[00:58] <sjefen6> I have been trying to install ubuntu-server i386 on an old Pentium 4 dell desktop, but during the installation the instalation stops and I am left with a blue screen, with a gray line without any text. Is there anything im doing wrong?
[01:03] <ewook> no. but 1, your cd / cd-reader / cable / or other falty hardware might create the scenario
[01:04] <sjefen6> xp installed flawlessly some hours ago. no memory-defects, and im testing the install media atm
[01:05] <ewook> xp does not always die upon install, it can show the problems later on resulting in bluescreens or simply falty behaviour.
[01:05] <sjefen6> the cd-rom is valid
[01:05] <ewook> do a memcheck
[01:05] <sjefen6> does that rule out the media and reader?
[01:06] <sjefen6> no memory-defects
[01:06] <ewook> it rules out the cd at least.
[01:06] <sjefen6> ran memcheck some hours ago
[01:07] <ewook> what release are you trying?
[01:07] <sjefen6> ubuntu-9.10-server-i386
[01:08] <ewook> do you have a spare cd to try the LTS also?
[01:08] <sjefen6> I can try...
[01:08] <ewook> not needed, but.. what dell model was it?
[01:09] <ewook> rather, the LTS is not needed if you don't wanna.
[01:11] <sjefen6> optiplex
[01:11] <sjefen6> GX260
[01:12] <sjefen6> torrent of downloards should be more easy accessable
[01:13] <ewook> there seems to be a few threads regarding GX260 and installation issues with 9.10, a rather few seems to mention issues with multiple kinds of hardware, weird.
[01:13] <ewook> you could go 9.04 if you wish to, LTS isn't the hottest cookie around right now :)
[01:14] <sjefen6> agree
[01:14] <sjefen6> but where can i find torrents? http is dead slow for me...
[01:16] <ewook> oh. there should be an alternative download selectable at the site.
[01:16] <ewook> hrr, not of 9.04 that I can find tho.
[01:16] <ewook> just click once on the download now, and then select alternative, and then more options pops up
[01:17] <ewook> my my. Haven't downloaded releases for a while, I'm quite lost as well :)
[01:17] <xperia> hello to all ! i have a question about some strange connection on my ubuntu webserver.
[01:18] <xperia> i have in my logs this connection here
[01:18] <sjefen6> the stupis redirects masks the adress to the mirrors, so im unable to browse the mirror for torrents
[01:18] <xperia> http://paste-bin.com/view/40ce54e0
[01:19] <ewook> sjefen6: http://www.ubuntu.com/getubuntu/downloadmirrors#bt
[01:19] <xperia> can somebody tell me if my ubuntu server was used for a dos attack ?
[01:19] <xperia> in the logs i have nearly all 7 to 8 min a connection established to a ssl port
[01:20] <xperia> and that for nearly a whole day
[01:20] <xperia> is this normal ?
[01:20] <ewook> xperia: that might be something more for the more apache-oriented folks, but it looks like you're proxying connections to mee.
[01:20] <bogeyd6_> xperia these are just http connects
[01:21] <bogeyd6_> assuming you arent running a proxy it might just them scanning you
[01:21] <xperia> no i dont run a proxy
[01:22] <ewook> true as bogeyd said. then it's just a great deal of connect-attempts. how's the other traffic? any other noticeable impact on the machine?
[01:24] <xperia> at the moment i see just only this redirects allways. i had this allready one time and i think people from apache told me
[01:24] <xperia> that this is related to my php page
[01:25] <xperia> if a php page has some code for redirection this logic could be used also for request redirecting
[01:25] <ewook> yes.
[01:26] <xperia> need maybe to fix my php page a litlle. will look now what the php people can told me ad maybe give me a solution
[01:26] <xperia> okay thanks for the help !
[01:27] <ewook> good luck.
[01:34] <sjefen6> looks like it is something wrong with performing the hdd operations
[01:34] <sjefen6> (9.10 in expert mode)
[01:40] <ewook> any errorcodes?
[01:40] <sjefen6> no
[01:40] <sjefen6> are there any debug mode i can turn on?
[01:40] <sjefen6> oh
[01:40] <ewook> flip around the tty's
[01:40] <sjefen6> looks like im gone tru 9.04 first
[01:41] <sjefen6> too late
[02:01] <sjefen6> humpf, 9.01 installed but does not boot correctly
[02:01] <sjefen6> corrupted memory?
[02:04] <ewook> sniffs like your disk isn't that well
[02:07] <sjefen6> lets try another disk then...
[02:13] <sjefen6> can same brand and model have the same issue? trying with 9.10 first
[02:15] <ewook> as in optiplex model? or what model are you referring to?
[02:16] <sjefen6> hdd
[02:17] <sjefen6> im using 10 gb fsc drives
[02:17] <sjefen6> fujitsu limited
[02:18] <ewook> well, they could be in the same state if old. or, perhaps a cable issue.
[02:18] <sjefen6> 9.10 locked again
[02:18] <sjefen6> changing cable
[02:19] <ewook> my oh my. ubu 5.10 isos still in my 'closet'.
[02:24] <sjefen6> new cable, same error
[02:25] <ewook> interesting.
[02:25] <ewook> or, perhaps just frustrating
[02:25] <sjefen6> looks like this box aint going to run ubuntu-server?
[02:26] <ewook> same error, with 9.10 that is?
[02:26] <sjefen6> 9.10 yes
[02:26] <ewook> hrm. did you try it with the 8.04.3 or the 9.04?
[02:27] <sjefen6> i tried once with 9.04, first hdd and first cable
[02:30] <ewook> and the result was that it installed. but didn't boot correctly?
