[00:04] <Bookman> Is there a LPI exam prep available online that is free/low cost?
[01:46] <ae> ubuntu hardy is currently using mysql server 5.0.x, intrepid+ are using 5.1.x - will hardy ever move up to 5.1? how can I find out about these policies?
[01:50] <thiagocrepaldi> hello, dears, i just installed slapd (openldap server) and executed sudo dpkg-reconfigure slapd to configure ldap server. it's done and working. However, i can't find slapd.conf anywhere. Where does ubuntu store it ? I am using a brand new ubuntu 9.04 server edition
[01:51] <mathiaz> thiagocrepaldi: 9.04 doesn't use slapd.conf. It uses the cn=config backend by default instead.
[01:51] <mathiaz> ae: hardy will probably never move up to 5.1.
[01:52] <mathiaz> ae: you could try to get a backport at https://help.ubuntu.com/community/UbuntuBackports.
[01:52] <ae> mathiaz: Thanks - I thought that might be the case but couldn't be sure.
[01:52] <docko> i have a software raid1, a fail event has been detected on /dev/sdb1 and /dev/sdb2. then /dev/sdb just disappeared, how is it possible?
[01:52] <ae> mathiaz: ahhh backports... good call
[01:53] <thiagocrepaldi> mathiaz, how can i see all configguration now ?
[01:53] <mathiaz> thiagocrepaldi: do a ldapsearch on the cn=config tree
[01:54] <mathiaz> thiagocrepaldi: https://help.ubuntu.com/9.04/serverguide/C/openldap-server.html
[01:55] <thiagocrepaldi> mathiaz, i am in this link already, but it's configuration to authenticate my console with ldap cited there do not work
[02:03] <thiagocrepaldi> additionally,  ldapscripts throws errors when i try to execute them. It's logs are too superficial, so it's not useful to debug =/
[02:47] <fbc-mx> My hosting company is giving me a choice of quad core opteron or xeon. Which would give me the best performance under ubuntu server?
[02:49] <KurtKraut> fbc-mx, if anyone that offer you a answer, it will be based on pure witchcraft.
[02:49] <KurtKraut> fbc-mx, I tend to prefer the processor with the biggest cache built in.
[02:50] <fbc-mx> KurtKraut, Intel Xeon DP E5405 or AMD Opteron 2344 HE?
[02:51] <KurtKraut> fbc-mx, I suggest you reading about them on Wikipedia.
[02:52] <fbc-mx> KurtKraut,  I have however the articles do not tell me about performance under ubuntu.
[02:55] <KurtKraut> fbc-mx, there is nothing specifically on Ubuntu that would differ performance on those processors.
[02:58] <fbc-mx> KurtKraut, ok the wikipedia article tells me that the xeon has 2x6mb L2 cache and the AMD has L1-Cache: 64 + 64 KB (Data + Instructions) per core L2-Cache: 512 KB, fullspeed per core L3-Cache: 2048 KB, shared.  If I had to decide just by L2 cache it would appear the xeon has more, but the way they break it out makes it confusing. Which on would you pick?
[03:00] <qman__> fbc-mx, if the Xeon is i7-based (Nehalem) it will outperform the Opteron
[03:00] <qman__> otherwise, they will be fairly similar in performance per clock
[03:01] <qman__> historically, Xeons usually have the performance edge, where Opterons have the power efficiency edge
[03:01] <qman__> though there are plenty of exceptions
[03:03] <fbc-mx> qman__,  it says 5400 "harpertown" series . does nto appear to be nehalem series. So if it's not nahelem based the AMD should outperform right?
[03:03] <qman__> no, they will be very similar in performance
[03:03] <qman__> the nehalems are the only ones with any significant performance advantage
[03:03] <qman__> go with number of cores, amount of cache, and clock speed as your judges in this case
[03:04] <fbc-mx> qman__, ok, then since they are similar in performance then my only consideration would fall back on compatability.. opteron or xeon more compatible with ubuntu?
[03:04] <qman__> they're both equally compatible
[03:05] <fbc-mx> qman__,  wow.. hard decision to make then... it sounds like a flip of a coin situation.
[03:05] <qman__> pretty much
[03:06] <qman__> I'd go with the opteron, but that's just because I'm an AMD fanboy, there's not an obvious choice
[03:07] <qman__> one probably is faster than the other, but not enough that you'll notice
[03:07] <KurtKraut> fbc-mx, ... remember what I said about witchcraft?
[03:07] <KurtKraut> :P
[03:07] <qman__> they're definitely in the same bracket
[03:07] <qman__> normally the deciding factor would be power consumption and cost
[03:07] <fbc-mx> hmm.. thanks guys... it appeared to be an exercise in futility
[03:08] <qman__> but since you've alluded that those are irrelevant, that makes the choice harder
[03:10] <fbc-mx> Heads XEON, tails Opteron.
[03:10] <fbc-mx> OK, it's XEON.
[03:12] <twb> i7 would make me nervous for a Linux system right now
[03:13] <twb> Since I haven't actually seen a Linux system handling e.g. the PCIE replacement
[03:23] <fbc-mx> twb, well, its for a dedicated server colocated server.. I'm not gonna be putting cards into it. Just using the stuff on the motherboard.
[03:23] <twb> fbc-mx: PCIE is what the previous generation uses between the CPU and the I/O controller.
[03:23] <twb> fbc-mx: in i7, it has been replaced with a new thing (quicksomething?)
[03:24] <twb> QuickPath Interconnect
[03:26] <twb> I mean, maybe the kernel doesn't need to understand such things, and they will Just Work.  But I'd still want to SEE it working before I laid down a couple grand for an i7 server.
[04:45] <amites> Any idea what I'm missing to access a directory? The steps I'm taking are add user to group and set directory permissions? http://pastebin.com/m2f2b58bf
[04:45] <jmarsden> amites: as I just said in #ubuntu, log out and back in so your user is really in the www-data group
[04:46] <jmarsden> But why are you doing that in the first place?
[04:46] <jmarsden> You shouldn't need to chmod the public_html directory for it to work...
