[00:06] <giovani> StrangeCharm: Areca makes decent  stuff
[00:06] <giovani> they make a 16-port PCIe/PCIx card
[00:23] <Debolaz> Hrmm, what structure is recommended to use on an OpenLDAP directory if you have say 10 servers or so with different users but some in common?
[00:24] <giovani> Debolaz: well they're all going to be in the LDAP directory I hope
[00:24] <giovani> I'm not sure what type of answer you're looking for though
[00:26] <Debolaz> giovani: Never really used LDAP before, except for playing a bit.. And while there's a plethora of tutorials available on how to set up the system to use LDAP, there's very little information about how the data should be structured.. No recommended standards, etc..
[00:27] <giovani> Debolaz: that's far too huge of a topic to answer on irc
[00:27] <giovani> there are 700-page books dedicated to LDAP design
[00:28]  * Debolaz has a 500 page book about OpenLDAP that has nothing about this in it. 
[00:28] <giovani> maybe it was the wrong book :)
[00:29] <Debolaz> Hehe, well, I'm shooting in the blind here. :)
[00:29] <Debolaz> At least it was very helpfull in setting up and making OpenLDAP work with the various services. :)
[00:29] <giovani> great
[00:29] <giovani> I mean typically you're gonna have an ou
[00:29] <giovani> for users
[00:30] <Debolaz> I don't suppose you have any good book recommendations?
[00:30] <giovani> nope -- never bought an LDAP book -- but we have a lot of them at work -- not sure which would be of quality
[00:30] <giovani> just learned by example
[00:35] <Debolaz> Oh the joys of modernizing a server park.
[00:35] <Debolaz> Well actually, it is quite fun.
[00:38] <Debolaz> Started employing virtualization to better take adventage of the hardware we've got.
[00:57] <BrixSat> hello
[00:57] <BrixSat> any one able to help on dns?
[01:00] <jmarsden> BrixSat: Don't ask to ask, just ask.  What is your real question?
[01:01] <BrixSat> :p
[01:01] <BrixSat> sorry
[01:01] <BrixSat> http://pastebin.com/m4d9df416
[01:01] <BrixSat> why tis?
[01:12] <Level15> hi. why does aptitude choose not to upgrade packages as the kernel and bind?
[01:14] <ScottK> Probably because new packages need to be installed.  Instead of aptitude upgrade, try aptitude dist-upgrade.
[01:15] <Level15> ScottK: thanks, will try that
[01:45] <BrixSat> http://pastebin.com/m4d9df416
[02:01] <jmarsden> BrixSat: rndc not working is probably a configuration issue of some kind... they key has to be the same in both the bind config file and the rndc one...
[02:03] <BrixSat> jmarsden where do iset the key?
[02:05] <jmarsden> man rndc-confgen    and look under /etc/bind/ for all the config files.
[02:05] <BrixSat> :)
[02:08] <jmarsden> I'm pretty sure it should "just work" out of the box on Ubuntu, unless you manually configured something related to it, though.
[02:08] <BrixSat> i have used ispcp :p
[02:10] <jmarsden> I don't know what it is, I don't think that is a Ubuntu package in main or universe... but apparently it broke rndc for you.
[02:10] <BrixSat> it is like cpanel a control panel for hosting
[02:10] <BrixSat> but it was working fine
[02:11] <BrixSat> i broke it then i mixed with the nameserver and apache config
[02:11] <jmarsden> OK... so look for key statements in your named.conf and in /etc/bind/rndc.conf, maybe?  If both exist they need to specify the exact same key...
[02:12] <BrixSat> :S they are the same
[02:13] <BrixSat> my domain is www.opoeta.net
[02:14] <jmarsden> Check permissions on the files, maybe?  Or remove the key stuff from named.conf completely (back it up somewhere first!), and then do    sudo rndc-confgen -a   and see if that fixes it?
[02:14] <BrixSat> ok
[02:16] <BrixSat> nothing :S
[02:18] <jmarsden> OK... does ispcp do things with the files under /etc/bind, or does it install a non-standard bind and put config files somewhere else?  If it uses the ones under /etc/bind/, can you pastebin /etc/bind/named.* and /etc/bind/rndc.key so I can see them?
[02:18] <BrixSat> it uses system bind
[02:19] <jmarsden> OK.
[02:19] <BrixSat> yes i can
[02:19] <BrixSat> http://pastebin.com/f471177ea
[02:20] <BrixSat> pastebin.com/f301dd305
[02:21] <jmarsden> So... you think that "Q/xVOF1uUAv89CFuy02pag==" and "pkVeO3RyPavpE4z4nsY04g==" are the same??
[02:22] <jmarsden> Those two keys are different :)
[02:23] <BrixSat> :|
[02:24] <jmarsden> Edit one or the other file to make them the same, and restart bind
[02:24] <BrixSat> wich one has the pk....
[02:25] <jmarsden> /etc/bind/rndc.key has the one starting with pk
[02:25] <jmarsden> The one you pasted at http://pastebin.com/f301dd305
[02:26] <BrixSat> i have seen and i feel so stupid
[02:26] <jmarsden> well, at least it should be an easy fix :)
[02:27] <BrixSat> same :S
[02:27] <jmarsden> You restarted bind after making the change?
[02:27] <BrixSat> yes
[02:28] <BrixSat> http://pastebin.com/md38f809
[02:30] <jmarsden> I'm not sure what is happening, then, maybe remove the whole key and control stuff from the end of /etc/bind/named.conf, so it looks like http://pastebin.com/m69a98d12
[02:30] <jmarsden> Or... well, does rndc status work now?
[02:31] <jmarsden> The error you pasted was from the old bind stopping, not from the new one...
[02:31] <jmarsden> Also check /etc/bind/named.local and /etc/bind/named.options in case they somehow have key stuff in them ??
[02:33] <jmarsden> Make that /etc/bind/named.conf.options and /etc/bind/named.conf.local
[02:36] <BrixSat> ok
[02:38] <jmarsden> Incidentally, why are you needing to use an "ISP" control panel when you only seem to have a single domain?
[02:40] <BrixSat> i will have other domains soon
[02:42] <jmarsden> OK... in general adding a "control panel" from outside Ubuntu is adding extra complexity and a source of issues that you don't need... if you *really* need a web based admin UI, you could try ebox, which at least is an official Ubuntu universe package.
