[00:00] <erichammond> -X preserve extended attributes
[00:01] <erichammond> I'm a huge rsync fan, but there were enough special cases, that I start to wonder if I got them all.
[00:02] <erichammond> A generic tool like this should do the right thing for all edge cases users are likely to come up with.
[00:09] <mdlueck> Has anyone ever heard of start_daemon not backgrounding processes? Suddenly one program trying to use that to start its daemon find the script hanging at the call to start_daemon and the daemon is responsive.
[00:10] <RoAkSoAx> zul, found why I couldn't run the hook and now I can, and it was something related to the apt cache, but now im doing something different to test the hooks and works including the staging
[00:11] <zul> RoAkSoAx: sweet
[00:11] <RoAkSoAx> anyways i'll work on it/test and then bother you for review :)
[00:21] <zul> coolio
[00:54] <sabgenton> what is the command to rebuild device.map
[00:54] <jpds> soren: update-grub ?
[00:54] <jpds> sabgenton: ^--.
[00:55] <smoser> erichammond, i'm just familiar with rsync, and its one command, not 2. so i don't have to check both sides of the pipe
[00:55] <smoser> but, regarding sparse, and acl, and hard links you're definitely right.
[00:55] <erichammond> smoser: That's fine. Just make sure you pack in enough options to really duplicate the contents.
[00:55] <smoser> yeah
[00:55] <sabgenton> grub 2 is a pain it didn't make the autodected boot entrys right
[00:56] <sabgenton> now I have grub entrys that are useless
[00:57] <sabgenton> have to make one in /etc/grub.d/40_custom I guess
[00:59] <sabgenton> oh wait I think it looks at fstab
[00:59] <sabgenton> and it was wrong at the time
[00:59] <sabgenton> will try again
[02:52] <zoran119> hi
[02:53] <zoran119> i have a ubuntu server (8.04 lts) running as a vm on hyper-v and almost every day the clock gets stuck in a 5 second loop...
[02:53] <zoran119> any idea what could be causing the problem?
[02:56] <cef> zoran119: are you running ntpdate or ntpd int hat vm? vm's usually don't need to sync their time, just the host
[02:56] <cef> zoran119: it's the only thing that comes to mind at the moment
[03:40] <maxagaz> how can I can kill such process: root     19294  0.0  0.0      0     0 ?        S    Feb28   0:00 [nfsv4-svc]
[03:41] <twb> maxagaz: that's a kernel thread.
[03:41] <maxagaz> knowing that kill -9 doesn't work
[03:41] <twb> The square brackets mean "I'm the kernel"
[03:41] <twb> (IIUC)
[03:41] <maxagaz> twb, okay
[03:41] <maxagaz> twb, what can I do then ?
[03:42] <maxagaz> twb, it was set by someone else...
[03:42] <twb> rmmod nfs or something, might get rid of it.
[03:42] <twb> Why do you need to get rid of it?
[03:42] <maxagaz> twb, I suspect it to be responsible of lags on the machine
[03:43] <maxagaz> rmmod nfs
[03:43] <maxagaz> ERROR: Module nfs is in use
[03:46] <twb> maxagaz: why do you suspect that?
[03:47] <maxagaz> actually, it seems better since I've unmounted to directory in /tmp
[04:30] <twb> I have a network of 8.04 clients and an 8.04 server.  Users and groups (above 1000) are centralized in an OpenLDAP RFC 2307 database.
[04:30] <twb> Now, I need all users to have access to floppies, i.e. to be in the floppy (25) group.
[04:30] <twb> How do I do that?  The floppy group isn't in LDAP, it's in flat files on each host.
[04:35] <persia> twb: Fiddle /etc/nsswitch.conf and stick it in LDAP?
[04:37] <twb> Thinking laterally, I wonder if I could just use a udev rule to "chgrp users /dev/fd0".
[04:38] <twb> Any old bodge will suffice, but for my own sanity I'd like it to be the least-worst bodge.
[04:38] <persia> Well, of course.
[04:38] <persia> udev rule either means manual management or a local package.  LDAP swich means manual management *and* LDAP config.  Which is less headache?
[04:40] <twb> I should add that the client machines are all netboot from a single image
[04:47] <axisys> i have two disks mirrored .. looking for procedure to upgrade to next release, for example lucid when it is available in apr, and failover back to older release..
