[00:00] <twb> It appears to print each setting that is unchanged from the default
[00:01] <aetaric> booting a thin-client ubuntu 10.04, it starts eth0 up, and then my cisco router(dhcp server) spews errors about malformed option 61. any ideas about how to solve?
[00:01] <twb> What's option 61?
[00:02] <aetaric> it's the client ident
[00:02] <econnell> client-identifier
[00:02] <twb> As in send host-name ?
[00:03] <econnell>    It is expected that this field will typically contain a hardware type
[00:03] <econnell>    and hardware address, but this is not required.  Current legal values
[00:03] <econnell>    for hardware types are defined in [22].
[00:03] <twb> So cisco appears to be more prissy  than the RFC
[00:03] <aetaric> pretty much
[00:04] <aetaric> this only happens when i boot a thin client.
[00:04] <twb> FWIW, netbooting Works For Me, with 10.04 desktops and 8.04/dnsmasq as the server, and intel PXE ROMs.
[00:04] <twb> aetaric: this is happening before pxelinux loads, right?
[00:04] <aetaric> no
[00:04] <aetaric> this is post pxelinux
[00:05] <twb> OK, so pxelinux.0 loads -- does it get as far as displaying a menu and loading the kernel and ramdisk?
[00:05] <aetaric> yes
[00:05] <twb> OK, so the bad DHCP request is somewhere within Ubuntu itself?
[00:05] <aetaric> right
[00:06] <twb> How did you generate the netboot image -- live-build?
[00:07] <aetaric> i followed the guide
[00:08] <aetaric> ltsp-build-client
[00:08] <aetaric> from https://help.ubuntu.com/community/ThinClientHowto
[00:08] <aetaric> still works all the way up to 10.04
[00:09] <aetaric> it worked when my tftp and nfs server was also a dhcp server
[00:09] <twb> Ah, LTSP
[00:10] <twb> All I can suggest is that you 1) packet sniff and examine the bogus flow; and 2) try building a smaller image
[00:10] <aetaric> is there a newer/better way to go about this?
[00:10] <twb> You might be able to replace ISC dhclient with some other dhclient.
[00:11] <twb> aetaric: well, AFAICT the problem is that your client is sending something that Cisco have decided not to accept, so ultimately you have to either fix the Cisco or workaround it in the client
[00:11] <aetaric> kk.
[00:11] <aetaric> have another dhclient you recommend?
[00:12] <twb> Not really
[00:12] <twb> aptitude search ?tag(client)?tag(dhcp) or so
[00:13] <twb> http://paste.debian.net/102350/
[00:28] <Frenk_> whats the irc command to read the chanell topic eg /topic
[00:29] <Frenk_> read /topic
[00:29] <Frenk_> anyone suggestions?
[00:31] <Frenk_> [ if you're joining for the first time, or are new to irc, the first thing you'll want to do is read the channel topic (/topic) ] but how do i do that?
[00:33] <hallyn_> Frenk_: just typing '/topic' doesn't work for you?
[00:34] <hallyn_> Frenk_: which irc client are you using?
[00:34] <Frenk_> Insufficient arguments for command.
[00:34] <Frenk_> webchat.freenode.net
[00:35] <Frenk_> qwebirc
[00:35] <Frenk_> #topic
[00:35] <Frenk_> nope
[00:36] <shauno> it's often displayed at the top of the channel window, or when you first enter.  /topic is meant to return it tho  (obviously not your fault if it doesn't ..)
[00:37] <The_Tick> it's your fault for using a bad client
[00:37] <hallyn_> Frenk_: maybe /topic #channelname?
[00:48] <Frenk_>  
[00:48] <Frenk_>  !topic
[00:49] <twb> Just have him ask the vendor about it
[00:51] <aetaric> twb: does it when i remove isc, install a replacement, and rebuild the ramdisk
[00:51] <aetaric> so, it might be a kernel issue
[00:51] <twb> aetaric: 11:10 <twb> All I can suggest is that you 1) packet sniff and examine the bogus flow; and 2) try building a smaller image
[00:52] <aetaric> yeah
[00:52] <aetaric> define smaller image...
[00:52] <twb> less things
[00:52] <aetaric> it loads it via nfs :p
[00:53] <twb> So it gets as far as NFS?
[00:53] <aetaric> nope
[00:53] <aetaric> kernel can't get an address to connect to nfs
[00:53] <aetaric> i have it load the kernel and initrd.img via tftp
[00:54] <twb> So the problem is in the ramdisk
[00:57] <aetaric> ok, how do i change how the ramdisk gets a dhcp lease?
[00:59] <twb> You can't without a lot of fucking about
[00:59] <twb> It'll be easier to replace the cisco IMO
[01:00] <twb> Or at least call Cisco and ask them how to tell it to ignore option 61 or so
[01:07] <pting> is there a command to preseed user/group ids? i want to sync up all the system user/group ids
[01:07] <twb> pting: that's called "centralized authentication"
[01:07] <twb> Oh, sorry, *system* accounts
[01:08] <twb> I don't know any way to get those synced, other than 1) put them in your central auth (e.g. LDAP); or 2) ensure that you install packages in the same order on every host.