[02:43] <jmarsden> But for less than a couple of hundred domains, some simple scripts and doing things at the command line is probably better still -- that way you know what is happening and can troubleshoot when anything breaks.  "Control panel" apps hide that from you...
[02:45] <BrixSat> ebox :/
[02:46] <jmarsden> If you don't like it, fine, use the command line :)
[02:46] <BrixSat> :)
[02:46] <BrixSat> ispcp is good for selling
[02:47] <BrixSat> webhost space
[02:47] <BrixSat> i got it to work
[02:47] <jmarsden> OK.  i hope ispcp doesn't break it for you again...!
[02:47] <BrixSat> no it was working :p
[02:47] <BrixSat> after i mixed with A record~
[02:47] <BrixSat> it stopped
[02:48] <BrixSat> and then i rebooted and never more could put it to work
[02:48] <jmarsden> OK... well, as long as you have it fixed now, I suppose all is well...
[02:51] <BrixSat> hostname: Unknown host when i try to reconfigure the iscp :S
[02:52] <jmarsden> As I said earlier:  <jmarsden> OK... in general adding a "control panel" from outside Ubuntu is adding extra complexity and a source of issues that you don't need...
[03:27]  * Debolaz needs a control panel for managing web hosting customers
[03:31] <Debolaz> Would ebox work for something like that?
[04:17] <monostone> I have correctly installed and configured dnscache + tinydns. I have tested dnscache via a tail -f logfile, and it is correctly cacheing content. Now, I am setting up a development/test environment, so I am using internal IP's, but i can NOT manage, to get tinydns to resolve the domains on my local network (commands dnsq a host IP and dig @IP host return correct responses, but dnsqr host, dig host and dig @IP -x IP all fail). My
[04:17] <monostone> resolv.conf has 127.0.0.1 defined as nameserver, i even added a search domain.internal, but nothing. I am forced to add the entry to the hosts file, but i want to avoid this, any ideas?
[05:57] <monostone> how do permissions cascade with the directory entry in apache? I set <Directory /> and <Directory /var/www> as denied for all, then i configured <Directory /var/www/site1> as allowed by all, but i get permission denied. don't permissions cascade? with the last directory rule taking precedence?
[06:01] <qman__> IIRC, set up that way, it should allow access to site1
[06:01] <qman__> did you also check filesystem permissions?
[06:01] <qman__> that can result in a 403 too
[06:08] <monostone> qman__, aha, right, i completely overlooked that
[08:41] <chindu> Am wondering if you can you update from 9.04 server to an lts build - say 10.04 when it's released?
[08:41] <twb> chindu: there are two upgrade paths
[08:42] <twb> chindu: LTS to LTS, or <release n> to <release n+1>
[08:42] <twb> So you would need to upgrade to 9.10, and then from 9.10 to 10.04
[08:43] <chindu> Thanks twb. Are there any compelling reasons to stick with 8.04 over 9.04? I'm going to be deploying in an Esx 3.5 environment.
[08:43] <chindu> or vice versa
[08:44] <twb> Yes: 8.04 is LTS, so it gets 5 years of support on the server.
[08:44] <twb> That means you can avoid upgrading for five years.
[08:44] <twb> (Upgrading is a great way to get new bugs.)
[08:46] <chindu> Lol. So in an esx environment will I see any real benefit from the newer releases then or just cause myself more grief? May as well stick with LTS if no performance reasons to go to newer releases...
[08:46] <twb> My rule for upgrading is: don't do it unless you can say what the benefits are gonna be
[08:47] <twb> e.g. "my webdevs insist they need a newer version of PHP5"
[08:47] <chindu> Am planning to run Postgres on one and the other will run Apache.
[08:49] <chindu> I've not delved too much into the versions available but am I right in thinking that you don't get the newer product versions if you aren't on the latest but they keep up-to-date with all security fixes?
[08:52] <twb> In principle, packages in the main category receive security updates in (say) hardy-security, backported to that version, until the package is EOLd.
[08:52] <twb> *until the release is EOLd, rather
[08:53] <twb> http://bazaar.launchpad.net/%7Enijaba/ubuntu-maintenance-check/trunk/ contains a little script to tell you when a package will stop receiving security updates
[09:00] <chindu> @twb cool - thanks very much
[10:18] <ttx> soren: do you plan to roll a new eucalyptus package today ? I could use the latest fixorz in my testing
[10:20] <soren> ttx: I can.
[10:21] <ttx> soren: I guess I'll be blocked by the various "403 Forbidden" bugs if I progress in testing ?
[10:21] <soren> ttx: Probably.
[10:22] <soren> Let me roll a new package..
[10:23] <soren> ttx: Blimey.
[10:23] <soren>  158 files changed, 7029 insertions(+), 4175 deletions(-)
[10:23] <ttx> uh
[10:24] <soren> uh, indeed :)
[10:24] <ttx> its the merge with their dev branch ?
[10:24] <ttx> I don't think we changed that much in ours :)
[10:27] <soren> That's the diff from bzr746 to bzr808.
[10:27] <soren> So from the most recent snapshot to a current one.
[10:28] <ttx> ok
[10:50] <soren> ttx: It doesn't compile.
[10:50] <ttx> ...
[10:50] <soren> ttx: It /seems/ to have grown a new dependency.
[10:51] <soren> ttx: http://people.canonical.com/~soren/eucalyptus.txt
[10:52] <soren> ttx: Ah, yes.
[10:52] <soren> ttx: libc3p0-java is already a runtime dependency, but now it's probably become a build-time dep as well.
[10:53] <soren> how things can move in that direction is beyond me, though.
[10:53] <ttx> your analysis seems exact.
[10:55] <soren> I can see how it can need stuff at build-time, but then not at runtime, but not the other way around.
[10:56] <soren> Anyhow, I'm rebuilding now. If it works, I'll upload.
[10:58] <Debolaz> Hmm, where can I read about what's new in 1.6?
[11:03] <kaushal> hi
[11:03] <kaushal> is the native package tomcat5.5 in Ubuntu 8.04 server have issue ?
[11:04] <kaushal> somehow it takes lot of time to stop/start it
[11:04] <kaushal> i install apt-get install sun-java5-jdk and tomcat5.5 on it
[11:05] <kaushal> am i doing it correctly ?
[11:05] <kaushal> is there a already raised bug
[11:05] <_ruben> have you checked the bugtracker?
