[00:01] <sparc> Hey there.  I was following the automated deployment guide, but it doesn't seem to have a list of boot parameters I can provide to Linux, on boot.
[00:02] <sparc> I need to do things, like set a static ip, and choose a network interface
[00:02] <sparc> is there a place I can look for that documentation?
[00:03] <zeknox> sparc: ubuntu.com
[00:03] <screen-x> sparc: is this so you can do a diskless netboot?
[00:03] <sparc> the (first stage?) installer isn't able to download my kickstart file, without
[00:03] <fluvvell> Danny78, my parting shot -  use this template smb.conf instead     http://pastebin.com/aCsxbnEA
[00:03] <sparc> screen-x: yes, that sounds right.  although the server does have disks
[00:04] <sparc> i just want my kickstart to start, but it's stuck not being able to find an ip with DHCP
[00:04] <screen-x> sparc: no DHCP server?
[00:04] <sparc> nope, our servers are ip'd statically
[00:04] <Danny78> fluvvell:  I checked win and now I can see buddha!  now need to make it permanent.  Will look at your suggested conf and see any differences
[00:04] <screen-x> doesn't mean you can't have a DHCP server ;-)
[00:05] <fluvvell> I'll be back on in about 4 or 5 hours, then probably tomorrow.   OK Thats good,  try mine as well, dont forget to restart samba each time you change it.
[00:06] <fluvvell> It takes a couple of minutes to be seen on a winbox.  You may also try rebooting the winbox if all else fails.
[00:06] <fluvvell> cheers
[00:06] <Danny78> fluvvell:  thanks again for your help!
[00:06] <fluvvell> np
[00:18] <sparc> It's funny, you can specify the ethernet interface to use with: netcfg/choose_interface=eth1
[00:19] <sparc> it looks like it's following a pattern
[00:19] <sparc> maybe there are more netcfg/* options
[00:20] <twb> sparc: er, yeah.  preseeding works that way
[00:21] <twb> sparc: the installation guide has an appendix that covers preseeding in detail
[00:24] <sparc> twb: hah! excellent, thank you
[01:21] <sparc> twb: aah yeah, i just saw that, this is great
[01:21] <sparc> it mentions i can create a preseed/run file
[01:22] <twb> What of it
[01:22] <sparc> to run things before it fetches the preconfiguration files from the network
[01:22] <sparc> is that something i would put on the ftp server, or inside the initrd?
[01:23] <sparc> https://help.ubuntu.com/10.04/installation-guide/amd64/preseed-contents.html#preseed-network
[01:23] <sparc> ^-- for reference
[01:23] <sparc> uvirtbot: shush bot :)
[01:24] <sparc> maybe that's in the guide too and i missed it
[01:24]  * sparc will google dilligently
[01:42] <twb> sparc: you should be reading the sparc one if you're on a sparc
[01:43] <twb> sparc: the document describes places you can put the preseed file; usually it's either placed on the CD, on an HTTP server, or inside the ramdisk
[01:43] <sparc> naah, i'im on x86_64
[01:43] <sparc> luckily
[01:43] <twb> sparc: very very new d-i instances can also get it from TFTP, from the same place as the PXE server.
[01:43] <twb> SPARC was a good architecture, man
[01:44] <sparc> if only we had some sun here
[01:44] <sparc> we're all IBM and HP
[01:44] <sparc> on x86
[01:44] <sparc> i'm sure the sun jumpstart is well thought-out
[01:44] <sparc> i'm on 10.04, so maybe i can put it on tftp
[01:45] <sparc> well, i'd put it on an HTTP server, if i had a network interface configured
[01:45] <twb> I doubt 10.04 is new enough, but feel free to try it
[01:45] <twb> sparc: typically one passes "auto" to the ramdisk, so it brings the network up before prompting
[01:45] <sparc> my issue, if that it's not configured by the time, it wants a kickstart file
[01:46] <twb> sparc: uh, preseed and kickstart do the same job, in different ways.  I don't think you can use both at once.
[01:46] <sparc> the docs say that we have to, don't they
[01:46] <sparc> very confusing
[01:46] <twb> The installation-guide document is written by Debian, it probably doesn't mention kickstart at all.
[01:47] <sparc> they tell us to put the preseed file in the same place as the kickstart
[01:47] <twb> That doesn't make ANY sense
[01:47] <sparc> yeah, there's a big section on kickstart in it
[01:47] <twb> What are you reading
[01:48] <sparc> The automated deployments white paper on Ubuntu.com, and the 10.04 installation-guide
[01:48] <twb> URL for the former?
[01:48] <sparc> https://help.ubuntu.com/10.04/installation-guide/amd64/preseed-contents.html#preseed-network # for the preconfiguring network section
[01:48] <sparc> and..
[01:50] <sparc> http://www.canonical.com/about-canonical/resources/white-papers/automated-deployments-ubuntu
[01:50] <sparc> i do like the automated deployment guide
[01:51] <twb> *I* can't help much with kickstart, because it's a RHism that was backported to Ubuntu/Debian
[01:51] <sparc> it just assumes that you're gonna use DHCP to get an ip for the server
[01:51] <sparc> after pxe boot
[01:51] <twb> preseeding (as covered in installation-guide-amd64) is the native equivalent
[01:51] <sparc> that's okay, i think i have my kickstart config already amde
[01:51] <twb> sparc: you can pass something like ip=... to force it to use static instead
[01:51] <sparc> i ported mine from redhat
[01:51] <sparc> twb: that would be very nice
[01:51] <twb> The installation-guide also mentions somewhere the options you can pass for that stuff
[01:52] <twb> apt-get install it and grep for it
[01:53] <sparc> it does mention boot params, but the network isn't included
[01:53] <sparc> https://help.ubuntu.com/10.04/installation-guide/amd64/boot-parms.html
[01:53] <sparc> just saying. i did try to do my homework before asking
[01:53] <sparc> maybe DHCP is just the only way to do it
[01:54] <twb> https://help.ubuntu.com/10.04/installation-guide/amd64/preseed-using.html#preseed-auto ?
[01:54] <twb> I *think* I've done it before
[01:54] <twb> Maybe I'm thinking of live-boot/casper
[01:54] <twb> You can preseed anything that debconf asks, so you might need to find the value in the d-i documentation
[01:54] <sparc> "This relies on there being a DHCP server that will get the machine to the point where autoserver can be resolved by DNS"
[01:54] <twb> sparc: that's only for url=autoserver
[01:55] <sparc> hmm ok
[01:56] <sparc> redhat has the ip= netmask= network= gateway=
[01:56] <sparc> options, which are pretty nice
[01:56] <twb> Do those work?
[01:56]  * sparc tries
[01:59] <sparc> Hmm, i guess this is what commerical support is for
[01:59] <twb> https://help.ubuntu.com/10.04/installation-guide/example-preseed.txt shows static config as an example
[01:59] <twb> e.g. #d-i netcfg/get_nameservers string 192.168.1.1
[01:59] <twb> becomes netcfg/get_nameservers=192.168.1.1
[01:59] <twb> ...when passed directly by the bootloader to the kernel/ramdisk
[02:00] <twb> *some* of them have shorthand, e.g. fb=false is shorthand for something
[02:00] <sparc> OH
[02:00] <sparc> so i can lift some of these from the example file
[02:01] <twb> Sure
[02:01] <sparc> and not have to put a special preseed file, somewhere in the ramdisk or on the tftp server
[02:01] <twb> Or RTFS
[02:01] <twb> sparc: well, you can only pass about 32 options from the bootloader
[02:01] <twb> So more than that you'll have to go elsewhere
[02:01] <sparc> that's okay, it shouldn't take too many to do the network
[02:01] <sparc> sweet, i'll give that a go, thank you twb
[02:02] <twb> Anyway, how can you be PXE booting without DHCP?
[02:03] <screen-x> does pxe support bootp?
[02:04] <sparc> i do have a dhcp server, but i'd rather just use it for the pxe booted clients
[02:04] <twb> I don't remember
[02:04] <twb> (@screen-x)
[02:04] <patdk-lap> bootp doesn't support pxe
[02:04] <twb> Unless you're on an untrusted network, I don't see why you're avoiding DHCP
[02:04] <sparc> i guess i can recognize installer dhcp clients by the "d-i" identifier?
[02:04] <twb> And if you're on an untrusted network, you need to turn off ARP, too
[02:04] <twb> sparc: you can, yes.
[02:05] <screen-x> what happens if you turn of arp? You have to manually map ip addresses to mac addresses? is that in etc/ethers?