[04:47] <amites> I logged out and logged back in - working now, thank you
[04:47] <qman__> your own public_html shouldn't need to belong to www-data
[04:48] <qman__> it should belong to the user
[04:48] <amites> apache runs as www-data
[04:48] <qman__> yes
[04:48] <amites> I use that dir to host all the sites on the server
[04:48] <amites> not just personal
[04:49] <qman__> standard practice is to put non-personal sites in /var/www/site-name
[04:49] <qman__> putting them in your own public_html is generally a bad idea
[04:50] <qman__> because that inherently breaks the automagic user site
[05:38] <NCommander> Does anyone know if Ubuntu supports dhcpv6 out of the box?
[05:55] <twb> What is DHCPv6?
[05:55] <twb> Do you mean DHCP over IPv6?
[05:56] <twb> Ah, apparently "DHCPv6" is a recognized term for that.
[05:57] <twb> Debian has an ongoing goal to support IPv6 throughout, so Ubuntu should have/get DHCPv6.  I don't know whether it's supported in any existing releases.
[05:58] <twb> http://packages.ubuntu.com/dhcpv6 indicates that WIDE DHCPv6 has been in Ubuntu for some time.
[06:06] <jmarsden> twb: rmadison shows it as being present only as a source package, though, which seems odd.
[06:07] <twb> Shrug
[06:07] <twb> I've never tried rmadison with Ubuntu
[06:07] <twb> Maybe universe packages aren't listed?
[06:08] <twb> jmarsden: nope, it's simply that the wide-dhcpv6 package doesn't produce a binary of that name.
[06:08] <jmarsden> Ah, OK.
[08:33] <error404notfound> using "mail" command, how can i send out an email with a tgz attached?
[08:41] <twb> You need something that can talk MIME
[08:41] <twb> I don't know if mail can.
[08:41] <jmarsden> error404notfound: It would be easier to install mpack and use that
[08:41] <jmarsden> mpack file.tar.gz user@example.com
[08:41] <error404notfound> hmmm, twb yup, MIME...
[08:41] <error404notfound> jmarsden, i see....
[08:41] <error404notfound> lemme check
[08:42] <jmarsden> sudo apt-get install mpack    # then  mpack file.tgz user@example.com    should work
[08:42] <twb> I still need to debug MIME attachments in my mbox.py plugin for roundup...
[08:43] <_ruben> i tend to use mutt instead of mail when i need to send attachments
[08:43] <twb> _ruben: even from a cron job?
[08:43] <_ruben> yes
[08:43] <twb> Hm.
[08:44] <_ruben> its pretty much a drop-in replacement for nail, which is a mail variant that does support attachments .. -a filename
[08:44] <twb> Yeah, I just never considered that mutt would work without a pty
[08:44] <_ruben> echo blaat | mutt -s subject -a file to@some.one
[08:44] <twb> If SHELL=/bin/bash, you can <<<blaat, btw
[08:44] <twb> herestrings ftw
[08:45] <error404notfound> i get sendmail: fatal: root(0): message file too big :D
[08:45] <jmarsden> So try it with a smaller file :)
[08:45] <_ruben> mail isnt best suited for transfering large files
[08:45] <twb> A nickel says error404notfound is doing this from PHP
[08:46] <error404notfound> twb, nope, commandline
[08:46] <twb> Darn
[08:46] <JanC> you can always try telnet/nc  :P
[08:46] <twb> socat!
[08:46] <error404notfound> twb, postdrop: warning: uid=0: File too large
[08:46] <error404notfound> see, sending file from root account
[08:46] <twb> I'm usually lazy and use tar c | ssh tar x instead of nc
[08:47] <twb> Who cares if it takes a couple more days
[08:49] <JanC> error404notfound: that sounds like the mail is too large for your postfix settings ?
[08:49] <error404notfound> JanC, lemme see...
[08:49] <error404notfound> anyone rememer how much large attachments gmail support? :P
[08:50] <error404notfound> got set one value which would last for ages
[08:51] <JanC> error404notfound: it looks like 'postdrop' issued the error/warning, not gmail?
[08:51] <error404notfound> JanC, yes, but i wana make both allowed attachment sizes equal as i use postfix +sasl + gmail
[08:59]  * soren goes for coffee
[10:09] <Jeeves_> Hey, where did kvm-pxe in hardy go?
[10:28] <soren> Jeeves_: It was never there.
[10:29] <soren> etherboot (5.4.3+dfsg-0.1ubuntu2) intrepid; urgency=low * Add kvm-pxe package containing the PXE ROM's kvm needs to pxeboot. -- Soren Hansen < soren@ubuntu.com>   Fri, 04 Jul 2008 15:22:12 +0200
[10:31] <Jeeves_> soren: ok, etherboot it is :)
[12:17] <soren> Ng: Ok, test building new Eucalyptus snapshot. If it works, I'll upload. ETA 15 in.
[12:17] <soren> min, even.
[12:18] <Ng> soren: excellent :)
[12:37] <zul> morning
[12:53] <heath|work> I'm having to type my password twice in order to sudo. This didn't start happening until I installed and configured an LDAP server on the system. Any ideas as to what might be happening?
[12:54] <pmatulis> heath|work: sounds like your PAM config is not optimized
[12:55] <heath|work> hmm. would that be a common-auth issue?
[12:56] <pmatulis> heath|work: look inside /etc/pam.d/sudo
[13:01] <heath|work> ah that would make sense. I didn't even see that one
[13:03] <heath|work> lol... well that says @include common-auth @include common-account
[13:03] <pmatulis> heath|work: yup
[13:04] <pmatulis> heath|work: maybe pastebin common-auth
[13:04]  * soren takes a break
[13:05] <heath|work> pmatulis, http://pastie.org/617305
[13:05] <soren> Ng: Eucalyptus is building, by the way: https://edge.launchpad.net/ubuntu/+source/eucalyptus/1.6~bzr746-0ubuntu1
[13:06] <pmatulis> heath|work: did you edit this file by hand?
[13:06] <heath|work> common-account is a single line: account required	pam_unix.so
[13:06] <heath|work> pmatulis, no virtualmin
[13:06] <heath|work> is it jacked up?
[13:07] <Ng> soren: cool thanks, I'll watch that
[13:07] <pmatulis> heath|work: yeah
[13:07] <heath|work> oh great.