[04:48] <persia> twb: That's why "local package".  You deploy something that adds an entry in /etc/apt/sources.list.d/ and includes your rule (or split in two packages if you like), and then you can update all the machines by tweaking that package, rather than by changing all the netboot images.
[04:48] <axisys> i solaris i break the mirror, while system is running, upgrade the in inactive disk - the detached one from mirror, and reboot from that ..
[04:48] <axisys> s/i solaris/in solaris/
[04:49] <axisys> if the upgrade breaks any os, i switch back to the orig disk
[04:49] <axisys> so one reboot for each is the only outage
[04:49] <axisys> is there something similar available w/ ubuntu server ?
[05:59] <Roxyhart0>  hi there i am trying to install moodle, i just download the files .tgz but when i try to uncompress using the command tar -zxvf moodle-weekly-19.tgz i got error that the file is not format gzip. any idea?
[06:04] <ivoks> file moodle-weekly-19.tgz
[06:06] <Roxyhart0> yes, i just see i can do apt-get moodle, but i am not sure if it can work fine. anyone have experience installing moodle?
[06:06] <ivoks> i don't see why it wouldn't work
[06:09] <Roxyhart0> im trying now thanks
[06:15]  * persia has recently seen a number of .tgz files that were actually uncompressed tarballs.
[06:19] <ivoks> probably web server is doing decompressing
[06:20] <twb> ivoks: if you mean when clicking in firefox, it'll be firefox, not the server sie
[06:20] <twb> *side
[06:20] <ivoks> you think?
[06:20] <twb> ivoks: because of Accept-Encodings: gzip
[06:20] <twb> Firefox thinks the file coming back is gzipped only because it said it supported gzip
[06:20] <ivoks> hm... yeah
[06:21] <twb> You won't see that fail with e.g. wget
[06:21] <twb> ivoks: oh, and also because the server returns application/x-gzip or whatever for .tar.gz
[06:21] <ivoks> true
[06:22] <ivoks> lucid is looking macish
[06:22] <persia> Actually, in at least one case, the decompression had happend prior to it being posted on the website (I spent some time investigating with the person posting the artifact).  Probably the same cause, but don't always blame your browser.
[06:23] <twb> I was just commenting on server vs. browser
[06:23] <ivoks> does w3m supports it?
[06:23] <twb> w3m sends Accept-Encodings: gzip ..., but midori doesn't
[06:23] <ivoks> probably, /lib/libz.so.1
[06:24] <twb> I know this because occasionally my proxy (polipo) will feed the gzipped file to midori and I get garbage onscreen
[06:37] <skrite> can someone reccomend an easy how to for getting postfix to allow a user to send via smtp from a computer not on our lan?
[06:37] <skrite> having an awful time with it
[06:38] <twb> skrite: you want SMTP/SSL on the submission port (587).
[06:38] <twb> I don't know the details
[06:39] <twb> That or become an open gateway and hope for the best until you get RBLd.
[06:39] <skrite> twb, i think i need port 25
[06:39] <skrite> RBLd?
[06:39] <twb> blacklisted
[06:39] <skrite> ah, ok
[06:40] <skrite> then yes !
[06:40] <twb> http://en.wikipedia.org/wiki/Message_submission_agent
[06:41] <skrite> thanks
[07:36] <skrite> this is insane
[07:37] <skrite> i have gone through the examples on the ubuntu wiki about a half dozen times, and still not getting it to work
[07:37] <skrite> just doesn't seem like this should be this difficult
[07:37] <|Mike|> what
[07:38] <|Mike|> struggling with sasl ? :D
[07:39] <skrite> yeah, still struggling
[08:07] <jiboumans> morning
[08:10] <soren> o/
[08:11] <skrite> hey
[08:17] <skrite> do i need to have the full domain name as my host name?  like  machine.mydomain.com ?
[08:17] <lucid_interval> skrite: do you mean for postfix?
[08:17] <skrite> yes
[08:18] <lucid_interval> skrite: you need to tell postfix which hostnames to "receive mail for" - i.e. consider as local machine
[08:18] <lucid_interval> this can be one or more hostnames - can include your FQDN (machine.example.com)
[08:19] <skrite> all the ones listed as mydestination?