[02:05] <screen-x> oh dear, too many questions for one line
[02:05] <twb> screen-x: you distribute MAC to IP mapping out-of-band
[02:06] <patdk-lap> heh, na, your manually map the router mac and ip, and that is all it needs to know :)
[02:06] <screen-x> both ideas make sense
[02:06] <twb> Or you switch to IPv6, where ARP is replaced by ICMPv6 and ICMPv6 requires IPSEC to guarantee trust
[02:06] <patdk-lap> :)
[02:15] <Danny78> anyone know of a tiny cpu monitor to add to taskbar?
[02:17] <screen-x> Danny78: Gnome has a system monitor applet, which can show CPU usage
[02:17] <Danny78> I changed my server name to buddha but somewhere it is still being referenced as Samba24...  where could that be?
[02:17] <screen-x> DNS?
[02:17] <Danny78> screen-x:  thanks will look for it
[02:18] <Danny78> screen-x:  how do I flush DNS?  or is that not right thing to do?
[02:19] <screen-x> Danny78: where are you seeing the old name of your server?
[02:20] <Danny78> in GADMIN Samba, security tab--  oh, wait, nevermind
[02:20] <Danny78> that's old data
[02:21] <pmatulis> anyone have experience access Microsoft Access database file over Samba?  initially it's fine but then performance degrades by 90%
[02:23] <patdk-lap> performance gets better when you uninstall office :)
[02:29] <twb> Does access use sqlite3 as its backend yet?
[02:32] <Danny78> do I remember right---  is fstab where you automount win shares?
[02:32] <Danny78> cd /etc
[02:32] <Danny78> sudo vim fstab
[02:32] <Danny78> oops
[02:32] <screen-x> Danny78: automount is a separate thing, that mounts dirs when they are accessed.
[02:33] <screen-x> fstab lists FSs that are mounted at boot time
[02:33] <Danny78> screen-x:  so is fstab a proper way to mount windows shares at boot?
[02:33] <screen-x> Danny78: yep
[02:34] <Danny78> screen-x:  here's my problem though.  I'm not familiar with finding network devices from the command line
[02:35] <twb> Danny78: ip link
[02:35] <twb> Or lspci -nn | grep -i ethernet, for 802.1 devices
[02:38] <screen-x> Danny78: fstab line should be something like //server/share  /mount/point cifs options
[02:38] <screen-x> the options are explained in man mount.cifs
[02:40] <Danny78> twb:  that's totally new to me
[02:40] <screen-x> "ip" is a very useful command
[02:43] <twb> ip(8) is the Right Thing.
[02:44] <Danny78> How do you specify section 8 when opening a man page?
[02:44] <Danny78> just as it's written?
[02:45] <twb> Danny78: man 8 foo
[02:49] <Danny78> brb, rebooting to test fstab
[02:57] <twb> How tightly intertwingled are postfix and dovecot?
[02:58] <twb> I have a policy of running each service in its own LXC container, but I'm not sure there's much point separating postfix and dovecot.
[02:58] <lamont> twb: they're pretty well integrated, but totally separate packages... postfix doesn't care if dovecot is there or not, dovecot mostly just wants an MTA under it
[02:58] <twb> lamont: I mean, will they be significantly less happy if dovecot's running on one host, and postfix is running on another?
[02:58] <twb> (With, I guess, a shared /var/mail)
[02:59] <lamont> the LDA part of dovecot probably wants to run in the postfix container, the IMAP beast could be separated into its own container - both need access to the mailboxes, of coures
[02:59] <lamont> postfix is going to invoke dovecot as the local delivery agent, or dovecot is going to need its own MTA under it.
[03:00] <twb> Hmm, that's a good point -- local delivery won't be supported on my satellites (msmtp-mta), only on the gateway/smarthost (postfix).
[03:00] <lamont> that is, another MTA instance, regardless of which MTA it is
[03:01] <twb> OK, so I should be able to put dovecot-imapd in one container, and postfix + <other bits of dovecot> in the other container?
[03:01] <fluvvell> twb, I think Danny78 was meaning finding other network servers (win shares) from the command line, as opposed to finding network devices (which is what he asked).
[03:01] <lamont> twb: I suspect so.  to be fair, I haven't really played with dovecot much yet.  Gonna have to fix that now that I have more mobility
[03:02] <twb> fluvvell: oh
[03:02] <twb> lamont: okey dokey
[03:02] <lamont> twb: I would be interested in your findings
[03:03] <twb> I'll try to remember to tell you :-)
[03:37] <twb> lamont: incidentally, do you have an opinion on where mailman should run (i.e. which container)?
[03:39] <EvilSushi> so what is syslogf
[03:39] <lamont> mailman will either want to drop things in postdrop, or talk to port 25
[03:39] <EvilSushi> syslog*
[03:39] <EvilSushi> how does it differ from logging on a centos box?
[03:40] <twb> lamont: eh, 25 will be open on the smarthost on the LAN side, so a separate container sounds OK.
[03:40] <twb> EvilSushi: from a user perspective, basic rsyslogd is identical to sysv syslogd
[04:02] <overrider> My RAID1 Ubuntu 10.04 worked fine until last night when i came to check it and it booted into initramfs console or busybox. Any ideas how i can make it boot again? Re-powering is no help, and taking one disk or the other off is no help either. Before it goes into initramfs it shows 'Could find dev/uuidxxxxx' or something along those lines.
[04:04] <patdk-lap> overrider, what kind of raid1? 3ware? software? ....
[04:10] <overrider> patdk-lap: no just software raid using mdadm, setup during the ubuntu install process
[04:11] <patdk-lap> hmm, dunno then :(
[04:11] <overrider> :-)
[04:11] <patdk-lap> my friend just had the same issue, but that is cause it refused to load the 3ware driver, very strange, but doing a modprobe 3w-9xxx fixed him up
[04:11] <overrider> bah...
[04:12] <overrider> its weird - this machine has been running great for a long time. Not sure whether maybe there was a power loss in some inconvinient moment or something that screwed it up...
[04:30] <sven_oostenbrink> Is is possible to have a "mount during boot, but dont wait 10 minutes, and if it fails, no biggie" option in fstab? I have a server that upon reboot needs to mount an cfis filesystem that may or may not be there.. Its no biggie if it cant mount it during boot, I can manually mount it after.. if its not there, the server hangs for 10 minutes, and and after that stops because it cant mount... How can I fix this?
[04:35] <twb> sven_oostenbrink: in Ubuntu?  I don't think so.
[04:35] <twb> Other than a hack in e.g. rc.local
[04:36] <twb> Oh, a network filesystem -- you could try -osoft
[04:37] <sven_oostenbrink> twb: is it possible to interrupt the mount? Im waiting on the @*#($@* server now, its waiting for nothing, its not there...
[04:37] <twb> Unfortunately, no
[04:37] <twb> Back before upstart, you could interrupt arbitrary init jobs, but even Debian has removed that feature now :-///
[04:38] <sven_oostenbrink> twb: why... ? and more, why is mount waiting a friggin 10 minutes for the mount? After 10 seconds it can be damned sure that there is nothing there.. the rest of the timeout is just to mess with people or what?
[04:38] <twb> Because you can't support ^C with parallelized jobs
[04:39] <twb> It will probably boot fastter if you ctrl+alt+del, then *unplug the network cable* for the next boot
[04:39] <twb> IME remote mounts will time out quickly if they know there's no network at all.
[04:40] <sven_oostenbrink> twb: well, after the timeout I tried ifconfig, and I had no network.. if I have no network, and its still hangs on the timeout.. I guess unplugging the network won't make much of a difference, will it?
[04:40] <twb> I guess so
[04:50] <patdk-lap> Phoenixz, heh, I want that option for nfs, I can't find it at all either :(
[04:50] <patdk-lap> but it doesn't hang for 10min for me, it hangs forever
[05:57] <Danny78> Ubuntu server is still unable to mount windows shares on bootup.  here is my fstab:  http://paste.debian.net/104272 and here is my smb.conf: http://paste.debian.net/104273
[06:05] <twb> Danny78: maybe mountall doesn't realize they need to be mounted AFTER the network is up
[06:06] <twb> Danny78: /mnt is not the appropriate place for permanent mounts
[06:06] <twb> Danny78: your probably want /srv
[06:06] <twb> Danny78: also, does root have access to your .smbcredentials file?  Are you using ecryptfs for ~danny?