[01:08] <twb> Neither of which are very nice
[01:08] <pting> twb, i see... i was thinking of using puppet to pre-create all the groups and users before installing the packages
[01:09] <twb> I suppose that could work
[01:09] <pting> heh, not sure if that's the brightest of ideas though
[01:09] <twb> Note that purging packages will often try to delete the system account
[01:09] <pting> i would hate to install a package i forgot to create users for beforehand
[01:10] <twb> So if someone purges one package and installs another, the new one might get the UID that you wanted for your system account
[01:10] <twb> But that's probably an unlikely occurrence
[01:19] <pting> twb, thanks, i'll just have to pick one.. i'm most interested in ensuring services like mysql and www-data are the same to read from nfs shares and such
[01:21] <twb> pting: many of them are standardized in base-files
[01:22] <twb> Sorry, base-passwd
[01:23] <twb> www-data is, at least
[01:23] <pting> o.O, so i could add puppet code to append to /usr/share/base-passwd/passwd.master and grouip.master before installing packages?
[01:23] <twb> #define DEFAULT_PASSWD_MASTER	"/usr/share/base-passwd/passwd.master"
[01:23] <twb> #define DEFAULT_GROUP_MASTER	"/usr/share/base-passwd/group.master"
[01:24] <twb> Dunno, try it in your test lab :P
[01:24] <pting> o.O i definitely will, thanks
[01:33] <pting> twb, so what's the recommended ubuntian way of assigning manual user/group ids? start from the max id for the segment and go descending from there?
[01:34] <pting> like start from system user id 999 and go down?
[01:36] <twb> pting: update-passwd seems to apply to global IDs (0-99)
[01:37] <pting> twb, hah, no!!! that solution seems like the right way to go though
[02:01] <Noobster> is there any way to use apt-cache to tell you the version #?
[02:01] <pting> twb, looks like it still works for system user ids.. i added mysql as 999, both as uid/gid... ran update-passwd and it was added
[02:01] <Pici> Noobster: version of what? A package?
[02:01] <Noobster> yes
[02:02] <Pici> Noobster: apt-cache show packagename  has a Version line.
[02:02] <Noobster> thanx!
[02:03] <Pici> apt-cache policy will show you what versions exist in the repositories  you have in your sources.list
[02:05] <twb> pting: eeeexcelent
[02:08] <pting> twb, so it looks like update-passwd doesn't change uid/gids if they already exist
[02:08] <twb> That would be about right
[02:09] <twb> It's intended to set up defaults, not clobber users' changes to those defaults
[02:09] <pting> ya, that's probably for the best =)... i'll just write some checks to notify me if these user ids are out of sync on my servers
[03:28] <dbowlby> hello all!
[03:28] <dbowlby> I have a ubuntu server running KVM
[03:28] <dbowlby> I'm using bridging
[03:28] <dbowlby> for some reason, after setting it up, I can't communicate to the internet (through the default gateway)
[03:28] <dbowlby> http://pastebin.com/siC29Sgj
[03:29] <dbowlby> ^my netstat -r and cat of /etc/network/interfaces
[03:29] <dbowlby> my netstat -r and cat of /etc/network/interfaces is in the pastebin
[03:29] <twb> dbowlby: pastebin "ip r" format, please.
[03:29] <twb> For bonus points, all of "ip l; ip a; ip r'
[03:29] <dbowlby> http://pastebin.com/XMv3pU20
[03:30] <dbowlby> bonus points: http://pastebin.com/BM1gpnqP
[03:30] <twb> Well, the first problem is that you've configured BOTH br0 and eth0 as 192.168.1/24
[03:31] <twb> Meaning that on a per-packet basis, it'll pick either br0 or eth0 at random
[03:32] <dbowlby> ok, I can see that
[03:32] <dbowlby> so br0 should be a different network?
[03:32] <dbowlby> I want my VMs to be on the 192.168.1.0 network
[03:32] <twb> ifdown eth0; then remove "auto eth0" from interfaces
[03:32] <twb> If you're bridging across eth0, you shouldn't be referring to eth0 directly
[03:34] <dbowlby> if I bring eth0 down, I loose connectivity to the box
[03:34] <twb> Then remove "auto eth0" from interfaces and simply reboot
[03:34] <dbowlby> ok, let me try that
[03:35] <twb> Currently you're trying to run the same network both over raw eth0 and over the bridge br0
[03:35] <twb> You need to only have one of those -- probably the latter
[03:35] <dbowlby> lol probably?
[03:35] <twb> Note that (obviously) you'll need to connect to .10 instead of .230
[03:36] <dbowlby> as feared, now I can't access the box
[03:44] <dbowlby> twb, thanks, I finally got my console hooked up and found that the damn reboot failed
[03:44] <twb> Sorry
[03:45] <dbowlby> twb, now I can access my box and it appears to be working properly
[03:45] <dbowlby> twb, no I'm sorry hehe
[03:46] <dbowlby> this would definitely explain my connectivity issues to the VMs :)
[03:48] <twb> What I described is pretty much what *I* use
[03:48] <twb> Maybe your router (on .27) doesn't know how to get to .10?
[03:49] <twb> Or maybe there's another host on .10 and they're fighting
[03:49] <dbowlby> twb, nah it's working now
[03:50] <dbowlby> twb, I commented out the auto
[03:50] <dbowlby> on eth0
[03:51] <dbowlby> twb, the whole reason I was looking at this was because I was having insane ping times to my VMs
[03:51] <dbowlby> twb, then I noticed I couldn't get out through my gateway
[03:52] <twb> dbowlby: I bloody told you to comment out "auth eth0"
[03:52] <dbowlby> twb, after I took out that auto eth0 I could ping outside my gateway
[03:52] <dbowlby> twb, I know, and I did, but I rebooted and the box didn't come back up
[03:52] <dbowlby> twb, I didn't have a console, so didn't know why
[03:52] <dbowlby> twb, after I hooked up the console, I found the system at a crash screen
[03:52] <twb> Is this a server, or is NetworkManager installed?