[02:31] <sjefen6> yes
[02:31] <sjefen6> thew boot resulted in an console that looked like tty4
[02:31] <sjefen6> (tty4 when using 9.10)
[02:32] <sjefen6> and i think it flashed manny lines with memory corrupted
[02:46] <sjefen6> oh
[02:46] <sjefen6> looks solved
[02:46] <sjefen6> no
[02:46] <sjefen6> false alarm
[02:51] <sjefen6> starting to look like the factory memory was at foult
[02:57] <sjefen6> hanged at "Validating initscripts" tty4:"debootstrap:Killed"
[03:07] <lwizardl> hi
[03:08] <lwizardl> from a networking of servers aspect what routers are the better ones to get that are able to be found anywhere ?
[03:09] <twb> lwizardl: are you asking for hardware recommendations?
[03:09] <lwizardl> yeah
[03:10] <twb> The best kind of router is one that you can install a flexible FOSS distro on, such as Ubuntu or OpenWRT or OpenBSD
[03:10] <twb> For a SOHO, I have had success with OpenWRT on the Asus WL-500g Premium.
[03:11] <lwizardl> ok because i've mostly used Linksys routers and 90% of the ones i have owned either got flashed by me with either dd-wrt or openwrt
[03:11] <twb> (Most low-end routers will run Linux internally anyway, but it'll be locked down and crippled.)
[03:12] <twb> If you're comfortable with (say) OpenWRT, then I'd grab the OpenWRT HCL and pick a well-supported unit from that lit.
[03:12] <twb> *list
[03:13] <lwizardl> this is the first time i've been working with openwrt but i have used dd-wrt for like the last 3 years or so. and before that i was forced to keep the linksys fw on the devices because they wasn't owned by me
[03:13] <lwizardl> they was isp owned
[03:14] <twb> The plural form is "were", not "was".
[03:15] <twb> e.g. "he was confused; they were confused"
[03:20] <lwizardl> yeah but you understood what i was saying :)
[03:21] <lwizardl> so basically continue doing what I have been doing using any router that supports a better fw like using openwrt
[03:21] <lwizardl> should be fine
[03:33] <twb> OpenWRT is a distro, not a "fw"
[03:33] <twb> Oh, perhaps that meant "firmware", which I guess makes sense.
[03:39] <lwizardl> yeah firmware
[03:47] <ruben23> hi what is the permission of this on octet------>drwxrwsrwx  3 root root 4096 2009-11-13 00:45 vicidia
[03:55] <twb> ruben23: run stat(1) on it
[03:55] <twb> IIRC it's setgid.
[03:56] <twb> "info coreutils" should cover unix permissions
[03:59] <thewrath> i use this to enable a site in ubuntu a2ensite
[03:59] <thewrath> how do i disable it
[04:01] <twb> I don't know what a2ensite is.
[04:12] <thewrath> that is to enable a site
[04:12] <thewrath> *enable a site config
[04:13] <twb> Presumably within apache2.
[04:13] <thewrath> yes
[04:23] <thewrath> got it
[06:15] <netdur> hi, I have php script (not web page) that parse few large log files (around million of lines), I have four desktop available, I wonder how can I make ubuntu server distribute computing between those desktops...
[06:16] <netdur> (desktop hardwares but ubuntu server)
[06:16] <netdur> runs*
[07:04] <maxagaz> hi
[07:05] <maxagaz> how to copy /my/path/myfiles to /opt/my/path/myfiles, creating sub-paths if they don't exist in /opt ?
[07:13] <khelvan> dunix, Anonymous2, if you guys are still around - it appears that my girlfriend was kicking the server, and that was shutting it down the hard way. I'm not 100% sure on that being the cause, but there seem to be no other issues at the moment (I still have yet to give it a monitor and run memtest though).
[07:20] <khelvan> Can anyone tell me why my 10/100/1000M card in my Ubuntu Server sometimes negotiates 100M, sometimes 1000M, with the same 10/100/1000M Netgear switch, and nothing changing other than a reboot? Is there a way to force it to 1000M, and does doing so have any disadvantages?
[07:24] <twb> khelvan: do you have ssh on the switch?
[07:28] <khelvan> twb - No, I don't think so. It's a mindless Netgear GS608NA.
[07:29] <twb> About the only thing I know of to use is ethtool / mii-tool
[07:29] <twb> But I think that just reports the current behaviour
[07:40] <qman__> maxagaz, cp -r
[08:52] <ageNT_666> âñåì ïðèâåò, äàéòå òîëêîâóþ ññûëêó ïî ïîëíîé óñòàíîâêå abills
[09:45] <jiboumans> morning
[09:48] <jumbers> My host provides me 2 IPs to my dedicated server and there is 1 NIC installed on the box. How would I be able to set up my network interfaces to take advantage of the second IP?
[09:49] <AlexC_> jumbers, using 'ip' command. Something like: ip addr add x.x.x.x/x brd x.x.x.x dev eth0 label eth0:1
[09:50] <AlexC_> brd being broadcast
[09:51] <AlexC_> jumbers, you can also add this to your /etc/network/interfaces line, just add a new line but prefix it with 'up '
[09:52] <FireCrotch> jumbers: http://www.cyberciti.biz/tips/ubuntu-linux-creating-ethernet-alias-for-eth0-network-device.html
[09:52] <FireCrotch> that link explains the procedure quite well :)
[09:52] <jumbers> Aha, that's what I was looking for, a permanent setup
[09:53] <AlexC_> no, don't use that
[09:53] <AlexC_> (in ref to FireCrotch link)
[09:53] <FireCrotch> AlexC_: And why exactly do you object to my link?