[06:08] <Danny78> twb:  yes, using encryptfs...  computer memory full, any way to increase sqwap...  really slow right now
[06:08] <twb> Danny78: well, that's why
[06:08] <twb> Danny78: root can't read your smbcredentials until you log in and provide a decrypted mount on ~danny
[06:10] <twb> Danny78: you can allocate more swap using a swapfile, although it would be better to isolate and kill off the broken proces
[06:10] <twb> *process
[06:12] <gobbe> if you want to mount windows-shares on boot you need to add credentials to fstab
[06:12] <gobbe> to work perfectly
[06:13] <twb> Or some file like /etc/cifs.secrets, I guess
[06:14] <gobbe> that's other option
[06:16] <twb> Since fstab is usually world-readable
[06:17] <Danny78> processor is ldle but everything is super slow
[06:36] <Danny78> I had to do a hard reset, I missed what you said, if you could please repeat
[06:40] <Phoenixz> my ubuntu server is hanging during boot.. I see fsck checks that are ok, then nothing.. I had some network mounts, which were hanging (after some 1o mins I would have a cifs filesys mount failure, drop to shell), so I removed them from fstab (I commented them),  but after a reboot, its still hanging.. What could be causing this? I have no shell, no log data, nothing.. :(
[06:51] <kaushal> Hi
[06:52] <kaushal> I am running ipsec on Ubuntu server 8.04
[06:52] <kaushal> I want to enable debug logging
[06:52] <kaushal> Please suggest/guide
[07:00] <delimiter> how can I upgrade to lucid from hardy ? do-release-upgrade only offers maverick even with prompt=tls in /etc/update-manager/release-upgrades
[07:15] <Phoenixz> my ubuntu server is hanging during boot.. I see fsck checks that are ok, then nothing.. I had some network mounts, which were hanging (after some 1o mins I would have a cifs filesys mount failure, drop to shell), so I removed them from fstab (I commented them),  but after a reboot, its still hanging.. What could be causing this? I have no shell, no log data, nothing.. :(
[08:06] <talat> Hi i want to learn ubuntu clod can do one vm use 2 physical computer hardware ?
[08:07] <twb> Ha!  "clod"
[08:07] <talat> twb: pardon clod = cloud
[08:08] <twb> talat: offhand, I suspect you'll run into problems because the cloud also wants to use virtualization
[08:17] <Hypoglybetic> Oh, theres only a 1/3rd as many people in here as #ubuntu.  You're all probably 3x smarter too ...
[08:17] <twb> Hypoglybetic: I daresay you are working to ameliorate that difference.
[08:18] <Hypoglybetic> maybe
[08:19] <Hypoglybetic> I have a couple questions about ub-server and would like to be pointed in the right direction.
[08:20] <twb> !ask > Hypoglybetic
[08:23] <Hypoglybetic> I'm very new to Linux.  I want to setup a basic web server and a media server for my house as well as host virtual machines.  I need a little help installing the correct packages.  I'm sitting at the Server install menu (dns, lamp, mail, openSSH, etc).
[08:24] <talat> twb: i try creat a cluster but i want to my vm can use more than one computer. I need 64 core but my computers 24 core. in cloud can be do ?
[08:24] <twb> "media center" as in somewhere to dump videos, or "media center" as in a box hooked up to your TV, playing videos?
[08:24] <Hypoglybetic> My question is could someone help me get the basics setup because I don't want to corrupt my files (repository?) again.
[08:24] <twb> talat: I don't understand the question.
[08:25] <Hypoglybetic> twb, Dump/host files, use PS3/Laptop/Windows  7 to watch the movies
[08:25] <twb> talat: obviously you cannot make a 24-core CPU into a 64-core CPU.
[08:25] <twb> Hypoglybetic: in that case you can probably just pick "file server" at that prompt
[08:26] <talat> twb: i have 4 24-core CPU computer. I want to use this computer for my 64 core vm. Do you understand ?
[08:26] <twb> Hypoglybetic: the keyword you want to google for is "samba"; it's the open source implementation of Windows-style file sharing.
[08:26] <twb> talat: nope
[08:26] <twb> talat: what is your native language?
[08:26] <talat> twb: Turkish
[08:27] <talat> twb: one minute
[08:27] <twb> !tr
[08:27] <Hypoglybetic> twb, do Iwant to select LAMP server as well? (currently openSSH, Samba File, and Virtual Machine Host are selected)
[08:27] <twb> talat: you can also ask them, they might understand better than me
[08:27] <twb> Hypoglybetic: you do not need "LAMP"
[08:27] <twb> Hypoglybetic: "LAMP" means "web server"
[08:28] <Hypoglybetic> twb But I do want a web server, do you suggest XAMPP? I used XAMPP on my windows machine.
[08:29] <twb> Hypoglybetic: I do not know what XAMPP is.
[08:29] <twb> Looks like a bag of random apache/mysql/php/perl for Windows.
[08:30] <Hypoglybetic> twb, yes, similar to what LAMP is (i'm trying to google things on my own, but I'm still paranoid of messing/misreading something)
[08:30] <twb> Hypoglybetic: don't expose it to the internet, and you'll be OK
[08:31] <Hypoglybetic> twb, wait, what? how do I make it work if it isn't on th internet to download/install things?
[08:33] <twb> Hypoglybetic: it is possible to set up a firewall such that you the internet is allowed to respond to your requests, but not INITIATE connections to your host.
[08:33] <twb> Indeed, this is the normal way a "hard shell, gooey insides" corporate firewall works.
[08:33] <twb> Obviously it means no one will be able to access your website, but if you don't know what you're doing, this is a GOOD THING.
[08:33] <Hypoglybetic> Jordan_U, are you awake?
[08:34] <talat> twb: ok last question. if you dont understand this can i ask turkish user. i Will create a virtual machine which can be use multi server hardware on my cloud. I dont want  onlt run on one server. this can be possible ?
[08:35] <Hypoglybetic> I think I understand talat.  He wants to do something like ESXi does.  You have multiple machines and you "pool" their resources so if one machine fails none of the actual guest OS's fail.
[08:35] <twb> talat: cloud allows you to run the same OS image on multiple hosts.  So you can have two hosts doing half of the work each.
[08:35] <twb> talat: cloud IS NOT like plan9, where you can directly pool resources to make a single "super computer"
[08:36] <twb> Note that I am not a cloud expert -- I don't even use it.
[08:36] <talat> twb: ok thank you for help
[08:36] <Hypoglybetic> talat do you know of VMware ESXi ?
[08:37] <talat> Hypoglybetic: i dont know
[08:37] <Hypoglybetic> talat, look into it -- I think you may find it useful
[08:37] <twb> talat: sorry I could not help -- your English is better than my Turkish :-)
[08:38] <sven_oostenbrink> Every time my ubuntu (server) reboots, it hangs after fsck of all filesystems.. a check in /var/log/messages shows JBD: barrier-based sync failed on dm-6:8 - disabling barriers.. Anyone knows how I can fix this and get my server to boot  normally again?  Google shows this message not to be important, but on every freeze, this is the last message seen.. each freeze has alt-sysreq-* intact, so its not a panic, or complete freeze.. any ideas?
[08:38] <talat> twb: :)
[08:40] <twb> sven_oostenbrink: you'll need a live cd, first of all
[08:40] <twb> sven_oostenbrink: what's happening is that mountall(8) gets an exit status of 4 from fsck(8), so it just stops
[08:41] <twb> sven_oostenbrink: you need to do... SOMETHING... from the live cd, so that fsck doesn't fail.
[08:41] <twb> (and yes, I hate this feature.)
[08:43] <sven_oostenbrink> twb: mount -a seems to mount all okay
[08:44] <twb> mountall isn't mount -a
[08:44] <twb> It's a stupid upstart thing
[08:44] <sven_oostenbrink> twb: a server not starting is a feature?
[08:44] <twb> sven_oostenbrink: in the sense that developers always claim bugs are "features"
[08:44] <twb> IMO it should *at least* shut down, not just hang
[08:45] <sven_oostenbrink> twb:  I tried to upgrade from 9.10 to 10.04, see if that would fix it.. upgrade all okay, minus kernel.. (ahw crap..) linux-server debends on linux-image-server (....); however: package linux-image-server is not configured yet.. Any ideas on how to fix this one?
[08:46] <sven_oostenbrink> twb: IMHO it should just fail to a shell with an error message so I actually would have a chance to fix things.....
[08:46] <twb> can you pastebin the full transcript ?
[08:46] <twb> sven_oostenbrink: yeah, that'd be favourite.
[08:47] <sven_oostenbrink> twb: http:.//pastebin.com/2h0vTitv
[08:47] <sven_oostenbrink> twb: sorry, http://pastebin.com/2h0vTitv
[08:48] <sven_oostenbrink> twb: wait.. thats not right
[08:50] <sven_oostenbrink> twb: how can I route output and error to a file again?
[08:50] <sven_oostenbrink> twb: 2>&1 ?
[08:50] <biston> good morning. for a box running ubuntu-server running webmin+apache+mysqld+vncserver+vmware(all vms turned off) isn't the load avg 0.09, 1.39, 2.82 too high ?