[13:12] <pmatulis> heath|work: to begin, common-auth should only have lines beginning with auth
[13:12] <pmatulis> someone else may want to corroborate
[13:15] <pmatulis> heath|work: having to enter your p/w twice may indicate that authentication with ldap is failing, you probably don't want to use ldap with sudo right?
[13:15] <heath|work> pmatulis, correct. The only concern there is making sure virtualmin can still perform system tasks
[13:17] <heath|work> This looks very understandable. I will backup what I have and start here: https://help.ubuntu.com/community/LDAPClientAuthentication
[13:17] <pmatulis> heath|work: ok
[13:17] <heath|work> pmatulis, thanks for the help
[13:19] <pmatulis> heath|work: i haven't tried it myself but it seems you could just take the auth lines from common-auth (minus the ldap line) and replace the include common-auth line with them
[13:19] <pmatulis> in /etc/pam.d/sudo
[13:20] <pmatulis> heath|work:
[13:21] <heath|work> I can try that first. It looks like virtualmin just dumps everything in the common-auth file instead of breaking it up across the normal three
[13:21] <pmatulis> heath|work: note that you should have a root shell in another terminal when you're messing with pam to prevent locking yourself out
[13:22] <heath|work> will do.
[13:41]  * soren dogfoods and does an upgrade
[14:03] <heath|work> pmatulis, after cleaning everything up following the guide it appears all is well
[14:04] <pmatulis> heath|work: can you paste your /etc/pam.d/{sudo,common-auth} again
[14:07] <heath|work> pmatulis, I'm going to put the sudo back to the orig next and check it out: http://pastie.org/617355
[14:08] <heath|work> I'm not sure why my password one has anything about smb in there, but I left it
[14:08] <pmatulis> heath|work: you must have samba installed
[14:08] <heath|work> or auth for that matter
[14:09] <heath|work> That's just it, I don't. Not even the client
[14:09] <pmatulis> heath|work: maybe virtualmin up to its tricks again?
[14:11] <heath|work> I guess so. I choose virtual cause from what I read it doesn't jack up your configs with proprietary crap. You can uninstall virtualmin and still have a working system. (unlike others I have tried)
[14:11] <pmatulis> heath|work: ok.  your pam configs look good to me
[14:12] <heath|work> and sudo is cooperating! :)
[14:12] <heath|work> pmatulis, thanks again for the help.
[14:12] <pmatulis> heath|work: np, pass it on
[14:12] <heath|work> that's my motto. Pay It Forward
[14:14] <heath|work> It even resoled not being able to log in virtualmin as administrator! I had to enable the root account (which I hate) in order to use virtualmin.
[14:15] <pmatulis> heath|work: probably b/c that user is not in ldap
[14:16] <heath|work> yep
[14:16] <heath|work> so to disable the root password again: sudo passwd -d root ?
[14:16] <pmatulis> heath|work: doesn't look right
[14:17] <heath|work> from the man page: 'Delete a user's password (make it empty). This is a quick way to disable a password for an account'
[14:17] <pmatulis> heath|work: ok, i'm used to another way
[14:17] <pmatulis> (usermod -p '!' root
[14:17] <pmatulis> )
[14:18] <heath|work> weird, I will look into that
[14:46] <garymc> Anyone got a softphone they could do a quick voip test to my server for me?
[14:54] <zul> hi mathiaz
[14:57] <zul> mathiaz: puppet issues have been fixed with the testsuite and your bug has been fixed
[15:00] <mathiaz> zul: great - so the MIR has been acceptedÉ
[15:00] <mathiaz> zul: ?
[15:00] <zul> ill let kees now today
[15:06] <ttx> mathiaz: I pushed your MIR for image-store-proxy and pinged the MIR team about it... You might want to ping again at one point in your Tuesday though
[15:07] <mathiaz> ttx: yop - I'll take over for today :)
[15:07]  * mathiaz loves 20 hour long days
[15:08] <ttx> mathiaz: I pushed a new ServerTeam/Roadmap btw, let me know if it doesn't make sense
[15:10] <mathiaz> ttx: great - did you generate the first table by hand?
[15:11] <ttx> mathiaz: yes. I wish there was a way to find those blueprints. Or to tag them.
[15:11] <mathiaz> ttx: we should be able to use LP for that
[15:12] <ttx> yes, we should.
[15:12] <mathiaz> ttx: https://blueprints.launchpad.net/~ubuntu-server/+specworkload
[15:12] <mathiaz> ttx: ^^ that may be the best option we have - but it times out for me now
[15:13] <oversize> hi, i am new to kvm and try to read through al that whats avail to read. now there are documents talking about python-vmbuilder while others (as the ubuntu server guied) use virt-install, are these the same with different names (versions), which should i get to know when i am just starting with kvm? thx
[15:13] <mathiaz> ttx: anyway - we should look into that later or for the next release cycle
[15:58] <smoser> man /join #ubuntu-meeting
[15:58]  * smoser joins #ubuntu-meeting, for anyone who missed that :)
[16:02] <zul> sure you did
[16:17] <agussman> According to this my ec2 instance might be vulnerable to "proxy request attacks": http://tiny.cc/BrgSh .  How do I a) test if I am vulnerable and b) prevent apache from forwarding url?
[16:18] <szczym> helo i have eth0 confugured as static at external ip in /etc/network/interfereces , now i added next net card with link to localnet/dhcp how i could initialize new card ? just add new entry to /etc/network/interfereces ?
[16:19] <genii> szczym: For eth1, yes
[16:19] <szczym> thanx genii
[16:19] <genii> np
[16:20] <szczym> auto eth1
[16:20] <szczym> iface eth1 inet dhcp
[16:20] <szczym> ?
[16:20] <genii> szczym: Looks good
[16:20] <szczym> thanx
[16:20] <genii> szczym: You can then bring it up manually with something like:  sudo ifconfig eth1 up && sudo dhclient eth1
[16:20] <szczym> how i could find out if i have dhcp client running ?