[09:54] <FireCrotch> it's the proper way to set up an interface alias
[09:54] <AlexC_> FireCrotch, it uses the older methods of doing it by setting up aliases. You should do it with the newer 'ip' suit of commands
[09:54] <AlexC_> all your managment should be through 'ip', not 'ifconfig'
[09:55] <AlexC_> ifconfig is old and doesn't provide you with everything
[09:57] <FireCrotch> Meh, it's just a different way of accomplishing exactly the same thing
[09:58] <AlexC_> different and old with less flexibility
[09:58] <FireCrotch> different and stable and perfectly fine
[10:00] <jumbers> So adding a virtual interface is evil? I don't see how/why
[10:01] <AlexC_> jumbers, no no, using 'ifconfig' to do it. Using IP is ideally recommended
[10:01] <AlexC_> s/IP/ip
[10:02] <jumbers> I planned to use the second part of that page; Adding the interface to /etc/network/interfaces
[10:02] <AlexC_> which does the same
[10:02] <AlexC_> use the 1 line I gave, added to /etc/network/interfaces
[10:03] <TeTeT> cjwatson: just installing lucid UEC. I installed the front end first with clc,walrus,cc and sc. Then the node. The node told me that it cannot find walrus.
[10:03] <AlexC_> for example, underneath 'gateway ....' add 'up ip addr add x.x.x.x/x brd x.x.x.x dev eth0 label eth0:1'
[10:03] <AlexC_> and also another duplicate line, but replace 'up' with 'down', and that will be removed when that interface goes down to
[10:03] <TeTeT> cjwatson: I checked ps aux | grep avahi on the frontend and the avahi line for walrus looks different from the others. It says avahi-publish -s 192.168.1.120 instead of avahi-publish -s TestCloud
[10:09] <AlexC_> jumbers, just far easier to manage
[10:17] <maxagaz> how to copy a file to a destination and create the destination if it doesn't exist ?
[10:25] <lau> hello am trying to monitor lighttpd 1.4.19 with munin 1.2.5-2 on hardy
[10:26] <lau> do you know if http://www.linuxweblog.com/blogs/sandip/20090902/munin-stats-apache-and-lighttpd is still accurate ?
[10:42] <jussi01> can someone tell me the standard way to add someone to svn access?
[10:44] <cjwatson> TeTeT: I'm on holiday, please file a bug
[10:55] <TeTeT> cjwatson: ok, enjoy your vacation
[12:10] <X-Sleepy-X> Is it possible to display the servers temperature on a web site with apache2?
[14:09] <kwork> i have two dns servers, i would like to test if master/slave servers all have correct records, can anyone recomend tool for what ?
[14:11] <Sam-I-Am> dig
[14:11] <kwork> its n+1 domains
[14:11] <Sam-I-Am> script + dig
[14:11] <kwork> so there is no tool out there for that ?
[14:11] <Sam-I-Am> dig will do an axfr from your dns servers if you permit it
[14:12] <Sam-I-Am> you could easily compare the axfr from one server to the other
[14:12] <kwork> i was just hope-ing there is something better then dig + diff
[14:13] <jiboumans> kirkland, ttx, how/when is this list triaged? https://bugs.edge.launchpad.net/ubuntu/+source/eucalyptus
[14:13] <kirkland> jiboumans: i just sort of do it when i get a chance
[14:14] <kirkland> jiboumans: though it should be done as part of our daily triage work
[14:14] <ttx> jiboumans: I wanted to have 1.6.2 and confront that list to it, but if it's delayed we should have a look at it sooner
[14:14] <kirkland> jiboumans: i'll triage the 3 undecideds right now, if you like
[14:14] <Sam-I-Am> kwork: most other things probably do roughly dig + diff heh
[14:14] <ttx> kirkland: I'm trying the proxool patch from Chris rigth now
[14:14] <ttx> unless you're at it already
[14:14] <Sam-I-Am> time to wander off to work
[14:15] <kirkland> ttx: no sir, please do
[14:15] <kwork> Sam-I-Am,  as of i m lazy as hell, i was hope-ing someone will point me towrads something that will do dig+diff
[14:15] <kwork> towards
[14:15] <kirkland> ttx: i just got up, about to make some coffee
[14:18] <kirkland> ttx: we have eucalyptus-nc upstart scripts ;-)
[14:18] <ttx> kirkland: I saw that, thx
[14:19] <kirkland> ttx: it worked well for me
[14:19] <kirkland> ttx: i had keybuk and slangasek review it too
[14:19] <kirkland> ttx: there's one thing we couldn't agree upon, though
[14:19] <kirkland> ttx: in my initial implementation, i had it start on started eucalyptus
[14:19] <kirkland> ttx: thus, on the node, you could just do "sudo start eucalyptus"
[14:20] <kirkland> ttx: and it would "figure out" that you have only a node controller, and it would start that
[14:20] <kirkland> ttx: however, it took kind of a hack in the upstart conf to make that work
[14:20] <kirkland> ttx: and i was frowned upon
[14:20] <kirkland> :-)
[14:20] <ttx> ok
[14:20] <ttx> I'm ok with it starting with eucalyptus-nc
[14:20] <kirkland> ttx: cool, fair enough
[14:20] <ttx> since it's always on separate machines
[14:25] <ttx> kirkland: patched proxool in my PPA, test coming up
[14:25] <kirkland> ttx: cool
[14:26] <kirkland> ttx: i'm trying to package qemu-kvm-0.12, which is in RC state
[14:26] <kirkland> ttx: upstream asked for some feedback before they GA
[14:41] <smoser> good morning all.
[14:41] <smoser> kirkland, does it let me use -curses  with lucid :)
[14:41] <ttx> kirkland: in case you missed my last message before the split, looks like it's starting up now
[14:41] <kirkland> smoser: is that a question?  if so, no, not in my experience so far :-(
[14:42] <ttx> kirkland: the cglib2.1 trick is still needed though
[14:42] <kirkland> smoser: please talk to cjwatson about that
[14:42] <smoser> yeah, sort of a qustion, but it should have ended with "if it does i'll use it" :)
[14:42] <kirkland> ttx: i didn't see your message
[14:42] <smoser> i will. cjwatson , ping
[14:42] <kirkland> smoser: yes, i very heavily depended on that in Karmic
[14:43] <kirkland> smoser: my testing is much more difficult in Lucid because of it
[14:43] <smoser> in mail i sent you last night it shows at least how to get a console up on ttyS0
[14:43] <smoser> which is a big help if you cant get a "real console"
[14:44] <X-Sleepy-X> If I want to execute a command every minute how would I achieve that with cron?