[08:51] <twb> script -c 'foo bar'
[08:51] <twb> pastebinit typescript
[08:51] <sven_oostenbrink> twb: http://pastebin.com/Zz3w3d0X
[08:51] <biston> with top -d 1 i only see init using resources, the rest is 0
[08:52] <sven_oostenbrink> gis
[08:52] <sven_oostenbrink> biston: first number is avg of last minute, last is avg 15 mins, so the current load is practically 0
[08:52] <twb> sven_oostenbrink: try again after "export TERM=vt100 LC_ALL=C"
[08:52] <biston> oh, alright, thank you sven_oostenbrink
[08:53] <twb> biston: you understand that load averages are AVERAGES over time?
[08:53] <twb> biston: i.e. your ten-minute average was 2.82; but now it's just 0.09
[08:53] <biston> i do now
[08:53] <twb> biston: i.e. whatever you did in the last ten minutes, stopped the load
[08:53] <biston> it was a virtual machine, got my load avg up to 9
[08:54] <twb> Shrug
[08:54] <sven_oostenbrink> biston: also, system load is over all processors.. you could more or less say that a dual core with load 2.00 has both processors working at 100%
[08:54] <twb> If it's a bloat-tacular VM, not surprising
[08:54] <twb> sven_oostenbrink: good point
[08:55] <sven_oostenbrink> twb: did the export, now no error, but seems to hang on Setting up console-setup (1.34ubuntu15)
[08:55] <sven_oostenbrink> twb: its there for over a minute now....
[08:55] <sven_oostenbrink> biston: howmany cpus?
[08:55] <twb> cpus and/or cores
[08:55] <twb> sven_oostenbrink: oh christ, that one
[08:55] <sven_oostenbrink> biston: what twb says
[08:55] <twb> sven_oostenbrink: I forget how to stop that
[08:55] <sven_oostenbrink> twb: uh oh.. bad?
[08:55] <twb> sven_oostenbrink: it drove us insane
[08:56] <biston> checking, i forgot which machine this one is
[08:56] <biston> 2 cores, 1 cpu
[08:56] <sven_oostenbrink> twb: its driving me insane....
[08:56] <biston> a core 2 du e8400, 3.00 ghz
[08:56] <sven_oostenbrink> twb: if you remeber an ything... please please pretty please..
[08:56] <twb> sven_oostenbrink: echo ACTIVE_TERMINALS= >>/etc/default/console-setup ## Try that
[08:56] <twb> Sorry, ACTIVE_CONSOLES
[08:56] <sven_oostenbrink> biston: load of 9 on a dualcore would say...  450% system load
[08:56] <twb> it's a shot in the dark
[08:57] <sven_oostenbrink> twb: CTRL-C the current install?
[08:57] <twb> Whatever
[08:57] <twb> IIRC it fucks up the entire system when you hit that console-setup issue
[08:57] <twb> What it's trying to do is mess with the VGA font and such
[08:57] <twb> fucks up as in hangs it, not as in wipes the disks or whatever
[08:58] <sven_oostenbrink> twb: eh, so echo ACTIVE_CONSOLES= >> /etc/default/console-setup ... any spaces anywhere?\
[08:58] <twb> No
[08:58] <twb> sven_oostenbrink: hopefully that tells console-setup not to run
[08:59] <twb> sven_oostenbrink: it might be easier to just blow away the system and do a clean install
[08:59] <sven_oostenbrink> twb: it says /etc/default/console-setup: 50: ACTIVE-CONSOLES=: not found
[09:00] <twb> sven_oostenbrink: no, ACTIVE_CONSOLES
[09:00] <twb> sven_oostenbrink: edit the file and remove the bad line; it's a _ not a -
[09:00] <sven_oostenbrink> twb: yeah, got that
[09:00] <sven_oostenbrink> twb: still hanging
[09:01] <twb> I give up
[09:04] <sven_oostenbrink> twb: :( Anyone else who might know? Google isnt of much help here
[09:06] <twb> Shrug
[09:06] <twb> 19:59 <twb> sven_oostenbrink: it might be easier to just blow away the system and do a clean install
[09:07] <biston> can anyone recommend a vm host app other than vmware to run on ubuntu server 10.4 ?
[09:08] <twb> biston: Ubuntu recommends KVM.
[09:08] <biston> alrite, i'll try kvm and see if i still get same loads. thank you
[09:09] <sven_oostenbrink> XEN uses paravirtualization which brings in a whole load of other crap..  last time I tried it it managed to corrupt all filesystems on my machine..
[09:10] <sven_oostenbrink> twb: Im so close to finishing this, cant give up now!  ... you have no recollection of what you did when you had this?
[09:11] <twb> sven_oostenbrink: got drunk IIRC
[09:12] <biston> thank you both so much for your help, i'll be off now. good luck with ur problem twb
[09:20] <sven_oostenbrink> twb: can I not remove console-setup?
[09:21] <twb> sven_oostenbrink: I'm not stopping you
[09:22] <sven_oostenbrink> twb: That doesnt sound very encouraging...
[09:24] <sven_oostenbrink> twb: http://pastebin.com/baKjxU8V
[09:24] <twb> sven_oostenbrink: ICBF helping further, sorry
[09:24] <sven_oostenbrink> twb: removed console-setup, I can now proceed with kernel, but now I have this
[09:24] <sven_oostenbrink> ICBF?
[09:24] <sven_oostenbrink> who's that?
[09:25] <twb> I Can't Be Fucked
[09:25] <sven_oostenbrink> right :)
[09:25] <sven_oostenbrink> anyone else?
[09:35] <sven_oostenbrink> twb: well, just said fuckit, reboot... lets see howfar it gets
[11:08] <ilrovato> hallo
[11:08] <ilrovato> someone online?
[11:11] <ilrovato> How do I install ubuntu 10.04 server using a PXE server?
[11:11] <ilrovato>  The online guides speak netboot installation via the package "netboot", but which distribution is installed? General or server?
[11:24] <twb> ilrovato: 1) get a PXE server; 2) put the netboot kernel and ramdisk on it; 3) pxeboot.
[11:24] <twb> Ubuntu only has one distribution.
[11:25] <twb> If you have an existing Ubuntu server, you might try installing di-netboot-assistant, which will hold your hand for much of the process.
[11:26] <ilrovato> mine pxe server is a debian squeeze
[11:27] <twb> Well, di-netboot-assistant works there, too
[11:27] <ilrovato> i will wach di-netboot-assistant
[11:28] <xampart> any idea if i can restrict a local user on a specific channel with ircd hybrid?
[11:59] <twb> xampart: I don't know; you might have better luck asking #oftc or #freenode (whichever one runs hybrid; I forget)
[11:59] <twb> Note that, as a rule, ircd won't be able to tell which users are local.  So you'd probably have to "default deny" all accounts, then allow a whitelist of registered users to connect to other channels
[12:11] <xampart> twb: k
[12:12] <\sh> maswan: I fixed my problem with the FlexFabric nics (be2net) , replaced ipconfig with udhcpc inside initramfs and it works like a charm
[12:25] <twb> Ugh
[12:26] <twb> syslog-summary (which logcheck suggested), has an env trampoline
[12:26] <twb> WORSE, it trampolines to an obsolete version of python, that's not installed!
[13:08] <azertyui> hello
[13:08] <azertyui> there
[13:08] <azertyui> i got server with 6 disk one of the disk is failure, physically there is no physicall apperence, what the command on ubuntu show the disk in failure ?
[13:11] <compdoc> one of your hard drives is failing?
[13:11] <compdoc> or has it failed?
[13:12] <patdk-wk> next is it in some kind of raid? lvm? ..., hardware or software raid?
[13:12] <azertyui> failed
[13:12] <azertyui> soft raid
[13:12] <xampart> cat /proc/mdstat
[13:13] <patdk-wk> assuming it's software raid
[13:13] <azertyui> what should display ?
[13:14] <xampart> something like this "md127 : active raid1 sda2[0] sdb2[1]"
[13:14] <compdoc> smartctl  can show you if the other drives are healthy
[13:15] <azertyui> it gives raid1 raid0 raid6
[13:15] <azertyui> correct xampart
[13:15] <azertyui> and also got [faulty]
[13:16] <xampart> so it should read there also the faulty drive (sdX)
[13:17] <azertyui> my question is how to determine the disk in fault ?
[13:18] <xampart> from the " blocks [2/2] [UU]" line. if it's [U_] then, in my case it would be sdb
[13:19] <xampart> unless i've understood it totally wrong
[13:19] <azertyui> ok
[13:19] <azertyui> thanks
[13:20] <xampart> but smartctl gives good output
[13:21] <azertyui> ok thanks xampart
[13:31] <Amivit> I have a server that is freezing periodically and suspect either a defective harddrive in the array, or the actual raid controller is unstable. How do I find out the fastest? Btw, the raid array is listed as "optimal" in bios. I have no log entries indicating any fault and it even freezes booting from live OS's from different kernals.
