[00:06] <patdk-lap> I used to use pam_mount, that was fun :)
[00:14] <clusty> hey
[00:15] <clusty> i am building my very first raid and was curious how bad is it with the TLER issues with consumer hdd's ?
[00:15] <clusty> is it that common for hdd's to take 7+ seconds to try to fix an error themselves ?
[00:42] <qman__> an encrypted drive set to automount rather defeats the point of an encrypted drive
[00:56] <terinjokes> Hey guys, running 10.04 here, and I have APT::Periodic::Update-Package-Lists set to "1", but I'm noticing that the package index is never updated until I explicedly refresh with `apt-get update` (as such, I'm not getting my update notifications sent to me)
[01:44] <giovani> terinjokes: just confirm where you've set it -- i.e. pastebin the file, and you know that's only supposed to update the lists once a day, right?
[01:49] <terinjokes> giovani: correct, but it's not being updated... one second
[01:50] <terinjokes> giovani: http://pastebin.com/jHraTu31
[01:53] <giovani> terinjokes: are you trying to do automatic upgrades, or just notifications of packages that can be upgraded?
[01:54] <giovani> the latter is done automagically via some cron jobs in the desktop install, unsure about the server install -- the former is handled by a specific package (which installs cronjobs) for the server and desktop afaik
[01:55] <jmarsden> terinjokes: sudo apt-get install unattended-upgrades    # for unattended upgrades... and read its docs too.
[01:55] <giovani> yeah, that would be the package for the former
[01:56] <terinjokes> giovani: i *just* want notification of the new packages
[01:57] <giovani> terinjokes: and how do you plan to get notification? do you have a custom setup for this? or are you using a pre-built mechanism?
[01:57] <terinjokes> jmarsden: i read those docs, but my understanding of unattended-upgrades is to install the upgrades, which I don't want
[01:57] <terinjokes> giovani: i already have notifications setup... works great (ie, as soon as i run `apt-get update` i get it...)
[01:57] <jmarsden> terinjokes: I think you can configure it not to actually do the upgrades, but just let landscape-client tell you what is available to be updated at login...
[01:58] <terinjokes> i've disabled most of the MOTD scripts
[01:59] <terinjokes> (so if it was there, then this might explain why they don't happen)
[01:59] <giovani> terinjokes: ok, well without the cronjobs provided by some package (or written by hand) I don't think that variable you set does anything
[02:00] <terinjokes> giovani: is there some package that setups these crons? would it be best to have unattended-updates installed, and just tell it not to update anything?
[02:00] <giovani> terinjokes: I honestly don't know, but I'd bet that they're in that package, and, jmarsden seems to know more about it -- it likely can be told not to upgrade
[02:01] <jmarsden> terinjokes: I think yes, just install unattended-upgrades and then *don't* set APT::Periodic::Unattended-Upgrade "1";
[02:02] <jmarsden> I've not used that that way, but I'm pretty sure it will do what you want.
[02:02] <wieshka> hi there - i have serios problem - my last try to install ubuntu-server with raid 1 + lvm made unsupported partitioning on my HDD's, so now reinstall normally cant detect my HDDs, so i cant redone partitioning for my HDD ?
[02:03] <wieshka> what are my options to clear partition tables/format HDD using busybox ?
[02:03] <wieshka> or maybe ubuntu live ?
[02:03] <wieshka> all i want to do, is fully clear partitioning data from my HDD's
[02:04] <wieshka> so i can do reinstall with /boot partition out of raid
[02:04] <wieshka> and looks like i will prefer installing Lilo
[02:04] <terinjokes> jmarsden: thanks, i'll give that a shot
[02:05] <jmarsden> terinjokes: You're welcome
[02:06] <jmarsden> wieshka: If you just want to destroy the partition tables you can do something like dd if=/dev/zero of=/dev/sdb bs=1024 count=1024
[02:06] <jmarsden> that will wipe the first 1MB of your /dev/sdb drive...  is that what you are looking for?
[02:06] <wieshka> jmarsden: it shouldnt take a long long process ?
[02:06] <jmarsden> wieshka: No, just 1MB of writes...
[02:06] <wieshka> jmarsden: ok i will try - basicly i want my hards clear
[02:07] <wieshka> to set up new installation
[02:07] <wieshka> i dont what i messed up
[02:07] <jmarsden> zeroing the whole drive would take more time, but you don't need to do that just to be sure the installer sees the drive as being "new".
[02:07] <wieshka> but my install neither boots up, neither shows correctly in partition detection
[02:09] <wieshka> done for both disks, now rebooting to see result
[02:09] <wieshka> jmarsden: how long it should take with 500 GB disk to fully write it with zeros ? :)
[02:10] <wieshka> it like a formatiing - i am correct ?
[02:10] <mase_wk> wieshka: depends how your doing it ?
[02:10] <mase_wk> -?
[02:10] <jmarsden> I don't remember, I do that kind of thing overnight.  Well, writing zeroes to every sector does not technically format the drive.
[02:10] <jmarsden> But it makes pretty sure noone will read whatever info was on it before :)
[02:11] <mase_wk> heh
[02:11] <mase_wk> if you do something like dd bs=2048 if=/dev/zero of=/dev/foo  it will be reasonably quick
[02:11] <mase_wk> you could set a larger bs too
[02:12] <mase_wk> but you will still need to format it after
[02:12] <wieshka> thx jmarsden - now i have clean two disks :)
[02:12] <wieshka> so what is the best suggestions for building RAID 1
[02:12] <jmarsden> wieshka: You're welcome :)
[02:13] <wieshka> make /boot out of the raid ?
[02:13] <fidelix> Guys, can you please take a look here? http://ubuntuforums.org/showthread.php?t=1548415
[02:13] <wieshka> becouse my disks restursn HDIO_IDENTITY fail (hdparm cant detect by name or something like that)
[02:13] <wieshka> so if my boot is inside of raid
[02:13] <wieshka> boot after install fails
[02:14] <wieshka> so how good is my idea about partitions: make 100m /boot, give it a boot flag, then make 30 Gb swap, and create one big partition on whole disk as raid, and after that as LVM
[02:15] <wieshka> so then i can make volume group and do with my partitions on running system what i want
[02:16] <wieshka> or the best way is do like this ? https://help.ubuntu.com/10.04/serverguide/C/advanced-installation.html
[02:16] <wieshka> ?
[02:18] <jmarsden> There is no "should"; the Server Guide doesn't do LVM with RAID1, just RAID1, for simplicity, I think.  If you want LVM, try it your way and it sounds like it should work.
[02:18] <wieshka> basicly i want to setup ubuntu server and then KVM with 5 vhs on it
[02:19] <wieshka> i wanted to make aprtition for each vh for better performance
[02:19] <wieshka> but also i can make a one big partition and use vh as images
[02:19] <jmarsden> Sounds like a reasonable use case for LVM to me.
[02:20] <wieshka> ok - i will give a try with /boot partition out of raid array, and whole disk as raid 1 + LVM
[02:20] <wieshka> is it better to make swap out of raid or i can make it under LVM ?
[02:20] <wieshka> sorry - first time serious install of ubuntu-server
[02:20] <wieshka> on multiple disks
[02:21] <wieshka> i have old install/use of ubuntu-server on single disk, but now i am building new one, so trying to make reasonable good install
[02:21] <wieshka> so - i am begginer - open for suggestions :)
[02:22] <jmarsden> wieshka: You won't be swapping anyway really, on a modern machine with gigabytes of RAM, so it probably doesn't matter :)  I'm not doing "real" virtualization server builds so I'm not really the best person to get advice from on that.  My guess: try the swap outside the LVM.
[02:23] <wieshka> outside of LVM and also outside of raid array ?
[02:23] <wieshka> hmmm, basicly i have 16 gb ram - so is there need for a swap at real ?
[02:24] <wieshka> 16 x 1 Gb
[02:24] <jmarsden> Perhaps not :)  Yocu ould leave a small swap partition outside both RAID and LVM "just in case" it is ever needed.
[02:24] <wieshka> hmm, not a bad idea :)
[02:27] <wieshka> so  summary: i will split each drive (2) in 3 partitions, 100 mb for /boot, 10 Gb for swap (maybe someday i will need it), and whole disk, as raid partition, after that i configure raid 1 as sda3 + sdb3, and set them as LVM volume group, and also i will make ~20 Gb logical volume on LVM as mount point /
[02:27] <wieshka> i have to gibe boot flag option to /boot partition ?
[02:28] <wieshka> and do i have to set this /boot partition on one disk or on each one ?
[02:28] <wieshka> i supouse one only
[02:28] <wieshka> so all this stuff/summary should work - so i am going to give a try :)
[02:29] <jmarsden> wieshka: You can try setting /boot on both drives, then if one drive fails you *might* be able to boot from the other one... although in the past getting that to work has needed some minor messing around, at least for me.
[02:30] <wieshka> what messing around
[02:30] <wieshka> you think, this configuration will not work on fly - install & go ?
[02:31] <wieshka> hmm, /boot must be as primary & at the begining ?
[02:31] <jmarsden> It should, and maybe these days it will, when I tried it was at least a couple of years ago, Ubuntu RAID support has changed since then.
[02:31] <wieshka> & with boot flag
[02:32] <jmarsden> I would suggest that, if you need the "failover boot" to work, then test it before you have a real disk failure, so you *know* it will work.
[02:32] <wieshka> so i made on each drive 100 mb primary partition as ext4, with mount point /boot and bootable flag
[02:32] <wieshka> jmarsden: i will test that :)
[02:32] <wieshka> at first i need working install :)
[02:33] <jmarsden> wieshka: True :)
[02:33] <wieshka> does swap need botable flag ?
[02:33] <wieshka> i supouse no
[02:34] <wieshka> and does raid 1 partition needs boot flag ?
[02:37] <jmarsden> I really doubt it.
[02:41] <wieshka> heh, install do not allow me to make to /boot
[02:41] <wieshka> so maybe set them in array as md1 and then set on them /boot ?
[03:20] <wieshka> jmarsden: still here ?
[03:20] <jmarsden> Yes... did it work?
[03:20] <wieshka> and again my boot stops ....
[03:21] <wieshka> with fsck error
[03:21] <wieshka> ata_id[650]: HDIO_IDENTITY failed for /dev/sda
[03:21] <wieshka> also for sdb
[03:21] <wieshka> and next line is, that with my LVM everuthing is ok (check statuss)
[03:22] <wieshka> and no more action
[03:22] <wieshka> at the same point again ....
[03:22] <wieshka> fuck.
[03:23] <jmarsden> I have not seen that error msg before... does googling for it show anything interesting?
[03:24] <wieshka> lets try - i removed one of my disks
[03:24] <wieshka> what it will say now
[03:24] <wieshka> i even tried to setup with lilo bootloader now
[03:24] <wieshka> insted of grub
[03:24] <wieshka> as i heard - it haves better compability with raid
[03:25] <wieshka> maybe - i should install system on single disk, and after that make RAID 1 over my existing installation
[03:25] <wieshka> at start as degraded RAID
[03:25] <wieshka> i dont even now anymore- tomorow i have ti put server in data center ....
[03:26] <jmarsden> Hmmm.  a test install to a single disk would be good, so you at least know the hardware is OK and will let you install to it and boot from it.
[03:28] <jmarsden> Doing a "first time serious install of Ubuntu server" with a constraint of "tomorrow I have to put the server in the data center" sounds like you are trying to learn too much too fast...
[03:29] <wieshka> jmarsden: i am this problem already 3 days
[03:29] <wieshka> on this
[03:29] <wieshka> heh :D :D :D
[03:30] <texastwister> Came in late to the conversation guys, what's the issue?
[03:30] <wieshka> i swicthed disk slots and now it works - whata hell
[03:30] <wieshka> asks login - give me moment i want to really test that RAID 1 worked or it is a degraded mode
[03:30] <jmarsden> wieshka: Ah, so maybe one of the "disk slots" is bad in your hardware, or the drive was not fully plugged in, or something?
[03:31] <wieshka> but if it shows up in BIOS and also durring install - connection & hardware must be good
[03:31] <wieshka> i dint changed hot swap ports
[03:31] <wieshka> i them only switched
[03:31] <jmarsden> texastwister: Software RAID1 install to wieshka's new server, it has not been booting... until just now, at least!