[08:20] <lucid_interval> skrite: what is mydestination?
[08:20] <lucid_interval> skrite: yes. whatever is listed as mydestination
[08:21] <lucid_interval> but your hostname is ONE of those values.
[08:21] <skrite> ok
[08:21] <lucid_interval> skrite: and you MIGHT list all of the values under mydestination in /etc/hosts
[08:22] <lucid_interval> skrite: in the simplest of cases, your hostname = machine.example.com, mydestination will be just that
[08:24] <skrite> ok. i am a little confused here.  The machine name with the prefix of the domain.   like our machine is  server.mydomain.com  so when i log in, my prompt looks like server #   however, the mail domain is mail.ourdomain.com
[08:24] <skrite> the hostname of the machine is server
[08:25] <lucid_interval> skrite: then your case is a little more involved. mydestination should include ALL the hostnames in the SMTP context that this postfix instance will consider local
[08:26] <skrite> and the same with the /etc/hosts ?
[08:26] <lucid_interval> skrite: strictly speaking it has NOTHING to do with /etc/hosts or DNS or hostname
[08:26] <skrite> ok
[08:27] <lucid_interval> skrite: if an incoming SMTP connection has a To/CC/BCC email address that is _NOT_
[08:27] <lucid_interval> skrite: one of the values in mydestination, that is considered non-local and the
[08:27] <lucid_interval> skrite: request is effectively a relay request. All email addresses that end in values in mydestination
[08:28] <lucid_interval> skrite are considered "local" and therefore _NOT_ relay requests
[08:28] <skrite> whoa, so i don't want to include them if i am trying to set up an account client that is not within the LAN?
[08:29] <skrite> not on the local network or local machine, i mean.
[08:29] <lucid_interval> skrite: don't understand what you're asking. I was referring to what is considered a relay
[08:29] <skrite> the user has a username and password and a Maildir on the machine, but needs to connect from a mail client not on the same machine
[08:30] <lucid_interval> skrite: request versus request for "local" delivery. There are separate constraints / checks
[08:30] <lucid_interval> skrite: around whether relaying is ALLOWED or not - e.g. AUTH
[08:30] <skrite> ok
[08:30] <skrite> think i get it
[08:31] <lucid_interval> skrite: I get your case now. User has a _LOCAL_ account. But connects with a mail client
[08:31] <skrite> right
[08:31] <lucid_interval> skrite: that will make SMTP requests from a remote network location
[08:31] <skrite> yes
[08:31] <lucid_interval> skrite: presumably some of those requests will be to send email to _OTHER_ users on this local
[08:32] <skrite> actually, no.
[08:32] <lucid_interval> skrite: mail domain (to be "locally" delivered) while other SMTP requests will be to relay emails
[08:32] <lucid_interval> skrite: to other mail domains.
[08:32] <skrite> the mail account is actually for a machine that will only need to send via relay. to other domains
[08:32] <lucid_interval> skrite: the mydestination piece specifies what is considered "local" versus "foreign" (non-local)
[08:33] <lucid_interval> skrite: ok. so case 1 may have ZERO requests. you still need to instruct postfix as to what is for "local"
[08:33] <skrite> ok, so what it considers a relay or not
[08:33] <lucid_interval> skrite: delivery versus "relay". If NOTHING is ever lcoally delivered to users on this box, you should
[08:34] <lucid_interval> skrite: be setting up this postfix as a "smart relay" to relay all mail to the main SMTP host.
[08:34] <lucid_interval> skrite: in that case, this client can also (probably should also) relay via _THAT_ main SMTP host
[08:36] <skrite> ok testing
[08:43] <skrite> what log can i look in to see why smtp is not happening right? mail.log does not show anything when i try to send via smtp
[08:47] <sherr> skrite: what mail server?
[08:49] <skrite> the smtp server
[08:49] <lucid_interval_> skrite: did you finish testing?
[08:51] <skrite> well, sort of, all i am getting from a client is that unable to send check with your network administrator, yadda yadda
[08:52] <skrite> check with the network administrator? i AM the network administrator !
[08:52] <lucid_interval_> can you try sending email _locally_ - to see postfix is accepting incoming mail for local delivery?