[16:21] <szczym> i know, just type dhclient ;)
[16:21] <szczym> thanx
[16:22] <orudie> hi. i need help with vhosts
[16:24] <orudie> i moved my site from /var/www to /var/www/mysite.com , and i am having trouble making it to work
[16:26] <qman__> pastebin your site configuration file
[16:28] <Keizer> Sup homies
[16:28] <genii> I'm not sure "somewhere.com" as a directoryname in which to put content works that great
[16:28] <soren> We're not homies.
[16:29] <soren> We're geeks.
[16:29] <soren> :)
[16:29] <Keizer> What is the best way to go about setting up a VPN connection on ubuntu server that only goes out that interface if those resources are requested?
[16:29] <Keizer> Been staring at the screen too long so my eyes are unhappy
[16:29] <soren> Keizer: What is the use case?
[16:30] <qman__> what exactly do you mean
[16:30] <Keizer> The RHEL network init scripts are awesome but I've decided to rawk out ubuntu
[16:30] <Keizer> qman__: I need to setup a VPN connection and I don't want all traffic on the machine going out that interface
[16:30] <qman__> that's a simple flag in openvpn
[16:30] <qman__> on or off
[16:32] <qman__> ;push "redirect-gateway"
[16:33] <qman__> uncomment to funnel all traffic through VPN, leave commented to split it
[16:33] <orudie> qman__-> about to paste it
[16:35] <orudie> qman__-> http://pastebin.com/d3970b9da
[16:35] <qman__> orudie, ok, first thing, is mod_rewrite installed and enabled?
[16:35] <qman__> a2enmod rewrite   I think
[16:36] <orudie> umm
[16:37] <orudie> what does it do ?
[16:37] <orudie> redirects from domain.com to www.domain.com ? if yes than its enabled
[16:37] <qman__> you are using rewrite in your site configuration
[16:37] <qman__> so, you need the rewrite module enabled in apache
[16:37] <qman__> it's not there by default
[16:38] <orudie> its enabled
[16:38] <qman__> ok
[16:41] <qman__> well, I don't see anything wrong with it, unless there's something wrong with the rewrite rules, I don't know enough about them to say
[16:41] <qman__> does apache throw any errors, or can you simply not access the site?
[16:42] <Keizer> Thanks qman__
[16:44] <Keizer> http://www.youtube.com/watch?v=R-Yjlx4MRUQ <--- laff
[16:47] <Keizer> qman__: OpenVPN is the defacto VPN correct? I'm used to OpenBSD VPN configuration
[16:47] <qman__> Keizer, it's pretty common, at least, I don't really know of any other full VPN solutions
[16:50] <Keizer> OpenSwan was one back in the day
[16:50] <sub> Thought everyone just used ipsec for site-to-site and PPTP for client access
[16:58] <Keizer> sub: Right
[17:13] <jjohansen1> smoser: hey you around for the kernel meeting
[17:14] <smoser> here
[17:15] <smoser> jjohansen1, here
[17:15] <erichammond> smoser: Over in #ubuntu-kernel
[17:17] <rtg> jjohansen1, moving.
[17:17] <jjohansen1> here
[17:17] <erichammond> her
[17:17] <erichammond> e
[17:17] <smoser> [17:18] <smoser> i'm planning on later today publishing the karmic kernels to eu-west-1, and getting ec2-version-query updated so that it will be the current karmic kernel
[17:18] <smoser> so that will mark bug 418130 as done
[17:18] <rtg> smoser, when will you be ready for the config updates?
[17:19] <rtg> shall I wait until after A6?
[17:19] <jjohansen1> rtg: we should test the config updates before committing them
[17:19] <smoser> for bug 428692, if you want to try something we can. and i'd love it if it had less drivers (ie, more like -virtual than -server)
[17:19] <rtg> jjohansen1, I've pushe3d them, so feel free
[17:20] <jjohansen1> rtg: right I will start a kernel build
[17:20] <smoser> good deal. that'd be great if we could get that for alpha6, but if not, i dont think its the end of the world.
[17:20] <smoser> the other kernel related bug is bug 429169
[17:21] <rtg> smoser, its getting pretty late for A6
[17:21] <smoser> rtg, i agree. so no big deal. we use the previous ones.
[17:21] <rtg> smoser, so, what you have in the archive right now is what's going out with A6
[17:22] <smoser> fine.
[17:22] <lukehasnoname> clear
[17:23] <rtg> erichammond, any issues from you?
[17:23] <erichammond> I think everything is being tracked in LP
[17:24] <smoser> so for 429169, after the MIR takes place (bug 427658) there will be a 'linux-ec2' metapackage in the archive that depends on the latest linux-image-... ?
[17:24] <smoser> for 429619, i'd like to have that so vmbuilder can just know it needs to get 'linux-ec2' rather than knowing what version that is.
[17:24] <smoser> sorry if this is all obvious to everyone else
[17:25] <rtg> smoser, yeah, I need to do a meta package. I'll get that today
[17:25] <Steve[work]> morning everyone
[17:25] <smoser> rtg. great.
[17:26] <rtg> smoser, incidentally, the MIR is complete and the package is accepted
[17:26] <Steve[work]> I have a minbor issue with the multipath-tools package on 9.04 64bit.  It picks up the paths and seems to be handling it properly, however will not create a /dev/dm-0
[17:27] <rtg> jjohansen1, anything else ec2 related?
[17:27] <jjohansen1> nothing I can think of
[17:27] <Steve[work]> I do see a /dev/mapper/*LONGSTRINGOFSTUFF* entry however
[17:27] <rtg> jjohansen1, when do you leave for ATL ?
[17:27] <jjohansen1> friday
[17:28] <rtg> ok, I suggest you suspend these meetings Friday-Fri since you're on the road.
[17:28] <jjohansen1> rtg: good idea, I will send a mail out, also include moving them to #ubuntu-server
[17:29] <rtg> k, is that a wrap for today?
[17:29] <erichammond> I thought these were already canceled on Fridays because of the Karmic status meeting?
[17:29] <rtg> erichammond, could be.
[17:30] <jjohansen1> I think we should, unless people want to move the time
[17:30] <rtg> well, for the next 2 fridays its a moot point
[17:31] <jjohansen1> rtg: true, really wed,thurs,friday next week too
[17:31] <rtg> which is why I suggested Friday through Friday.
[17:31] <rtg> cover it in an email.