[14:45] <X-Sleepy-X> The command I want to execute is: acpi -Bt 2>&1 | tee temp.txt
[14:46] <kirkland> smoser: ah, i see that now
[14:46] <kirkland> smoser: you sent it to my gmail :-)
[14:46] <smoser> where do you want things sent?
[14:46] <kirkland> smoser: work related?  kirkland@canonical
[14:48] <smoser> kirkland, so, i just booted a karmic image, and dont see any gettys on f1 or f2
[14:49] <kirkland> smoser: rsyncing ....
[14:50] <Daviey> kirkland / smoser: It is documented on the wiki.
[14:50] <kirkland> Daviey: whereabouts?
[14:51] <Daviey> https://help.ubuntu.com/community/SerialConsoleHowto
[14:52] <Daviey> kirkland: before you showed me kvm curses, i was using a serial console as a fall back to ssh.
[14:52] <smoser> Daviey, thats what i came up with.. the upstart method.
[14:53] <smoser> although i didn't do 'vt102' arg to sgetty, which might have been helpful.
[14:53] <kirkland> smoser: okay, i can get to gettys on f*
[14:53] <kirkland> smoser: at least launched from a kvm SDL window
[14:53] <kirkland> smoser: alt-f2, alt-f3, etc.
[14:53] <smoser> yeah, i see them if i run in sdl window
[14:54] <kirkland> smoser: drop the ctrl
[14:54] <smoser> but not in curses
[14:54] <kirkland> smoser: ah, right, yeah, i've never gotten that working
[14:54] <smoser> this is karmic kvm booting karmic image (uec image, but thats not too important at this point)
[14:54] <kirkland> smoser: i don't think the key stroke is being captured/passed properly
[14:54] <smoser> hmm..
[14:55] <smoser> kirkland, well, in curses mode i get to the kvm console with 'alt-2'
[14:55] <smoser> then 'sendkey alt-f2'
[14:55] <smoser> how do i get to kvm console window in SDL ?
[14:57] <kirkland> smoser: alt-1
[15:02] <smoser> kirkland, so if i omit '-curses', (ie, i get sdl)
[15:02] <kirkland> smoser: right
[15:03] <smoser> Using karmic host, karmic guest.
[15:03] <smoser> then i can go to kvm console (with 'ctrl-alt-f2') and type 'sendkey alt-f1'
[15:03] <smoser> oops
[15:03] <smoser> above 'sendkey alt-f2'
[15:03] <smoser> then hit 'ctrl-alt-f1' and i can see a login prompt on , with 'tty2' label
[15:04] <smoser> if i do the same with '-curses' i can't see a login prompt
[15:04] <smoser> in karmic guest it doesn't tell me a graphic mode (like it does with lucid guest), but there is no login
[15:04] <smoser> same process (using sendkey from the console) but console keymaps are 'alt-fx' rather than 'ctrl-alt-fX'
[15:05] <smoser> does that all make sense?
[15:05] <smoser> i swear i've seen all this work at some point
[15:07] <kirkland> smoser: what's the end goal?
[15:07] <kirkland> smoser: to get a tty login on tty2?
[15:11] <smoser> be able to login at console via -curses
[15:15] <ttx> kirkland: so it looks like it should work with the proxool fix. I can upload that fix when I get Chris's explanation on potential regressions with it
[15:17] <acalvo> hi
[15:18] <acalvo> why sometimes, depending on which server I ping, I get a response from servername.local and sometimes from the ip? (usually form servername.local is faster than the IP)
[15:19] <zul> jdstrand: https://launchpad.net/bugs/497790 <-- seen that one?
[15:19] <jdstrand> zul: I did 2 minutes ago, yes ;)
[15:19] <zul> pretty interesting ;)
[15:21] <ttx> kirkland: i'm not exactly sure yet of the best way to handle the cglib situation
[15:22] <ttx> kirkland: we can get rid of the new package and keep the old one, or tweak the new one so that it doesn't interfere with the old one
[15:23] <jdstrand> zul: https://wiki.ubuntu.com/SecurityTeam/Roadmap#AppArmor%20Confinement
[15:23] <jdstrand> zul: I added the bug reference to there
[15:24] <jdstrand> zul: I don't think the security team will be able to manage adding that to the squid package in enforcing mode (but you are welcome to). We could add it to apparmor-profiles though
[15:24] <zul> i think it would be nice to have though
[15:25] <zul> but it would be totally up to you guys
[15:25] <jdstrand> zul: it would-- it has been on our roadmap. we just don't have the resources this cycle to turn it on by default
[15:25] <zul> jdstrand: ok sounds good
[15:26] <jdstrand> zul: if you aren't interested in putting it into enforcing mode yourself, then I'll add it to apparmor-profiles
[15:26] <jdstrand> (or don't have time, or however that should be phrased to be non-offensive ;)
[15:26] <jdstrand> zul: ^
[15:26] <zul> jdstrand: i think it would be a good idea because it would also take care of the request to have squid chroot
[15:27] <jdstrand> I aboslutely agree-- we just don't have the resources this cycle
[15:28] <jdstrand> zul: well, *I* don't. I suppose it is possible kees or mdeslaur might want to take it on, but I think they have their plates full too
[15:28] <zul> jdstrand: heh ok ill bug them about it after christmas
[15:33] <kirkland> smoser: from an installed system, or from an installer ISO?