[11:05] <kaushal> _ruben: is it launchpad ?>
[11:05] <kaushal> ?*
[11:05] <_ruben> yes
[11:06] <soren> What's the problem, exactly? It takes a while to start it?
[11:06]  * soren is hardly surprised, it being java and all
[11:07]  * _ruben neither
[11:07] <kaushal> soren: do i need to install sun-java6-jdk or sun-java5-jdk ?
[11:07] <kaushal> for 5.5.25-5ubuntu1.1 ?
[11:07] <soren> kaushal: Do you need either of them specifically?
[11:07] <kaushal> yeah
[11:07] <kaushal> i need sun-java5-jdk
[11:08] <soren> Then you've just answered your own question, as far as I can see.
[11:08] <soren> If you /need/ sun-java5-sdk, then yes, of course you should install it.
[11:09] <kaushal> soren: i did install sun-java5-sdk but as i said when i try to stop tomcat service it doesnot stop
[11:09] <soren> When did you say that?
[11:10] <kaushal> 15:34 < kaushal> somehow it takes lot of time to stop/start it
[11:11] <soren> "Takes a lot of time to stop/start" is hardly the same as "doesnot stop"
[11:11] <soren> Which is it? Does it stop or doesn't it?
[11:11] <kaushal> it doesnot stop
[11:11] <soren> Ok. In that case, I'd file a bug.
[11:12] <kaushal> soren: ok
[11:12] <kaushal> Let me try installing sun-java6-jdk and see if it works
[11:14] <kaushal> soren: where do i need to report a new bug ?
[11:14] <soren> kaushal: On Launchpad.
[11:15] <soren> https://bugs.launchpad.net/ubuntu/+source/tomcat5.5/+filebug
[11:15] <soren> kaushal: Remember to include exactly which version of the packages you're using and which version of Ubuntu you're running on.
[11:16] <ttx> might be bug 299866
[11:17] <soren> Could be.
[11:18] <soren> ttx: New eucalyptus uploaded.
[11:18]  * soren wanders off for lunch
[11:22] <kaushal> ttx: https://bugs.launchpad.net/ubuntu/+source/tomcat5.5/+bug/433906 is it ok ?
[11:24] <ttx> kaushal: add information about which JVMs you have installed on the system, that will help to reproduce the issue
[11:24] <ttx> otherwise that's a good start.
[11:24] <kaushal> sure
[11:25] <kaushal> ttx: please have a look at now https://bugs.launchpad.net/ubuntu/+source/tomcat5.5/+bug/433906 ?
[11:26] <ttx> kaushal: ok
[11:28] <kaushal> ttx: is that Bug correct ?
[11:29] <kaushal> just curious to know
[11:29] <ttx> kaushal: yes, it's filed correctly. Will have a look when I get a chance
[11:29] <kaushal> sure
[11:29] <kaushal> ttx: Thanks
[11:51] <soren> ttx: Get 'em while they're fresh: https://edge.launchpad.net/ubuntu/+source/eucalyptus/1.6~bzr808-0ubuntu1/+build/1253738
[11:54] <kaushal> ttx: when i start tomcat i see it as here http://paste.ubuntu.com/275171/
[11:54] <Debolaz> soren: Do you know where I can read about what's coming in 1.6?
[11:54] <kaushal> its not started as tomcat55 user
[11:54] <kaushal> any ideas ?
[11:54] <kaushal> I see it as root user
[11:55] <kaushal> may be thats the issue ?
[11:55] <soren> Debolaz: I don't know where there's a list, no.
[12:01] <kaushal> ttx: you around ?
[12:33] <kaushal> soren: hi again
[12:33] <kaushal> shall i briefly discuss the issue ?
[12:33] <soren> Um... No :)
[12:34] <kaushal> when i start tomcat it starts as root user
[12:34] <kaushal> so is there a way to fix it ?
[12:35] <ttx> kaushal: it starts as root and spawns a process that runs as tomcat55.
[12:36] <kaushal> ok
[12:36] <ttx> soren: bug 431847 is a euca2ools issue. Works with ec2-api-tools
[12:37] <ttx> I'll mark it as such.
[12:37] <kaushal> what about tomcat 5.5 which does not stop ?
[12:37] <kaushal> ttx: any way you will be looking at it
[12:37] <ttx> kaushal: I didn't look into the issue yet. I'm kinda busy on other things.
[12:37] <kaushal> sure
[12:37] <zul> morning
[12:37] <kaushal> np
[12:37] <soren> zul: o/
[12:37] <ttx> zul: yo
[12:39] <soren> ttx: I'll push a new euca2ools to my ppa in 15 minutes. Can you retest with that, please?
[12:39] <zul> hey soren and ttx
[12:39] <ttx> soren: sure.
[12:42] <zul> ahhh euca is filling up my inbox
[12:45] <soren> zul: with love?
[12:46] <zul> umm...yeah...you can say that
[12:57] <soren> ttx: http://people.ubuntu.com/~soren/euca2ools_1.0+bzr20090921-0ubuntu1_all.deb
[12:58] <ttx> soren: ack
[13:00] <ttx> soren: same issue.
[13:00] <ttx> euca-describe-images has the same problem, fwiw
[13:01] <soren> ttx: Yes, I'm not seeing much improvement either. Did you use the fresh eucalyptus packages?
[13:01] <ttx> soren: I'm still on 1.6~bzr746-0ubuntu4
[13:01] <ttx> but I can try that.
[13:02] <soren> I am too. I'm upgrading in a few minutes.
[13:09] <ttx> soren: no change.
[13:09] <ttx> dpkg -L euca2ools
[13:10] <ttx> oops
[13:10] <soren> :(
[13:14] <ttx> hmmm
[13:16] <ttx> soren: I made it work by fiddling with TZ as well, let me narrow that down
[13:17] <ttx> yes, choosing UTC as the TZ and rsetarting eucalyptus-cloud works around it
[13:19] <oversize> Is it possible to alter the location of the images in libvirtxml.tmpl like this: <source file='/var/libvirt/domains/$hostname/$disk.filename' /> !? and then pass that template to vmbuilder? I am asking cause its not working, and i dont know if i just dont understand, or if it just wont work at all :) thx
[13:20] <soren> oversize: How is it not working?
[13:20] <ttx> soren: anything <= UTC works. UTC+1 fails
[13:20] <soren> oversize: You can change anything in the templates. That's the whole idea.