[17:31] <jjohansen1> will do
[17:32] <erichammond> And perhaps update http://kernelcalendar.notlong.com
[17:32] <jjohansen1> erichammond: yes, thanks
[17:37] <erichammond> I guess the EC2 kernel status meeting is adjourned.
[17:37] <jjohansen1> yes sorry,  EC2 kernel meeting adjourned
[18:08] <kansan> i'm trying to shut down mysql server and i'm getting a [fail] message ... any ideas on how i can fix?  also is it ok to reboot a server with "sudo reboot" ... if mysql is running and apache is running too?
[18:10] <qman__> kansan, are you sure mysqld is running? in any case, a sudo reboot will attempt to stop them cleanly, then kill them if it can't
[18:11] <kansan> qman__, yes its running
[18:11] <qman__> and 'sudo /etc/init.d/mysql stop' fails?
[18:14] <kansan> yes
[18:14] <qman__> was the current instance of mysqld started with the init script, or manually
[18:15] <qman__> you could try to give it a friendly exit signal manually, 'sudo kill -1 1234' where 1234 is the PID for the process
[18:16] <qman__> that sends a HUP, -15 sends a TERM, and -9 sends a KILL
[18:28] <szczym> helo i setup webdev server, i would like to write and read all in /var/www/ - its all belonging to www-data:www-data so i added my self to group www-data and still cant write there, what is wrong ?
[18:30] <qman__>  /var/www belongs to root on all my systems
[18:30] <qman__> did you change it?
[18:30] <mathiaz> kees: hi!
[18:31] <qman__> also, /var/www is 755, so only the owner can write to it, not group owner
[18:31] <qman__> you'd have to change that to 775
[18:31] <mathiaz> kees: do you have some time to conduct a review of the image-store-proxy (bug 429894)?
[18:32] <szczym> that is example dir where i cant write drwxrwxrwx 2 www-data www-data 4096 2009-09-08 07:37 includes
[18:32] <mathiaz> kees: it's a python daemon that is required for the image store on UEC
[18:32] <szczym> higher dir also belongs to www-data
[18:34] <qman__> since the permissions are 777 you don't even need to be in www-data to write there
[18:34] <szczym> yes but i can`t ...
[18:35] <szczym> btw i just added new user via useradd and i do it as that user
[18:35] <szczym> that user belongs to www-data but not to sudoers
[18:37] <qman__> you don't need to be in sudoers or www-data to write there
[18:37] <qman__> the problem must be elsewhere
[18:40] <szczym> ls
[19:04] <kees> mathiaz: sure, was going to look at that today
[19:04] <mathiaz> kees: great - thanks!
[19:08] <kees> mathiaz: could it be made to run its tests during the build?
[19:09] <mathiaz> kees: I could look into that - I just received the code yesterday
[19:10] <mathiaz> kees: gutsavo said the test suite almost covered everything
[19:10] <mathiaz> kees: one of the goal would be to get it on the -server iso for alpha6
[19:10] <kees> mathiaz: yup, which is why I'd love to see it enabled in the build.
[19:10] <mathiaz> kees: would that be a blocker for main promotion?
[19:11] <mathiaz> kees: I can commit to enable it in the build for karmic
[19:11] <mathiaz> kees: may be not in time for alpha6 though
[19:12] <kees> mathiaz: let me make note in the bug
[19:16] <kansan> having trouble loading up mysql on ubuntu hardy; heres my error message when attempting to start mysql: http://pastie.org/617720
[19:18] <Hypnoz> does this file exist /etc/mysql/debian.cnf
[19:19] <Hypnoz> and permissions look ok etc
[19:20] <kansan> Hypnoz, huh?
[19:20] <Hypnoz> not sure how else to phrase that. Does that file exist?
[19:24] <kansan> no it doesnt
[19:24] <kansan> Hypnoz, how do i create it
[19:24] <Hypnoz> I think its supposed to be created when mysql is installed
[19:24] <Hypnoz> but it looks like mysql is checking that file before it starts
[19:26] <kees> mathiaz: I can't do promotions (you'll need an archive admin for that), but I have approved it in the bug report.
[19:27] <mathiaz> kees: thanks!
[19:27] <kees> np :)
[19:28] <kansan> Hypnoz, so what to do
[19:28] <Hypnoz> did you have anything in your mysql database or is this a new install?
[19:29] <mathiaz> kees: should the status of bug 429894 be set to In Progress?
[19:31] <kees> mathiaz: probably, yes
[19:31]  * mathiaz updates the status
[19:35] <mathiaz> soren: where is the image store UI located? in the eucalyptus-cloud package?
[19:36] <mathiaz> soren: I'm looking where to image-store-proxy should be added as a dependency
[19:43] <mathiaz> soren: https://code.launchpad.net/~mathiaz/eucalyptus/image-store-proxy-recommends/+merge/11819
[19:43] <mathiaz> soren: I've added python-image-store-proxy as a recommends on eucalyptus-cloud
[20:23] <kirkland> mathiaz: howdy
[20:24] <mathiaz> kirkland: yo!
[20:24] <kirkland> mathiaz: do you still need that package reviewed?
[20:24] <kirkland> mathiaz: sorry i'm just getting to it now
[20:24] <mathiaz> kirkland: nope - it's already in the archive
[20:24] <mathiaz> kirkland: MIR accepted and promoted to main
[20:24] <kirkland> mathiaz: sorry about the delay, man
[20:24] <axisys> where can I get the netboot image for ubuntu 7.10 sparc version
[20:24] <mathiaz> kirkland: np
[20:24] <axisys> ?
[20:24] <kirkland> mathiaz: i've been swamped
[20:24] <kirkland> mathiaz: okay, question for you ....
[20:24] <axisys> i dont see it here http://old-releases.ubuntu.com/releases/7.10/
[20:25] <mathiaz> kirkland: sure
[20:25] <kirkland> mathiaz: how would preseed the debconf questions for moodle image built by vmbuilder?
[20:25] <mathiaz> kirkland: hm - I don't know how vmbuilder support preseeds?
[20:25] <mathiaz> kirkland: does it support preseeds?
[20:26] <kirkland> soren: ?
[20:26] <kirkland> hmm, okay another question ...