[15:33] <kirkland> smoser: that "just works" with my installed systems
[15:34] <kirkland> ttx: good to hear that the proxool fix gets us moving
[15:34] <kirkland> ttx: tweaking the new one seems the best way forward
[15:34] <ttx> kirkland: not so sure about that
[15:34] <kirkland> ttx: for the big picture
[15:34] <kirkland> ttx: hmm
[15:34] <ttx> I think having proxool 0.8.3 in eucalyptus-commons-ext starts to make sense
[15:35] <ttx> since lots of packages now expect libcglib-java to have replaces libcglib2.1-java
[15:35] <kirkland> ttx: heh :-)
[15:35] <ttx> (that's the situation in debian)
[15:35] <ttx> i'll write up a summary so that we can discuss it in the call
[15:36] <kirkland> ttx: sounds good
[15:38] <jiboumans> ttx++ # on the ball
[15:38] <jiboumans> ttx, kirkland: make sure to add the point to the agenda for transparency
[15:50] <ttx> sent -- hopefully it will make the options clearer
[16:08] <rputnins> hello!
[16:09] <rputnins> does anyone has installed 8.04 server on software raid5?
[16:09] <rputnins> I know that officialy it's not possible but I think there must a way
[16:10] <rputnins> after installation and reboot I get grub rescue console
[16:11] <zul> smoser: ping
[16:15] <ruben23> hi is there i way that when i login using ssh to my ubuntu-server i dont have to input password, which i used oftenly..
[16:15] <rputnins> any ideas about the raid5?
[16:15] <bogeyd6> rputnins there is a way
[16:16] <rputnins> good! how?
[16:16] <bogeyd6> https://help.ubuntu.com/community/Installation/SoftwareRAID
[16:16] <bogeyd6> ruben23 if you set up keys for ssh you dont need a password
[16:18] <bogeyd6> !fakeraid | rputnins
[16:19] <rputnins> bogeyd6 - I have tried this manual, the is nothing about raid5! Only that it will not work for / partition
[16:21] <rputnins> ubottu: thanks, I havent tried the fakeraid yet, I will give it a try
[16:26] <smoser> zul, here
[16:26] <LinuxCommunity> im here
[16:27] <bogeyd6> here
[16:27] <zul> smoser: for ec2 are you still suppose to be getting a message to use tasksel when you first login? (this is for lucid)
[16:28] <smoser> i had not made a decision that you should not, i've never looked at that stuff.
[16:28] <rputnins> ubottu: the fakeraid install is for desktop only, am I wright? How to install server on software raid5?
[16:30] <smoser> zul, does it not show that any more ?
[16:31] <zul> smoser: no but I might logged in to early
[16:31] <zul> smoser: it doesnt
[16:33] <smoser> i dont have a personal big interest in that showing up (especially since its broken, it says 'use tasksel' but doesn't say 'run apt-get update first')
[16:33] <LinuxCommunity> sudo apt-get dist-upgrade    and sudo apt-get update
[16:34] <zul> smoser: ok i was just checking to see if its a regression
[16:34] <zul> smoser: it doesnt matter to me either way though ;)
[16:34] <smoser> it probably is a regression. i suppose if you dont mind please follow a bug
[16:35] <smoser> its easiest to run 'ubuntu-bug ec2init' on the instance
[16:35] <zul> okies
[16:35] <smoser> or if not easiest, preferred.
[16:36] <rputnins> ubuntu 8.04 boot from software raid5 possible? Opinions!
[16:36] <LinuxCommunity> rputnins: it should be possible, i run 8.04 server
[16:37] <rputnins> do you have / partition on raid 5?
[16:37] <msantos> rputnins, i boot 8.04 off of a software raid1
[16:37] <LinuxCommunity> rputnins: no i don't but you should be able to run raid 5
[16:38] <rputnins> msantos: thanks, on raid1 I have seen dozens of tutorials, I need raid5
[16:38] <rputnins> LinuxCommunity: yeah, I know I should but how? :)
[16:39] <LinuxCommunity> rputnins: let me see if i can get you a good tutorial
[16:40] <rputnins> LinuxCommunity: installation was fine after the reboot I get: GRUB loading.
[16:40] <rputnins> erro: biosdisk read error
[16:40] <rputnins> erro: file not found
[16:40] <rputnins> grub rescue>
[16:40] <LinuxCommunity> ehh
[16:40] <LinuxCommunity> yea that's not good
[16:41] <msantos> rputnins, are you using LVM or just staight raid?
[16:41] <LinuxCommunity> rputnins: you might want to read this   https://bugs.launchpad.net/ubuntu/+source/linux/+bug/220493   not sure how relavent it is
[16:41] <rputnins> msantos: I am using just raid, I thought about lvm, you think I should try?
[16:42] <zul> smoser: done
[16:42] <msantos> it's trickier with LVM :)
[16:42] <smoser> danke
[16:42] <LinuxCommunity> this one to http://nerdonaplane.blogspot.com/2008/04/ubuntu-upgrade-hardy-harron-804-lts.html
[16:42] <msantos> what you can try is partitioning your disks into a /boot (straight raid1 or whatever) and the rest as a big raid 5
[16:43] <TeTeT> how do I increase the number of loop back mounts on Karmic? the loop module seems to be built in now
[16:43] <msantos> if your /boot is on raid 1, just be sure to install grub on all the disk or it's fun after an hd dies ;) (not that I had this happen to me <cough>)
[16:43] <rputnins> LinuxCommunity: msantos: OK, I am running the server install right now
[16:44] <msantos> rputnins, have to go, good luck on your install
[16:45] <LinuxCommunity> rputnins: i don't run raid. i still got my server running off old IDE. i would like maybe scsi or sata 3 but for what im doing theres no profit
[16:45] <rputnins> msantos: I see why I havent used LVM - I dont have it in manual partition part
[16:45] <LinuxCommunity> LVM isn't always the best way to go. most times it's just use whole disk
[17:03] <mathiaz> ttx: hey
[17:03] <mathiaz> ttx: doing some uec testing
[17:04] <mathiaz> ttx: it seems that the cluster name is not asked at the same debconf priority during the cc and the sc package installation
[17:04] <mathiaz> ttx: would that cause a problem when doing an SC only install?