[03:32] <wieshka> yeah, but i have to check is everything ok :)
[03:32] <jmarsden> wieshka: Indeed.  I suggest you test it a lot in your remaining time... at least it now boots :)
[03:32] <wieshka> i full y agree
[03:33] <wieshka> hdparm still outputs that he cant recognize this HDD model
[03:34] <wieshka> it is simple enterprise edition server HDD from seagete
[03:34] <wieshka> weird
[03:36] <wieshka> hmmmm, mdadm returns that disk are in active sync statuss
[03:36] <wieshka> what else test i should perform ?
[03:39] <wieshka> dmesh also says that raid1 is active with 2 of 2 mirrors
[03:40] <wieshka> also /proc/mdstat says that RAID is active
[03:45] <wieshka> wtf - now bootup hangs
[03:45] <wieshka> its not normal ....
[03:46] <jmarsden> No, that's not normal... sounds like you have some sort of weird intermittent boot issue.
[03:46] <wieshka> the disks should sync durring bootup or after that ?
[03:47] <jmarsden> You did a clean shutdown/reboot, right, you didn't just hit the reset button?
[03:47] <wieshka> i run command reboot from shell
[03:47] <wieshka> now i did reset ....
[03:47] <wieshka> lets se, what i have there
[03:47] <wieshka> now
[03:47] <wieshka> BIOS detects both drives ....
[03:47] <wieshka> LILO runned
[03:48] <wieshka> hmmm only one disk activity is flashing
[03:49] <wieshka> i am confused - now it again booted well
[03:49] <jmarsden> Does cat /proc.mdstat still look fine?
[03:49] <wieshka> eem .... no
[03:49] <wieshka> there is that HDIO identity again
[03:49] <wieshka> but later showed up then last time
[03:50] <wieshka> hmmm, wait ..... my disks activity LED is flashing for both disks
[03:50] <wieshka> maybe i should give them a try
[03:50] <wieshka> some sync or something ?
[03:50] <jmarsden> I think you could have some sort of weird hardware issue, I would research that HDIO_identity message some more.
[03:51] <jmarsden> If you want to give the disks a bit ot a test workout, you could use bonnie++ or some other disk performance test tool.
[03:52] <wieshka> hmmmm, my sata disks are connected via SAS on motherboard
[03:52] <wieshka> maybe i should switch to sata ?
[03:53] <wieshka> or i am miising some disconfiguration in BIOS ?
[03:53] <jmarsden> I have no idea... sorry :)
[03:53] <wieshka> ok, looks like i have to cancel tomorrows data center :D
[03:53] <wieshka> hmmm, HDDs activity still flashed
[03:54] <wieshka> lets try what happens, when i remove one HDD
[03:54] <wieshka> as degraded mode activated, it should work
[04:59] <wieshka> jmarsden: heh, single install works like a charm ....
[04:59] <wieshka> so now trying to set raid over existing installation :)
[05:30] <robertpayne> I have one server with all my domains pointed to it and my webserver setup on it.. Currently it's also my mail server but I'm looking to forward all the mail to a second server. Is the best way to do that to just update the dns record and point the mail commands to the second server's ip?
[05:32] <jmarsden> You would point all MX records the second server, yes.  Possible also update all A records for mail.DOMAIN.com to point there too, and pop.DOMAIN.com if you have folks using those naming conventions for their incoming and outgoing mail.
[05:33]  * twb pedantically points out that "example.net" exists for examples like that.
[05:34] <robertpayne> jmarsden: thanks. I'm the only one using it.. I have like 9 emails I'm using over imap right now and it's  bogging down the main server too much
[05:34] <jmarsden> Not much of a server if one user can bog it down :)  But OK.
[05:35] <robertpayne> jmarsden: Well I mean my email client is maintaining about 9 imap connections simultaneously it's not a beefy server
[05:35] <robertpayne> 512mb VPS
[05:35] <jmarsden> OK.  I run a mailserver with 150 domains and maybe 1200 or more email users, it also runs 100 or so small web sites :)
[05:35] <robertpayne> daaayum :P
[05:35] <robertpayne> How much ram though :P
[05:36] <jmarsden> 4GB.  Basic cheap rackmount server.
[05:37] <robertpayne> yeah webserver is running nginx reverse proxy to apache ( dynamic content only ) it does well the process count is what keeps going a bit red when I keep my email client open
[05:37] <twb> postfix and dovecot should place negligible load on the server
[05:37] <robertpayne> twb: that's what I'm running
[05:38] <robertpayne> It also could be that Apple Mail ( desktop client ) opens about 5 connections for each account to the server to speed up stuff
[05:38] <jmarsden> robertpayne: Then most likely, removing your email from that server will not significantly affect how bogged down it is... are you 100% sure it is the email slowing things down?
[05:39] <robertpayne> jmarsden: I Shouldn't say it's "slowing it down" or necessarily any sort of speed decrease.. I just simply am getting process count warnings from my host.. if I close my email client the process count drops by about 40-50
[05:40] <robertpayne> I'm also setting up a dev box as a backup to my production box ( because not doing so is retarded.. ) and thus was going to offload the mail to the dev box but back it up to the production box nightly
[05:40] <jmarsden> OK... odd that your host cares about process count... CPU load, or RAM usage, sure... but process count?
[05:41] <ruben23> guys any known apps that can do directory service for windwos client PC and windos client PC..?
[05:41] <robertpayne> jmarsden: heh not really sure if they care so much but their status panel does show alerts when it gets too high.. I know kinda dumb.. maybe something to do with VPS stuff who knows
[05:42] <twb> robertpayne: Samba 4 can act as an AD server
[05:42] <twb> Sorry, bad completion
[05:42] <robertpayne> maybe I'll just leave it on the main box.. like I said not really an issue just figured I'd forward it to the dev box
[05:42] <robertpayne> and wanted to make sure modifying the DNS was the best way
[05:44] <ruben23> twb: i can do it with samba4
[05:44] <robertpayne> anyways i got to go for now thx for answers
[05:46] <jmarsden> robertpayne: You're welcome
[05:47] <mase_wk> ruben23: samba4 is not really production ready yet though
[05:47] <mase_wk> at least on my last try a few months back
[05:47] <robertpayne> I think ill just leave it on main box.. and use dev box as backup MX :)
[06:32] <pheelineerie> Hey guys, I could use some help making US get along with the internet
[06:32] <pheelineerie> it won't play nice with my wifi and i have been at it all weekend
[06:34] <Roxyhart0> sorry about this question...somebody know what mean the blue color in the zone file with bind and why some of them are in blue color and others not?
[06:39] <SpamapS> pheelineerie: can you be more specific?
[06:40] <SpamapS> Roxyhart0: maybe you can paste your zone file at http://paste.ubuntu.com and we can look at it ourselves?
[06:40] <pheelineerie> SpamapS: well for example iwconfig says this: IEEE 802.11bg ESSID:off/any. Mode:Managed   Access Point: Not-Associated
[06:41] <SpamapS> pheelineerie: that seems to mean that you're not associated to any SSID's
[06:41] <pheelineerie> trying things with apt-get results in a bunch of Failed to fetch errors
[06:42] <pheelineerie> yes, and I've been trying to set it but i can't get it to take my network name
[06:42] <SpamapS> pheelineerie: are you sure your wifi chipset is fully supported?
[06:42] <pheelineerie> i would think so, because it was working with ubuntu 10.04 before i installed server
[06:43] <mase_wk> perhaps the desktop kernel has patches applied to it which the server kernel does not
[06:43] <SpamapS> pheelineerie: so 'iwconfig eth0 essid "yournet"' doesn't work?
[06:44] <SpamapS> mase_wk: unless you explicitly select a different kernel, they both should get the generic.
[06:44] <mase_wk> k
[06:44] <mase_wk> i didn't know if you had installed the -server kernel
[06:44] <SpamapS> Tho I think desktop makes it really easy to install non-opensource drivers, IIRC
[06:44] <pheelineerie> SpamapS: i've tried that command but with wlan0 instead of eth0... should i try with eth0?
[06:44] <SpamapS> pheelineerie: no
[06:44] <SpamapS> pheelineerie: And the result was.. ?
[06:44] <pheelineerie> it sees the networks when i do the scan thing
[06:45] <pheelineerie> SpamapS: the result was nothing, i restarted interfaces and still the same problem
[06:45] <pheelineerie> SpamapS: i'm putting the name of my wireless network in quotes, but there is a space in it, does that matter?
[06:46] <pheelineerie> SpamapS: and when i say the result was nothing, i mean it just accepted it and went to the next command line with no feedback
[06:46] <SpamapS> pheelineerie: when you restarted interfaces, you probably wiped out the essid.
[06:46] <pheelineerie> SpamapS: o.
[06:46] <pheelineerie> i thought i was supposed to. lol
[06:46]  * pheelineerie feels dumb
[06:47] <SpamapS> pheelineerie: run it, then run 'iwconfig wlan0' again, and see if it shows something different
[06:47] <SpamapS> pheelineerie: the idea is to set your essid, then the card should associate to the correct AP
[06:47] <SpamapS> pheelineerie: once you're associated, then the dhcp client that is already running on wlan0 should get an IP.
[06:47] <pheelineerie> well now it says ESSID:"Kansas 2" but it still gives me "Failed to fetch" etc
[06:48] <SpamapS> I'm honestly quite ignorant at the proper convention for using iwconfig on server..
[06:48] <SpamapS> but for a manual test, this should work
[06:48] <SpamapS> pheelineerie: what still says "failed to fetch" ?
[06:49] <pheelineerie> well i've been testing to see if it was online by trying "sudo apt-get update"
[06:49] <pheelineerie> SpamapS: and it gives me a bunch of "Temporary failure resolving 'us.archive.ubuntu.com' type errors
[06:50] <SpamapS> pheelineerie: right because it hasn't received a network configuration from your DHCP server yet...
[06:50] <SpamapS> pheelineerie: is wlan0 listed in /etc/network/interfaces?
[06:50] <ohzie> I hosed my samba conf file. I'm not sure how, but I'm content with going back to the default conf file. How do I do this?
[06:52] <SpamapS> ohzie: you can move the damaged conf file out of the way, and reinstall samba
[06:52] <SpamapS> ohzie: as in, 'apt-get remove samba && apt-get install samba'
[06:52] <SpamapS> ohzie: if you don't want to do that, you can also extract the file from the .deb, but thats sort of complex
[06:53] <ohzie> SpamapS: No that solution is perfect and makes sense.
[06:53] <Roxyhart0> somebody know how i can create a domain aleas with bind?
[06:53] <pheelineerie> SpamapS: right now interfaces has the following lines: auto lo / iface lo inet loopback / # auto wlan0 / iface wlan0 inet dhcp / wireless-mode managed / wireless-essid 'Kansas 2' / wireless-key 20830835u0fjf etc
[06:53] <ohzie> When I tried reinstalling, the conf file was still there
[06:53] <ohzie> I don't know why it never occurred to me to just delete the damned thing.
[06:53] <ohzie> God, I feel stupid right now.
[06:54] <SpamapS> ohzie: Confusion is the only way we grow. :)
[06:55] <SpamapS> pheelineerie: hrm.. well that seems very logical.. not sure what to tell you
[06:55] <ohzie> SpamapS: Yeah I've been using linux since Red Hat 6. I should have known better. =P
[06:55] <ohzie> It's just late and I am exhausted. Thank you so much for putting up with me.
[06:56] <SpamapS> ohzie: its our pleasure, thanks for using Ubuntu. :)
[06:58] <ohzie> I haven't actually used anything Redhat based since FC4, and I recently threw Centos 5 on a server at work. For anyone wondering? Centos is so different from Debian/Ubuntu that I am helpless most of the time. It's really horrible, and I can't stand it, and the default user you create during the install is not added to the sudoers file.
[06:58] <ohzie> so i intend to replace that with pure Debian or Karmic as soon as I can
[06:59] <chrislabeard> Hey guys I'm trying to use webdav and for some reason its not working
[07:00] <chrislabeard> I have set everything up but when I try to use my mac to connect to it it says it doesn't exist. If I got to the domain in a browser it asks for password and user
[07:05] <ohzie> SpamapS: So I removed samba, deleted my smb.conf, and then reinstalled samba, and it didn't install a conf file.
[07:05] <ohzie> I'm not sure how to react to that.