[08:53] <skrite> yes, that is working
[08:53] <lucid_interval_> then check from the remote client sending email that should be considered local - for local delivery on the postfix box
[08:53] <lucid_interval_> the final part is selectively enabling relaying based on AUTH (since you don't want to be an open SMTP relay)
[08:53] <skrite> ok
[08:55] <skrite> you mean test from like user1@mydomain.com to user2@mydomain.com?
[08:56] <lucid_interval_> skrite: from can be anything. to should be something like user2@mydomain.com - where mydomain.com is in the mydestination list
[08:56] <lucid_interval_> in fct you should test all teh different values in the mydestination list
[08:56] <skrite> no, i am getting a login failed
[08:59] <lucid_interval_> skrite: so here is where the AUTH part comes in. Did you finish setting up the STARTTLS?
[08:59] <skrite> yes, i think so.
[08:59] <lucid_interval_> skrite: Are you planning on using saslauth directly with postfix or authenticate through dovecot?
[09:00] <skrite> i am good with either, really, dovecot is our pop client and is working
[09:01] <lucid_interval_> skrite: ok. when you do a 'telnet localhost 25' on the postfix box and then type 'ehlo localhost' at the SMTP prompt,
[09:02] <lucid_interval_> skrite: do you see STARTTLS and AUTH?
[09:03] <skrite> I see STARTTLS but  the lines with AUTH are   AUTH PLAIN LOGIN and AUTH=PLAIN LOGIN
[09:04] <lucid_interval_> skrite: cool. that works.
[09:04] <lucid_interval_> skrite: take a look atthis page to setup the sasl through dovecot:
[09:04] <lucid_interval_> skrite: https://help.ubuntu.com/community/PostfixDovecotSASL
[09:04] <skrite> ok
[09:05] <skrite> thanks for all this, by the way
[09:06] <lucid_interval_> skrite: cool. you're welcome. lots of ppl have helped me too - not least by writing all this awesome FOSS
[09:07] <skrite> :)
[09:11] <\sh> hmm....libjs-query as default install pkg on server flavour? did I miss something? ;)
[09:13] <\sh> bah...python-apt recommends libjs-jquery..wth
[09:23] <sabgenton> in grub2
[09:23] <sabgenton> how does grub-mkconfig know what partion to assign to  root=bla
[09:23] <sabgenton> ?
[09:24] <soren> ttx: I don't know if I'm all alone in this, but I really don't care about eucalyptus bug mail. At all. Yet I get tons of it, because ubuntu-virt is subscribed to it.
[09:25] <soren> ttx: How would you feel about a ubunt-uec team that is subscribed to all things UEC?
[09:26] <soren> ttx: That team, in turn, could be a member of ubuntu-virt (and not the other way around).
[09:26] <ttx> soren: makes sense, maybe send mail to ubuntu-virt members to propose the disassociation ?
[09:27] <soren> ttx: The team consists of all Canonical folks and me and Rick.
[09:28] <soren> ttx: I think it's a matter of jfdi.
[09:28] <ttx> soren: I just don't want people (if any) relying on it to get euca bugmail to be surprised by the move
[09:28] <soren> ttx: Ok, how about this:
[09:28] <soren> ttx: You create ubuntu-uec..
[09:29] <soren> ttx: ..subscribe that team to all the relevant packages (eucalyptus, euca2ools, axis2c, rampart, all the java stuff, etc.)..
[09:29] <soren> ttx: ...I'll e-mail all the members of ubuntu-virt telling them that if they really do care about UEC bug mail, they should join that other team.
[09:30] <soren> ttx: You may want to keep all the java stuff separate.
[09:30] <ttx> soren: ubuntnu-server is already subscribed to those packages... so I don't really see the need for a ubuntu-uec team, tbh
[09:30] <soren> ttx: Oh.
[09:30] <soren> ttx: You are quite right.
[09:31] <soren> Everyone in ubuntu-virt is already ubuntu-server members.
[09:31] <soren> Oh, but they get that bug mail differently.