[17:04] <ttx> mathiaz: istr it had a sane default
[17:05] <ttx> mathiaz: haven't tested that yet
[17:06] <mathiaz> ttx: I think it will fail on a single SC installation
[17:06] <mathiaz> ttx: cluster-name is only asked for in eucalyptus-cc.config
[17:06] <mathiaz> ttx: eucalyptus-sc.postinst just db_get it
[17:07] <ttx> mathiaz: you sound on the right track :)
[17:11] <jiboumans> guys, pitti just mailed out an update for the workitem tracker; bugs are now counted as work items. so time to do another quick pass at your blueprints and make sure duplicate items are removed
[17:14] <rputnins> no luck with that raid
[17:15] <rputnins> :(
[17:28] <jiboumans> cya tomorrow guys
[17:44] <rputnins> hello folks! any ideas on howto setup ubuntu 8.04 server on softrware raid 5?
[17:44] <rputnins> to be able boot from raid5?
[18:26] <Doonz>  Hey guys i never installed a webserver on my box but when i goto http://server it tell me that the webserver is working
[18:26] <Doonz> how can i see what web server may have been accidently installed?
[18:26] <jpds> Doonz: Check: sudo netstat -ltnp
[18:27] <Doonz> it doenst show anything
[18:29] <Doonz> so who wants to give me a cookie..... it was using some chached page on my browser :/
[18:29] <jpds> Doonz: The site should of given your a cookie.
[18:29] <Doonz> im a noobie
[18:29] <Doonz> sorry for bothering
[18:30] <jpds> Doonz: no problem, remember, the stupid quesitons are only the unasked ones!
[18:30] <Doonz> thanx ... back to my virtual machine
[18:32] <xperia> hello to all. i have a question about phpmyadmin. for some reason the password and the username that i use in mysql dont work in phpmyadmin. i cant pass the login screen in phpmyadmin. does anybody know what the problem could be ?
[19:04] <zul> jiboumans: server-landscape-client-refresh is basically done fyi
[19:14] <kirkland> smoser: image names changed?
[19:14] <kirkland> http://uec-images.ubuntu.com/karmic/current//karmic-server-amd64.tar.gz
[19:14] <kirkland> smoser: that was karmic-uec-amd64 yesterday ....
[19:14] <smoser> not yesterday
[19:14] <smoser> but recently
[19:14] <kirkland> smoser: k, recently :-)
[19:15] <smoser> it was to accomodate '-desktop'
[19:15] <kirkland> smoser: i updated https://help.ubuntu.com/community/UEC/BundlingImages
[19:15] <kirkland> smoser: i'm not sure where else that string is
[19:15] <smoser> yeah. i dont know. i wondered about it when i did it.
[19:15] <kirkland> smoser: it's not a big deal
[19:15] <kirkland> smoser: but we'll have some docs to update
[19:16] <smoser> yeah, i'm not certain if the new name is good, though, or if it should include 'uec' in it.
[19:18] <kirkland> smoser: hmm, if the seeds that are generating those images are basically identical (or a superset) of what you get with a default server or desktop image, leave the name as is
[19:18] <kirkland> smoser: if the seed is somehow different
[19:19] <kirkland> smoser: specifically, if it's a subset of what in the desktop/server default install, then I think adding a uec appendage would be nice
[19:19] <smoser> it is somewhat different. it has ec2init for example.
[19:28] <M1TE5H> can anybody help to configure startx
[19:35] <genii> M1TE5H: This is probably the wrong channel to ask about X, since Ubuntu Server doesn't have it
[19:35] <M1TE5H> ok thanks 4 reply can give me any other channel list
[19:37] <ScottK> M1TE5H: #ubuntu
[19:42] <kirkland> jiboumans: mdz: \o/  eucalyptus running from packages
[19:43] <kirkland> jiboumans: mdz: specifically, i was able to a) install from lucid cd, b) upgrade using PPA packages, c) add iscsi to the system (hack for now), d) bundle and run an image
[19:44] <RoyK> http://www.robweir.com/blog/2009/12/relevancy-of-odf-10.html
[19:45] <mdz> kirkland, great news
[19:45] <mdz> kirkland, ready to upload to lucid?
[19:46] <kirkland> mdz: i think so ... though users will need to either install iscsitarget on their CLC (for now)
[19:46] <kirkland> mdz: or we need to wait for the code drop from Eucalyptus that disables the iscsi paths
[19:46] <mdz> kirkland, is that not trivial to fix?
[19:46] <kirkland> mdz: i'll look at their code and see
[20:03] <mathiaz> kirkland: which ssh public key uec component is supposed to go where?
[20:03] <mathiaz> kirkland: IIUC the CC public key needs to be on the NCs?
[20:04] <kirkland> mathiaz: correct
[20:04] <mathiaz> kirkland: what else is needed?
[20:04] <kirkland> mathiaz: CC issues commands to NCs
[20:04] <kirkland> mathiaz: I think CC also talks to its SC
[20:05] <kirkland> mathiaz: and i think CLC talks to Walrus and CCs
[20:05] <kirkland> mathiaz: nurmi is online and active in #eucalyptus, you might ask him there
[20:14] <mathiaz> kirkland: the home directory of the eucalyptus user is /var/lib/eucalyptus?
[20:15] <kirkland> mathiaz: ack
[20:19] <nuclearshare> hello
[20:19] <nuclearshare> is anyone here?
[20:59] <ttx> kirkland: how is it going ?
[21:00] <kirkland> ttx: testing is good
[21:01] <ttx> cool
[21:01] <kirkland> ttx: i'm trying to figure out what of yours needs to be uploaded
[21:01] <kirkland> ttx: i'll upload eucalyptus
[21:01] <kirkland> ttx: as i have another change or two necessary
[21:01] <kirkland> ttx: could you upload the rest?