[07:07] <lifeless> ohzie: did you purge it ?
[07:07] <ohzie> lifeless: Don't know that one.
[07:07] <ohzie> !purge
[07:08] <ohzie> He doesn't know anything about purge.
[07:08] <lifeless> ohzie: dpkg has two separate states for software which you have removed.
[07:08] <lifeless> ohzie: the first state, 'removed' has the binaries/docs/libraries etc gone, but keeps the config files.
[07:08] <lifeless> ohzie: the second state, 'purged' removes the config files too.
[07:08] <lifeless> ohzie: installing software which was only removed, does not touch the config files, because they are already present :)
[07:09] <lifeless> ohzie: so, if you purge samba, then install it again, it should do what you want.
[07:09] <ajmitch> 'apt-get purge samba'
[07:10] <ajmitch> just make sure you've got backups of any important samba files
[07:14] <chrislabeard> Can the webdav you set up on ubuntu work with macs, pcs?
[07:16] <ajmitch> wwwwwwwwgouge
[07:16] <ajmitch> grr
[07:17] <ohzie> ajmitch lifeless: Thank you very much. :D
[07:19] <pheelineerie> Question: do i HAVE to configure Server on the command line? Does it come with any kind of gui at all? i know they can be downloaded but i can't get it online to do so, and i'm going insane :'(
[07:22] <ohzie> pheelineerie: you do have to get it online to install the commandline tools.
[07:23] <ohzie> pheelineerie: "aptitude install ubuntu-desktop" has hilarious results on ubuntu server. The GDM theme has purple on it?! I don't know where it came from, but I like it a lot more than the default gdm theme in normal ubuntu.
[07:23] <pheelineerie> ohzie: but the problem is the fact that i can't get it to connect to my wifi, so i can't install any packages
[07:25] <ohzie> pheelineerie: I've never done wifi from the command line. I wonder if ebox has a network-manager applet?
[07:25] <pheelineerie> ohzie: it is weird and annoying and it's making me feel very dumb. i installed ubuntu-server and it just came up to the command line. i don't know of any other way to do anything?
[07:26] <ohzie> pheelineerie: Don't feel dumb. Feel inexperienced, and realize that learning how to do this will grant you that experience. It will make you feel less dumb. =P
[07:26] <pheelineerie> ohzie: i don't know what ebox is, but i just typed network-manager and it said command not found. i don't know if that answers the question
[07:27] <ohzie> network-manager is a service that controls networking in my desktop systems.
[07:27] <ohzie> I don't know if server uses it, to be honest.
[07:27] <ohzie> Let me see.
[07:27] <ohzie> !!!!!
[07:27] <ohzie> Found what i was looking for.
[07:28] <ohzie> Everything i use wifi on, I use "knetworkmanager" which is a graphical thing.
[07:28] <ohzie> but there is a terminal one
[07:28] <ohzie> can you try the command "nmcli" ?
[07:28] <ohzie> I've not used it so I don't know if it's installed or what the results are
[07:28] <ohzie> but I know that it is a terminal-based client for network manager.
[07:28] <ajmitch> you may as well just use wpa_supplicant directly & setup the interface in /etc/network/interfaces
[07:29] <pheelineerie> ajmitch: that's what i've been trying to do all day, and i do believe i have completely and utterly failed
[07:29] <pheelineerie> sigh
[07:30] <pheelineerie> ohzie: nmcli: command not found
[07:30]  * ajmitch wouldn't expect any of the network-manager stuff to be present
[07:31] <pheelineerie> ohzie: is your nick in reference to otzi, the frozen guy? cause that would be awesome
[07:31] <ohzie> pheelineerie: No.
[07:32] <pheelineerie> ohzie: oh.
[07:32] <ohzie> pheelineerie: Although, upon googling 'Otzi' I do approve of names referencing him, and I will probably use his name from now on in Supreme Commander.
[07:32] <pheelineerie> ohzie: :D
[07:39] <pheelineerie> oh my god. i think i am an idiot
[07:40] <pheelineerie> if i do "sudo nano filename" on a file that doesn't exist.... it creates a new file, doesn't it?
[07:40] <pheelineerie> i think instead of appending something to interfaces.conf or whatever file it was supposed to be, i just created a new, useless file and put the stuff there
[07:41]  * pheelineerie smacks forehead
[07:47] <ohzie> ajmitch: I did sudo aptitude purge samba4, and it ran, and I reinstalled, and I still don't have a conf file.
[07:48] <ajmitch> ohzie: it may be in samba-common
[07:48] <ajmitch> (or samba4-common)
[07:49] <ohzie> Is there a way to ask samba what conf file it grabbed when it started up?
[07:52] <WalterN> https://help.ubuntu.com/10.04/serverguide/C/advanced-installation.html
[07:53] <WalterN> when I get done with the first 9 steps, it complaints that there is no file system defined for root
[07:53] <WalterN> and wont let me continue
[07:54] <ohzie> ajmitch: Yes, the conf file is included in samba-common and wasn't  removed because it is a dependancy for freenx and smbfs. Thank you so much for helping me look in the right place! =D
[07:58] <WalterN> :/
[08:00] <WalterN> oh wait
[08:02] <ohzie> WalterN: Do you have a partition that has a mounting point of "/"
[08:02] <ohzie> ?
[08:03] <nllptr> Is there a standard directory structure for a multi-site dedicated server
[08:03] <nllptr> ?
[08:03] <WalterN> silly me
[08:04] <WalterN> dont go to continue at the bottom, select RAID at one of the top options after arranging the partitions for RAID
[08:04] <WalterN> 0.o
[08:04] <ohzie> nllptr: Most companies will put the WWW folder in the user's home directory, or if each site is running in a VM it will just be under /var/www like normal, but inside the VM.
[08:05] <ohzie> WalterN: lol
[08:05] <nllptr> ohzie: if the sites are larger and have multiple developers should the sites be put into a group's home directory instead?
[08:06] <ohzie> Okay, so I made the changes to the default conf file, and now everything is hosed again. Ugh.
[08:06] <WalterN> hmm
[08:06] <ohzie> nllptr: I would have them in VMs because I am lazy annd not that good with apache.
[08:06] <ohzie> and*
[08:07] <ohzie> I am better with VMs than I am with apache, so I would do that option
[08:07] <ohzie> but I know some apache ninjas who could probably work it out either way and it would be amazing.
[08:07] <ohzie> so I can't really answer what is the best option...but I can tell you what I'd do and why! :D
[08:09] <WalterN> here is the server box I made last week... http://i36.tinypic.com/ivx355.jpg?
[08:09] <WalterN> installing Ubuntu for that as a server :)
[08:10] <nllptr> ohzie: are VMs any less efficient?
[08:10] <nllptr> ohzie: I can't seem to find any good documentation on setting up hosting in a VM on google :S
[08:12] <ohzie> nllptr: Currently I only use virtualbox, that won't work for your purposes. You're going to want to look at Xen.
[08:12] <ohzie> Xen is, I've heard, awesome and amazing.
[08:12] <ohzie> I haven't touched it because I haven't needed to yet
[08:12] <ohzie> but I hope to in the near future.
[08:14] <WalterN> I love my server box
[08:18] <mase_wk> nllptr: if you have the hardware support VT-x i'd recommend looking at KVM to.
[08:19] <mase_wk> Not that Xen is bad in anyway, but we've just migrated from Xen to KVM as the management is fairly trivial
[08:19] <_chris_> moin all :)
[08:20] <mase_wk> moin
[08:20] <WalterN> well, there goes the install
[08:20] <twb> mase_wk: are you using libvirt?
[08:20] <mase_wk> yarp
[08:20] <nllptr> mase_wk: not sure the hardware will support it. The site we're running is a non-profit teenage-run teenage portal site for teenagers running on donated computers
[08:20] <WalterN> two RAID1 arrays for SWAP and /, one RAID5 for extra storage pool, 3x 2tb drives :)
[08:21] <mase_wk> nllptr: yeh if the hardware doesn't support it Xen is useful
[08:22] <WalterN> does it build the arrays before it installs?
[08:23] <WalterN> as in, do I need to wait 8 hours while it builds the RAID5 of 4tb?
[08:25] <Roxyhart0> hi somebody know how to create a domain aleas with dname? where i need to put the record as is not working
[08:26] <joschi> Roxyhart0: well, obviously you need to put the DNAME resource record in the zone for the domain you want to set it for.
[08:27] <Roxyhart0> i did but i think im doind something wrong as is not working..what should be the sintax?
[08:28] <Roxyhart0> i got WHCL. IN DAME WHCL.COM
[08:28] <joschi> Roxyhart0: foo DNAME target.example.com.
[08:28] <Roxyhart0> i got WHCL. IN DAME WHCL.COM.
[08:28] <lifeless> Roxyhart0: DAME != DNAME
[08:28] <Roxyhart0> yes sorry dname
[08:35] <joschi> Roxyhart08: the period after WHCL is probably wrong, unless it's the zone file for the zone 'WHCL.'
[08:36] <joschi> Roxyhart08: you should also describe 'not working' in a little more detail...
[08:36] <Roxyhart08> yes the ast one is the file zone
[08:36] <Roxyhart08> when i do nslookup fot WHCl tell me palm.whcl.com (ns) doesn find whcl
[08:40] <WalterN> ok, guess I dont need to wait for it to build
[08:40] <WalterN> good
[08:40] <WalterN> installing
[08:42] <nllptr> We're going to start completely fresh and set up our four servers from scratch. Any tips on what we can do so it'll be easy to track down problems later on? Any conventions we should follow?
[08:42] <nllptr> We're using one server for a shared hosting service for ourselves. Right now everyone just uses sudo to modify virtualhost settings
[08:42] <nllptr> Any better way?
[08:44] <twb> nllptr: etckeeper is an easy one
[08:53] <ohzie> How do I dump the output of a command to a text file?
[08:54] <ohzie> Nevermind, I found it.
[08:54] <ohzie> I'm going to put my errors and my conf file in a pastebin
[08:54] <ohzie> because I am so blown away, I don't know what to do. :[
[08:56] <robertpayne> Does ubuntu automatically logrotate the syslog file or do you have to setup that manually?
[08:56] <Jeeves_> robertpayne: That depends if there is a config for the specific file
[08:57] <robertpayne> Jeeves_: I don't see anything in logrotate.d so I'm going to assume no
[08:57] <Jeeves_> robertpayne: Which file?
[08:58] <robertpayne> Jeeves_: er do you mean a config file for log rotate? There's no mention in /etc/logrotate.conf or /etc/logrotate.d/*
[08:58] <Jeeves_> robertpayne: No, which file should be rotated?
[08:59] <robertpayne> Jeeves_: /var/log/syslog /var/log/mail.log
[09:00] <ohzie> http://paste.ubuntu.com/475330 - I've been trying to figure this out for a while and I'm at a loss.
[09:03] <Jeeves_> robertpayne: There should be a config for those files
[09:03] <Jeeves_> robertpayne: /etc/logrotate.d/rsyslog
[09:03] <robertpayne> Jeeves_: ohh it might be hidden crap haha
[09:04] <robertpayne> Jeeves_: nope doesn;t exist hmm
[09:05] <Jeeves_> robertpayne: Do you have rsyslog installed?
[09:05] <Jeeves_> Or are you on a older Ubuntuversion?
[09:05] <robertpayne> Jeeves_: sec
[09:05] <robertpayne> Jeeves_: running 10.4 server
[09:06] <WalterN> .04
[09:06] <robertpayne> Jeeves_: and no I don't have rsyslog installed
[09:07] <robertpayne> Jeeves_: looks like I have a /etc/cron.daily/sysklogd that runs savelog on all log files in /var/log .. cycling them every 7 days
[09:08] <robertpayne> actually take that back it only rotates syslog
[09:13] <robertpayne> ahhh alright! Yea it's syslogd that is rotating all the ones that logrotate isn't
[09:19] <huats> morning
[10:00] <parapan> I have ubuntu-server 10.04 installed .....what mail service do you recommend installing ? which is the most easy one to configure / install hassle free :D ?
[10:00] <WalterN> parapan: good question, I'd like to know at some point too
[10:00] <WalterN> though its not quite installed yet
[10:00] <WalterN> XD
[10:00] <twb> apt-get install default-mta
[10:01] <twb> It's covered in apt-get install ubuntu-serverguide
[10:02] <parapan> twb: is this postfix by any chance ?? the default mail transfer agent ??/
[10:03] <twb> It is.