[09:31] <ttx> soren: yes, that allows them to filter... that's why I prefer to warn about the move before doing it
[09:31] <soren> I'm just fed up to the point where I was about to leave ubuntu-virt, but I really don't want to do that. I feel very close to that team :)
[09:32] <ttx> soren: it all boils down to the purpose of that team
[09:33] <ttx> soren: as the bug supervisor for a number of virt packages, that puts it outside the regular ubuntu-server processes as far as bug review is concerned
[09:33] <ttx> so we discussed moving the virt packages into ubuntu-sever as well
[09:33] <soren> It was meant to consist of people who cared about virtualisation in Ubuntu. It was never really decided whether this membership was supposed to come with special privileges (write access to bzr branches or whatnot), or if it was just a handy way to get more bug mail.
[09:36] <ttx> soren: if everyone agrees the purpose of the team is about key virt packages and not their downstream application (UEC), then yes, your move makes sense
[09:36] <ttx> uec packages are already taken care of in ubuntu-server, so they wouldn't get lost
[09:38] <skrite> lucid_interval_, YOU ROCK !   successful test after setting up with dovecot auth, thanks !
[09:39] <lucid_interval_> skrite: awesome :-)
[09:40] <lucid_interval_> skrite: now if you want to know a LOT more about postfix - check out this presentation: http://www.arschkrebs.de/slides/Postfix_Configuration_and_Administration-handout.pdf
[09:41] <skrite> cool, thanks. certainly will do.
[09:41] <lucid_interval_> skrite: and if you ever want to setup a CLIENT postfix server that is allowed to relay based on a TLS certificate: check out this presentation: http://www.state-of-mind.de/assets/postfix_tls.pdf
[09:42] <skrite> got it
[10:44] <jayvee> so I'm running Ubuntu 9.10 here
[10:45] <jayvee> what's the best way to get Xen working?
[10:45] <jayvee> there isn't any xen-enabled kernel in the repo
[10:45] <jayvee> is xen considered edwardian or something?
[10:48] <soren> jayvee: Xen is not really supported in Ubuntu. Try KVM instead.
[10:48] <jayvee> That's what I've been using, but I'm quite unimpressed with its performance.
[10:48] <jayvee> This copy of Ubuntu has been installing for the past 30 minutes.
[10:49] <soren> Then you're probably not using kvm.
[10:49] <jayvee> module kvm_amd is loaded, and /usr/bin/kvm is running
[10:49] <jayvee> -enable-kvm
[10:49] <soren> jayvee: How are you running kvm? From a command line?
[10:50] <jayvee> within libvirt
[10:50] <soren> jayvee: Through libvirt?
[10:50] <soren> jayvee: Ok.
[10:50] <soren> jayvee: Put the output of "ps -ef | grep kvm" on pastebin
[10:52] <jayvee> soren: http://pastebin.ca/1823647
[10:53] <soren> jayvee: now "sudo ls -l /proc/26559/fd
[10:53] <soren> "
[10:53] <soren> (without the spurious newline, sorry)
[10:53] <jayvee> This is a desktop install, btw. It's going to be a group of NoMachine servers.
[10:54] <soren> A desktop install? With 512 MB of RAM?
[10:54] <soren> That's going to be slow even on real hardware.
[10:54] <|Mike|> soren: are you from sweden btw?
[10:54] <soren> |Mike|: No.
[10:54] <soren> |Mike|: Denmark
[10:54] <jayvee> soren: http://pastebin.ca/1823649
[10:54] <|Mike|> Close tho, your nick sounds pretty familair, as if i seen it before *somewhere*
[10:54] <jayvee> I shut it down and started it, so the PID changed. :)
[10:55] <soren> jayvee: Well, any further diagnostics is unnecessary: 512 MB of RAM desktop installs will always be slow.
[10:55] <jayvee> hmm, runs perfectly fast in VirtualBox
[10:55] <jayvee> I'd use ubuntu-vm-builder to deploy this stuff, except it keeps crashing.
[10:56] <soren> jayvee: With 512 MB it runs fast in VirtualBox? Seriously?
[10:56] <jayvee> yep
[10:56] <ttx> |Mike|: soren is a famous golfer.
[10:56] <iLLiZT> Nothing runs fast on 512 MB
[10:56] <jayvee> heck, I even bump it down to 256 MB once it's finished installing
[10:57] <iLLiZT> What's wrong with you?
[10:57]  * soren smacks ttx around
[10:57] <|Mike|> ttx: lol
[10:57] <iLLiZT> :)
[10:57] <ttx> |Mike|: it's true !