[21:01] <kirkland> ttx: common, cglib, proxool, whatever?
[21:01] <kirkland> ttx: i have a fix for the iscsi thing, to turn it off
[21:01] <ttx> kirkland: proxool was uploaded
[21:02] <kirkland> ttx: and common?
[21:02] <ttx> kirkland: you don't need the eucalyptus-commons-ext in my PPA, but you need one 0.5.0
[21:02] <ttx> is it still stuck in DEPWAIT ?
[21:03] <ttx> I'd thought that component-mismatches would not be blocking
[21:03] <ttx> maybe it's just a matter of retrying the build of the one already uploaded
[21:04] <kirkland> ttx: okay, i'll handle that
[21:04] <ttx> or we might need some shortcut
[21:04] <ttx> the mir is filed, if we need to fast-track it in some way
[21:04] <ttx> but i'm just not sure that's why it's blocked :)
[21:05] <ttx> kirkland: cool, looks like we might end up releasing 1.6.2 before you go for holidays, after all
[21:07] <kirkland> ttx: i'm trying to do so
[21:07] <kirkland> ttx: kees any chance you can help us fast track a simple MIR?
[21:07] <kirkland> kees: it's blocking eucalyptus 1.6.2
[21:07] <kirkland> kees: which is urgently needed in Lucid
[21:07] <kirkland> ttx: can you paste the MIR url?
[21:08] <ttx> sure
[21:09] <ttx> bug 497390
[21:09] <ttx> might also need bug 497455 (runtime dep)
[21:09] <kirkland> ttx: let's elevate the priority from Wishlist to High
[21:09] <ttx> kirkland: be my guest
[21:10] <kirkland> ttx: done
[21:10] <ttx> kees: would a pending MIR on a new build-dep result in a DEPWAIT for the package ?
[21:10] <ttx> kirkland: did you try to restart the build ?
[21:10] <kirkland> ttx: not yet
[21:11] <ttx> I can do that.
[21:11] <kirkland> ttx: i just did
[21:11] <ttx> ok
[21:15] <kirkland> ttx: ordered 3 more vostros
[21:15] <ttx> heh
[21:19] <ttx> kirkland: stuck in depwait again, so I guess universe build-deps are blockers :)
[21:21] <kirkland> ttx: yeah, i suspect
[21:21] <ttx> hmmmm
[21:22] <ttx> kirkland: you know what, I'm not 100% sure that proxool is needed for the build
[21:22] <ttx> I'm trying to see who added it
[21:24] <kirkland> ttx: interesting, okay
[21:24] <ttx> arh
[21:24] <ttx> nah, I'm confusing things
[21:24] <ttx> it's not a b-d of euca, but a b-d of euca-commons-ext
[21:24] <kirkland> right
[21:26] <ttx> and the b-d was added by chris, 99.9% sure needed to compile hibernate-proxool
[21:26] <ttx> so no easy way out
[21:26] <ttx> :)
[21:26] <kirkland> okay, so we need 3 MIRs approved
[21:26] <kirkland> ttx: before eucalyptus 1.6.2 can be built for Lucid, then, right?
[21:27] <ttx> the woodstox one might not be blocking
[21:27] <ttx> let me check if its a b-d
[21:27] <ttx> no it's not
[21:27] <ttx> And I wonder if that part of their codechanges even landed
[21:28] <ttx> so it's proxool that's needed. and avalon-framework is a b-d of prxool
[21:30] <sirenita> hi.. i have a problem with an instalation of two instances of oracle database in the same server
[21:31] <sirenita> somebody can herlp me
[22:04] <sjefen6> getting message "the attempt to mount a file system with type ext4 in LVM VG Unknown...., LV root at / failed" during the installation of ubuntu server 9.10 on a dell optiplex gx260. What is wrong?
[22:48] <owh> If I use channel bonding to combine three Ethernet interfaces, using adaptive transmit load balancing (mode 5), is there a way that I can force all ssh traffic to one specific slave?
[22:49] <sub> no. why would you want to? what would the behavior be if that specific slave interface is down?
[22:51] <ruben23> hi guys, is it possible to ssh without always making input of passwd
[22:53] <owh> sub: I want to because I have 3 WAN interfaces, one of which has unlimited traffic which I need to use for backups. I do not want to use the other 2 interfaces which have higher throughput, but limited traffic.
[22:53] <owh> I don't particularly care if it dies if the interface is down.
[22:55] <sub> Ethernet bonding != IP load balancing
[22:59] <owh> sub: That's interesting, since that's not what I've been reading. Specifically the documentation tells me that mode 5 is "Adaptive transmit load balancing: channel bonding that does not require any special switch support. The outgoing traffic is distributed according to the current load (computed relative to the speed) on each slave. Incoming traffic is received by the current slave. If the receiving slave fails, another slave 
[22:59] <owh> Unless of course I cannot read, which is a possibility :)
[23:00] <sub> do each of your 3 WAN interfaces reside on different subnets?
[23:00] <sub> ie: have different IPs?
[23:01] <owh> They are three completely separate connections from different ISPs. There is a modem between me and the link for each of the links. Each modem has a separate WAN IP address and normally a workstation would connect to the modem and get a LAN address from the on-board DHCP server.
[23:02] <owh> Think of it as combining three DSL connections. It's not that, but the analogy holds.
[23:03] <sub> no, it doesn't, because each packet out is likely to have different IPs
[23:03] <sub> thanks to NAT
[23:04] <owh> That is correct, but under mode 5, I understand that a connection is routed through a specific slave once it has been established.
[23:05] <owh> Hmm, perhaps I should be looking at mode 2.