[03:52] <dbowlby> twb, after rebooting it, it worked great :)
[03:53] <dbowlby> twb, it's ubuntu server with the virtualization stuff installed
[03:53] <twb> Please confirm that "dpkg -l network-manager" has "un" in the lefthand column
[03:55] <dbowlby> no packages found
[03:55] <twb> Good enough
[03:55] <twb> I don't know what's wrong
[03:55] <twb> I assume bridge-utils is installed?
[03:56] <dbowlby> it says ii
[03:56] <twb> What does "brctl show" have to say?
[03:56] <dbowlby> bridge name     bridge id               STP enabled     interfaces
[03:56] <dbowlby> br0             8000.406186e288dc       no              eth0
[03:56] <dbowlby>                                                         vnet0
[03:56] <dbowlby> virbr0          8000.000000000000       yes
[03:57] <dbowlby> my pings are still all over the place to this darn windows vm
[03:57] <dbowlby> even from the host itself
[03:57] <dbowlby> goes from 1ms all the way up to 400ms
[03:59] <twb> That might simply be because the VM is sleeping while idle
[03:59] <dbowlby> twb, a continuous ping?
[04:01] <dbowlby> seems to be better
[04:01] <dbowlby> aight, well, hehe, thanks a lot twb!
[04:02]  * twb shrugs
[04:04] <dbowlby> twb, I think you're on to something with the idle, don't really see the delay with my rhel vms
[04:32] <axisys> how do I install ubuntu server non-interactively ?
[04:34] <twb> grep sambashare /etc/group
[04:34] <twb> sambashare:x:111:cyber
[04:34] <twb> grep sambashare /var/lib/dpkg/info/*
[04:34] <twb> ...no hits.  What created this group?
[04:34] <axisys> https://help.ubuntu.com/10.04/installation-guide/amd64/appendix-preseed.html is it the recom way to auto install ubuntu ? or there is something easier ? may be some project in sf.net or freshmeat?
[04:35] <twb> axisys: that's the best way, yes
[04:35] <twb> axisys: FAI can hold your hand, but I don't like it much
[04:35] <axisys> twb: FAI ?
[04:36] <twb> apt-cache search ^fai-
[04:37] <axisys> twb: it is in maverick.. but not in lucid (LTS)
[04:38] <twb> Answer: the samba postinst
[08:11] <Noobster> I have started my tutorial on installing Samba3 as a PDC for Win7 on Ubuntu 10.04 x64 LTS. It is still in beta but if it helps anyone..  http://tinyurl.com/SMB4PDCWin7
[08:12] <Noobster> oops Samba4
[08:25] <twb> What is the difference between x, * and ! in the passwd and shadow databases?
[08:26] <twb> AFAICT x means "ask shadow", and ! and * in shadow are identical
[08:38] <\sh> twb: * <- disabled account x <- encrypted password in /etc/shadow (and eventually is ! another indicator with the same meaning of x)
[08:38] <\sh> s/disabled/deactivated/
[08:39] <twb> I'm a bit confused as to why some accounts are ! and some are *
[08:39] <twb> I *suspect* it's just a matter of some being there when "shadowconfig on" first runs, and some being created after -- but I don't want to assume that when my dodgy thing sets them all to *
[08:41] <\sh> twb: man 5 shadow ;) "If the password field contains some string that is not valid result of crypt(3), for instance ! or *, the user will not be able to use a unix password to log in, subject to pam(7)."
[08:41] <twb> I don't trust it
[08:41] <twb> For example passwd(8) locks using a ! explicitly
[08:42] <twb> So so passwd --unlock presumably works by removing the !
[08:42] <twb> These are all system accounts so it PROBABLY doesn't matter
[08:44] <twb> Yep
[08:44] <twb> adduser --disabled-password uses a *, but --disabled-login uses a !
[08:44] <twb> So obviously su and/or sshd treat them differently
[08:53] <beric> Hello. How can I get rid of that fancy grub menu when booting the installation CD? I'm using SOL and it clutters the console.
[08:53] <twb> beric: remaster it
[08:53] <twb> beric: and it's not a grub menu, it's isolinux
[08:54] <beric> thanks..
[08:58] <beric> well that's quite a lot of work :( , keep it simple people.
[08:59] <twb> You could just netboot
[08:59] <twb> optical media are so passé
[12:43] <pmatulis> hard to accept
[12:56] <pmatulis> couldn't confirm.  install went well
[13:29] <tgywa> When I do: apt-get install php5-xdebug I got: php5-xdebug: Depends: phpapi-20090626.. how do I fix it?
[14:04] <raubvogel> Doesn't ubuntu come with a default (mail) aliases file?
[14:20] <pmatulis> with the latest and greatest (10.10) do i really need to do 'sudo /etc/init.d/networking restart'?  seems nothing else works?
[14:41] <raubvogel> pmatulis, i think they are switching to the redhat-like "service some-application start/stop/restart"
[14:44] <soren> pmatulis: To achieve what?
[14:45] <pmatulis> soren: trick question?
[14:45] <soren> pmatulis: 14:20 < pmatulis> with the latest and greatest (10.10) do i really need to do 'sudo /etc/init.d/networking restart'?  seems nothing else works?
[14:45] <soren> pmatulis: No, no tricks involved :)
[14:45] <pmatulis> soren: yes, i want to restart networking
[14:46] <soren> pmatulis: If you want to restart networking, "really needing to do 'sudo /etc/init.d/networking restart'" seems perfectly reasonable, doesn't it?
