only have one 200GB drive...which is a hardware raid - Five 40GB drives = 200GB
[21:12] <RoyK> I thought you were talking about software raid?!?
[21:12] <pythonirc101> I've a software raid (5x 1TB) and a hardware raid (40GB x 5)
[21:12] <RoyK> the root partition on an ubuntu server won't need more than a mirror of those
[21:13] <RoyK> 40GB is sufficient for most
[21:13] <RoyK> most, being very large fileservers doing nasty stuff
[21:14] <JanC> 5 x 40 GB = 200 GB  doesn't sound like (real) RAID to me  ;)
[21:15] <pythonirc101> ssd
[21:15] <pythonirc101> ocz
[21:16] <pythonirc101> so grub wont install on a software raid i take it?
[21:16] <pythonirc101> can one install the os on a software raid?
[21:16] <RoyK> pythonirc101: that's a waste, really
[21:16] <RoyK> pythonirc101: just use two of them in a mirror
[21:17] <RoyK> or three in a three-way mirror
[21:17] <RoyK> or rather, give up using SSDs for the root
[21:17] <RoyK> it won't help much - most of the root fs is cached upon bootup anyway
[21:17] <RoyK> use cheap spinning drives
[21:17] <pythonirc101> RoyK: If I mirror two drives for the OS, that'll leave me with only 3 drives for a software RAID...not good
[21:17] <RoyK> use SSDs for something fun like caching
[21:18] <pythonirc101> I need capacity
[21:19] <RoyK> well, I'm going to bed and refuses to listen to idiots about this topic - one last message - use small, spinning drives for the root, two in a mirror, use large drives for data, raid-5 or striped mirrors, and if you can, use ZFS so you can allocate SSDs as caching devices
[21:20] <pythonirc101> I'll try, thanks.
[21:24] <sw0rdfish> RoyK, hei man
[21:24] <sw0rdfish> du er veldig aktiv I dag :)
[21:26] <sw0rdfish> you there, mate?
[21:44] <raubvogel> Question on https://help.ubuntu.com/10.04/serverguide/C/kerberos-ldap.html: don't you have to create the principals for the slave KDC?
[22:36] <RoAkSoAx> kirkland: this server side (powerwaked) ARPMonitor to auto wakeup machines is sweet!!
[22:58] <kirkland> RoAkSoAx: neat :-)
[22:59] <zooko> Dear people of #ubuntu-server: I'm used to debian/ubuntu packages coming configured to a working state.
[23:00] <zooko> But cron-apt apparently requires me to un-comment-out dozens of lines in /etc/cron-apt/config. What gives?
[23:10] <adam_g> kirkland: hey dude ive been using tmux for a year or so, id be happy to checkout any of the profiles you're working on
[23:10] <kirkland> adam_g: smoser told me you were a tmux user ;-)
[23:10] <kirkland> adam_g: nice, so have you used byobu at all?
[23:15] <adam_g> kirkland: not so much, lately i haven't been doing enough remote work to need anything like it.  i use a tiling window manager locally that works similarly
[23:15] <kirkland> adam_g: right; okay
[23:16] <kirkland> adam_g: well, perhaps we'll catch up in Dublin and you can show me some tips and tricks
[23:17] <zooko> Sorry to ask a question and then disconnect.
[23:17] <zooko> The question was: why does cron-apt come with a config file with everything commented-out, instead of with useful default values?
[23:29] <RoAkSoAx> kirkland: By the end of this week I hope to have a fully functional powerwaked server that will grab ips/mac's for powerwake cache, listen for ARP Broadcast in the network, and wakeup those machines which need to be woken up