[10:03] <parapan> then, next question .....in order to adminster the server ...will webmin be the right choice ?
[10:04] <parapan> ubottu was recommanding ebox - looks like a similar package for doing the same thing ...
[10:06] <twb> !webmin >parapan
[10:07] <twb> He may only be a bot, but he's smarter than some people here...
[10:07] <WalterN> o.0
[10:07] <parapan> twb > I know that ...but some other guys were saying that webmin works OK for 64 bit version of ubuntu server ....
[10:08] <parapan> that's the reason for asking on the ubuntu-server channe;
[10:08] <twb> Then those "other guys" are idiots
[10:08] <parapan> =))
[10:08] <parapan> I'll let them know when talking again =))
[10:08] <twb> I grant you that if you deploy webmin you probably won't notice it eating a hole in the floorboards for a few years
[10:09] <twb> But then the house will collapse and you will die of pneumonia
[10:09] <WalterN> lol
[10:09]  * WalterN glances at 1.5gallon jug of acid
[10:09] <parapan> looks extreme ....
[10:20] <alvin> Over the weekend, some servers have had a high load. This message can be found on the monitor: INFO: task blocked for more than 120 seconds (see bug 276476 and bug 494476). Aside from the bugs, the immediate problem is now: it's impossible to log into the servers. They are running virtual machines, and those are working, but I can't use the console or ssh. Any other methods/key combo's that would allow me to log in?
[10:38] <kim0> Hola ubuntu server folks .. In preparation for 10.04.1 .. check this out http://foss-boss.blogspot.com/2010/08/ubuntu-server-10041-virtual-release.html
[10:38] <twb> alvin: walk over to it and type at the local keyboard
[10:38] <alvin> twb: I can go to a tty and type my name, but that's it. Hit enter and nothing happens
[10:46] <twb> Bounce it, then
[10:48] <ttx> kim0: cool, nice
[10:48] <kim0> everyone here feel free to retweet and blog the hell out of it :)
[10:49] <twb> Boo, hiss
[10:49] <twb> Down with this new-fangled technology
[10:49] <wieshka> any suggestions, why my SAS RAID complains about my SATA disks with this error while boot ?
[10:49] <wieshka> ata_id[680]: HDIO_GET_IDENTITY failed for '/dev/sda'
[10:49] <wieshka> ata_id[664]: HDIO_GET_IDENTITY failed for '/dev/sdb'
[10:49] <kim0> your disk is having an identity crisis :)
[10:49] <wieshka> the same reports hdparm -i /dev/sdx
[10:51] <wieshka> kim0: it is a Enterprise seagete HDD - not no brand from garage ... :)
[10:52] <twb> kim0: the "just show me" link shows an empty page in w3m
[10:52] <kim0> twb: well it's a google map (all javascript) .. so
[10:52] <twb> Oh, it's using google
[10:53] <alvin> I'm going to halt production, kick every user, reboot the servers and wait for better I/O scheduling.
[10:53] <twb> It doesn't know nat064.c.c.a is in .vic.au
[10:54] <twb> alvin: ionice(8) ?
[10:54] <alvin> iotop is broken in Lucid, so I don't know what the real problem is, but ionice will not prevent this.
[10:54] <alvin> It's because I use LVM snapshots from time to time
[10:54] <twb> Fair enough; I don't know what your actual problem was
[10:55] <twb> I was thinking along the lines of "these stupid VMs are flooding I/O and breaking my getty", so you ionice the VMs
[10:55] <alvin> Well, me neither because I can't log in, but I suppose it's because I use LVM snapshots. The VM's are not the cause this time.
[10:56] <alvin> If you take a snapshot, and the snapshots gets full, the kernel will start having problems. That is what I think happened.
[10:56] <twb> alvin: that's not my experience (on 8.04)
[10:57] <twb> IME all that happens is I/O to the snapshot breaks
[10:57] <twb> alvin: suggest booting in single, examining if any snapshot are full
[10:57] <alvin> twb: It's possible that this is new
[10:58] <alvin> Well... if there are snapshots, I'll have to use the recovery CD and remove them first, or the system won't boot. It's a mirror, and booting mirros while snapshots are present breaks Lucid booting.
[10:59] <alvin> almost 12:00. Let's start the process.
[10:59] <twb> alvin: that's royally fucked
[10:59] <alvin> I'll boot with recovery cd in single and see whether they are full first
[11:07] <alvin> The good news is: the servers booted without 'not finding root'. The weird news is: there is a snapshot on both servers, but without origin. Never seen such a thing.
[11:12] <twb> Are you using RAID1+0?
[11:12] <twb> Rather: what kind of RAID are you using?
[11:12] <alvin> twb: No, just RAID1 (mdadm)
[11:13] <twb> I've never seen it before either
[11:13] <alvin> Usually, this bug 563895 occurs, but maybe it didn't happen this time because the snapshot didn't know it's origin. I couldn't mount it either.
[11:14] <twb> Ew, that's still present in lucid?
[11:14] <alvin> sometimes, mdadm forgets about the array (bug 599135), but not on this server.
[11:15] <alvin> It's new in Lucid. It wasn't there in karmic
[11:15] <twb> Hum, there was a different on in 8.04, then.
[11:15] <twb> udev preferred snapshots over origins when mounting by label or uuid
[11:17] <alvin> Yes, I'll disable the whole snapshot thing immediately. I'll take backups from live systems from now on.
[11:17] <twb> Man, that's gonna screw me
[11:17] <twb> My entire backup infrastructure relies on LVM snapshotting
[11:18] <alvin> It did here, but performance was too heavily affected. I tried with less snapshots, but this crash was caused by 1 snapshot, so it 'll have to go.
[11:20] <twb> Oh.
[11:20] <twb> What I do is make the snapshots and then reap them as part of the backup process
[11:20] <alvin> That would be good if there's not too much I/O when you're taking a backup.
[11:21] <twb> Shrug
[11:21] <twb> It's an rsync job over (say) a 200GiB /home
[11:21] <alvin> I noticed that when I took a snapshot, and then used rsync, the kernel would start showing 'blocked tasks' (rsync, pdflush,...)
[11:22] <twb> Mine probably says that and I never noticed
[11:22] <alvin> Maybe it has something to do with large files/ext4. That's the situation here. Maybe /home is more doable.
[11:22] <twb> You're backing up >4GiB files?
[11:23] <alvin> yes
[11:23] <DigitalDeviant> Question: How do i set permissions on the public_html folder so everytime it belongs to the user/group, everytime i upload its always user/user
[11:23] <twb> Maybe play with rsync --in-place or something, I dunno
[11:23] <twb> I vaguely recall there's a "work better" option for large files in rsync
[11:23] <alvin> I can't find the blocked tasks messages. rsync crashed too
[11:23] <twb> DigitalDeviant: upload *as* that user
[11:23] <alvin> twb: Hmm, I should read that man page then.
[11:23] <soren> --inplace?
[11:24] <twb> DigitalDeviant: oh, sorry, I misread.
[11:24] <twb> DigitalDeviant: either change the user's default group (probably a bad idea), or make the parent directory setgid.
[11:24] <alvin> 'update destination files in-place'. Thanks. Might work
[11:24] <twb> soren: whatever
[11:24] <DigitalDeviant> twb what is setgid ?
[11:24] <twb> !RUTE > DigitalDeviant
[11:25] <DigitalDeviant> thanks for the info!
[11:25] <alvin> DigitalDeviant: Also look for documentation about 'user private groups'. Can come in handy.
[11:26] <DigitalDeviant> ok!
[11:26] <twb> They're the default
[11:26] <DigitalDeviant> Thanks guys
[11:27] <twb> (And IMO they're basically an admission that POSIX DAC groups aren't much use.)
[11:28]  * twb bumps "grok grsecurity" up on his TODO list from "never" to "probably never"
[11:33] <xampart> having a problem. i have /boot on sdb1 and the hdd is failing I/O errors. uname -r "2.6.32-24-server". i have an old backup (initrd.img-2.6.32-21-server) on sda1. what is the proper way to change /boot from sdb1 to sda1?
[11:33] <twb> xampart: grub legacy?
[11:35] <xampart> twb: so eg. just changing /etc/fstab -entry from /dev/sdb1 on /boot type ext4 (rw) to sda1 and reboot wouldn't work?
[11:35] <twb> xampart: are you using grub legacy>>
[11:35] <twb> s/>>/?/
[11:36] <robertpayne> If you use useradd and specifiy -p XXX to set the password is it supposed to be a crypted format or plain text? It's never working for me I have to go ahead and manually add user
[11:37] <twb> Don't use useradd; use adduser.
[11:38] <twb> As the manpage says, useradd -p takes a pre-crypt(3)ed password.
[11:38] <robertpayne> twb: ah.. would explain thx
[11:39] <twb> Note that crypt(3) is cryptographically weak.
[11:39] <robertpayne> what does useradd use if you don't use the -p? Is it crypt(3) as well?
[11:41] <xampart> twb: using the ubuntu-server 10.04 default (grub2?)
[11:42] <twb> robertpayne: by default it'll either prompt you to supply a password, or configure the account to be locked and without a password.
[11:42] <robertpayne> twb: interesting.. is the password protection any better than useradd ( not crypt(3) )?
[11:43] <twb> robertpayne: my point was that if you do it interactively it should at least use md5
[11:43] <twb> Whereas if you hack your own crypt(3) interface via perl (as I've done in the past), it'll use sucky old crypt
[11:44] <twb> If you look in /etc/shadow and the password begins with $1$ (or some other digit), it's OK.
[11:45] <robertpayne> bwuaaha wow
[11:45] <robertpayne> mine are plain text
[11:45] <robertpayne> yea that is BAD
[11:45] <twb> You think that's bad, you should see NIS
[11:46] <robertpayne> NIS?
[11:46] <twb> !NIS
[11:46] <twb> Stupid bot.
[11:47] <robertpayne> lol.. I'm fairly new to all this server stuff so learning as fast as I can
[11:47] <twb> In NIS if you have network access you're trusted to access the shadow file.
[11:47] <robertpayne> :S
[11:47] <twb> Because it was designed back when individuals couldn't afford computers, let alone laptops
[11:50] <robertpayne> there is no way to feed a crypted ( md5 or better ) password to useradd or adduser?
[11:51] <twb> !xy problem
[11:51] <twb> robertpayne: why do you want to?
[11:51] <joschi> robertpayne: see parameter -p of useradd
[11:51] <robertpayne> automated script installing a vhost
[11:51] <robertpayne> joschi: I did but as twb has pointed out to me it's crypt(3) format.. not very strong
[11:51] <twb> Are you preseeding it?
[11:52] <robertpayne> preseeding? as in prompting for input and then generating?
[11:52] <twb> crypt(3) isn't strong, but $1$-format encrypted strings should be compatible with the inverse function (decrypt(3)?)
[11:52] <twb> robertpayne: no, as in a preseeded installation
[11:53] <twb> Oh, wait, do you mean vhost as in an apache vhost?
[11:53] <robertpayne> yes sorry
[11:53] <twb> Mea culpa
[11:53] <robertpayne> twb: It's no biggie if I just run the commands after.. just nice to be able to do it all automated.. I'm setting up a htpasswd with the same info for access to awstats
[11:54] <robertpayne> the user I'm adding is chrooted pretty heavily to their home dir
[11:54] <twb> If this is for remote users to upload files, I'd be using assymetric crypto (SSH keys) rather than symmetric crypto (passwords)
[11:57] <robertpayne> twb: yea know what you mean... wish my ftp client ( Transmit ) supported SSH keys instead of passwords
[11:57] <twb> http://mywiki.wooledge.org/FtpMustDie
[11:57] <robertpayne> twb: I'm not using ftp.. sftp only
[11:58] <twb> Then get a better client
[11:58] <robertpayne> hehe :P
[11:58] <twb> Hell, OS X should have OpenSSH's CLI utilities installed by default
[11:59] <robertpayne> yeah it does
[11:59] <robertpayne> actually Transmit might support it.. I've just never gotten it to work
[12:10] <Daviey> Hey All... if you use ubuntu server, please add your location to http://maps.ubuntu.com
[12:10] <Daviey> Thanks!