[23:06] <sub> dest mac address will always be your gateway's mac in that case. i don't know if it will continue to route anything through the same interface on mode 5 because it specifically says it's based on load and also interface bonding is a different layer than mode 5, but admittedly this isn't something i'm 100% familiar with
[23:07] <owh> Let me ask the question in a different way. I have 3 WAN links, two high speed, with limited data, one lower speed with unlimited data. I need to route all ssh traffic across the unlimited data link and want to provide the best throughput for LAN clients who use the server as their gateway. How do I accomplish this?
[23:07] <sub> i've done etherchannel and LACP, but always at the ethernet layer and never had to worry about NAT/routing issues
[23:08] <owh> LACP assumes that the other end can deal with what's coming in, almost like having a multi-link ppp support at both ends. That is not the case here. The links are completely separate, think of them as three domestic ADSL connections.
[23:09] <owh> I am not particularly interested in one user having all available bandwidth, just the best experience for all users.
[23:09] <sub> i realize that, I was explaining my background with bonding
[23:09] <owh> Cool.
[23:10] <owh> Just making sure I'm on the same page too :)
[23:10] <sub> in fact, i wrote this: http://subversity.net/link-aggregation-on-a-redhat-centos-server-an :|
[23:10] <sub> well
[23:10] <sub> because of NAT and routing, you likely want something to do your load balancing at the IP layer
[23:11] <sub> for forcing the SSh traffic over a certain interface you could probably use iptables, I'm not familair with the rules enough, I just know that you can forward it out a specific interface
[23:11] <sub> as for load balancing the remaining two links, i'm not really sure
[23:11] <owh> Just to make sure that I'm making myself clear, if one user is downloading a big file, I don't want or need them to have their download spread over three links - in fact I cannot see how that would ever work unless I setup a VPS in the cloud, make a VPN to that and use its pipe to get the data - not really what I had in mind.
[23:13] <owh> sub: Yes, that article is not what I'm looking for. (For that I have a gigabit port on the server -- more than enough for 20 users in this environment :)
[23:13] <sub> I know it's not what you're looking for, once again
[23:13] <owh> Basically we have a crappy link, it's slow but unlimited. We're locked into another two years of contract and need to increase bandwidth.
[23:13] <sub> Yeah, I get that, and my answer was:
[23:14] <sub> as for load balancing the remaining two links, i'm not really sure
[23:14] <sub> for forcing the SSh traffic over a certain interface you could probably use iptables, I'm not familair with the rules enough, I just know that you can forward it out a specific interface
[23:14] <owh> So, then it appears that you're saying that bonding isn't the right hammer for this nail.
[23:15] <owh> (As in, when all you have is a hammer, every problem looks like a nail.)
[23:15] <sub> correct
[23:15] <owh> Back to the drawing board.
[23:15] <sub> bonding is generally used when you need more throughput to the LAN, or redundancy to the LAN
[23:16] <sub> look into FORWARD rules for iptables
[23:16] <sub> should be easy enough to forward tcp/22 traffic over a specific links, I can't answer for the load balancing though
[23:16] <owh> Bear with me for a moment if you will. You can tell me to bugger off if you've lost interest :)
[23:19] <owh> The three WAN links have different WAN addresses. Internally they can be configured to have the same LAN IP addresses. If I bond the interfaces and for a moment ignore ssh, would mode 2 not work? I understand that it keeps the same client/server pair on the same slave unless it fails. Or am I missing something?
[23:20] <sub> how are you planning on giving them the same LAN IP address?
[23:20] <jiboumans> kirkland++ ttx++ nice job on eucalyptus
[23:21] <kirkland> jiboumans: we're almost there
[23:21] <kirkland> jiboumans: gotta get the MIRs approved
[23:21] <kirkland> jiboumans: kees is helping us with that
[23:21] <owh> The three modems can be configured to give the same address via DHCP. Alternatively, if the bond defines the IP address, then I can configure each modem to see that as a valid LAN address.
[23:21] <jiboumans> kirkland: anything contentious there?
[23:21] <kirkland> jiboumans: just the urgency/late-breaking nature of all-things Eucalyptus
[23:21] <jiboumans> kirkland: of course =/
[23:22] <jiboumans> better in alpha2 than beta1 though :)
[23:22] <sub> having three modems that give out the same IP address is not the same as giving each modem the same IP address, not that it matters because that would just be an IP conflict and would result in a broken network
[23:23] <owh> But the modems are not connected to the LAN, they are only connected to my ethernet sockets.
[23:23] <sub> i guess i don't really understand your question
[23:23] <sub> they're in bridging mode?
[23:23] <owh> No
[23:23] <sub> do the modems have internal IP addresses?
[23:24] <owh> Yes
[23:24] <sub> can you make a diagram of what you are describing? i'm not really sure where you're going with this
[23:24] <owh> (Remember, there is a forth interface that connects to the LAN.)
[23:24] <sub> also, have you looked into something like pfSense?
[23:25] <sub> oh
[23:25] <sub> so you're saying the modems are connected directly to your gateway
[23:25] <sub> ok
[23:25] <jiboumans> zul++ # for doing the landscape code drop
[23:25] <sub> i don't think you're going to get where you want by using bonding
[23:26] <jiboumans> i'm very happy looking at this chart: http://piware.de/workitems/server/lucid-alpha2/report.html
[23:26] <owh> sub yup. Still need a picture?
[23:27] <sub> no, i think i've got it now
[23:27] <sub> i still say look into iptables
[23:28] <owh> I will. Thank you for your comments.
[23:30] <sub> i've had the best luck with that sort of thing using pfsense, check out http://doc.pfsense.org/index.php/Multi_WAN_/_Load_Balancing and see if it can't get you where you need to go
[23:30] <sub> #pfsense on this network as well
[23:34] <owh> sub: This would mean that I'd have to introduce another piece of hardware to run pfsense on, unless I misunderstand.
[23:47] <owh> In case anyone cares, what I appear to be looking for is called multiwan, thanks to sub for pointing me at pfsense, which lead me down the trail.