[14:47] <pmatulis> soren: i just figured we might be using 'service' thingy by now but it's broken
[14:47] <soren> pmatulis: Oh, you prefer to do "service networking restart" instead?
[14:47] <soren> Is that what you're saying?
[14:47] <pmatulis> soren: right
[14:48]  * pmatulis is here b/c he is configuring a bridge (which is broken on bootup btw)
[14:48] <soren> pmatulis: Oh.
[14:48] <soren> pmatulis: Right, I see why it doesn't work.
[14:49] <soren> pmatulis: Yeah, that's probably what you're roing to have to do for now.
[14:49] <pmatulis> soren: ok
[14:49] <pmatulis> soren: any idea why bridge will not come up on boot?
[14:50] <soren> pmatulis: In a standard 10.10? Not sure.
[14:50] <soren> pmatulis: Can I see your /etc/network/interfaces?
[14:51] <pmatulis> soren: one moment
[14:51] <soren> sure
[14:54] <pmatulis> soren: http://pastebin.ca/2019364
[14:54] <patdk-wk> heh, networking hasn't been moved to upstart yet?
[14:54] <patdk-wk> seems not, but network-interface and network-interface-security have
[15:00] <soren> patdk-wk: Depends on what you mean by "networking".
[15:00] <patdk-wk> the /etc/init.d/networking script :)
[15:01] <soren> patdk-wk: That would not work. It's a shell script.
[15:02] <soren> pmatulis: What does "ifquery --list --allow auto" output?
[15:02] <soren> (with sudo)
[15:04] <pmatulis> soren: i presume i should run that when br0 is not working?
[15:04] <pmatulis> soren: b/c i just ran the networking script after boot (which does bring it up)
[15:04] <soren> pmatulis: I don't think it matters.
[15:05] <soren> With --list it just lists stuff.
[15:05] <pmatulis> soren: then 'lo, eth0, br0'
[15:05] <soren> ok
[15:07] <pmatulis> soren: btw, this is also broken in karmic (and probably lucid)
[15:07] <soren> It works for me, though.
[15:07] <pmatulis> soren: really?  on which release?
[15:08] <soren> Lucid and maverick, at least.
[15:08] <soren> Let me double check to be sure.
[15:08] <pmatulis> soren: interesting.  i wonder if the fact that my machine is a kvm guest
[15:08] <soren> Is this a server? Is network-manager installed?
[15:08] <pmatulis> soren: is to blame
[15:08] <pmatulis> soren: cli only, yeah
[15:08] <soren> kvm guest> Shouldn't matter.
[15:10] <pmatulis> soren: i have to resort to rc.local and then start all network-dependent services.  it's horrible  :(
[15:10] <soren> :(
[15:10] <soren> pmatulis: slangasek spent quite a bit of time getting this working.
[15:11] <soren> pmatulis: Succesfully, I believe.
[15:11] <soren> Hmm...
[15:11] <soren> I just tried a configuration similar to yours.
[15:11] <soren> Failed.
[15:11] <soren> Weird.
[15:11] <soren> Let me look at it for a couple of minutes.
[15:11] <pmatulis> soren: nice.  thanks
[15:12] <soren> Hehh... misspelled "static" :)
[15:12]  * soren reboots it aga in.
[15:13] <soren> pmatulis: Worked brilliantly.
[15:13] <soren> pmatulis: How did you install this kvm guest?
[15:13] <pmatulis> soren: i suck
[15:14] <pmatulis> soren: the karmic one was with the server iso.  this 10.10 one was with something called vmbuilder
[15:14] <soren> Never heard of it :)
[15:14]  * soren hides
[15:14] <pmatulis> soren: ha ha ha
[15:14] <soren> Well, it doesn't do anything that should affect this.
[15:14] <soren> ...that I can think of, anyways.
[15:15] <soren> It's kind of hard to debug from afar.
[15:15] <soren> I'd file a bug against bridge-utils.
[15:19] <pmatulis> soren: alright
[15:50] <cap_00> anyone know how to check user access logs for a samba file server? i'm trying to track down which of my users might have deleted a folder
[15:51] <Psi-Jack> Alright. So here's a question. For some reason, everytime I ifup/ifdown lo:X, which are VIPs for several clustered servers in 10.04.1, open-iscsi restarts. I need this reaction to stop. open-iscsi seems to be effected by anything in upstart resulting in the 'networking' event.
[15:52] <Psi-Jack> Looking at /etc/init.d/open-iscsi, Required-Stop has networking in it, but does upstart even care about lsb-init that way at all?
[15:52] <RoAkSoAx> it doesnt
[15:54] <Psi-Jack> Okay, so why is it open-iscsi is being effected by ifup/ifdown for loopback network devices? Everytime it does this knee-jerk reaction, it completely breaks my gfs2 /home mount that's on iscsi because it completely stops and restarts open-iscsi causing a stale and broken lock. Only fix for that is a total cluster restart.
[15:57] <Psi-Jack> Aha!
[15:57] <Psi-Jack> It's in /etc/network/if-down.d/open-iscsi
[15:58] <Psi-Jack> It ONLY checks if the interface is lo, and not lo:*
[16:02] <cap_00> good job
[16:02] <cap_00> ok, i found the logs, is there someway to search through multiple logs for a term?
[16:03] <_ruben> grep term log1 log2 log3
[16:03] <cap_00> ugggh.... can i search an entire dir?
[16:04] <Pici> grep 'term' /path/*
[16:05] <thomas_s> cap_00: add the -R option to grep if there are subdirs you need to search (and then, use the following command: "man grep" )
[16:05] <cap_00> sweet
[16:05] <cap_00> can i search for a term with a space?