[12:10] <Daviey> (note, it doesn't store personal data)
[12:17] <WalterN> nice
[12:17] <WalterN> marked
[12:18] <WalterN> not technically the server edition, but it is ubuntu and using it as a server... close enough?
[12:31] <robertpayne> Daviey: done one added to new zealand :)
[12:33] <WalterN> MT, USA
[12:33] <kim0> WalterN: close enough yeah ;)
[12:33] <kim0> Daviey: thanks for spreading the news :)
[12:33] <WalterN> should turn the map orange in no time
[12:34] <kim0> fingers crossed
[12:34] <WalterN> hehe
[12:35] <WalterN> I love Ubuntu
[12:35] <WalterN> :)
[12:35] <WalterN> http://i36.tinypic.com/ivx355.jpg? is the box Ubuntu is stuffed in
[12:35] <kim0> UK is hyper active :)
[12:36] <Daviey> robertpayne: awesome!
[12:36] <Daviey> WalterN: Well "yes", whilst technically not the server edition... it's fundamentally similar, and you are using it as a server
[12:36] <Daviey> kim0: no problem :)
[12:36] <robertpayne> Granted my server is in the US :( and I'm a US citizen but I'm lviing here now
[12:37] <WalterN> Daviey: I just like shiny things to click on when possible :)
[12:37] <twb> Poor bastard
[12:37] <WalterN> lol
[12:37] <Daviey> kim0: Good to see :)
[12:37] <twb> (re "US citizen")
[12:38] <patdk-lap> that map defently needs a better geoip database
[12:38] <WalterN> its not too far off of me
[12:39] <twb> It *should* be backed onto the same database as lucid's d-i uses
[12:39] <WalterN> well, considering it montana USA
[12:39] <Daviey> patdk-lap: It's probably the location your IP is set to be :/
[12:39] <Daviey> blame your ISP patdk-lap :)
[12:40] <robertpayne> WalterN you're in montana usa?
[12:40] <WalterN> yeah
[12:40] <robertpayne> what town?
[12:40] <WalterN> close to bozeman
[12:40] <robertpayne> haha awesome.. I grew up in Lewistown :)
[12:40] <patdk-lap> daviey, I just tried from several different isp's :)
[12:40] <WalterN> I just moved here from Oregon
[12:40] <robertpayne> WalterN: my brother lives in bozeman
[12:40] <WalterN> and I dont know *anybody* still
[12:41] <WalterN> which is fine by me, for the most part
[12:41] <WalterN> lol
[12:41] <WalterN> well, not *just*, moved 8 months ago now?
[12:42] <patdk-lap> heh, seems one of mine actually hit the correct location
[12:42] <patdk-lap> the comcast one
[12:43] <patdk-lap> none of the others did, amc, cogent, antietum, hopone
[12:44] <Pici> One of mine did.  The other one thought it was in atlanta.  Oh wll.
[12:44] <WalterN> actually mine is darn close
[12:44] <WalterN> maybe 5-10 miles off
[12:44] <WalterN> for montana thats nothing
[12:44] <WalterN> lol
[12:44] <Pici> Heh.
[12:46] <ttx> smoser: ping
[12:46] <robertpayne> WaltnerN: heh 10 miles in montana is like 2 footsteps in oregon
[12:47] <WalterN> depends on the place in Oregon
[12:48] <patdk-lap> oregon is the one state I never hear about, and don't know anyone in
[12:48] <WalterN> heh
[12:49] <WalterN> I'm going back there to the coast for my sisters wedding the 21st
[12:59] <kinygos> how do i get a command output to pause while i read it?  apologies for the n00b question
[12:59] <kinygos> (hi btw)
[13:00] <kinygos> to clarify, the output just flies off the top of the screen, and all i can see is the last page
[13:00] <soren> It depends a bit on the command.
[13:00] <soren> ...but probably just add "| less" to the end of the command line.
[13:01] <kinygos> bugger :( it's output from a python script execution
[13:01] <soren> Sounds like you just want "| less" at the end of the command line, really.
[13:02] <kinygos> just tried it, it didn't seem to work...just ran the script twice bizarrely enough...
[13:02] <kinygos> is there a way to redirect the output to a text file?
[13:03] <soren> "> filename" at  the and of the command line.
[13:03] <kinygos> awesome...thanks soren :)
[13:03] <soren> sure
[13:10] <lau> hello, I created a second instance of mysql using /var/lib/mysql2 and conf files in /etc/mysql2
[13:10] <lau> it is working ok with sudo mysqld_safe --defaults-file=/etc/mysql2/my.cnf
[13:11] <lau> but I tried to cp /etc/init.d/mysql /etc/init.d/mysql2 and update the latest
[13:11] <lau> when I service mysql 2 start I got some avahi or apparmor error and it looks like it is reading files in /etc/mysql not /etc/mysql2
[13:11] <lau> any idea ?
[13:12] <Jeeves_> lau: /etc/apparmor.d
[13:12] <Jeeves_> There's a file there for mysql
[13:12] <Jeeves_> that states which files can ben read, written and executed
[13:14] <Roxyhart0> hi there i got a couple of zone (domain) defined in my bind, one is call whcl and the onther one is call whcl.cl. however when i do nslookup from a client (windows) it give me thwe server whcl.cl coulnd find whcl. Somebody know why could be this problem?
[13:15] <lau> Jeeves yes I updated it already and sudo mysqld_safe --defaults-file=/etc/mysql2/my.cnf is working ok
[13:15] <lau> my problem occurs when I tried to copy the init script and update teh reference to /etc/mysql => /etc/mysql2
[13:16] <lau> if I /etc/init.d/mysql2 start it looks like mysqld_safe is reading its conf in /etc/mysql nor /etc/mysql2
[13:17] <lau> my /etc/apparmor.d/usr.sbin.mysqld http://dpaste.com/226611/
[13:18] <lau> and mysql2 init script http://dpaste.com/226612/
[13:33] <kinygos> leave
[13:34] <swub> Who?
[13:39] <Onga_The_Ghastly> OLA BRUDERZ
[13:39] <Onga_The_Ghastly> Sziasztok testverek
[13:41] <Onga_The_Ghastly> ikonia bruder ola
[13:54] <a_ok> which fs is better GFS or OCFS2?
[13:54] <Onga_The_Ghastly> GPS
[13:56] <a_ok> gps?
[13:58] <a_ok> Onga_The_Ghastly: Was that a joke or a serious answer?
[14:03] <Onga_The_Ghastly> a_ok brotha joke
[14:04] <a_ok> k
[14:09] <smoser> ttx, here.
[14:10] <ttx> smoser: yo
[14:11] <ttx> smoser: last time I look you still had one "INPROGRESS" item in alpha3, was wondering if it was DONE or should be POSTPOSNED
[14:11] <ttx> smoser: the other question I had is about the environment in euca2ools: https://bugs.launchpad.net/ubuntu/+source/euca2ools/+bug/556528/comments/2
[14:12] <ttx> IIUC --config should do the trick, and diverging from upstream to have env take precedence sounds overkill ?
[14:13] <ttx> smoser: upstream implementation is fully exclusive (doesn't stack config files) so if we were to support env variables first we would also probably need to implement config stacking
[14:13] <smoser> item now done. i  had commited code, just hadn't made to archive.
[14:14] <ttx> smoser: right: that's what I was thinking... that's why I didn't POSTPONE it myself on Friday
[14:15] <smoser> if --config works, then i guess i'm ok with that.
[14:15] <ttx> smoser: well, if it doesn't, then that's a bug :)
[14:16] <smoser> i guess i have no objections to "use --config" as an answer.
[14:18] <ttx> smoser: please quickcomment on the bug that you're ok with it
[14:18] <ttx> i'll close as wontfix
[14:18] <FDX> Guys, care to take a look at this specific reply? http://ubuntuforums.org/showthread.php?p=9695293#post9695293
[14:18] <FDX> Postfix seems to be delivering the message, but it does not exist.
[14:19] <smoser> ttx, so the one isue i have with "use --config" is that we suggest "use eucarc"
[14:19] <smoser> and that just doesn't work.
[14:20] <ttx> smoser: explain ?
[14:20] <smoser> never mind;
[14:20] <joschi> FDX: usually, when postfix says it has delivered a mail, it's been delivered. period.
[14:20]  * ttx gets confused
[14:20] <joschi> FDX: maybe it's just not at the directory you're expecting
[14:21] <FDX> hmmm... what about the last line? Aug  8 22:09:58 anbient postfix/qmgr[16053]: 1DDEE1DD01CC: removed
[14:21] <smoser> not important. it does seem to me that eniornment variables should  trump user configuration files, but its a nit.
[14:21] <FDX> Maybe its delivering and removing afterwards
[14:21] <lamont> FDX: that's postfix removing its copy of it
[14:22] <FDX> lamont, where would postfix store it then?
[14:22] <joschi> FDX: no. that just means that the mail has been removed from the queue
[14:22] <joschi> FDX: (see `man qmgr` ;)
[14:22] <lamont> it stores the interim copy in /var/spool/postfix
[14:22] <ttx> smoser: I kinda agree -- but then ideally you should be able to use one env variable to change just one bit of conf
[14:23] <ttx> smoser: the way it's written it's exclusive, it just uses the first thing it finds
[14:23] <smoser> yes, they should merge.
[14:23] <FDX> Thanx joschi, i'll take a look
[14:23] <ttx> smoser: so to put env in first in a suable way, we'd have to change that logic as well
[14:24] <ttx> smoser: I don't want to diverge that much, we could ask them to change that in a later version though
[14:24] <smoser> well, yes. in general i wouldn't suggest diverting, but getting it upstream.
[14:25] <ttx> right, changing precedence would be good in a new version
[14:26] <Onga_The_Ghastly> ikonia bruder? he is sleeping now?
[14:26] <Onga_The_Ghastly> Wake Up Puppetboy!
[14:26] <hggdh> and using environmentvariables to override configuration is common
[14:32] <joschi> FDX: you could also raise the verbosity of postfix's processes by adding -v in master.cf to the process binaries (e. g. smtpd -v)
[14:36] <FDX> joschi, that will certainlly be useful
[14:51] <Carleas> Do I need to set up SMTP with Postfix?  Can I run it without that?
[14:53] <Carleas> And can I set up Dovecot without having SMTP in Postfix?
[15:03] <MagicFab> I suppose this has already been circulated here (adding your city to "Ubuntu Server users" map):
[15:03] <MagicFab> http://foss-boss.blogspot.com/2010/08/ubuntu-server-10041-virtual-release.html
[15:03] <MagicFab> Could someone pls. add it to the topic of the channel ?
[15:05] <hggdh> Pici: ^ please
[15:12] <FDX> OK Guys, i guess i managed to do it. But now i'm getting RELAY ACCESS DENIED
[15:19] <nijaba> smoser: ping
[15:21] <smoser> nijaba, here, although i think i've alrady responded to you
[15:23] <nijaba> smoser: oh, really?  where?
[15:23] <nijaba> smoser: got it, thanks
[16:05] <Carleas> During configuration, Postfix asks for a FQDN.  All Ubuntu documentation says to give the MX record name, but that seems to contradict what Postfix is asking for.
[16:08] <Carleas> So which is it, FQDN or mx record?
[16:08] <joschi> Carleas: usually the (public resolvable) hostname (FQDN) of your system
[16:08] <joschi> Carleas: it's probably the string debconf sets for $myhostname
[16:08] <Carleas> Do I need the trailing dot?
[16:09] <joschi> no
[16:09] <Carleas> OK.  Thanks, Joschi.
[16:12] <FDX> Guys, why would postfix bounce the email back to the sender if it delivered it locally with success?
[16:17] <sveinse> Hi. I ran a aptitude dist-upgrade on a Lucid machine, and now it wont boot. When booting i see "error: no such disk", then a reboot and then I get a console output running fsck and then it asks for a password. But it does not accept any keypress unfortunately.
[16:19] <sveinse> What is the best approach for getting into recovery where I can repair grub and/or the partitions?
[16:19] <rooks> how to upgrade kernel on ubuntu server? how to select it for upgrade?
[16:19] <sveinse> The machine i running Ubuntu server amd64 with lvm2
[16:22] <raubvogel> rooks: apt-get kernel-upgrade?
[16:23] <raubvogel> Er, I meant dist-upgrade
[16:25] <raubvogel> sveinse: did you try booting from a livecd to see if the partitions are still ok?