[20:26] <kirkland> is it legit to "run" an image once, to fix up a few things, before making it an appliance image?
[20:27] <mathiaz> kirkland: hm - I would guess so
[20:27] <mathiaz> kirkland: It's like rebundling an image in EC2
[20:27] <kirkland> smoser: have you done this before?
[20:27] <mathiaz> kirkland: it's just that it's very reproducable
[20:27] <soren> kirkland: It doesn't, no.
[20:28] <mathiaz> kirkland: given the time frame I would do this for the moodle appliance
[20:28] <kirkland> mathiaz: gotcha
[20:28] <mathiaz> kirkland: and since vmbuilder doesn't support preseeds
[20:28] <smoser> i've somewhat rebundled.
[20:28] <smoser> i dont knwo what sort of doc we have for it.
[20:28] <mathiaz> kirkland: it seems that's the only option for now
[20:28] <kirkland> mathiaz: gotcha
[20:28] <smoser> and personally i expect that its not perfectly bug free
[20:29] <kirkland> okay, next question ... suggestions on what to do about this FQDN question?
[20:29] <smoser> what is the FQDN question , for those of us who haven't been playing along at home
[20:29] <kirkland> smoser: it's the one question I can't just choose a sane default for, in the moodle image
[20:29] <kirkland> smoser: FQDN = fully qualified domain name
[20:30] <kirkland> smoser: moodle's debconf uses this for the apache config, seeding the database, etc.
[20:30] <kirkland> ideally, the user would choose this appropriately on first login to the system
[20:30] <smoser> i'm guessing you've already tried and failed with localhost.localdomain or for some reason that isn't sufficient, right?
[20:31] <kirkland> smoser: can't really point a web browser at that
[20:31] <kirkland> smoser: after you do this, you just point a web browser at the hostname, and you work through a web-based setup utility
[20:31] <kirkland> smoser: to customize your instance
[20:32] <mathiaz> kirkland: hm - so you're trying to provide a default answer to the debconf question when the admin logs in?
[20:32] <kirkland> mathiaz: there are about 5-6 questions
[20:32] <kirkland> mathiaz: most of which we can "just pick"
[20:33] <kirkland> mathiaz: do you want to use http or https?  we'll select http for this appliance
[20:33] <kirkland> mathiaz: do you want to be able to access this from outside of localhost?  yes, of course
[20:33] <smoser> but if you select localhost.localdomain, does it not listen then on all ip addresses ?
[20:33] <kirkland> mathiaz: what is your FQDN, this is a critical question...?  ermmm, i was picking "moodle.local" and installing avahi-daemon
[20:34] <kirkland> smoser: i don't think so... i'll try again, though
[20:34] <kirkland> smoser: i think the apache conf sets it up to match only on this name
[20:34] <mathiaz> kirkland: don't you have the remote hostname when the user logs in via ssh?
[20:34] <mathiaz> kirkland: hm - well nevermind
[20:35] <smoser> if it doesn't listen on all ips and you have to give it one (the *right* one) then i have no solution
[20:35] <kirkland> smoser: okay, thanks for playing
[20:35] <mathiaz> kirkland: well - my point being that the end user already knows what the FQDN of the appliance is
[20:36] <mathiaz> kirkland: since he uses it to ssh into the vm
[20:36] <mathiaz> kirkland: could it be possible to get that information automagically?
[20:36] <kirkland> mathiaz: that's what I'm asking :-)
[20:36] <mathiaz> kirkland: right - I don't how to do that.
[20:37] <mathiaz> kirkland: so just ask the question on first login
[20:37] <kirkland> mathiaz: okay
[20:37] <mathiaz> kirkland: well
[20:37] <mathiaz> kirkland: hm - if you need the fqdn for the apache virtualhost
[20:37] <mathiaz> kirkland: you can configure the default virtualhost to server the moddle install
[20:38] <mathiaz> kirkland: then you can also configure the db connection to use localhost
[20:38] <mathiaz> kirkland: for what else is the FQDN needed?
[20:38] <kirkland> mathiaz: i'm not sure
[20:39] <smoser> wait
[20:39] <mathiaz> kirkland: if it's just needed for the DB connection and the apache virtual host then you're good to go
[20:39] <smoser> i'm sorry
[20:39] <smoser> i misunderstood the question
[20:39] <smoser> i thought you were wanting to provide an IP address at *build* time
[20:39] <smoser> is that wrong ?
[20:40] <smoser> if you want to find out the IP address of the running system at runtime, then thats easy. and also the fqdn at that time.
[20:44] <kirkland> smoser: well, i can put a script in the image to run on first boot
[20:44] <smoser> i'm confused . how were you trying to find the info ?
[20:45] <smoser> or when did you want it ?
[20:45] <kirkland> smoser: okay ... ideally, i think, the user would deploy this image
[20:45] <smoser> you can't possibly know it at image build time (other than by using localhost or some canned IP and then doing some magic at boot to make that IP work like the real one)
[20:45] <kirkland> smoser: and just point a web browser at a hostname or ip address
[20:45] <kirkland> smoser: and just configure it from the web
[20:45] <kirkland> smoser: never actually logging in over ssh
[20:46] <kirkland> mathiaz: okay, here's a question i think you probably can answer ...
[20:46] <kirkland> mathiaz: i'd like to write a script to set some debconf values, and then pass it to dpkg-reconfigure
[20:47] <kirkland> mathiaz: that should be pretty doable, i think
[20:47] <kirkland> hmm, that's kinda how d-i works....