[13:21] <soren> ttx: Fantastic.
[13:22] <ttx> soren: hm. wait, it seems a little less predictable now.
[13:22] <soren> Even fantasticer.
[13:22] <ttx> GMT-2 fails :)
[13:23] <oversize> the images get created in the current dir where i issued the command.  i have set templates in /etc/vmbuilder.cfg to a path where my modified libvirtxml.tmpl resides. And this has the pasted <source xx >
[13:23] <soren> ttx: Oh. euca-run-instances... errr... did not fail the way it used to.
[13:23] <soren> $ euca-run-instances emi-67941325
[13:23] <soren> RESERVATION	None	None
[13:23] <ttx> ah-ah.
[13:24] <soren> Oh!
[13:24] <soren> hah, seems to be configuration problem.
[13:24] <soren> com.eucalyptus.util.EucalyptusCloudException: image too large [size=10240MB] for instance type m1.small [disk=2048MB]
[13:25] <ttx> forget what I just said about GMT-2. It's something else
[13:26] <ttx> so it seems predictable (and an understandable bug)
[13:27] <oversize> oh ... think i got it. did not modified the path in libvirtxml_fsimage.tmpl ... but why the both files? is there some doc on this template thing? i have only found https://help.ubuntu.com/community/JeOSVMBuilder
[13:27] <soren> oversize: There's different templates depending on the type of image you're building.
[13:28] <soren> oversize: If you're building Xen images, you should fiddle with libvirtxml_fsimage.tmpl
[13:28] <oversize> ok, i am building kvm images:  vmbuilder kvm ubuntu ....
[13:32] <soren> then libvirtxml_fsimage.tmpl shouldn't be relevant.
[13:33] <soren> ttx: Odd.
[13:33] <soren> ttx: com.eucalyptus.util.EucalyptusCloudException: Not enough resources: vm instances.
[13:35] <ttx> soren: but the client comm seems alright now
[13:35] <soren> Something's screwy here, that's for sure.
[13:36]  * soren reboots the thing
[13:39] <oversize> do i have to have the libvirt/libvirtxml.tmpl structure under the path specified with templates= ?
[13:43] <soren> Yes.
[13:43] <soren> oversize: You give vmbuilder an entirely new set of templates to play with.
[13:43]  * soren takes a good, long break
[13:43] <oversize> ok explains alot :D
[13:43] <oversize> thx
[15:33] <aubre> I upraded my karmic this morning and now when I do aeuca-describe-availability-zones verbose all I see is AVAILABILITYZONE,  the name of the cluster, and 127.0.0.1 . When I try to run an instance I get this :RESERVATION	None	None. Using the web interface I can see the page where I can modify the characteristics of the types of instances.
[15:34] <aubre> and I have rebooted both my front end and my node controller
[15:42] <zul> hmmm...45,000 iteriations i think that bug is fixed
[15:45] <XiXaQ> I am trying desperately to get my cpufreq scaling_governor to permanently use "performanc", but I'm unable to. I've added "/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor=performance" (and cpu1) in both /etc/sysctl.conf and /etc/sysfs.conf. I've also added "echo 'performance' > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor" to /etc/rc.local.
[15:46] <XiXaQ> when I reboot, it's set to "performance", but then after a while, it's back on "ondemand". Can someone please help me figure this out? It's crashing all my services, so it's very, very important to me.
[15:47] <XiXaQ> can I add an inotify to /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor and overwrite any changes?
[15:48] <XiXaQ> that would be a very bad solution as well, but it's better than having to pull the virtual powercord on all my kvm guests once a day.
[15:51] <ttx> soren: is the cloud installer supposed to autoregister the cluster and walrus with the cloud controller ?
[15:52] <soren> ttx: That's the idea, yes.
[15:52] <ttx> soren: because I don't think it did, in my recent install.
[15:53] <soren> ttx: Shame. I /may/ have messed something up in the merge. There was a /lot/ of conflicts in that init script.
[15:53] <ttx> soren: i'll retry that.
[15:54] <kinnaz> should come and help you guys with this eucalypt testing
[15:54] <kinnaz> seems nice thing to mess with
[16:00] <soren> ttx: I'm calling it an afternoon. I'll be around this evening if anything pops up.
[16:00] <ttx> soren: ok
[16:01] <vxnick> hi all - is 64bit MySQL stable enough on Ubuntu 9.04?
[16:04] <XiXaQ> 9.04 is so unstable here, it's completely useless. I have an uptime record of 26 hours for my kvm guests here.
[16:05] <_ruben> XiXaQ: tried disabling cpu scaling in the bios?
[16:06] <XiXaQ> of course.
[16:06] <XiXaQ> seems to be a script in ubuntu to bypass the proper configuration files from time to time, but nobody seems to know anything about it.
[16:08] <_ruben> then again, i've only experienced the opposite (in a way): ondemand not throttling down my cpu
[16:08] <XiXaQ> right. I've heard people complaining about that as well, on desktops.
[16:08] <_ruben> was a desktop machine used as server, yes :)
[16:08] <XiXaQ> but that doesn't cause all of their systems to stop functioning.
[16:09] <giovani> XiXaQ: heh, there can't be a mysterious script
[16:09] <giovani> either there's a daemon running
[16:09] <giovani> or there's something in cron
[16:09] <giovani> look at both
[16:09] <XiXaQ> I have. I can't find anything.
[16:10] <_ruben> any related tools/daemons installed at all? powersaved/powernowd/cpufrequtils/etc
[16:10] <smoser> soren, are you planning on releasing a new vm-builder for karmic?
[16:11] <XiXaQ> _ruben, I haven't installed any.
[16:11] <smoser> i'm asking, because bug 427288 is marked as 'fix commited' for vm-builder/karmic, but its not commited anywhere there.
[16:11] <smoser> (only in vm-builder trunk)
[16:11] <XiXaQ> _ruben, well, I installed cpufreq-set.
[16:11] <XiXaQ> _ruben, that's cpufrequtils...
[16:12] <XiXaQ> _ruben, but I installed that because of the problems, so that's not the cause.
[16:14] <XiXaQ> giovani, "both"?
[16:14] <kaushal> ttx: hi
[16:18] <XiXaQ> kondemand, of course, it must be that one.