[16:25] <hggdh> Daviey: re. bug 602540 -- when will 'cat' return? at end-of-file? ...
[16:25] <sveinse> raubvogel: I'm doing that right now
[16:26] <sveinse> raubvogel: But I think they are. I believe grub is the culprit
[16:27] <padhu1> sveinse: culprit?! :-?
[16:27] <sveinse> raubvogel: So jumping into that assumption: How do I use the Ubuntu install CD to rescue? It seems the rescue a broken system tries to start the installation all over
[16:28] <raubvogel> sveinse: AFAIK, it loads enough crap but in the RAM. Then it will ask you to mount disks
[16:28] <raubvogel> Another option is just to boot off livecd as if you just want to run it and then mount/check partitions
[16:29] <sveinse> raubvogel: Ah. So I can chroot from there...
[16:29] <raubvogel> Exactly
[16:36] <sveinse> Ok. The data's there. update-grub and install-grub installed successfully. But I still can't boot properly. I.e. I dont get a grub window, only the message "error: no such disk" and then it reboot.
[16:37] <hggdh> Daviey: why not just do 'head -c64 /dev/urandom | ...'?
[16:37] <sveinse> Ehm. Important note: This is a virtual machine on a VMW datacenter
[16:37] <rooks> raubvogel, thanks
[16:38] <Daviey> hggdh: What is the difference?
[16:38] <hggdh> Daviey: (1) /dev/urandom does not block; (2) we do not need a *lot* of binary data, just enough for 8 characters;
[16:39] <hggdh> Daviey: of course, /dev/urandom is not cryptographically secure (or strong, in the sense of /dev/random)
[16:39] <Daviey> hggdh: agreed
[16:40] <hggdh> but frankly, I do not see the difference. And using /dev/random to generate password is overkill -- and may affect other programmes that depend on it, like SSL
[16:40] <Daviey> hggdh: it's only for the initial password.. so it's not really a big deal
[16:41] <hggdh> Daviey: then we can certainly stop using /dev/random...
[16:42] <Daviey> hggdh: switching to head -c64 /dev/urandom | LC_ALL=C tr -dc 'a-zA-Z0-9' | head -c8
[16:43] <hggdh> Daviey: additionally, I am not sure 'cat' will ever end
[16:44] <mdeslaur> Daviey: you're assuming 64 chars is enough to extract 8 alphanum chars
[16:44] <mdeslaur> Daviey: you run the risk of getting a one-char password :P
[16:44]  * Daviey screams... shall i add a if $PASS == wc; repeat ? :D
[16:44] <hggdh> mdeslaur: yes indeed. To play safe we could use 128. A quick test here (about 500 runs) did not show much impact, if at all
[16:45] <mdeslaur> hggdh: why limit it at all?
[16:45] <mdeslaur> oh wait, it's draining the pool
[16:46] <hggdh> mdeslaur: because when I run 'cat /dev/urandom | tr -dc ...| head -c8 it get stuck
[16:46] <hggdh> there you go
[16:46] <Daviey> Speed difference, http://pb.daviey.com/I7uS/raw/ :)
[16:47] <hggdh> there you go. Even if /dev/urandom does not block, we would still be draining the entropy, which is not good
[16:48]  * hggdh uses the moment to grumble about how slow his laptop is (as compared to daviey's)
[16:49] <Daviey> heh
[16:50] <hggdh> I can only get 6ms if I just run the 'head /dev/urandom'
[16:50] <smoser> couldn't that block ?
[16:51] <smoser> indefinitely
[16:51] <smoser> how do you know that reading 64 bytes of data will result in 8 that are a-zA-Z
[16:51] <hggdh> /dev/urandom does not, it just goes to a less-random PRNG
[16:51] <smoser> i guess not block, but get you less than 8 chars and a exit code from 'head -c8' of non-zero
[16:52] <lau> any idea why when this script is louanched http://dpaste.com/226612/ it starts /etc/mysql/my.cnf conf ?
[16:52] <hggdh> smoser: you do not. It is a bet. We can increase the number of bytes read from /dev/urandom, but we cannot read a whole lot
[16:52] <sveinse> Anyone here with issues regarding grub and lucid server?
[16:52] <hggdh> smoser: because we would still be draining the random pool
[16:53] <smoser> well, urandom doesn't, as you said.
[16:53] <sveinse> My disks are fine, when doing rescue and chroot, grub seems to behave nicely and installs without fuzz. Yet, it does not when booting!
[16:53] <smoser> but if you check the return code of the 'head -c8' somewhere (which, you should), then you're going to occasionally get failure there as you're not going to get 8 chars.
[16:53] <hggdh> no, urandom does not block -- if it deplects the pool, it goes to a different (and weaker) PRNG
[16:54] <smoser> ah. i didn't realize it read from the real pool if there was data there.
[16:54] <hggdh> smoser: we could, then, check & re-run the conversion as many times as needed to get 8 displayable chars
[16:54] <hggdh> Daviey: ^
[16:54] <mdeslaur> ok, I think we should use pwgen
[16:54] <sveinse> of course this happened on a production server just before leaving... :(
[16:54] <mdeslaur> so we don't drain the entropy and we make sure we have the required length
[16:54] <Daviey> mdeslaur: hehe
[16:54] <hggdh> heh. pwgen it is ;-)
[16:55] <smoser> bike shed
[16:55] <Daviey> pwgen does mean we reinforce a standard 'library', which i approve of...
[16:55] <smoser> in deed
[16:55] <Daviey> means features and fixes are in one place :)
[16:55] <sveinse> I am able to change /etc/default/grub to whatever I want, but it does not seem to have any effect
[16:55] <hggdh> Daviey: +1
[16:56] <hggdh> FWIW
[16:56] <smoser> sveinse, after modifying that you have to run update-grub
[16:56] <sveinse> smoser: Sure. Did that
[16:56] <smoser> and you have grub 2, right ?
[16:56] <Daviey> hggdh / mdeslaur / smoser: Thanks :)
[16:56] <sveinse> smoser: 1.98 is grub2, right
[16:58] <sveinse> Well I have apt up and running in the chroot, so I can revert to whatever I want
[16:58] <smoser> sveinse, i just tested here, that updating GRUB_CMDLINE_LINUX_DEFAULT= and running update-grub modified /boot/grub/grub.cfg
[16:58] <smoser> and i can attest elsewhere that update-grub does run in chroot
[16:59] <sveinse> smoser: great, thanks. Let see here..
[17:01] <sveinse> It seems grub.cfg here is changed as well. Removed quiet and splash, and they are gone
[17:01] <sveinse> But it could seem the problem is deeper: I never get any grub menu to begin with
[17:03] <sveinse> However it boots into *something* which does fsck and then prompts me for a password
[17:03] <cabrey> Hey all. I've got a question regarding a vpn server. I have setup a pptpd server and everything connects fine with the exception that traffic doesn't seem to tunnel through the connection. For example, I just get an error message when trying to use a browser. is there a fix or setting to help me out?
[17:07] <sveinse> It seems the error occurs after installing grub-pc 1.98-1ubuntu7
[17:08] <smoser> although, sveinse yeah, you dont have grub installed on a disk that is getting loaded. at least not one that is running correctly.
[17:09] <sveinse> smoser: eh? sorry? please explain
[17:09] <smoser> well, i'm confused as to why you're in a chroot, if you operating on a "real system".
[17:09] <sveinse> grub-install /dev/sda is safe from a chroot env, isnt it?
[17:10] <smoser> it probably is supposed to be.
[17:10] <smoser> but it would appear maybe its not working
[17:10] <smoser> i would make sure that /dev, /proc, and /sys are mounted in the chroot
[17:10] <sveinse> I'm in a chroot because I'm sitting in Ubuntu install CD Rescue. That the only way I can access the harddrive data and installation
[17:11] <sveinse> My grub wont boot properly. And I cannot seem to be able to get up any grub menu either, so something is fishy
[17:11] <smoser> ie, i might do: for x in /dev /proc /sys; do mount --rbind ${x} /chroot/${x}; done
[17:12] <smoser> sveinse, what *does* happen  when you try to boot ?
[17:13] <sveinse> A brief message saying "error: no such disk". Then a reboot and later I see a blank screen. After I while it fscks all the disks on the system (including the lvm2 ones) and then prompts me for a password (not login). The keyboard does not work
[17:14] <smoser> :-(.
[17:14] <smoser> i'm really sorry, but it would appear i'm not able to help at the moment then.
[17:14] <sveinse> All drives are intact when inspecting them from a live CD
[17:14] <sveinse> pity.
[17:14] <smoser> you may get more help from #grub, but that might be more a deveopment channel
[17:15] <sveinse> Well, perhaps if you knew how to roll back to another older grub or perhaps grub(1)
[17:15] <sveinse> Well, I'm grateful for trying at least. Thanks
[17:16]  * sveinse chained to my desk until this production server is up again...
[17:24] <kim0> Daviey: Thierry mentioned you're working on creating a 10 line openstack/nova installation script. How close are we to that :)
[17:24] <Daviey> kim0: Not me, sorry sir.
[17:25] <kim0> ah ok
[17:36] <sveinse> OK. The cause of the server's failure is not grub
[17:37] <sveinse> All my older kernel version also fails, so I would guess that it's failing during initrd booting
[17:43] <Jason> is anyone running vsftpd that could help me "correctly" configure a user?
[17:46] <sveinse> What do I do when initrd fails during boot on a production server?
[17:48] <cwillu_at_work> sveinse, cry
[17:48] <sherr> sveinse: what happens exactly? Error message?
[17:48] <sveinse> I get fsck check, then "init: ureadahead-other main process (593) terminated with status 4"
[17:48] <cwillu_at_work> that's shouldn't be a fatal error
[17:49] <sherr> I think I see those non-fatally ...?
[17:49] <sveinse> Then it asks for "Passwd: "  It does not respond to any kbd input
[17:49] <sherr> Boot up minus "quiet"?
[17:49] <sveinse> sherr: I am
[17:49] <cwillu_at_work> sveinse, the password won't be echo'd to the screen
[17:50] <sherr> PS/2 or USB kbd directly attached?
[17:50] <sveinse> *what* passwd? Root which is never set on Ubuntu?
[17:50] <cwillu_at_work> boot up without "quiet splash"
[17:50] <cwillu_at_work> the splash screen might be interfering
[17:50] <cwillu_at_work> but ya, that's a bit odd
[17:50] <cwillu_at_work> what happened right before you rebooted?
[17:50] <sveinse> This is a VMW datacenter virtual server, so no physical kbd
[17:50] <cwillu_at_work> (i.e., a server doesn't break itself)
[17:51] <sherr> Try another boot kernel?
[17:51] <sveinse> Kbd works elsewhere; in grub and in rescue cd
[17:51] <cwillu_at_work> sveinse, and the vm window is grabbing the keyboard?
[17:51] <sveinse> Yes, all of the 4 installed on the server. All of them are comprimised
[17:51] <cwillu_at_work> sveinse, "compromised" means something else
[17:51] <cwillu_at_work> generally means you've been hacked
[17:52] <sveinse> cwillu_at_work: heheh. Sorry. I doubt that (being hacked that is)
[17:52] <cwillu_at_work> try booting up in "single" or the recovery mode
[17:52] <sveinse> Recovery mode behaves exactly the same way. I can try single
[17:54] <sveinse> The drive's data, including the lvm2 is intact, as I can access it from the live CD
[17:55] <sherr> All 4 kernels/initrd broken? Something else must be wrong or different? Hardware change/error?
[17:56] <sveinse> Nope. Single behaves the same way.
[17:56] <sherr> Any errors reported at boot before "init: ureadahead-other .."?
[17:56] <sherr> Can you capture boot messages?
[17:57] <sveinse> When booting without splash and quiet, I see the kernel output. Then I see init running a few seconds. Then it changes fontset and blanks the screen. From there I see fsck running through the disks, then the ureadahead-other. Then the Passwd: input
[17:58] <sherr> Good question on passwd ... what password?
[17:58] <sherr> I am afrid I always set a root password personally. I know it is not the "ubuntu way"/
[17:59] <sveinse> Well IT tells me they have upped the datacenter version, but since init is able to run for 5 secs, do you really think its HW/emulation related?