[16:06] <Pici> cap_00: enclose your search term in quotes.
[16:07] <cap_00> hmmm guess it doesn't log who delete's something :(
[16:11] <cap_00> ok, so new question, how do i turn on better loggin for smb? i had a backup, but i'd like to know who deleted a folder on the fileserver
[16:12] <i0nic> what would be the best way to put 2 files into a archive to send to another server, tar?
[16:13] <i0nic> its just that tar makes 405byte files into a 10k archive
[16:14] <Daviey> JamesPage / hggdh: Is that hudson instance still supposed to be server iso only?
[16:14] <hggdh> Daviey: not really, it has already been overloaded ;-)
[16:14] <JamesPage> Daviey: no its gained a few friends...
[16:15] <JamesPage>  I'll do some work on the views to make it a bit more accessible.
[16:15] <Daviey> That is fine... might be a good idea to change the title :)
[16:15] <Daviey> JamesPage: Rocking ;0
[16:15] <JamesPage> Daviey: ack
[16:15] <Daviey> JamesPage: Grab me at some point this week, regarding me setting up a node.
[16:15] <JamesPage> OK; its pretty smooth now
[17:05] <wng-> So I had to add a few rules to UFW to allow an external client to connect to my server for a while, now im trying to remove the rules
[17:05] <wng-> is there a way to delete any rule that originates from a certain ip with ufw?
[17:10] <jdstrand> wng-: just put 'delete' in front of the rule you added. alternatively, you can use 'sudo ufw status numbered ; sudo ufw delete <number of rule to delete>
[17:11] <wng-> jdstrand: I know that, what I'm trying to do is delete all the rules at once, theres like 50 of them
[17:11] <jdstrand> wng-: no. either by the full rule or by the number
[17:11] <wng-> so I can't pass a range of numbers?
[17:11] <wng-> eh, i guess i can
[17:15] <njin> hello, can someone look at bug 579572 ? thanks
[17:16] <njin> At wich package can I assign this bug '
[17:16] <njin> ?
[17:22] <ScottK> njin: linux
[17:22] <ScottK> (kernel issue)
[17:23] <njin> ScottK: thanks
[17:23] <ScottK> You're welcome.
[17:34] <ne7work> hello all please someone tell me from where can I start to learing how to make programs and witch languages I need or only c++?
[17:47] <datz> ne7work: c++ wouldn't hurt
[17:58] <smoser> ne7work, if you're just learning, you might find some languages more easy to get started in than c++.
[17:59] <datz> maybe start with some scripting languages
[17:59] <smoser> ubuntu heavily favors python for many things.  http://wiki.python.org/moin/BeginnersGuide/NonProgrammers has lots of pointers.
[18:24] <SpamapS> woot new version of upstart!
[18:31] <pmatulis> SpamapS: anything for server peoples to be excited about?  :)
[18:33] <zul> kirkland: how did you get it to work?
[18:33] <zul> kirkland: the iso importing
[18:34] <kirkland> zul: Fedora "just worked", as soon as I gave it a full DVD ISO
[18:34] <kirkland> zul: i was trying to import the Fedora netinstall yesterday
[18:34] <kirkland> zul: which did not work
[18:34] <zul> kirkland: thats no fun
[18:34] <kirkland> zul: over night, i downloaded the DVD
[18:34] <kirkland> zul: yeah, I know :-)
[18:34] <kirkland> zul: I just fixed wake-on-lan
[18:34] <zul> kirkland: through the cli?
[18:34] <kirkland> zul: that was *slightly* more fun
[18:35] <kirkland> zul: what about the cli?
[18:35] <zul> kirkland: you used the cli to import the cd?
[18:36] <kirkland> zul: no, the web interface
[18:36] <zul> kirkland: what if you do it through the cli
[18:36] <kirkland> zul: dunno;  you tell me :-P
[18:36] <zul> hmmm...ok
[18:37] <kirkland> zul: i'll try it here, if you give me the CLI invocation
[18:43] <kirkland> zul: ?
[18:43] <zul> kirkland: looking around
[18:43] <kirkland> zul: k
[18:43] <zul> ill let you know
[18:47] <smoser> SpamapS, when did you think your patch pilot session started?
[18:53] <pmatulis> can i have a cron job PATH modified by PAM environment settings (/etc/security/pam_env.conf)?  the crontab manpage says cron supports pam_env but it doesn't have any effect
[18:53] <pmatulis> http://manpages.ubuntu.com/manpages/maverick/man5/crontab.5.html
[18:53] <pmatulis> (confusing)
[18:54] <SpamapS> smoser: noon PST
[18:54] <SpamapS> smoser: which is in 1 hour ;)
[18:54] <SpamapS> pmatulis: debug stanza is pretty cool :)
[18:55] <SpamapS> pmatulis: other than that, I'm not sure
[18:55] <smoser> SpamapS, believe it or not, i wasn't actually being a jerk.
[18:55] <pmatulis> SpamapS: halleluiah (spelling?)
[18:55] <smoser> given different time zones, who knows what a half day is.  ie, mine started 8:30 is US/Eastern, it is really past.
[18:55] <SpamapS> smoser: we've discussed this before.. its much easier for you to tell us when you're not being a jerk ;)
[18:55] <smoser> yeah i know.
[18:56] <smoser> well i guess you should jsut assume jerk unless you hear other wise.
[18:57] <axisys> is the journaling added in ext4 or is it also available on ext3 ?