[16:18] <giovani> XiXaQ: both running deamons, and cron jobs
[16:29] <slicslak_> i was running sudo aptitude safe-upgrade over ssh and foolishly forgot to use screen first.  the connection is lost, I was at the step where it was asking confirmation to overwrite config files.  Can I resume?  i am no longer using services for production, so i don't care particularily about /etc/ config files.  could i just reboot?  is it far enough along in the upgrade process?
[16:35] <soren> smoser: Yes, I expect to upload a new vm-builder today, actually.
[16:36] <smoser> ok. good.
[16:36] <smoser> we'll definitely need another rev, ideally before beta
[16:36] <soren> Today.
[16:36] <soren> :)
[16:40] <oversize> soren: the $disk thing in libvirtxml.tmpl is actually a Disk() from disk.py?
[16:46] <smoser> soren, i meant another after today.
[16:47] <smoser> as I'd like to see bugs 423497, 431103, 429169, 429106 fixed.
[16:47]  * zul goes for lunch
[16:49] <kaushal> ttx: checking in again you around ?
[16:49] <ttx> kaushal: yes
[16:49] <kaushal> ttx: did you had a chance to look at that issue ?
[16:50] <ttx> kaushal: no, and there is little chance that I will, at least until Karmic BetaFreeze
[16:50] <kaushal> Bug 433906
[16:50] <ttx> that said, now that the bug is filed anyone else will be able to have a look
[16:50] <kaushal> sure
[16:52] <XiXaQ> I think I figured it out: /etc/init.d/ondemand sets the scaling_governor to "ondemand", no matter what you've configured. It's hardcoded... What a strange thing to do.
[16:55] <XiXaQ> it's even documented: # Short-Description: Set the CPU Frequency Scaling governor to "ondemand"
[17:00] <till_> hey -- i was wondering who i could talk to about: https://bugs.launchpad.net/ubuntu/+source/php5/+bug/420639
[17:40] <zul> till_, yes?
[17:40] <domas> hi!
[17:41] <domas> how does one run a server without having linux-image-debug packages? :)
[17:41] <till_> zul: i was wondering what needs to be done so it would be fixed in karmic :)
[17:42] <zul> till_: me finding some time this week and actually doing it
[17:43] <smoser> soren, what do you think about bug 431255
[17:43] <till_> zul: can i bribe you? ;)
[17:43] <till_> or well, whatever i could do to help
[17:44] <zul> till_: no bribes are needed
[17:45] <till_> zul: that's awesome
[18:08] <jpds> jdstrand: Re: #433316 - apport is running.
[18:10] <jdstrand> jpds: not sure why that isn't working then, but can you provide the other info in the bug?
[18:11] <jdstrand> s/info/info requested/
[18:12] <jpds> jdstrand: Just have. :)
[18:13] <jdstrand> jpds: fyi, this should be in #ubuntu-devel (it isn't server related)
[18:22] <Hypnoz> has anyone been able to pxe boot jaunty livecd or any other livecd?
[18:24] <blistov> bens@karmic-cc:~$ ec2-run-instances emi-39CA160B -n2
[18:24] <blistov> FinishedVerify: Not enough resources: vm instances.
[18:24] <blistov> what's this mean?
[18:25] <SyL> it means what it says, you don't have enough resources.
[18:25] <blistov> vm instances though?
[18:25] <SyL> if you do have enough resources, then I'm guessing the configuration is not working
[18:26] <blistov> Does eucalyptus not have any way of reporting WHICH resources I've not enough of?
[18:26] <SyL> euca-describe-availability-zones
[18:26] <blistov> bens@karmic-cc:~$ euca-describe-availability-zones
[18:26] <blistov> AVAILABILITYZONE        iamacluster     127.0.0.1
[18:26] <SyL> if you run that, it tells you what resources you have availible
[18:27] <blistov> Is this telling me I have ... no resources?
[18:28] <SyL> yes...
[18:28] <blistov> ...
[18:28] <blistov> What have I missed?
[18:28] <SyL> you would normally see 0/2 CPUs etc.
[18:28] <baffle> soren: Did you ever get any further with Cobbler/Ubuntu? I see you set up the page in 2008. :)  https://wiki.ubuntu.com/CobblerSpec
[18:29] <agussman> Quick question: When doing apt-get install, how do I specify a particular entry from my sources.list file for retrieving the package?
[18:29] <agussman> I'm running into a problem where "the same" package is available in different repositories and it's downloading the wrong one
[18:30] <agussman> I guess I could just reorder sources.list, but that seems like cheating
[18:30] <blistov> SyL: I've added a node to my cluster.  Is there something else that needs to be done to make it usable by the cluster?
[18:30] <sub> agussman: You could look into "apt pinning" (google that) or apt-get install package-name=version.number
[18:30] <sub> Oh for that I'd say try apt pinning
[18:31] <sub> Otherwise you can grab the package manually from the repo you want and install it using dpkg
[18:31] <agussman> sub: Thanks for the info, I'll google them
[18:34] <Hypnoz> When i /etc/init.d/networking restart, one of my interfaces doesn't come up, But if i do #ifup eth1 it does. Anyone know if there's a log file I can look into to get some more info?
[18:36] <sub> Hypnoz: Check /etc/network/interfaces for a line similar to "auto <iface>" where <iface> is the name of the interface that isn't coming up on its own
[18:38] <Hypnoz> http://pastebin.com/m22538205
[18:38] <acalvo> how can I reinstall apache2 and its config files (to have the default behavior) -- or uninstall completely apache2
[18:38] <Hypnoz> eth1 is the one that isn't coming up right
[18:38] <Hypnoz> apt-get remove --purge apache2
[18:39] <Hypnoz> remove only leave the config files, --purge cleans config files
[18:39] <acalvo> may I remove the directories too?
[18:39] <blistov> euca_conf --delete-nodes "node1"      does not work.  syntax is apparently incorrect.
[18:40] <blistov> anyone know the correct syntax?  (--help says euca_conf --delete-nodes "host")
[18:41] <Hypnoz> acalvo:  apparently apt-get purge <packagename> is the same. And I think doing that should remove the directories automatically
[18:41] <acalvo> well, it does not
[18:41] <acalvo> I've done it and I've already have it
[18:41] <acalvo> I guess is for all the libapache* packages that are installed
[19:31] <smoser> can someone please verify... i'm needing to fill out MIR for ec2-init (https://wiki.ubuntu.com/MainInclusionEc2-Init) . do i need to fill out MIR for its dependencies ?  do those MIR also need bugs ?