[20:47] <mathiaz> kirkland: yes - look at debconf-set-selections
[20:47] <kirkland> mathiaz: thanks
[20:48] <mathiaz> kirkland: right - I'm not sure that you'd actually need to run such a script
[20:48] <mathiaz> kirkland: depending on what the FQDN is used for, you could just install a moodle in a vm
[20:48] <mathiaz> kirkland: and rebundle it
[20:48] <kirkland> mathiaz: hmm, i don't think i can have vmbuilder install the moodle package in the image for me
[20:49] <kirkland> mathiaz: right, so i was going to use vmbuilder to create a jeos appliance
[20:49] <mathiaz> kirkland: right - if that's the case, then you need to install the package manually
[20:49] <kirkland> mathiaz: then run the image, install moodle
[20:49] <kirkland> mathiaz: in debconf there, I would set all of the selections to sane defaults
[20:50] <kirkland> mathiaz: then, i was going to leave a script to run at first boot
[20:50] <kirkland> mathiaz: that would autodetect FQDN, and debconf-set-selection to that detected value
[20:50] <mathiaz> kirkland: well - I'd investigate first what the FQDN is used for
[20:51] <mathiaz> kirkland: you may be able to by pass it completely if it's not used for anything else than the DB connection and the apache virtual host
[20:51]  * kirkland is looking at the source now
[20:53] <kirkland> mathiaz:  Moodle will be configured to use this system's Fully Qualified Domain Name
[20:53] <kirkland>  (FQDN). This must be correct, and you cannot easily change this after your
[20:53] <kirkland>  Moodle installation is complete.
[20:54] <kirkland> mathiaz: okay, so i don't have to provide that as input into moodle debconf...
[20:54] <kirkland> mathiaz: its just that the system's hostname *must* be set correctly when moodle's debconf runs
[20:58] <MK-ubuntu> hi, i was wondering
[20:58] <MK-ubuntu> on my vps I was assigned 2 IPs ..
[20:59] <MK-ubuntu> but, I cant seem to get the 2nd one to work.
[20:59] <MK-ubuntu> can you guys help me?
[21:00] <MK-ubuntu> anyone here at the moment?
[21:03] <MK-ubuntu> quiet everytime i come in
[21:09] <mushroomtwo> MK-ubuntu: lurk moar
[21:09] <MK-ubuntu> pardon?
[21:11] <MK-ubuntu> mushroomtwo, ?
[21:13] <apache> I have this entry on my Apache 2 access.log file but I have not 192.168.23.13 machine on my LAN. How is this possible?
[21:13] <apache> 192.168.23.13 - - [15/Sep/2009:19:41:25 +0200] "GET /virtual/ HTTP/1.1" 200 12294 "http://my.host.name/" "Mozilla/5.0 (Windows; U; Windows NT 5.1; es-ES; rv:1.9.0.14) Gecko/2009082707 Firefox/3.0.14 (.NET CLR 3.5.30729)"
[21:26] <genii> apache: Spoofing?
[21:28] <genii> apache: If you have some unsecured wifi router, someone can manually put an IP address for themself that is on same segment and specify the router as gateway, etc all manually so no dhcp records get left.
[21:39] <LeChacal> hello, i need to change the order in which certain modules are loaded so that one is loaded before the other, if i change the order of the lines in /lib/modules/2.6.28-15-server/modules.dep and /lib/modules/2.6.28-15-server/modules.order is that all i need to do?
[21:45] <genii> LeChacal: They are loaded in the order listed
[21:45] <LeChacal> genii: so i just change those two files and I am good, no where else?
[21:46] <genii> LeChacal: If you have some list in /etc/modules  maybe
[21:47] <LeChacal> geniiL i don't have anything in there, ok thank you I will try changing the orders
[22:00] <thrillyb> hello #ubuntu-server
[22:01] <domas> hello thrillyb
[22:02] <thrillyb> do people talk a lot on irc anymore these days?
[22:02] <thrillyb> or is it better to go to forums for help / discussion?
[22:05] <thrillyb> lots of idlers?
[22:06] <Slimu_> Hello
[22:06] <Slimu_> Can anyone help? my hardy server wont go back online after a pwoer outage
[22:06] <Slimu_> *power
[22:07] <thrillyb> how far does it get?
[22:07] <thrillyb> does it post video at all?
[22:07] <Slimu_> It boots, and starts apache
[22:07] <Slimu_> , and when i run ifconfig i can see the router
[22:08] <Slimu_> but the networking directory has gone!
[22:08] <thrillyb> oh!
[22:08] <Slimu_> I looked in /etc/networking, and got: no such file or directrory
[22:08] <Slimu_> erm - i mean /etc/network
[22:08] <thrillyb> hehe
[22:08] <thrillyb> was about to say
[22:09] <Slimu_> Im still learning, but i got it all up and running for over 3 months before the power cut
[22:09] <thrillyb> have you ever done a reboot successfully?
[22:10] <Slimu_> thats one of the 1st things I tried, but no joy
[22:11] <Slimu_> this is ODD: this time the /etc/interface directory is there, but the interfacces file is empty!
[22:11] <thrillyb> before the power outage were you able to reboot though?
[22:11] <Slimu_> yes
[22:11] <thrillyb> k
[22:11] <Slimu_> but i only rebooted after an update
[22:11] <thrillyb> have you tried rerunning anything like this guide: http://www.ubuntugeek.com/ubuntu-networking-configuration-using-command-line.html
[22:12] <Slimu_> I'll check that out now
[22:12] <thrillyb> if the device is there that should work
[22:12] <Slimu_> thanks for the pointer
[22:12] <thrillyb> but if you had a power surge maybe something got toasted in your nic card
[22:14] <kirkland> soren: what's the state of vmbuilder in LP and in karmic?
[22:14] <kirkland> soren: i'm generating images, but they're not bootable
[22:14] <kirkland> soren: at least not under kvm
[22:14] <mathiaz> soren: from karmic?
[22:14] <Slimu_> my router and server are on a surge protected cable and live behind a ups
[22:14] <mathiaz> kirkland: do you know that karmic is broken right now?
[22:14] <kirkland> mathiaz: yes, the image is jaunty
[22:14] <thrillyb> k, you should be good then! :)
[22:15] <Slimu_> brb
[22:15] <mathiaz> kirkland: hm ok. if you're building an jaunty guest, then I don't know
[22:17] <frojnd> Does anyone here connects with a phone using java MIDP 2.0 to a ssh and if so what java app do you use?
[22:21] <rickspencer3> kirkland: I have a buddy who is starting a business, and thinks he is going to need to wrap some of the aws commands into a web UI
[22:21] <rickspencer3> is there such a think already, or should he roll his own?
[22:22] <kirkland> rickspencer3: possibly elasticfox
[22:22] <rickspencer3> kirkland: what a name!