[19:01] <patdk-wk> journaling exists in ext3
[19:01] <patdk-wk> but extents only exist in ext4 and make it incompatable with ext2/3
[19:01] <patdk-wk> if you use them
[19:45] <LewisCawte> How do I configure my server install to connect to my Wireless network?
[19:45] <MrPicard> Hello, how would i be able to install LAMP and PHPmyadmin in one command?
[19:47] <patdk-wk> LewisCawte, edit /etc/network/interfaces
[19:49] <LewisCawte> patdk-lap: whats the config for it?
[19:50] <patdk-wk> no idea, I would never touch wireless on a server
[19:51] <LewisCawte> I'm more than likely going to keep it as what I'm using it for now, but I'm testing stuff in my room, so I don't want to move it closer to the router for now
[19:57] <zul> kirkland: ping can you do a small text file on how to on what you have done so far? i need to know how to confgiure the web interface
[19:58] <kirkland> zul: sudo apt-get install cobbler
[19:58] <kirkland> zul: sudo vi /etc/cobbler/modules.conf
[19:58] <kirkland> module = authn_testing
[19:59] <kirkland> zul: sudo stop cobbler
[19:59] <kirkland> zul: sudo killall cobblerd
[19:59] <kirkland> (there's a bug in the upstartization, all cobblerd aren't killed)
[19:59] <kirkland> (rather, 2 are started, not sure how/why)
[19:59] <LewisCawte> anybody have any idea of the config for wireless (USB adapter) on a Ubuntu Server install?
[19:59] <kirkland> zul: sudo start cobbler
[19:59] <zul> thanks
[19:59] <kirkland> zul: sudo service apache2 restart
[19:59] <kirkland> zul: then point browser to http://hostname/cobbler_web
[20:00] <kirkland> zul: login with testing/testing
[20:00] <kirkland> zul: daz it
[20:00] <kirkland> zul: but yeah, we should wiki up this schtuff
[20:06] <fluvvell> in server 10.04, what would all the ksoftirqd/n watchdog/n migration/n ksoftirqd/n  processes be for?   I have n=0-7 of each of these
[20:09] <SpamapS> kirkland: want me to fix the upstart job?
[20:09] <kirkland> SpamapS: please!  do you know what's wrong?
[20:10] <SpamapS> kirkland: yeah, expect fork is hard to get right... just start with -F and remove expect fork
[20:10] <kirkland> SpamapS: ah, sure, go for it
[20:10] <SpamapS> oh there's no expect fork
[20:11] <SpamapS> well thats actually the problem then ;)
[20:11] <SpamapS> but...
[20:11] <SpamapS> running with -F is better anyway
[20:11] <SpamapS> let me read the daemon.. if it listens then forks, expect fork can actually be better
[20:14] <SpamapS> yeah, listens way after forking so expect fork is not helpful
[20:14]  * SpamapS pushes
[20:22] <SpamapS> kirkland: I think I'm going to propose a lintian check for upstart jobs.. not having a 'stop on' is very dangerous.
[20:22] <kirkland> SpamapS: okay
[20:22] <SpamapS> kirkland: even 'stop on runlevel [016]' still has a race condition tho. :(
[20:32] <RoAkSoAx> kirkland: btw.. is testdrive working now? (haven't had time to test it with the moving and all)
[20:32] <kirkland> RoAkSoAx: yup, thanks!
[20:32] <RoAkSoAx> kirkland: np ;)
[20:33] <kirkland> RoAkSoAx: when do you think powernap will be ready for natty?
[20:41] <stanman> hi, i've got an openvz box running hardy and simias/ifolder. I can't login to its webadmin and receive a: System.ObjectDisposedException in my ifolder.log. Anyone had this or knows a solution?
[20:44] <cap_00> this might be a weird question, are partitions the only thing that can be mounted?
[20:45] <cap_00> if i wanted to add a HD to my system and mount it to two different places in the system the only way i'd do that would be creating two seperate partitions?
[20:46] <RoAkSoAx> kirkland: I'm rescheduling everything right now (because of moving, graduation, etc etc). But I'm gonna start working in full with PowerNap this week, to hopefully finish modifying the algorithm for the integration of the Monitors.
[20:47] <RoAkSoAx> kirkland: before Xmas, which should also involve the monitors functional
[20:47] <cap_00> i just don't want to have to determine ahead of time how big each partition is on the drive, i'd rather have both mount points using the full space and let them fill up the drive at the same time
[20:48] <SpamapS> cap_00: no you can mount file system images
[20:48] <cap_00> hmmm and those can grow?
[20:49] <SpamapS> cap_00: you can also resize partitions
[20:49] <cap_00> yes, but i'd like to make it as easy to manage as possible
[20:50] <SpamapS> cap_00: once we have BTRFS, that will be easy. Until then.. partitioning everything onto one big / is probably still the simplest answer.
[20:50] <SpamapS> cap_00: you can use LVM to manage the physical disks if you do that
[20:50] <SpamapS> cap_00: with LVM you just ad the physical space to the pool, and then extend the FS onto it as needed.
[20:50] <cap_00> hmmm
[20:50] <SpamapS> cap_00: but, shrinking the FS is hard
[20:52] <SpamapS> cap_00: look into LVM.. I think it will do what you want.. but just know that shrinking one logical volume, and the filesystem that rides on top of it, is not simple.
[20:52] <SpamapS> cap_00: though it can be done
[20:57] <SpamapS> cap_00: you're saying you want to build a single RAID1 and then mount two different filesystems on top of that, right?