[19:31] <smoser> fyi, its non-main dependencies are python-boto, python-cheetah, python-configobj
[19:34] <m0ej0e> Anyone here know how to activate USB support on 8.04 LTS?
[19:37] <hexa> it is active by default?
[19:37] <KillMeNow> check your BIOS and make sure that USB support is turned on
[19:38] <hexa> ah wait nvm
[19:38] <KillMeNow> otherwise, what are you trying to do with USB?  just extra storage ?
[19:39] <KillMeNow> is it a virtualbox?
[19:39] <m0ej0e> It is Virtualboxed with "Virtualbox" and I need to activate it for extra storage.
[19:40] <mathiaz> smoser: yes - you need to write MIR for the universe dependencies
[19:40] <Pici> m0ej0e: The ose version of Virtualbox in the repositories does not support USB.  You'd need to grab the non-free version from sun/vbox's website to get that.
[19:40] <m0ej0e> Sorry I am very new so If I say something a bit iffy like "Virtualboxed" it is because I don't know a proper term. Sorry for that.
[19:41] <Pici> m0ej0e: And 'non-free' just means that its not open source, its still free to download.
[19:41] <KillMeNow> try this out:  http://www.lostechies.com/blogs/jason_meridth/archive/2008/04/27/using-the-gutsy-gibbon-ubuntu-7-10-non-ose-version-of-virtualbox-with-your-hardy-heron-ubuntu-8-04-install.aspx
[19:45] <till_> zul: hey, re: your email, what do you "mean"? :)
[19:45] <smoser> thanks mathiaz
[19:46] <zul> till_: can you split the bug report into two please?
[19:46] <zul> mathiaz: puppet has been seeded
[19:46] <till_> zul: ah, i can try :P
[19:46] <zul> smoser: yes you do
[19:47] <mathiaz> zul: great - have all the universe dependencies been promoted as well?
[19:47] <smoser> such fun. 6 MIR
[19:47] <zul> mathiaz: I believe so
[19:47] <till_> zul: is there a function, or should i just open another one and link to it?
[19:47]  * till_ is a first timer on launchpad
[19:47] <zul> till_: yes
[19:48] <smoser> zul, mathiaz i'd appreciate your comments in the ec2-init MIR linked to above
[19:48] <till_> ?
[19:48] <till_> =)
[19:48] <zul> till_: you would be opening two bugs one for the dangling symlink and one for the php...actually Ill do it myself
[19:49] <till_> sorry, i didn't exactly know what you needed.
[19:49] <zul> till_: no problem
[19:49] <zul> smoser: you dont need the build dependency stuff in the rationale
[19:49] <smoser> its in the template.
[19:50] <smoser> oh.
[19:50] <smoser> i see.
[19:50] <zul> its a formatting error i think
[19:50] <smoser> i'm not asking for promotion because its a build dependency
[19:50] <smoser> thats what it is for
[19:50] <zul> k
[19:51] <zul> smoser: I think the first part is good enough
[19:57] <vxnick> is there a quicker way to erase data from LVM during the install process?
[19:57] <giovani> vxnick: what do you mean, exactly? erase data from LVM?
[19:58] <vxnick> giovani: I'm deleting some old partitions which are LVM in order to install on to this HDD
[19:58] <vxnick> giovani: this is in the installer
[19:58] <giovani> ok
[20:00] <vxnick> it's only a 250GB drive, but taking absolutely ages to delete - says it's going to write random data to, which I don't really need in this case
[20:01] <m0ej0e> Pici Would it be ok if I PM you for some more information?
[20:02] <Pici> m0ej0e: You can, but I'm not sure I'll be able to help.
[20:04] <giovani> vxnick: well that's separate from deletion
[20:04] <giovani> vxnick: there may have been an option in the installer to not do a full wipe (there should be)
[20:04] <vxnick> giovani: I'll have a look, thanks for your help
[20:05] <vxnick> standard installer or advanced?
[20:07] <giovani> both should have an option to not do a full wipe
[20:07] <vxnick> gotcha, thanks
[20:08] <giovani> there's virtually no reason to do a wipe on a personal machine just because you're reinstalling -- in a business environment, or before you discard a drive, wiping is a very good idea
[20:25] <Belloto> hi ... I have a 8.04 ubuntu server and I mounted a windows shared folder using smbmount //192.168.1.5/shared /mnt/winfolder -o username=Belloto
[20:26] <soren> smoser: I don't think it's generally a very good idea to adopt other people's hysterical raisins. Just because Eric put it some arbitrary place doesn't mean that we have to.
[20:26] <soren> smoser: S71 might be a great place. I just want to choose it because it's a great place, not because "that's what Eric did".
[20:26] <Belloto> then I enter sudo password and user password, and I got it running properly ... but later I use to get "CIFS VFS: No response for cmd ..." errors when some programs try to access files at /mnt/winfolder
[20:27] <soren> smoser: An argument could be made for S99, since that's where rc.local runs, which has a bit of the same feel to it as the user-data script.
[20:27] <Belloto> so, I can browse files but for some reason samba is not running properly
[20:27] <smoser> i think it reasonable to have it run very late.  although, as I've stated, i think it should run very early
[20:27] <soren> smoser: Heheh :)
[20:27] <Belloto> any clues?
[20:27] <smoser> personally, i think the user should have the option to run things in the initrd
[20:28] <m0ej0e> Server 8.04 LTS need help getting USB devices running. OS sees a printer, but not a USB flash drive.
[20:28] <soren> smoser: Fascinating proposition.
[20:28] <smoser> but for now, that is not acheivalbe in karmic
[20:28] <soren> smoser: Right.
[20:28] <smoser> soren, fwiw, i have a "debug initrd" (currently only i can launch it) that allows hooks from the root filesystem.  its good for debugging.
[20:28] <m0ej0e> ls
[20:28] <soren> smoser: I can imagine :)
[20:29] <smoser> do you think there is much potential for regression by moving the init script to run later ?
[20:29] <soren> m0ej0e: file not found
[20:29] <soren> smoser: Yes.
[20:29] <soren> smoser: Do I care? I'm not sur.e
[20:29] <smoser> well, your comments would be appreciated in that bug
[20:29] <soren> smoser: Well, "much" is a strong word.
[20:29] <soren> smoser: sure, sure, I'm just rambling a bit here first :)
[20:29] <m0ej0e> soren: What do you mean?