[17:59] <sveinse> Theres not kernel oops I can see
[18:00] <sherr> What's "upped the datacenter version" mean?
[18:00] <sveinse> Sorry. Language thing. They have upgraded it to a newer version
[18:01] <sherr> Upgraded what?
[18:01] <sherr> It is very suspicious that, suddenly, no kernel boots?
[18:02] <sveinse> The datacenter. The server I struggle with is running on a VMWare Datacenter server
[18:02] <sherr> IS this some sort of VMware VM server?
[18:02] <sherr> Ahh .. yes :-)
[18:03] <sherr> Seems a bit suspicious ... I'd ask for details i.e. old version = ?, new version = ?
[18:03] <sherr> And quickly check the VMware forums perhaps, for screams .. :-(
[18:04] <sveinse> VMware vCenter Server 4.0.0 258672. I dont quite remember the exact old version, 3.5 something IIRC
[18:04] <sherr> Well .. 3.5 -> 4.0 seems like a "major" upgrade" to me. Maybe something broke.
[18:04] <sveinse> Well, the server ran perfectly up until I did a dist-upgrade where a linux-kernel were upgraded and it requested a reboot
[18:05] <sveinse> So the trap needs to be in boot only, else the server wouldn't run prior to my reboot
[18:06] <sherr> Yes, but if it was a "bad" kernel update, you could still use old kernels. This is why I am suspicous.
[18:06] <sherr> I'm guessing though, no idea really.
[18:06] <sveinse> Yes. But the initrd is always updated for all kernels, isn't it? So a bug there could be fatal, couldn't it?
[18:07] <sveinse> Well anyways, I'm stuck with a dead production server and cannot go home until that's fixed :(
[18:07] <sherr> Yes, you might be right. Would be good to be able to enter "single" user mode.
[18:08] <sveinse> Yeah.
[18:08] <sveinse> I do have access via a live CD. That way I can set the root pwd to something foo
[18:09] <sveinse> BTW: While working on ubuntu-arm I remember adding a init=/bin/init --verbose to the kernel option. Could this be used on the server?
[18:09] <sherr> Yes, maybe. Good luck.
[18:10] <sherr> I think you could try init=/bin/sh
[18:12] <sveinse> No, so I join you in the suspection of something serious :(
[18:13] <sveinse> I.e. it didn't work. No prompt with init=/bin/sh
[18:14] <sveinse> If I have to reinstall this server, this would be the third time. The second time was caused by an upgrade from karmic to luicid which crashed the server
[18:14] <sveinse> Is Lucid more unstable than the previous version, or is there something in regards of the VM which contributes this unstability I'm observing.
[18:15] <sveinse> Because I'm really loosing trust in either this server or Lucid. Which of these two, I dont know yet.
[18:15] <ivoks> no prompt?
[18:15] <sveinse> nope
[18:16] <ivoks> do caps and num lock blink?
[18:16] <sveinse> But this time I see all of the kernel output
[18:16] <sveinse> ivoks: I don't think I can know, because this is a virtual server
[18:17] <ivoks> ah, missed that part :)
[18:17] <sveinse> But pressing enter scrolls down the sceen, so the machine is responding
[18:17] <ivoks> hypervisor is...?
[18:17] <ivoks> oh, ok
[18:17] <ivoks> what kind of disks do you use?
[18:18] <ivoks> scsi or ide emulation?
[18:18] <sveinse> boot is standard ext2, while lvm2 on the rest.
[18:18] <sveinse> Hold on
[18:18] <ivoks> lvm?
[18:18] <ivoks> could it be that lvm isn't ready
[18:19] <ivoks> could you boot older kernel?
[18:19] <sveinse> scsi it seems. I have them in the kernel output. And it does work from the Debian install cd in rescue mode. All the data is intact
[18:19] <Jason> how do I add a user but not give him a home directory?  Is this possible?
[18:19] <ivoks> Jason: adduser --help
[18:19] <ivoks> adduser --system [--home DIR] [--shell SHELL] [--no-create-home] [--uid ID]
[18:19] <ivoks> ...
[18:20] <ivoks> without --system
[18:20] <sveinse> I can't boot older kernels because the lockup occurs after the ramdisk has started.
[18:20] <ivoks> sveinse: can you remove splash quiet from the kernel command line?
[18:20] <ivoks> sveinse: each kernel has its own ramdisk
[18:20] <sveinse> Even the rescue targets is not workng.  I have removed both the quiet and splash
[18:21] <ivoks> sveinse: ok, boot without splash and quiet
[18:21] <ivoks> and leave it like that
[18:21] <sveinse> done
[18:21] <ivoks> sveinse: after couple of minutes it will give up and drop you into busybox
[18:21] <ivoks> sveinse: then we can investigate fruther
[18:21] <ivoks> iirc, 5-10 minutes
[18:22] <sveinse> ok, couple of minutes... Havent tried that. Well hold on
[18:22] <ivoks> sveinse: / is on LVM, right?
[18:23] <sveinse> yes
[18:23] <ivoks> ok, just sit and wait for # :)
[18:23] <sveinse> hmmm. I got something new here. Could a cifs mount which is requesting mount at boottime lock up everything. Because that might be it...
[18:24] <ivoks> sveinse: is it in /etc/fstab?
[18:24] <sveinse> yes
[18:24] <ivoks> sveinse: if you set it up to wait uncoditionaly... then yes
[18:24] <sveinse> "0 0" should wait should it?
[18:24] <sveinse> not, I mean
[18:24] <ivoks> 0 0 should pass
[18:25] <ivoks> it something else then
[18:25] <ivoks> 's
[18:25] <ivoks> probably lvm
[18:25] <ivoks> i'll be back in couple of minutes
[18:25] <ivoks> smoke time
[18:25] <sveinse> thanks!
[18:27] <sveinse> but: If "0 0" should pass, why do I get a "mount error: could not resolve address for xxx" which shows that it actually tries to mount the cifs share?
[18:27] <sveinse> ...now while waiting for that prompt
[18:28] <SpamapS> sveinse: maybe mountall is picking it up anyway? It shouldn't.
[18:28] <zul> wohoo..
[18:29] <sveinse> I have a hunch that this is it you see
[18:30] <SpamapS>        This is a temporary tool until init(8) itself gains the necessary flexibility to perform this processing; you should not rely on its behaviour.
[18:31] <SpamapS> Nice man page.. :)
[18:31] <ivoks> sveinse: so, you see mount issues?
[18:32] <sveinse> yes, now I did
[18:32] <ivoks> sveinse: then it already mounted your / and can't pass that fstab line
[18:32] <ivoks> sveinse: try ctrl+c :)
[18:33] <sveinse> And the Password prompt could originate from the cifs mount prompt, because I saw that before the reboot
[18:33] <ivoks> sveinse: if that doesn't work, reboot with ctrl+alt+del and change kernel command line; remove splash and quiet and add init=/bin/bash
[18:33] <sveinse> However, it does not respond to kbd. That could be virtual datacenter issues and not ubuntu
[18:34] <sveinse> ivoks: I did try that, and by some reason it did not work.
[18:34] <sveinse> Well I will try to fix fstab, and retry
[18:34] <ivoks> sveinse: just comment out that line with cifs
[18:36] <ivoks> sveinse: did you supply password for that mount?
[18:36] <ivoks> sveinse: in fstab
[18:36] <sveinse> Yes. But it fails sometimes. Dunno why, but its there in fstab so I hope that is the cause
[18:37] <ivoks> sveinse: could be, if all kernels fail
[18:37] <ivoks> sveinse: it's obvious that your / gets mounted
[18:37] <ivoks> sveinse: cause it started upstart and tries to mount the share
[18:37] <ivoks> maybe mountall starts before network :)
[18:38] <sveinse> YES YES YES ! It works
[18:38] <SpamapS> mountall does start before net
[18:38]  * sveinse very happy!
[18:38] <SpamapS> but on net-device-up it gets sent USR1 which tells it to mount the network filesystems
[18:39] <sveinse> So I have a fstab line which makes the boot fail. Even with 0 0
[18:39] <ivoks> SpamapS: i know for sure that nfs doesn't work on karmic
[18:39] <ivoks> SpamapS: haven't tried with lucid
[18:39] <ivoks> net-device-up doesn't guarantee network is up, actually
[18:39] <SpamapS> ivoks: net-device-up is really broken
[18:40] <ivoks> sveinse: does your machine has dhcp/bridge network?
[18:40] <SpamapS> ivoks: hah, yeah, so we agree. ;)
[18:40] <sveinse> ivoks: No. A simple static IP4 address
[18:40] <SpamapS> ivoks: I am fairly certain that that signal should be moved from ifupdown to dhclient for dynamic interfaces.
[18:40] <ivoks> hm
[18:40] <ivoks> SpamapS: but then you still have this problem with bridged interfaces
[18:41] <ivoks> SpamapS: and bonded too
[18:41] <ivoks> bridge is started, mountall starts, but the IP isn't there ye
[18:41] <ivoks> t
[18:41] <SpamapS> ivoks: agreed, we basically have to write an ifmond that sends the up/down signals when it detects IP changes on interfaces.
[18:41] <ivoks> or
[18:41] <ivoks> remove networked filesystem from /etc/fstab
[18:41]  * ivoks runs and hides
[18:42] <SpamapS> ivoks: that still leaves the question of when to mount them.
[18:42] <ivoks> just before rc.local, in while loop
[18:42]  * ivoks runs and hides again
[18:42] <SpamapS> customizing the upstart conf file for every special machine seems like a boring, thankless job for every sysadmin. ;)
[18:43] <sveinse> This is the faulty entry: "#//nosrv051/SWRepository/Backup/nosrv111     /srv/backup/nosrv051  cifs,user=user,pass=pass,dom=dom       0 0"
[18:44] <ivoks> if we put 'mount-network-filesystem' in /etc/network/if-up.d/...
[18:44] <SpamapS> ivoks: thats where net-device-up is generated
[18:44] <sveinse> When this is uncommented, the server fails to boot as it tries to boot the mount.
[18:45] <SpamapS> ivoks: because those get called as soon as ifup ethX returns
[18:45] <ivoks> sveinse: try adding manual mount in /etc/rc.local
[18:45] <ivoks> sveinse: i'm interested would it work then
[18:45] <SpamapS> well, before, but basically if you've started dhclient, you run the ifup script
[18:45] <sherr> There's a "netdev" option for the fstab - I thought that affected mounting
[18:45] <sherr> I use NFS on Ubuntu. I hope it isn't broken.
[18:45] <ivoks> hm... netdev
[18:46] <ivoks> _netdev actually
[18:46] <sveinse> While the countless reboots, I did mostly see "Password:" which is the cifs password prompt (and thus the network is up). The last iteration is the only time I got a network failure for the mount (and where I understood what it was)
[18:46] <sherr> Yes. I always thought it was dangerous to mount CIFS in fstab ...
[18:47] <sherr> Plus having a password in fstab
[18:47] <ivoks> SpamapS: if-up.d always worked for me with dhclient
[18:48] <ivoks> SpamapS: those scripts were started after i got Ip
[18:48] <ivoks> SpamapS: not when i requested interface
[18:48] <SpamapS> ivoks: I think if we just add a 'default-route-up' event, it can be sent in /sbin/dhclient-script right after the 'route add default...' line.
[18:48] <sveinse> ah, I realize I lack the noauto option in the first place.
[18:49] <SpamapS> ivoks: hm, so maybe I'm wrong. ;)
[18:49] <ivoks> SpamapS: but as i said, dhclient isn't the only one problematic
[18:49] <SpamapS> ivoks: Basically services that configure interfaces need to send the signal, or a daemon needs to monitor interfaces/routing tables and send signals when things change.
[18:49] <sveinse> I have no need for automount of this during boot. Do you still want me to try it in rc.local?
[18:49] <ivoks> sveinse: no :)
[18:49] <sherr> I would skip CIFS in fstab and rc.local
[18:50] <ivoks> SpamapS: or we could rely on dbus to think for us
[18:50] <ivoks> SpamapS: :)
[18:50] <SpamapS> sherr: you can at least put the password in a root-readable-only file.. there are mount options to allow that..
[18:50] <sveinse> But guys: I am really happy for you assistance. Thanks a lot!
[18:50] <ivoks> np
[18:51] <ivoks> anyone with htc desire?