[13:31] <Amivit> So looks very much to be a hardware issue.
[13:31] <compdoc> best to look at the smart data - for reallocated sectors on any drive
[13:32] <twb> compdoc: or just queue a self-test
[13:32] <compdoc> who makes the controller?
[13:33] <compdoc> a customer has a failing drive that causes the server to lock up from time to time as the sectors go bad
[14:10] <gobbe> Amivit: is there anything on logs from the time of freezing?
[14:10] <gobbe> Amivit: i mean logs under /var/log
[14:11] <Amivit> Unfortunately nothing. Just a freeze. Cant queue self-test because that causes a freeze aswell.
[14:12] <Amivit> Basically anything requiring harddrive semi harddrive activity provokes the freeze. Maybe I should attempt moving the harddrive and pci raid controller to a different server?
[14:13] <RoyK> Amivit: what sort of controller?
[14:14] <Amivit> It is an adaptec 2410SA
[14:16] <RoyK> Amivit: just a hunch, but I guess it's a controller driver issue
[14:16] <Amivit> RoyK, maybe. But even when the server has been running without problems for 5+ years? :S
[14:16] <RoyK> Amivit: which version of ubuntu is this?
[14:16] <RoyK> oh
[14:16] <RoyK> so bad hardware, then
[14:16] <Amivit> the 2003 one :D
[14:17] <gobbe> i would bet that problem is with hardware
[14:17] <gobbe> it acts just like broken hardware, mysterious freezes without any logs
[14:17] <RoyK> Amivit: if it has ben working for that long and just stopped, I'd bet my dime on hardware problems
[14:17] <RoyK> bad hardware can do that quite easily
[14:17] <RoyK> :P
[14:18] <gobbe> yep
[14:19] <Amivit> I agree, thinking it must be that. I'm just not quite sure how I should tackle the situation snice there are a total of 2x 2410sa's and 8x harddrives.
[14:20] <Amivit> Perhaps start by disconnecting one 2410sa and see if I still get the freezes.
[14:20] <gobbe> yes
[14:20] <gobbe> that's good way to do it
[14:27] <Amivit> Once I identify the proper raidcontroller - I should probably move on to scanning each drive individually. Any recommendations as to which tool?
[14:29] <gobbe> it might be the controller
[14:29] <gobbe> or just disk
[14:29] <gobbe> controller should be able to tell if the disk is broken
[14:30] <RoyK> I doubt it's  a disk
[14:30] <Amivit> Exactly. Most likely is the controller in that case because when booting the raid is listed as "optimal" status.
[14:30] <RoyK> try a new controller first - just make sure you get one that can read the RAID format
[14:33] <Amivit> Yeah good idea. I'll see if I can use the first 5-10 minutes to identify exactly WHICH raidcontroller contains the problem (or IS the problem) and replace it with a fresh 2410sa, probably also the most time effective way to locate the problem since testing drives i could imagine will take a long time.
[14:33] <Amivit> Appreciate the help despite the fact the server is a 2003 :)
[14:36] <gobbe> :)
[14:36] <gobbe> well send you bill, don't worry ;)
[16:52] <fullstop> Hi!  When does /etc/fstab get processed at boot time?
[16:53] <shauno> usually mountall in /etc/init/
[16:54] <shauno> altho it's largely just processed by mount itself (mount -a will mount most of what's in fstab)
[16:56] <fullstop> shauno: I'm in a pickle.  I would like to mount drbd volumes from /etc/fstab..
[16:56] <fullstop> but I believe that it is attempting to mount before drbd is started.
[16:56] <RoyK> shouldn't be much of a problem
[16:56] <RoyK> oh
[16:56] <fullstop> I can't tell, though.  I never get an interactive shell and I have to boot from a usb stick and comment out the line in /etc/fstab
[16:57] <RoyK> paste that line, please
[16:58] <fullstop> /dev/drbd_vmfiles	/vmfiles	xfs	defaults,noatime,nodiratime,logbufs=8,allocsize=512m		0	0
[16:59] <fullstop> if I put it in there, mount / umount work fine as long as drbd is running.  drbd is started at boot time.
[17:00] <gobbe> drbd needs to be fully up, and gfs or other cluster-filesystem running
[17:00] <RoyK> fullstop: changing mount order from 0 to 2 might help
[17:00] <gobbe> before you can mount it
[17:00] <fullstop> RoyK: I'll give it a shot.
[17:01] <gobbe> i added my drbd-disk to rc.local to be sure that everything is up and running :)
[17:01] <fullstop> I'm not using any clustered filesystem.. just syncing to a 2nd system for redundancy
[17:01] <gobbe> so i run it manually from rc.local
[17:01] <fullstop> I thought of that..
[17:01] <fullstop> but there is something nice about having it in /etc/fstab  :)
[17:01] <gobbe> okei
[17:01] <gobbe> i use drbd in active/active
[17:02] <shauno> there's a couple of lines in /etc/init/mountall-net.conf that send USR1 to mountall to get it to try fstab again.  I'd be tempted to try adding that to the end of the script that's starting drbd at boo
[17:02] <fullstop> gobbe: what filesystem?
[17:02] <RoyK> gobbe: iirc rc.local is run after the normal init scripts - just a wild guess here about mount order 0 is a bad thing for something relying on services started
[17:02] <shauno> (er, boot)
[17:03] <gobbe> RoyK: yep
[17:03] <gobbe> fullstop: gfs2
[17:04] <shauno> I'm not entirely clear where mountall runs tho, or rather, when it stops running
[17:05] <fullstop> I'm not exactly sure why I never get a shell.. it should just fail to mount the device and move on.
[17:05] <RoyK> fullstop: do you still get that shell with boot order 2?
[17:06] <fullstop> RoyK: Which entry is that in fstab?  The last one?
[17:06] <gobbe> yes
[17:06] <fullstop> RoyK: I always thought that it had to do with what order they were checked for errors.
[17:07] <fullstop> rebooting now.. I'll know in a few minutes.
[17:07] <RoyK> erm - you're right
[17:07] <RoyK> but perhaps fsck fails if it's run too early?
[17:07] <fullstop> _netdev?
[17:08] <shauno> yeah.  the last two fields are dump and fsck, with the theory that you can fsck two disks in parallel, but you wouldn't want to fsck two slices on the same disk in parallel
[17:09] <RoyK> shauno: fsck fixes that itself
[17:09] <RoyK> that field should be set to 2 for most filesystems
[17:09] <RoyK> see fstab(5)
[17:09] <fullstop> Well, that didn't change a thing.  Unfortunately, it hangs before sshd starts so I have to reboot
[17:09] <fullstop> from a usb stick.  =/
[17:09] <shauno> I'm sure it was relevant last time I paid attention to it.  get off my lawn, etc :p
[17:10] <fullstop> I still pipe gzip decompression to tar from time to time just for the memories.
[17:11] <fullstop> New tar handles bzip2 and gzip without any extra flags.  tar -xf will do either.
[17:12] <gobbe> fullstop: you could do it like i fixed it ;)
[17:12] <RoyK> fullstop: iirc last I setup drbd, it worked well with the volumes in the fstab
[17:12] <RoyK> fullstop: which ubuntu version is this?
[17:13] <gobbe> my old server with other distro worked also fine with setup i used in 10.04, however i didn't get any way to mount my drbd at boot
[17:13] <gobbe> without doing it "manually" from rc.local
[17:16] <fullstop> RoyK: 10.04, but drbd was compiled from source.
[17:16] <RoyK> fullstop: what's wrong with the one in lucid?
[17:17] <fullstop> RoyK: I needed to build a custom kernel for this server
[17:17] <gobbe> i used the one in lucid
[17:17] <fullstop> I needed patches for scst...
[17:17] <fullstop> hopefully vlad and crew can get them into the mainline kernel.
[17:18] <RoyK> k
[17:18] <RoyK> isn't scst in kernel already?
[17:18] <fullstop> RoyK: no..
[17:19] <fullstop> RoyK: maybe you are thinking of iet?
[17:19] <RoyK> I thought scst was included - sorry - my fault
[17:20] <RoyK> I haven't used either, but scst looks far better
[17:20] <fullstop> I have been very happy with scst.  It's frustrating watching kernel politics sometimes, but I hope that they make it in.
[17:20] <RoyK> tell me about kernel politics :P
[17:20] <RoyK> it's a bunch of hard core nerds after all
[17:20] <fullstop> is there documentation on /etc/fstab and ubuntu options in there?
[17:21] <fullstop> I'm interested in giving _netdev a shot, but I'd like to see more about it first.