[20:30] <soren> m0ej0e: It was an attempt at a joke. (as a response to your "ls")
[20:30] <soren> smoser: do you happen to have an instance running right now?
[20:30] <m0ej0e> soren: Oh lol. I require more coffee it seems.
[20:31] <smoser> soren, MIR requests "Who is the package bug contact in Ubuntu?"
[20:31] <smoser> what would the answer be to that for a.) ec2-init b.) euca2ools.  c.) where do i find that generally?
[20:32] <smoser> soren, i do not. i've tried to be mroe careful as I have a bout 5 day worth of un-used amazon ec2 instances this month :)
[20:34] <soren> smoser: You're generally expected to know things like that when you're writing the MIR. ec2-init would be me (and you, probably). euca2ools would probably also be you and I, and hopefully upstream.
[20:34] <soren> smoser: Yeah, I accidentally left 9 running when I left St. Louis. Didn't turn them off until 3-4 days later. Doh.
[20:34] <soren> smoser: Launchpad will show you the bug contacts, though.
[20:34] <soren> smoser: Gimme a sec.
[20:35] <smoser> soren, i was wondering if that was more a question of "what developers get bugs eamiled to them for this package"
[20:35] <smoser> yeah, thats what i meant.
[20:35] <soren> smoser: https://edge.launchpad.net/ubuntu/+source/ec2-init
[20:35] <soren> smoser: Fourth box from the top in the right column.
[20:36] <smoser> and soren, if you want to do the MIR since you're more knowledgeable, feel free to take over ;-)
[20:36] <smoser> thanks
[20:36]  * soren whistles innocently and looks the other way
[20:41] <soren> smoser: For our images, S71 and S99 makes virtually no difference. http://pastebin.com/m6085f7f7
[20:42] <danage> anyone here working on home folder encryption? i have a serious bug to report but dunno how to fetch the necessary information and which pakage to attach it to (sorry for cross posting)
[20:42] <smoser> soren, do you know when apache runs ?
[20:42] <smoser> never mind.
[20:42] <soren> danage: It's much easier to just ask your question that to ask for someone who knows the answer to a question that has not yet been asked.
[20:43] <smoser> i'm thinking that it makes sense to run user data before likely-configured-by-user-data services ran
[20:43] <smoser> (and that, i thought, apache)
[20:43] <smoser> but who knows.
[20:43] <soren> smoser: I was thinking the same thing.
[20:44] <soren> smoser: The slight twist is that -- with our images -- the user-data script will by definition run before these services, because it's going to be the user-data script that installs them.
[20:44] <smoser> thats why i think the first hook is initrd. then, you run before everything....
[20:44] <smoser> soren, well, yes. that is true.
[20:45] <soren> danage: In other words: Just ask your question. If someone knows the answer, they'll respond.
[20:45] <smoser> thats actually very interesting though
[20:45] <danage> ok so here is the thing: i installed karmic alpha with home folder encryption. after an update (about a week or so), the system hangs at boot - it cannot load some encryption component. the system is unusuable and halts at a console.
[20:45] <smoser> it means that the earlier the user-data runs the higher likelyhood there is of doing an 'apt-get install server-that-shouldnt-run-till-later'
[20:45] <danage> i want to report that on launchpad, but a) what package and b) how do i extract the console output
[20:46] <soren> danage: Ok. File a bug against something ecrypts-like. If that's not where it belongs, whoever looks at it will know better where to file it and will take it from there.
[20:46] <soren> danage: Sorry, ecryptfs-like. I type quite porrly today.
[20:46] <soren> Err... poorly.
[20:46] <soren> Case in point.
[20:46] <danage> ok thanks soren, n/p i got the idea :)
[20:46] <soren> danage: As for the console output, take a photo.
[20:47] <soren> danage: Digital camera, camera phone... Whatever.
[20:47] <danage> soren: meh i feel stupid now, shoulda figured....
[20:47] <soren> danage: Oh, don't.
[20:47] <soren> danage: Virtually noone thinks of that. :)
[20:48] <soren> smoser: Oh.. hm... very interesting point.
[20:52] <soren> smoser: Ok, AFAICS, there's around 3 places we can put this: a) As soon as network (actually the meta-data service) is available, or b) S71-ish or c) S99-ish.
[20:52] <soren> smoser: (Interrupt as soon as there's something you disagree with)
[20:53] <soren> smoser: Benefits of c) (and to a lesser extent b)): You're reasonably sure that any dependency on other services will be fulfilled.
[20:54] <smoser> i would say there is the ability to do it before a.) though (requiring manual bringup of network device)
[20:54] <soren> right.
[20:54] <smoser> but i agree otherwise.
[20:54] <soren> I don't think it's practical at this point, really, but ok, sure, let's add that option as well for completeness.
[20:55] <soren> Are there other benefits?
[20:55] <soren> (of S99-ish run time)
[20:55] <soren> It matches rc.local.
[20:55] <soren> ...which is also place for local customisations of sorts.
[20:57] <soren> Downside: Changing configuration of services(*) takes longer since it will often require restarts that would otherwise have been avoided.
[20:57] <smoser> right.
[20:57] <soren> (*): This only really applies to SSH for our images, though, since that's the only service running in the images.
[20:57] <smoser> and means a severe security hole in said service cannot be completely plugged by a user-data script
[20:58] <soren> Same disclaimer applies.
[20:58] <soren> But yes, also a good point.
[20:59] <soren> Hm... Similarly, if you want to mount an EBS volume for storing your database or whatnot, it will require stopping services, perhaps moving files around, and mounting the volume, and starting the service again.
[20:59] <soren> It's very similar to the configuration issue, but slightly different.
[21:00] <smoser> i really think that at this point in cleverness of user data scripts, the best solution is probably rc.local time.
[21:01] <soren> At least different in the way that ec2-init has (rudimentary) support for reading EBS volume descriptions from the user-data field and makeing use of that much earlier in the process. If that were to become widespread, this difference becomes significant.
[21:01] <smoser> but, as you pointed out, the difference between that and S71 is small to none
[21:01] <soren> smoser: In our images, yes.
[21:01] <soren> smoser: If people have pre-bundled images with a bunch of services (a lot of which run at S90-ish or later), I'd say we want to go before that.