[18:51] <sherr> Yes. Just an allergy to this sort of thing for boot :-/
[18:51] <sherr> I have a desire.
[18:51] <sveinse> I have one
[18:51] <ivoks> which rom do you guys use? :)
[18:52] <sveinse> The vanilla from HTC for my part. Unfortunately I need exchange for my work mail
[18:52] <sherr> I am on Vodafone UK - it is "2.1 update1"
[18:52] <sveinse> FW: 2.1-update1. SW: 1.21.405.2
[18:53] <sveinse> This is in Norway
[18:53] <ivoks> i'm on a stock too
[18:53] <ivoks> but i have 2.2 :)
[18:53] <sherr> I'm not in an upgrade hurry really. Phone upgrade treadmill ....
[18:53] <sveinse> I noticed from the Hero upgrade, that HTC rolled the images out in different parts of the world at different dates.
[18:53] <ivoks> yep
[18:54] <sveinse> ivoks: What's your location?
[18:54] <ivoks> i took the rom for some other part of the world :)
[18:54] <sveinse> Ah
[18:54] <SpamapS> I have an HTC Magic .. still running the original donut OS that T-mobile pushed to me like a year ago.. because it included their little tmobile wifi hotspot backdoor program, so I get free wifi at all tmobile hot spots.
[18:55] <SpamapS> even tho I'm not on tmobile.. on att.. so I also get free wifi at att hot spots
[18:55] <ivoks> sveinse: http://forum.xda-developers.com/showthread.php?t=741775
[18:56] <sveinse> Havn't HTC protected the devices against rooting?
[18:57] <ivoks> boot loader is the one that can lock you
[18:57] <sveinse> I thought I read something about it somewhere...
[18:57] <ivoks> that's why this custom roms don't upgrade hboot
[18:58] <ivoks> battery life sucks :/
[18:58] <zul> smoser: where can i get uec images again?
[18:59] <sveinse> But this implies that eventually 2.2 will arrive through OTA, right?
[18:59] <ivoks> sveinse: yes, during august
[18:59] <sveinse> I can wait... :D
[19:00] <sveinse> Well. I need to head home. Thanks again guys. You saved my evening
[19:01] <ivoks> take care
[19:10] <sveinse> I'm back. I did a reboot of the server, and it failed once more. It seems the option "noauto" is ignored. So during booting it tries to mount the cifs mount regardless of the option
[19:10] <sveinse> I really had to uncomment the thing to make the server boot
[19:12] <sveinse> I just wanted you to know. I can probably come back tomorrow and talk about it (e.g. if a bugreport should be filed), but I need to go now
[21:08] <ball> Daviey: you know I'm going to have to install Ubuntu Server on something just to put a dot on that map! ;-)
[21:08] <WalterN> ball!
[21:08] <WalterN> you know you want to
[21:09]  * ball looks around the room for candidate hardware.
[21:13] <ball> Aha!  That might work.
[21:25] <tomsdale> I'm just thinking about a new webserver layout using lvm (inside a vm). If I'm already using lvm - does it make sense to you it for all of the partitions or are there mountpoints where it makes sense to have a primary partition?
[21:25] <tomsdale> s/you/use
[21:30] <tomsdale> I already got that /boot makes sense as normal partition.
[21:31] <ScottK> Any suggestions on what could cause a PID file to be removed?
[21:31] <ScottK> One possibility is I was confused about which server I was logged into when I stopped clamd.
[21:32] <guntbert> ScottK: shutting down the process is one, someone manually removing the pid file is another
[21:33] <ScottK> clamav/clamav.log:Mon Aug  9 10:54:29 2010 -> Pid file removed is what the logfile says.
[21:36] <ScottK> Just checked on my test server.  That is what it logs if you stop it.
[21:38] <guntbert> ScottK: so I guess is was above mentioned possibilty -- be glad that you didn't issue a halt :-)
[21:39] <ScottK> Yeah.  Well I have physical access to the box in question, so that would have been bad, but not horrible.
[21:39] <ScottK> Probably fingers on autopilot when typing hostnames.
[21:40] <guntbert> ScottK: :)
[21:40] <kirkland> smoser: ping
[21:41] <smoser> kirkland, hi
[21:41] <kirkland> smoser: hey
[21:41] <kirkland> smoser: at the kvm forum today;  qemu-0.13 hasn't GA'd yet
[21:41] <kirkland> smoser: talking with hallyn, we're thinking about uploading 0.12.5 for Maverick
[21:41] <kirkland> smoser: we can push 0.13 to a PPA when it releases, but probably not Maverick
[21:42] <kirkland> smoser: it's in RC right now, but avi hasn't merged it yet
[21:42] <kirkland> smoser: i was just checking if there was anything you were expecting in qemu-kvm 0.13 that you needed
[21:43] <smoser> i don't know exactly why i would wwant 0.13. except for that it will act in ways i'm not aware of :)
[21:43] <smoser> so i would defer to your and serge judgement.
[21:43] <smoser> i opened a good one today.
[21:43] <smoser> bug 615529
[21:44] <Kutakizukari> On have Ubuntu server running on Ubuntu 10.4 How can I have the server not start on I start the computer?
[21:44] <smoser> "the server" ?
[21:44] <ball> Kutakizukari: Are you running Ubuntu Server running on some sort of virtual machine?
[21:45] <ball> hosted by a (desktop) Ubuntu system?
[21:45] <Kutakizukari> just on my laptop Unbuntu 10.4
[21:46] <Kutakizukari> ball, yes
[21:47] <ball> Kutakizukari: ...and the virtual machine is starting when you boot the host, or do you not want the VM to boot into Ubuntu Server when you launch the VM?
[21:48] <Kutakizukari> ball, When I turn on my laptop, I don't want the server to auto start.
[21:49] <Kutakizukari> ball, just when I want to use it for development.
[21:49] <ball> That makes me think the VM is launching when you start (desktop) Ubuntu
[21:49] <ball> Kutakizukari: so I'm not sure it's really an Ubuntu Server question as such.
[21:49] <todd> What virualization software are you using?
[21:50] <Kutakizukari> todd, how do I find out?
[21:50] <todd> I'm sorry. I'm not sure I can help you with that.
[21:50] <Kutakizukari> k
[21:50] <todd> If you don't know then chances are that you're not running virtualization software at all?
[21:50] <todd> That just confuses me.
[21:51] <todd> Because if you're not running virtualization software it makes your first comment make even less sense.
[21:52] <Kutakizukari> todd, explain what is a virtualization software?
[21:52] <todd> Virtualization Software: http://tinyurl.com/2vz5sfn
[21:53] <Kutakizukari> I'm running desktop edtion Ubuntu 10.4 and then unstalled the server via command line.
[21:53] <todd> Oh now I get it.. you mean the Server kernel?
[21:53] <ball> "unstalled" is a nice word.  I may coopt that.
[21:54] <Kutakizukari> Installed sorry ball
[21:54] <todd> Kutakizukari: to make your machine not boot the server kernel automatically all you have to do is edit your grub configuration. This is not a server specific question.
[21:55] <todd> That being said.. Now that you know to edit the grub config you should be able to figure it out by googling.
[21:55] <Kutakizukari> k thanks
[21:55] <ball> I had no idea it was possible to install Ubuntu Server on top of (desktop) Ubuntu.
[21:56] <todd> My assumption: he installed ubuntu server then apt-get install'd ubuntu-desktop.
[21:56] <todd> His machine would still boot the server kernel by default causing issues with certain packages.
[21:56] <todd> Huge ASSumption, but the only thing I can come up with.
[21:57] <todd> If that's not it I have no idea what he's on about.
[21:58] <ball> todd: I was struggling to make sense of the question, so at least I'm in good company.
[21:59] <Kutakizukari> I have the desktop software installed and then sudo tasksel install lamp-server
[21:59] <todd> OH.. aha
[21:59]  * todd facepalms
[22:00] <ball> Does that install Apache, MySQL and PHP?
[22:00] <todd> Yes.
[22:00] <ball> (on top of (desktop) Ubuntu?)
[22:00] <todd> Yes.
[22:00] <Kutakizukari> ball, yes
[22:00] <todd> Kutakizukari: ignore the grub bit from earlier
[22:00] <Kutakizukari> https://help.ubuntu.com/community/Drupal
[22:01] <todd> Kutakizukari: if you don't want it to run all the time then purge lamp-server and install xampp.
[22:01] <Kutakizukari> but there is other docs on how to do the same
[22:01] <todd> The instructions for getting drupal up will be a bit different though.
[22:02] <Kutakizukari> I have everything working fine and even xampp in the past but xampp is no longer supported by ubuntu so they said lamp-server is
[22:04] <Kutakizukari> thank you for your help
[22:04]  * ball doesn't know what an xampp is.
[22:05] <Kutakizukari> ball, http://www.apachefriends.org/en/xampp.html
[22:05] <tomsdale> xamp - preconfigured amp package.
[22:05] <ximal> i can't get my cli terminal to be centered properly and text is hanging off the  right side of my monitor ... How can I fix this ?
[22:05] <tomsdale> used to use it on windows back in the days before enlighenment
[22:06] <ball> ximal: grab the title bar and drag it where you want it?
[22:06] <ximal> I tried to use my monitor's AUTO ADJUST feature ...
[22:06] <ximal> errr nope ball
[22:06] <ximal> i'm in cli
[22:06] <ball> ximal: Ah, you probably mean from the console
[22:06] <ximal> i install text only
[22:06] <ball> I have no idea then, if your monitor is set right.
[22:06] <ximal> err i only installed the cli text edition
[22:06] <ximal> well how can I install some form of gui or text
[22:07] <ximal> I tried install xorg or xserver etc but all i got was a frozen black screen when i do startx
[22:07] <guntbert> !xampp | todd don't recommend unsupported packages please
[22:08] <Kutakizukari> bad todd
[22:08] <todd> You couldn't tell me that without the bot switch?
[22:09] <ximal> well can someone please tell me what I could do to install maybe gnome or kde ?
[22:09] <ball> !donkey | wombat turnip
[22:09]  * ball shrugs
[22:09] <guntbert> todd: I could, but my mere opinion would not count much
[22:12] <qman__> ximal, what you need to do is fill the screen with text, then use the auto adjust
[22:12] <qman__> your monitor can't figure out where the edges are because nothing is on screen
[22:12] <tomsdale> is ext2 a good choice for /tmp - I guess I don't need journaling there.
[22:12] <guntbert> tomsdale: sure
[22:13] <ximal> how's that possible to fill it with txt ? nano ?
[22:13] <ximal> err use nano maybe ?
[22:13] <qman__> hold down a key for a while
[22:13] <ball> ximal: "banner" ?
[22:14] <ball> Does Ubuntu Server ship with a banner?
[22:14] <ximal> banner ?
[22:14] <ximal> ohh i could use the text browser
[22:14] <ximal> and bring up a photo
[22:14] <ximal> maybe ?
[22:14] <ximal> brb
[22:15] <ball> a photo in a text browser?
[22:15] <tomsdale> ball you mean a splash screen or the command 'banner'
[22:15] <ball> tomsdale: I meant the command 'banner'
[22:16] <tomsdale> nop, not in the default ball , sysvbanner is the package
[22:16] <tomsdale> that's for 10.04]
[22:16] <ball> tomsdale: Thanks
[22:17] <ximal> hmm
[22:17] <ximal> i thought you could
[22:17] <ximal> err I thought you could bring up photos in cli
[22:17] <ball> ximal: Perhaps you can on Linux, which might use a graphical console of some sort.
[22:17] <ball> I'm from a different world.
[22:17] <qman__> only with libaa or libcaca, I don't think lynx has that feature built into it though
[22:18] <qman__> ah, one good way would be to start up irssi
[22:18] <qman__> since it creates title bars
[22:18] <ball> qman__: Also, I like irssi.
[22:18] <ximal> ahh
[22:18] <ximal> brb
[22:18] <ximal> thanks for the irssi thought
[22:27] <mathiaz> SpamapS: ceph is in da' houze!
[22:37] <ximal> AKKKK ...
[22:38] <ximal> still no center even if i use open box
[22:38] <ximal> i'm gonna blow a dang gasket ... new bug found
[22:40] <ball> ximal: bug in software or in your monitor's firmware though?
[23:32] <SpamapS> mathiaz: ooohhh snap (re: ceph)
[23:32]  * SpamapS does the cabbage patch