[17:21] <RoyK> fullstop: man fstab/mount
[17:21] <RoyK> most of it is in mount(8)
[17:22] <fullstop> i was looking in fstab, didn't check mount
[17:22] <fullstop> does Ubuntu respect _netdev?
[17:22] <RoyK> fullstop: see mount(8)
[17:25] <fullstop> I swear, the more "enterprise" the server, the slower the boot.
[17:25] <fullstop> ta-da!  _netdev worked.
[17:26] <gobbe> ok :)
[17:27] <RoyK> fullstop: heh - tell me about it :)
[17:27] <RoyK> btw, we have some rather nice supermicro boxes at work that boots in a matter of minutes, still with hw raid and 2x12-core CPUs and a truckload of RAM
[17:28] <RoyK> some scientists are using those 24x7 with their models
[17:29] <compdoc> Im building a server with a supermicro board - theyre nice
[17:29] <RoyK> which mobo?
[17:29] <compdoc> x7sbl-ln2
[17:30] <fullstop> I believe that this machine actually has an asus MB.. 2x X5670 and 48GB of ram.
[17:30] <fullstop> and.. 14T usable storage.
[17:31] <fullstop> times 2, because there is an identical machine in the 2nd rack.
[17:32] <RoyK> fullstop: we're using these http://www.supermicro.com/products/motherboard/QPI/5500/X8DTH-6.cfm with 100TB net storage (158TB raw)
[17:32] <SpamapS> people still "build" servers?
[17:32] <RoyK> pretty damnd good I/O on those
[17:32] <compdoc> some do
[17:32] <EvilSushi> who?
[17:32] <fullstop> we do
[17:32] <compdoc> I do
[17:32] <RoyK> SpamapS: I was wondering the same - especially when supermicro ships readily built systems quite cheap
[17:33] <fullstop> We've been burned with under-performing servers from large 3 letter companies.
[17:33] <gobbe> :)
[17:33] <gobbe> fullstop: I B and M ?-)
[17:33] <RoyK> AST?
[17:33] <RoyK> :)
[17:33]  * RoyK wonders how long ago that was
[17:34] <RoyK> HAL
[17:34] <fullstop> gobbe: those may be the 3 letters... :-D
[17:34] <gobbe> fullstop: which kind of servers from ibm?
[17:35] <fullstop> I forget the exact number, but they were opteron servers purchased in ... 2007.
[17:35] <fullstop> IBM X series, I think.
[17:35] <gobbe> aah, those old opterons
[17:35] <RoyK> fullstop: there's quite a few of those :P
[17:35] <fullstop> The opteron part was fine.  It was the I/O which was awful.
[17:35] <gobbe> i have one customer with new x3950, with 4x12 cores and 512 GB ram :)
[17:36] <RoyK> I used to run this x3850 - worked well back then
[17:36] <gobbe> yep
[17:36] <gobbe> we have x3850 in our own vmware-cluster
[17:36] <fullstop> However, after flashing the firmware on the discs, we had a 40% improvement.
[17:36]  * Daviey quite likes supermicro servers
[17:36] <fullstop> and now we will never purchase another seagate again.
[17:37] <fullstop> They were rebadged IBM, but you could tell from the serial number that it was really a seagate drive.
[17:38] <RoyK> fullstop: most drive producers have issues from time to time - I don't think seagate is worse than any other
[17:38] <fullstop> RoyK: Look into their 750GB drives and the problems they had with them.
[17:39] <fullstop> RoyK: I understand that they have bad runs from time to time, but this was more than that.
[17:39] <RoyK> I don't disagree with the problems, but blaming all seagate drivers, is like saying hitachi deskstars are deathstars because of those bad drives 8 years ago
[17:40] <RoyK> fullstop: btw, 750GB drives? sounds like SATA to me
[17:40] <RoyK> 7k2?
[17:40] <fullstop> SATA, yes
[17:40] <gobbe> :)
[17:40] <gobbe> RoyK: i just thought to say hitatchi disks :-D
[17:40] <fullstop> RoyK: I have a hitachi at home, even though I had one of their deathstars.
[17:41] <RoyK> hit-hat-chi
[17:41] <fullstop> Anyway, the seagate drive problem happened shortly after they moved manufacturing to China...
[17:42]  * RoyK found a few 60MB drives at work labelled things like DOS 5.0 and win 3.0
[17:42] <SpamapS> RoyK: the day MS-DOS 6.0 arrived at my house I distinctly remember squealing with glee.
[17:43] <RoyK> lol ;)
[17:43] <RoyK> I remember DOS 4.0 wasn't very good
[17:43] <RoyK> supported large drives, though
[17:43] <RoyK> that is > 32MB partitions
[17:44] <plm> hi all
[17:44] <EvilSushi> what was the awesome windows version with the tiled windows
[17:44] <EvilSushi> windows 2 i think
[17:44] <fullstop> We had a bernoulli drive, once.
[17:45] <plm> people, where I find a updated page about servers supported by ubuntu. this site for example, about hp servers supported https://wiki.ubuntu.com/HardwareSupportMachinesServersHp  is very old
[17:46] <pmatulis> plm: http://webapps.ubuntu.com/certification/
[17:47] <plm> pmatulis: thanks
[18:05] <nealmcb_> Anyone know why openssl doesn't have cms compiled in?  https://bugs.launchpad.net/ubuntu/+source/openssl/+bug/380424
[18:11] <RoyK> nealmcb_: no idea, but apt-get source that package, change the configure line, build a new package,install it, and you're done
[18:18] <nealmcb_> RoyK: Yeah - I did that, and it works.  And now I see that it was more or less an experimental backport first added in openssl 0.9.8h  The version in lucid is 0.9.8k.  Somehow I thought it was there earlier....
[18:32] <robbiew> smoser: Daviey: can you have a look at https://blueprints.launchpad.net/ubuntu/+spec/cloud-server-n-desktop-images and let me know if the work items and milestones assigned make sense?
[18:38] <nealmcb_> Ahh - I see the question is now when openssl 1.0 is brought into Ubuntu, now that it is available in Debian.  Can we get it in natty?  Bug 675566
[18:42] <Daviey> robbiew, on it
[18:44] <robbiew> Daviey: thnx!
[19:01] <jdstrand> win 19
[19:03] <mdeslaur> jdstrand: you win
[19:13] <EvilSushi> can upgrading a server from 8.04 to 10.04 break anything?
[19:16] <compdoc> a backup would be a good idea
[19:17] <binBASH> Backup is overrated :D
[19:22] <gobbe> EvilSushi: it could, every upgrade has risks
[19:22] <gobbe> there's no fail proof upgrades, but LTS to LTS is almost fail proof :)
[19:30] <deadsmith> can anyone tell me about video support on an XServe2,1?
[19:30] <gobbe> what do you want to know about it?
[19:30] <gobbe> there's page on ubuntu-wiki about apple+ubuntu
[19:33] <deadsmith> gobbe:  Trying to figure out 1)  if there's any support past fbdev for Xorg, and 2) if not, what kernel and EFI args and Xorg config options are needed to get minimal support.
[19:33] <deadsmith> gobbe:  I see some stuff on MBP, but nothing about the XServe2,1.  I'll try harder...
[19:34] <resno> having a bit of trouble settig up a user to access a samba share.
[19:34] <resno> they can access their dirs, but cant access another users
[19:35] <deadsmith> gobbe:  to be clear, I've been through this: https://help.ubuntu.com/community/Xserve2-1 but there's no Xorg discussion, just that it may need to be limited to the fbdev driver.
[19:36] <gobbe> ok :/
[19:37] <resno> whats the best location to store files in linux?
[19:37] <resno> that arent spefiic for a user
[19:37] <RoyK> anywhere, really
[19:39] <resno> im having a problem setting up a samaba share. heres the setting im using: http://dpaste.com/307365/
[19:40] <resno> ive tried several configurations and still cant get the windows uses access
[19:40] <resno> err, cant get write access
[19:40] <pmatulis> resno: check filesystem permissions
[19:41] <resno> pmatulis: i figured thats where the issue was
[19:41] <pmatulis> resno: so why did you ask?
[19:41] <resno> ok, then i  create a usergroup to allow access there. but im not sure i did it right
[19:42] <resno> err, i created the usergroup and then chmod on that dir to allow access to the group
[19:42] <resno> pmatulis: ^
[19:45] <Danielc1234> hi all, just setup a new VM using ubuntu as OS. Now trying to install Zimbra as mailserver. It states that it is imperative to have all DNS working properly before installing. If our DNS is being controlled by godaddy.com then are there special settings I need to do on the server side?
[19:46] <compdoc> I think it just means you should be able to reach sites from your lan
[19:47] <gobbe> Danielc1234: well, you need to have mx-entry that points to your mailserver
[19:47] <compdoc> or have the world be able to reach you by name
[19:47] <pmatulis> Danielc1234: there are many variants of "mailserver".  what do you really want to achieve with it?