[21:04] <aubre> In 1.6, after I upgraded the version of eucalyptus on both my front-end and my node controller, when I do $euca-describe-availability-zones verbose  I get AVAILABILITYZONE	clustername	127.0.0.1 when I expect to also see the types of instances I can run. When I try to do a
[21:04] <aubre> euca-run-instances emi-xxxxxxxx -k key I getRESERVATION	None	None
[21:04] <danage> ok behold, here is my bug report: https://bugs.launchpad.net/ecryptfs/+bug/434238
[21:05] <aubre> I then try euca-describe-instances and nothing is returned.
[21:05] <smoser> soren, i dont know. i think its really arguing about things we dont' know a lot about.
[21:05] <soren> aubre: I don't think this channel has developed sufficient proficiency with Eucalyptus to generally be the best place to ask these questions, I'm afraid. I recommend #eucalyptus. Some of also hang out there. :)
[21:05] <aubre> soren oh ok sorry.
[21:06] <smoser> so if we have ec2-init running early, and user data running later, we'll have to have another init script, correct?
[21:06] <soren> smoser: all the more reason to research, analyse, and discuss rather than "just doing whatever someone else has done for a while, which seems to have been a completely horrible decision". :)
[21:06] <soren> Err..
[21:06] <soren> Seems to have NOT been a completely horrible decision :)
[21:07] <soren> smoser: Either that or some kind of conditional based on $0. I don't think that's worth it, though.
[21:07] <aubre> I'm happy I can package images, before when I tried to do it in 1.52 it kept invalidating my kernel.
[21:08] <soren> aubre: Oh, don't apologise. I understand why you ask here, and you really should be able to, but I just think -- at this point -- you'll be better helped in #eucalyptus.
[21:09] <aubre> soren: gotcha. I asked them, and I've done a bug report.
[21:09] <smoser> soren, i think i would prefer 2 init scripts.  ec2-init and ec2-user-data.  the first pulling the user data and writing it to a location that the second would consume.  since the first is already hitting the metadata service i'd like to alleviate the second from doing so. of course the second somehow has to get its "my-ami" for run-once-per-ami
[21:09] <aubre> soren: I just need to be a little more patient :)
[21:10] <soren> smoser: We should totally cache that as well.
[21:13] <smoser> soren agreed. i'm just hesitant in any changes at this point.
[21:13] <soren> smoser: Good man :)
[21:13] <smoser> because we keep regressing
[21:14] <soren> smoser: Perfectly fine. We'll make a note of it and make it happen for Lucid.
[21:20] <m0ej0e> How do you disable the floppy drive for the server?
[21:20] <soren> m0ej0e: You... have... a floppy drive?
[21:21] <giovani> m0ej0e: define "disable"?
[21:21] <zul> man after my own heart
[21:22] <jart> is it possible to put "HZ=1000" as a boot option in my grub config?
[21:22] <aubre> m0ej0e: have you tried disabling the floppy drive controller in the bios? or you could always disconnect the cable
[21:23] <aubre> m0ej0e: if you disable the floppy controller in the bios it shouldn't be detected when you boot the system
[21:24] <m0ej0e> I am running the server in virtualbox. Within Virtualbox I have disabled the floppy drive, and no the computer does not even have one on it.
[21:24] <giovani> m0ej0e: if you disabled it in virtualbox ... then it's not there
[21:25] <m0ej0e> Then why do I still get an i/0 error?
[21:25] <m0ej0e> *I/O
[21:25] <giovani> m0ej0e: because you haven't rebooted?
[21:25] <giovani> you shouldn't disconnect hardware like that while the system is running
[21:25] <giovani> it's not how-swappable
[21:25] <m0ej0e> giovani: I have rebooted.
[21:49] <drgonz> hey guys, just wanted to try and ask you guys for some help
[21:50] <addisonj_> grrr... i just started a new job, last sys admin did everything with arch linux...
[21:51] <drgonz> I just finished compiling a 2.6.30.7 kernel for a 8.04 LTS server in virtualbox, and I have a "end_request: I/O error, dev fd0, sector 0" and "o	SMBus base address uninitialized – upgrade BIOS or use force_addr=0xaddr" any advice?
[21:51] <addisonj_> call my crazy, but arch does not seem the ideal choice for stability and reliability in an enterprise server situation, that could just be my
[21:53] <giovani> drgonz: hahaha
[21:53] <giovani> you are likely either m0ej0e, or a classmate of his
[21:53] <giovani> since you just asked the same question
[21:53] <drgonz> hahahaha
[21:54] <giovani> and you both came from the same IP range
[21:54] <drgonz> hes my roomate
[21:54] <drgonz> hes standing right here too
[21:54] <giovani> why not follow the advice you were given in the paste
[21:54] <drgonz> what paste?
[21:54] <drgonz> im not his rig
[21:54] <giovani> "end_request: I/O error, dev fd0,  sector 0" and "oISMBus base address uninitialized – upgrade BIOS or use force_addr=0xaddr"
[21:55] <drgonz> yah, tried that, thats y im here
[21:55] <giovani> is the floppy device enabled in virtualbox?
[21:55] <giovani> you tried what?
[21:55] <drgonz> no, the floppy isnt enabled, and im tried to enter force_addr=0xaddr, but it does nothing
[21:55] <giovani> enter where?
[21:56] <drgonz> CLI
[21:57] <giovani> sigh
[21:59] <drgonz> so, u got any advice?
[22:01] <drgonz> or anyone else? could use some help for this project
[22:01] <drgonz> prof wants a no error boot
[22:03] <addisonj_> don't use virtual box :P
[22:04] <drgonz> lol
[22:04] <drgonz> i wish it was that easy
[22:06] <drgonz> ive tried google the errors, sifted through tonnes of pages, but no fixes, just the fact that its a problem
[22:12] <SyL> drgonz: why do you need to use virtual box?
[22:13] <drgonz> for the prof
[22:13] <drgonz> we need it so we can work on it either at the lab or at home
[22:13] <drgonz> or if he ants the image to check it
[22:14] <drgonz> we don't have a common drive to store/retrieve our image
[22:45] <soren> smoser: Hm... That vmbuilder upload will have to wait until tomorrow.
[22:45] <soren> First thing.
[23:42] <benc>  is there a tutorial on how to write init script for daemon?
[23:55] <KillMeNow> http://www.cyberciti.biz/tips/linux-write-sys-v-init-script-to-start-stop-service.html
[23:55] <KillMeNow> all kinds of tutorials