[20:57] <cap_00> ya
[20:57] <SpamapS> cap_00: lvm does that very easily
[20:58] <SpamapS> cap_00: you partition the whole disk to lvm.. then do  pvcreate /dev/sdXX ; vgextend name-of-volume-group /dev/sdXX ; lvcreate -n foo -L 500G name-of-volume-group ; mkfs.ext4 /dev/name-of-volume-group/foo
[20:58] <SpamapS> cap_00: and then to make another LV..   lvcreate -n bar -L 300G  ... get it?
[20:59] <SpamapS> cap_00:  you can even use it for migrating data between physical disks w/o downtime.
[21:07] <blistov> Waiting for br0 to get ready (MAXWAIT is 32 seconds).  What does this mean and why did am I getting it?
[21:07] <blistov> If I bring up the bridge manually, it is instantaneous.  If I let init.d/networking handle it, it takes about 20 seconds so most of my dependent service fail to start.
[21:35] <kirkland> RoAkSoAx: okay, i just bought a second watt meter, so i can help with testing
[21:45] <blistov> Ok, I'm going crazy now.  br0 is brought up by init.d/networking but hangs on "Waiting for br0 to get ready".  Then when I do a networking restart, br0 can't be added because it already exists.
[21:45] <blistov> So br0 isn't coming up properly, not going down properly.
[21:45] <blistov> standard bridge setup.
[21:46] <RoAkSoAx> kirkland: cool ;)
[21:51] <hallyn_> blistov: can you pastebin your /etc/networking/interfaces ?
[21:53] <blistov> hallyn_, http://pastebin.ca/2019758
[21:53] <timholum1> does anyone know if it is possible to install ubuntu cloud on a single server ( combine the front end with a node ) ?
[21:53] <blistov> hallyn_, this was all working yesterday, and I can not for the life of me figure out what's wrong now.  Built a new server just to test, and same issue.
[21:56] <RoAkSoAx> kirkland: did you get one of those fancy ones that can pass the data to a machine?
[21:57] <kirkland> RoAkSoAx: nah, just the cheapy $20 one
[21:57] <blistov> hallyn_, and when i restart networking i get this:  http://pastebin.ca/2019766
[21:57] <guntbert> blistov: I never used that kind of setup, but I'm surprised that you should be able to bridge to eth0 while it is down...
[21:57] <blistov> guntbert, this is the setup according to https://help.ubuntu.com/community/OpenVPN
[21:57] <blistov> And I'm sure I had this working yesterday based on this config.
[21:58] <blistov> Yes, I'm sure something must have changed, but I can't for the life of me figure out what.
[21:58] <blistov> I'm getting the nerd rage fog.
[21:58] <RoAkSoAx> kirkland: oh, I was thinking if at the end we can gather the data and produce charts to kinda "sell" PowerNap by providing power saving charts
[21:59] <guntbert> blistov: did you look at https://help.ubuntu.com/10.04/serverguide/C/openvpn.html ?
[21:59] <hallyn_> blistov: how about if you just 'ifdown br0; ifup br0' ?
[21:59] <kirkland> RoAkSoAx: yeah, would be cool
[22:00] <blistov> hallyn_, like I said, if i do it manually, it all works.
[22:00] <blistov> but something is screwy.
[22:00] <hallyn_> blistov: oh, yeah i think i've had that happen about 6 months ago
[22:00] <hallyn_> hm
[22:00] <blistov> The fact that it takes 20 seconds to come up when using init.d/networking but goes immediately if i do it manually, is suspicious
[22:02] <blistov> Gah.
[22:02] <blistov> So many cheap monitors I could throw right now.
[22:02] <blistov> Damn these social conventions!
[22:08] <m4xx> ntpdate keeps failing, i've verified that port 123 is unblocked and i am able to ping the ntp servers, yet it keeps returning no server suitable for synchronization found
[22:08] <m4xx> any ideas?
[22:11] <kirkland> RoAkSoAx: hmm, i just launched a test drive vm and it's only defaulting to 384M of memory
[22:11] <kirkland> RoAkSoAx: i have 8G in my laptop
[22:12] <kirkland> RoAkSoAx: i used to have smarter logic in there that chose the amount of memory to give the guest a little more dynamically
[22:16] <hallyn_> blistov: could you open a bug about that?  i think it needs to be investigated.
[22:16] <hallyn_> (sorry, had another 'outage' due to window manager rioting and striking)
[22:17] <patdk-wk> m4xx, try, ntpdate -qu pool.ntp.org
[22:18] <m4xx> "no server suitable for synchronization found"
[22:18] <patdk-wk> you defently have something blocking ntp or udp
[22:19] <patdk-wk> nslookup pool.ntp.org :)
[22:20] <m4xx> hah! udp was it ;]
[22:20] <m4xx> ty
[22:20] <m4xx> i was smart enough to not unblock 123 udp =x
[22:20] <dany> hello everyone!. Can someone tell me if there is an active directory server alternative that runs on linux?
[22:21] <dany> is LDAP a good replacement ?
[22:21] <patdk-wk> dany, windows 2008 in kvm? :)
[22:21] <patdk-wk> samba 4 is suppost to do AD I hear
[22:21] <patdk-wk> and AD is much much more than just ldap
[22:21] <m4xx> perhaps i can hit you with this one too? i've installed ubuntu server 10.04 then installed gnome-core. how can i add "extract here" to the right click menu? I thought it would be included with archive manager or file-roller but neither did the trick
[22:22] <patdk-wk> m4xx ask #ubuntu, no gui here :)
[22:22] <dany> ok ok
[22:22] <m4xx> any idea how i can do this?