[22:22] <rickspencer3> sounds like a plugin
[22:22] <kirkland> rickspencer3: heh
[22:22] <rickspencer3> I think he wants to render web pages
[22:22] <kirkland> rickspencer3: its a firefox plugin
[22:23] <rickspencer3> is there a project for that, I guess aws web interface?
[22:23] <kirkland> rickspencer3: hmm, i might have misunderstood the question ...
[22:23] <erichammond> rickspencer3: What exactly are you looking for
[22:23] <erichammond> ?
[22:23] <kirkland> rickspencer3: what sort of aws commands?
[22:23] <rickspencer3> kirkland: create instance, etc...
[22:23] <rickspencer3> describe instance
[22:24] <rickspencer3> I think just so you can do it via a web ui instead of command line
[22:24] <rickspencer3> hi erichammond :)
[22:24] <erichammond> AWS is a (W)eb (S)ervice so all of the functionality is available through web APIs.
[22:24] <erichammond> (programmable)
[22:25] <kirkland> rickspencer3: people have put frontends around it ... eg elasticfox, rightscale
[22:25] <rickspencer3> erichammond: right, but I think he wants to make his own that do specific tasks
[22:25] <Slimu_> thrillyb: DAMIT: I went through the link that you posted, but i still cant get it online!
[22:26] <rickspencer3> kirkland: erichammond - thanks!
[22:26] <rickspencer3> I'll check those out
[22:26] <kirkland> rickspencer3: no problem; come back if you get some specific questions ;-)
[22:26] <erichammond> rickspencer3: What I'm saying is that he would put a human-usable web interface and interact with the AWS API on the back end, probably through one of the many programming language APIs.  If you know what programming language, we can point you in the right direction there, too.
[22:27] <thrillyb> slimu_: what's your ifconfig say now?
[22:27] <rickspencer3> erichammond: right
[22:27] <Slimu_> thrillyb: it shows ip of 192.168.1.11, so ther router has assigned the ip address
[22:28] <thrillyb> slimu_: can you ping the router and vice versa?
[22:28] <Slimu_> thrillyb:the router conf page (on other desktop) shows the server as connected, but it wont let me ping
[22:28] <thrillyb> hey everyone else... anyone have any experience setting up directory service with ec2? is it reasonable to even try?
[22:29] <Slimu_> thrillyb: scap that. just refreshed the page and is says disconnected
[22:30] <MK-BB> on my vps I was assigned 2 IPs ..but i got the 1st ip working but cant seem to set the second one up. can u help?
[22:31] <thrillyb> erichammond: btw thanks for all your work with ec2 amis. I have used a handful and they're lovely
[22:33] <thrillyb> slimu_: that's pretty effed up sounding
[22:33] <erichammond> thrillyb: Glad to hear they're working for you.
[22:33] <Slimu_> thrillyb: :-(
[22:34] <thrillyb> slimu_: what does your /etc/interfaces look like?
[22:34] <Slimu_> one mo
[22:34] <Slimu_> its blank!
[22:35] <Slimu_> i dont remember whats supposed to go in there anymore
[22:35] <thrillyb> erichammond: we just updated to your 9.04s a month or so ago. currently trying to figure out a way to get multiple users on all our amis without updating them everytime we wanna add someone
[22:38] <thrillyb> Slimu_: read the example for some good help > file less /usr/share/doc/ifupdown/examples/network-interfaces.gz
[22:38] <Slimu_> thrillyb: ok. many thanks for that. sounds like i've found the problem
[22:39] <thrillyb> Slimu_: no problem! access to the internet is probably pretty important for your server :)
[22:40] <erichammond> thrillyb: What do you mean by "multiple users"?
[22:40] <Slimu_> thrillyb: yea, I've been rendering some animations and wanted to add them to my site, and found the problem while trying to ssh into it
[22:40] <erichammond> thrillyb: i.e., ssh accounts?  client web sites? ...?
[22:43] <thrillyb> erichammond: admins and such. yeah ssh accounts mainly
[22:44] <thrillyb> erichammond: i've been messing around with the idea of setting up LDAP to authenticate but the idea seems a lot simpler than the implementation
[22:44] <erichammond> thrillyb: You should be aware that the current Amazon 2.6.21 kernel has a bug which allows normal users to gain root access.  It is possible to close this by downgrading to their latest 2.6.18 kernel or upgrading to Canonical's 2.6.27 kernel.
[22:45] <erichammond> thrillyb: So the problem you're trying to solve is having users be able to log in to new instances started from your private AMI?
[22:46] <thrillyb> erichammond: exactly. centrally stored usernames, passwords, home directories etc
[22:47] <thrillyb> erichammond: and thanks for the note on the kernel vulnerability - that's frightening
[22:48] <erichammond> thrillyb: Got it.  Lots of ways to do this, I suppose, depending on your requirements.  Also consider: (1) passing in user-data script which creates all accounts at instance boot, (2) startup script which downloads list of users to create from S3, possibly using private URL.
[22:48] <erichammond> thrillyb: Adding users to existing instances could be done with chef/puppet/... or simple automated ssh scripts or cron jobs.
[22:49] <erichammond> thrillyb: But if the list changes a lot, central service sounds like a good approach.  You just have to keep it running and solve the problem of having all instances talk to the correct master, especially if it needs to be replaced.
[22:49] <erichammond> thrillyb: This approach may help with the last issue: http://alestic.com/2009/06/ec2-elastic-ip-internal
[22:51] <thrillyb> erichammond: i've thought about a cronjob/rsync type situation, it sounds more legit the way you described it
[22:53] <erichammond> thrillyb: I'm a fan of clean, simple, understandable systems. In many cases chef/puppet/et.al. is just complicating things.
[22:53] <thrillyb> erichammond: LDAP or any of it's brethren are probably over kill
[22:55] <thrillyb> erichammond: thanks for the guidance. i'll see give this a try.
[22:58] <kirkland> mathiaz: smoser: \o/ moodle configurable by ip address alone
[23:22] <ahe> kirkland: so does it work?
[23:25] <kirkland> ahe: i'm getting there....
[23:25] <ahe> cool
[23:26] <ahe> i'm always interested in modifications, escpecially if there is something you couldn't do before
[23:43] <dayo> how do i use something like `passwd -l` for users with ldap authenticated nfs home directories?