[19:48] <compdoc> mx records are the most important for mail
[19:48]  * resno walks away and finds own answers
[19:48] <Danielc1234> gobbe: so I would need to define a nameserver on this VM like mail.oursite.com ?  We are trying to use this VM as a primary source for all our emails, etc.
[19:48] <gobbe> Danielc1234: you need to have mx-entries in your DNS_server
[19:49] <gobbe> and infact this has nothing to do with ubuntu in this case
[19:49] <gobbe> you are using DNS hosted by godaddy, and trying to use software provided by vmware
[19:49] <compdoc> resno, I dont use masks. I just set the permissions on the folder I share, and have a uid and gid for a user
[19:50] <compdoc> does it not work?
[19:50] <Danielc1234> gobbe: well as you can tell I'm very new to this. However, isnt there a place on the server that I need to define the settings?
[19:51] <resno> compdoc: it works everywhere except this dir. which is owned by another user. i have a user named media, and i want everyone to be able to write and change that folder through samba
[19:51] <gobbe> Danielc1234: well, i don't know what settings you mean
[19:51] <compdoc> the VM guest only needs dns for outgoing, and an ip address that mail ports are forwarded to from the outside.
[19:51] <pmatulis> Danielc1234: 'primary source?'  that doesn't mean anything to me.  but the others seem to understand so let them help
[19:52] <Danielc1234> gobbe: wouldnt I need to define the nameserver within the server? Sorry guys, like I said I am very new to all of this....
[19:52] <gobbe> Danielc1234: if you don't know how mailservers work you should try to start from finding out that, before you try to create one
[19:52] <gobbe> Danielc1234: and like i said, this really isn't ubuntu-issue
[19:52] <gobbe> so this channel is wrong
[19:52] <Danielc1234> gobbe: okay, thanks all.....
[19:53] <gobbe> Danielc1234: but you don't need nameserver in your server if you have already one
[19:53] <compdoc> resno, I use 757 on the folder, and that usually works
[19:54] <compdoc> allows everyone access
[19:58] <niteria> Hi
[19:59] <gobbe> hi
[19:59] <niteria> I'd like to have /boot on /dev/md0
[19:59] <gobbe> that's ok if you use raid1
[19:59] <niteria> but I'm not sure what to tell grub
[19:59] <niteria> title ....
[19:59] <niteria> root (???) ?
[19:59] <niteria> root (md0) ?
[20:00] <niteria> i did update-initramfs -u
[20:01] <gobbe> grub is installed on both disks
[20:01] <gobbe> and you boot from hd
[20:02] <niteria> will grub know if I give it root (hd1,0) of type fd that it should use it as raid partition?
[20:03] <resno> compdoc: just change chomod 757 on it?
[20:06] <gobbe> niteria: grub boots directly from disks, not md
[20:06] <gobbe> niteria: so you need to have grub in both disks
[20:06] <compdoc> resno, if you ls -al the directory the folder is in, whats the attributes now?
[20:06] <pmatulis> niteria: firstly, what version of grub are you using?
[20:06] <compdoc> on the folder you want to share?
[20:08] <resno> drwxr-xr-x compdoc
[20:08] <niteria> pmatulis: default one from debian lenny
[20:09] <niteria> my kernel's root is already booting from it properly
[20:09] <compdoc> resno, I think thats 757, so Others (outside the group) cant write, but only read
[20:10] <niteria> and / is lvm partition on raid1
[20:10] <compdoc> err, 755, sorry
[20:10] <pmatulis> niteria: you're not running ubuntu?
[20:10] <compdoc> you have it set to 755, I mean
[20:11] <niteria> pmatulis: no, not ubuntu, but question is more about grub and raid
[20:11] <pmatulis> niteria: maybe go to #grub?
[20:11] <pmatulis> niteria: but basically you install grub on both disks
[20:12] <niteria> I have separate /boot partition
[20:12] <pmatulis> niteria: and ensure that the UUIDs for the arrays are properly set up.  where depends on the version of grub, hence my 1st question to you
[20:12] <niteria> which I want to have on /dev/md0
[20:13] <pmatulis> niteria: it's not on an array now?  i'm missing something
[20:13] <niteria> ok, from the beginning
[20:14] <niteria> I have 2 disks /dev/sda and /dev/sdc
[20:14] <niteria> they're the same
[20:14] <pmatulis> niteria: gotcha
[20:14] <niteria> on /dev/sda i have /dev/sda1 = /boot
[20:15] <niteria> /dev/sda2 + /dev/sdb2 = /dev/md1 (raid1)
[20:15] <niteria> /dev/sdb1 is /dev/md0
[20:15] <niteria> i want /dev/sdb1 + /dev/sda1 = /dev/md0
[20:16] <niteria> no it boots from /dev/sda1
[20:16] <niteria> now*
[20:16] <gobbe> niteria: like i told, you cannot boot from md0
[20:16] <gobbe> niteria: grub needs to be on disks
[20:17] <fullstop> You can have grub on both sda1 and sdb1 and make it _feel_ like you boot from md0
[20:17] <gobbe> niteria: so you install grub to both disks, if you use /boot, you create it to BOTH disks
[20:17] <RoyK> gobbe: no problem with that if /boot is on a separate partition
[20:17] <gobbe> RoyK: well, you want it to both disks still
[20:17] <gobbe> RoyK: or when you break the disk with /boot you cannot boot your machine
[20:17] <gobbe> :D
[20:17] <RoyK> gobbe: ubuntu installs fine on mirrored root disks
[20:17] <gobbe> yep
[20:18] <gobbe> and it creates it to both disks
[20:18] <gobbe> because there's no idea to install boot-things to only one disks
[20:18] <niteria> http://www.linuxsa.org.au/mailing-list/2003-07/1270.html here is similiar setup
[20:18] <gobbe> if that disks breaks down, you cannot boot
[20:18] <gobbe> so basicly you have boot in both disks
[20:18] <RoyK> it creates an md device and install grubs on both drives
[20:19] <gobbe> yep
[20:19] <RoyK> grub2 understands md
[20:19] <gobbe> i use that kind of solution
[20:19] <niteria> # grub --version
[20:19] <niteria> grub (GNU GRUB 0.97)
[20:19] <gobbe> niteria: which version of ubuntu you are running?
[20:19] <RoyK> iirc grub1 can't boot form an md
[20:19] <gobbe> no
[20:19] <niteria> I'm running debian lenny ;p
[20:19] <gobbe> you need to have it on both disks, like i have told
[20:20] <gobbe> niteria: eh..and you are asking here?
[20:20] <gobbe> niteria: this is meant for ubuntu-support
[20:20] <gobbe> grub needs to have boot on both disks, grub2 can boot from md0
[20:20] <niteria> they couldn't help me on #debian
[20:20] <gobbe> niteria: that's sad but we cannot either
[20:20] <ScottK> Also the way Debian Lenny and recent Ubuntu versions boot is significantly different so even if it was on topic, this wouldn't be a great place to get answers.
[20:21] <gobbe> yes
[20:21] <niteria> :(, sorry
[20:21] <gobbe> niteria: but if you look my advice you will be able to do it
[20:21] <RoyK> niteria: just reinstall the OS - it'll save you some time
[20:21] <gobbe> niteria: grub needs to have it on both disks, grub2 can do it
[20:22] <deadsmith> What's the right way to boot into text login?  Some docs say Ubuntu doesn't support runlevel 3?