[22:23] <dany> but i just want a sort of directory of users so that the mailboxes and shares and all that willl be automatically loaded
[22:23] <RoAkSoAx> kirkland: do you have any config files on which you are hardcoding the default MEM?
[22:23] <dany> is there a way to get that done?
[22:24] <RoAkSoAx> kirkland: cause the logic is still there
[22:24] <kirkland> RoAkSoAx: yes, /etc/testdriverc is hardcoding it
[22:24] <kirkland> RoAkSoAx: maybe we just need to comment that out
[22:24] <kirkland> RoAkSoAx: let me test
[22:24] <RoAkSoAx> kirkland: ok :)
[22:25] <dany> i just want to throw the bulky server 2003 in the bin
[22:25] <kirkland> RoAkSoAx: nope
[22:25] <kirkland> qemu: invalid ram size: -smp
[22:25] <kirkland> RoAkSoAx: okay, that needs a bug :-)
[22:25] <RoAkSoAx> kirkland: indeed ;)
[22:26] <qman__> dany, there isn't a drop-in AD replacement as of now
[22:27] <qman__> I don't know how far along samba4 is, but that's the closest thing
[22:28] <dany> ok thank you qman
[22:29] <dany> so Fedora Directory Server, OpenDS, or  Apache Directory Server  will not do ?
[22:30] <fluvvell>  what causes a SMB PACKET: SMBreadX (REQUEST), a client with a new 10.04 server and their old smb.conf file  has a bunch of these appearing on a tcpdump
[22:30] <fluvvell> Followed closely by SMB PACKET: SMBreadX (REPLY), then the whole thing repeats
[22:38] <i0nic> anyone seen this before, "agent admitted failure to sign using the key"
[22:40] <blistov> Why does openvpn and dhcp3-server try to come up before networking has started?
[22:40] <blistov> Is there any way to force them to wait for the interface to come up?
[22:43] <RoAkSoAx> kirkland: this is the code right? :http://pastebin.ubuntu.com/543829/
[22:43] <kirkland> RoAkSoAx: yeah
[22:44] <kirkland> RoAkSoAx: so MEM must be getting set to 384 somewhere else in the code
[22:44] <kirkland> RoAkSoAx: before the len(MEM)
[22:44] <RoAkSoAx> kirkland: apparently i removed the code... which seems extrange cause I remember a related issue when determing memory automatically... I guess I must have removed it... I;ll get it back
[22:46] <fluvvell> i0nic, ssh-agent ?
[22:47] <pting> is there a way to check a directory's permission to ensure it's 000 without first unmounting it?
[22:50] <RoAkSoAx> kirkland: did you report a bug yet?
[22:50] <RoAkSoAx> kirkland: neverming just got it
[22:50] <kirkland> RoAkSoAx: yup
[22:52] <RoAkSoAx> kirkland: fixed in the trunk ;) thansk for reporting bugs sir!! I think you are the biggest bug reporter :P
[22:52] <kirkland> RoAkSoAx: hehe
[22:53] <qman__> dany, those will do fine for linux clients, but windows, not so much
[22:53] <kirkland> RoAkSoAx: also, while you're in there, add a stanza in that if-block to set MEM=1024 if total memory > 2000000
[22:53] <qman__> the only system going for full windows compatibility is samba4
[22:58] <lenios> pting, what do you mean without first unmounting it?
[22:59] <lenios> ls -l should show it
[22:59] <RoAkSoAx> kirkland: done! pushed to trunk already
[22:59] <kirkland> RoAkSoAx: sweet;   you gonna upload to natty?
[23:01] <RoAkSoAx> kirkland: I could do it now if you like?
[23:01] <kirkland> RoAkSoAx: sure, i like bitesize uploads ;-)
[23:01] <kirkland> RoAkSoAx: makes it easier to spot regressions
[23:02] <clayd> can someone point me in the direction to som documentation explaining server set up for multi host, multi client web servers?  Like laughing squic, or one of the other big hosting groups would use.
[23:03] <RoAkSoAx> kirkland: indeed! Ok then, will do so now! otherwise it will prolly be sitting there till I finish with the improvements and GRUB thingy
[23:05] <pting> lenios, i want to ensure a directory is empty and the permissions on it is set to 000 before mounting; however, if it's already mounted, i want to check to make sure the permissions on it is 000 without having to unmount it first... if that makes sense
[23:06] <lenios> ls -l should do it
[23:07] <pting> the intent is to make sure the directory isn't writable unless it's mounted
[23:11] <RoAkSoAx> kirkland: done
[23:14] <lenios> unless it's mounted?
[23:14] <patdk-wk> normally easy
[23:14] <patdk-wk> umount, chown/chmod the path
[23:14] <patdk-wk> mount it
[23:14] <patdk-wk> Ive done that on a few systems
[23:14] <lenios> what do you chmod if it's unmounted?
[23:14] <patdk-wk> the path you would mount it to
[23:15] <patdk-wk> whole paths and crap can exist under a mount
[23:15] <patdk-wk> normally not accesable
[23:15] <lenios> i see
[23:15] <lenios> it's the same if it's mounted
[23:15] <patdk-wk> like, for an encrypted homedir can work like that on my laptop
[23:16] <patdk-wk> I have my .ssh path for login keys on it, but if I mount my home dir, it overlays it
[23:16] <patdk-wk> and I get my ssh private keys and everything else
[23:16] <patdk-wk> maybe it won't work on the mount directory itself
[23:17] <patdk-wk> but only subdirs of the mount dir
[23:50] <SpamapS> smoser: still around?