[20:22] <niteria> i did grub-install /dev/sda, grub-install /dev/sdc if that's what you mean
[20:25] <niteria> I suppose I could forget about raid1'ed /boot and just create identical partitions /dev/sda1, /dev/sdc1
[20:25] <niteria> but then I have a problem in /etc/fstab
[20:25] <RoyK> niteria: grub1 doesn't support boot from raid
[20:25] <RoyK> niteria: just reinstall the box with something that supports root raid
[20:26] <RoyK> niteria: I know it's a bitch, but it's simpler than spending hours trying to fix something that's hard or impossible to do
[20:26] <niteria> That's a bummer
[20:27] <RoyK> it's the easy way out
[20:27] <gobbe> yes
[20:27] <niteria> I don't understand why lenny uses grub1
[20:29] <RoyK> grub2 is quite new
[20:29] <RoyK> lenny isn't
[20:30] <gobbe> well
[20:30] <gobbe> lenny is newer then 10.10 ;)
[20:30] <gobbe> it's release 27.11.2010
[20:30] <gobbe> there arent many distros using grub2 still
[20:30] <gobbe> most of distros uses grub
[20:31] <RoyK> oh, lenny, sorry
[20:33] <RoyK> niteria: still, if your root filesystem isn't mirrored in the first place, even ubuntu might be hard to use for mirroring an existing one - I'd recommend reinstalling the box
[20:34] <gobbe> yep
[20:35] <niteria> my / is mirrored
[20:35] <niteria> and lvm'ed
[20:35] <RoyK> but not /boot
[20:36] <niteria> no, /boot is another partition
[20:36] <niteria> it's weird how lvm on raid1 / works properly with grub1
[20:36] <RoyK> that's just linux
[20:37] <RoyK> grub doesn't know about that
[20:37] <niteria> ok
[20:37] <niteria> i'm installing grub2 before I reinstall everything
[20:37] <RoyK> niteria: changing a partition into a mirror isn't trivial in linux
[20:38] <RoyK> it might be easier to reinstall
[20:38] <RoyK> with grub2 on top of that, I'd recommend a reinstall - it shouldn't take too long
[20:38] <niteria> i've changed / into mirror
[20:38] <niteria> so I kind of know the process
[20:38] <niteria> but I'm stuck on grub
[20:39] <niteria> and I've just found out I had no way of doing this with grub1
[20:39] <RoyK> just my hint - changing a fs into a mirror is like create a mirror with a dead drive, copy data, add the drive etc
[20:39] <RoyK> but I'm not sure how that works with grub
[20:39] <niteria> that's what I did with /
[20:40] <RoyK> yeah
[20:41] <niteria> I guess grub2 detects partition type and if it is fd it acts accordingly
[20:41] <niteria> fd = linux raid autodetect
[20:42] <RoyK> honestly, I don't know much about how grub2 works
[20:44] <pmatulis> grub2 will notice the raid and know what to do (use the UUIDs for the arrays essentialy)
[20:50] <niteria> okay, I hope ubuntu-server installator lets me create raid1 and lvm on top of it
[20:51] <niteria> becouse you've convinced me to use ubuntu
[20:51] <niteria> because*
[21:00] <smoser> SpamapS,
[21:00] <smoser> st_dev=os.stat("/").st_dev ; print(os.major(st_dev),os.minor(st_dev))
[21:00] <smoser> (8, 8)
[21:01] <SpamapS> SWEET
[21:15] <sparc> Hmm, when running automated installs, the error messages sometimes flash buy quickly, never to be seen again
[21:16] <sparc> can we get that sent to a remote syslog or something?
[21:17] <sparc> gotta scour the guide again
[21:24] <savid> My server is 9.04,  and I'd like to have git>=1.7 installed (1.6 is the latest available for 9.04).  Is my only choice to either do a full dist upgrade, or install from src?  Or is there some ubuntu-ish way of upgrading a single package?
[21:25] <MagicFab> savid, 9.04 has been EOL a few months already - I'd suggest full upgrade
[21:26] <savid> MagicFab: I'm generally paranoid about upgrading a production server :-|
[21:26] <MagicFab> savid, there is, however a PPA which should also work for you: https://launchpad.net/~git-core/+archive/ppa
[21:26] <savid> gives me the willies
[21:26] <sparc> best to have two, so you can fail over to the other, while you upgrade
[21:26] <MagicFab> savid, well, you are missing security updates since October which is not better
[21:28] <sparc> Should the Ubuntu installer be stopping on "Partitioning Method:" if I specified partitioning info in my kickstart?
[21:33] <sparc> oh neat
[21:34] <sparc> \o_ hi eric
[22:10] <zul> !clone
[22:13] <AlanBell> hi all, I have an Ubuntu server 10.04 running KVM and several ubuntu server 10.04 guests. When running a simple CPU benchmark on the server and guest the guest seems to outperform the host which puzzles me
[22:14] <AlanBell> it takes 28 seconds on the guest, but 40 seconds on the host to run this ->   time echo "scale=5000; 4*a(1)" | bc -l -q
[22:16] <AlanBell> can someone help me to understand how this is possible?
[22:19] <GatorAlli> Hello, is there any way for me to run a browser(firefox) on ubuntu server, and make a script that would have it load a webpage and return the rendered html?
[22:20] <AlanBell> GatorAlli: would the wget command do the job for you?
[22:20] <GatorAlli> :/
[22:20] <GatorAlli> no
[22:21] <GatorAlli> the html has to be rendered
[22:21] <AlanBell> rendered to what?
[22:21] <GatorAlli> *and the javascript
[22:22] <GatorAlli> give the browser an html file ("<html><body><script>dicument.write("Hello World!")</script></body></html>")
[22:22] <GatorAlli> and the browser should give me back : <html><body>Hello World!</body></html>
[22:24] <fallous> does it need to be html as the result?   webkit2png can parse a page and spit out a png rendering of it
[22:24] <GatorAlli> it has to be html... :/
[22:25] <GatorAlli> just like you open up firefox
[22:25] <GatorAlli> put in the url, load the page
[22:25] <GatorAlli> start firebug, and get the interpreted source
[22:25] <GatorAlli> it might be too much
[22:25] <GatorAlli> ill ask around on firefox ;)
[22:36]  * RoyK just got the xkcd paper version :)
[22:37] <air^> \o/
[22:37] <air^> RoyK: does it include the "dead pixels in the sky" strip?
[22:38] <air^> (as it probably doesn't work very well on paper :P )
[22:38] <air^> http://xkcd.com/395/
[22:48] <mrrrothhh> can I get the same feature or similar features of windows home server with ubuntu server
[22:49] <air^> depends on what the features you want are.
[22:50] <mrrrothhh> Place for backups (storage), store pictures, store home videos, a place to handle my linux iso torrnet downloads
[22:50] <mrrrothhh> and do raid 1
[22:50] <jforman> yes to all of those
[22:50] <mrrrothhh> is their a premade ubunt-home-server distro
[22:50] <mrrrothhh> or do i have to bulid it my self
[22:51] <mrrrothhh> I see this http://www.newegg.com/Product/Product.aspx?Item=N82E16859106058&cm_re=whs-_-59-106-058-_-Product <-- but 549 for a windows home server and no video output
[22:51] <mrrrothhh> Lenovo IdeaCentre D400 Intel Atom 230 1.60 GHz 1GB DDR2 Windows Home Server
[22:51] <jforman> mrrrothhh: machines with linux pre-installed can be found, but you'll understand the system a lot more by grabbing a ubuntu server ISO and installing that, then configuring each of your desired services
[22:52] <mrrrothhh> I have a dell p4 wigh 3 gigs of ram
[22:52] <mrrrothhh> p4 3.0 ghz 3 gigs of ram
[22:52] <jforman> that is quite powerful enough
[22:53] <mrrrothhh> what else should a home server have
[22:54] <air^> pr0n
[22:54] <air^> :D
[22:54] <air^> mrrrothhh: seriously, you gotta make that decision yourself.
[22:55] <mrrrothhh> oh, so samba, backup,  media, torrnet, and a nice webfrontend, and maybe email
[22:56] <air^> "a nice webfrontend".... for torrent?
[22:56] <jforman> mrrrothhh: you can normally find guides on how to setup by searching on google "ubuntu X" where X is what you would like (web server, mail, torrent, file server, etc)
[22:57] <mrrrothhh> ubuntu Home server
[22:58] <mrrrothhh> oh I see https://wiki.ubuntu.com/UbuntuHomeMediaServer <--
[23:04] <RoyK> air^: don't think so...
[23:06] <RoyK> air^: btw, the page numbers are in the 3-number system :P
[23:06] <air^> :D
[23:06] <air^> RoyK: I might have to buy that one as well..
[23:06] <RoyK> it's well worth it :)
[23:07] <air^> the only thing against it is, I never read paper...
[23:07] <air^> so it will just end up in a bookshelf :D
[23:07] <RoyK> it's just good to have that on the bus.......
[23:07] <RoyK> big, flat
[23:07]  * RoyK likes books
[23:09] <RoyK> I wish I could share this book by Tor Åge Bringsværd with non-norwegians - IIRC it's translated to German and Dutch, but not to English
[23:40] <Deeps> anyone have any suggestions for a tool that can be used to compare the contents of two directories and verify that they're identical? other than rsync, that is
[23:41] <air^> diff is not good enough?
[23:42] <Deeps> well, it's 4TB of binary data
[23:42] <air^> :)
[23:45] <air^> Deeps: calculate hashes from the files, then compare that?
[23:45] <Deeps> well yes, worst case scenario i could script something up, but i was wondering if there were any tools that did it already
[23:46] <Deeps> basically, got a sata card and the drivers for it aren't particularly great, had to backport the latest kernel module from 2.6.37 and it's still not great, resulting in errors like so: http://pastebin.com/ba7VqSxt every now and then, had about 8 of them with 460GB data transferred so far
[23:47] <air^> :/