[10:57] <soren> jayvee: Once it's done installing is a completely different story.
[10:57] <jayvee> agreed
[10:58] <jayvee> So you reckon it's a memory thing more than anything?
[10:58] <soren> jayvee: At that point, you don't have a multi-GB disk image as a ramdisk.
[10:58] <iLLiZT> KVM is quite awesome anyway
[10:58] <ttx> Google says he is, it /must/ be true.
[10:58] <iLLiZT> I'd rather go with KVM than Xen.
[10:58] <soren> jayvee: Oh, yeah. Totally.
[10:58] <iLLiZT> But Virtualbox is pretty nifty if you want hosted virtualization
[10:58] <henkjan> jayvee: i'm running a Xen dom0 on debian lenny
[10:58] <soren> iLLiZT: What is "hosted virtualisation"?
[10:58] <henkjan> jayvee: with ubuntu domu's
[10:58] <iLLiZT> soren: Virtualbox, VMware Workstation, etc.
[10:59] <soren> iLLiZT: How is that different from Xen or KVM?
[10:59] <|Mike|> eeuw.
[10:59] <soren> iLLiZT: What about KVM is not "hosted virtualisation"?
[10:59] <soren> I've never heard that term.
[10:59] <iLLiZT> KVM is hosted virtualization as well.
[10:59] <soren> What is /not/ hosted virtualisation, then?
[10:59] <iLLiZT> It's mainly what VMware uses to describe what's not a bare-metal hypervisor.
[10:59] <iLLiZT> VMware, Citrix, Microsoft etc.
[10:59] <soren> What does that even mean?
[11:00] <FireCrotch> soren: it's a VMWare buzz-word
[11:00] <iLLiZT> A bare-metal hypervisor is virtualization without an underlying OS.
[11:00] <FireCrotch> :)
[11:00] <iLLiZT> Yeah
[11:00] <soren> iLLiZT: ...which in and of itself is nonsense.
[11:00] <iLLiZT> Well, naturally they have some sort of OS
[11:00] <soren> There's always an OS.
[11:00] <FireCrotch> since the hypervisor itself is the OS basically
[11:01] <iLLiZT> But their OS is used only for virtualization.
[11:01] <iLLiZT> And nothing else.
[11:01] <soren> What difference does that make?
[11:01] <FireCrotch> Like I said... buzzword :)
[11:01] <iLLiZT> Less overhead, and like FireCrotch says; buzzword
[11:01] <|Mike|> soren: go get em, rawwwwwwww!
[11:01] <soren> bah
[11:02] <jayvee> Are there any virtualisation solutions that don't use disk images?
[11:02] <jayvee> chroots don't count ;)
[11:02] <iLLiZT> jayvee: You could boot vms from iSCSI.
[11:02] <soren> I'd exchange a bit of overhead for not having a /completely separate codebase/ to do virtually the exact same thing as a proper OS /any/ day.
[11:02] <iLLiZT> I don't mind using a server for just virtualization.
[11:03] <jayvee> iLLiZT: never touched iSCSI before
[11:03] <soren> jayvee: What's the problem with disk images?
[11:03] <jayvee> oh, nothing
[11:03] <jayvee> just curious
[11:03] <soren> What counts as a "disk image" in your book?
[11:03] <jayvee> just cloning a disk image at the moment, and it crossed my mind
[11:03] <blinkiz> Hello. Installed "redhat-cluster-suite" and now I can not boot my machine anymore. Last message at boot is "dlm: Using TCP for communications". Anyone know how to resolve this?
[11:03] <soren> Maybe that's the right question to ask.
[11:04] <jayvee> disk image = vmdk, vdi, img, etc.
[11:04] <soren> jayvee: Sure. KVM can use other stuff.
[11:04] <soren> jayvee: It can use a disk, partition, lv, whatever.
[11:04] <FireCrotch> so can Virtualbox, I think
[11:04] <jayvee> I'm thinking more folder on the hard drive without a filesystem
[11:05] <blinkiz> I know how to make it go away. Just boot the server with USB stick and remove all deps of redhat-cluster-suite. But heay, I want cman, clvm, nfs and stuff
[11:05] <iLLiZT> So, anyone install 3.1.4 yet?
[11:05] <jayvee> but that would require NFS
[11:05] <soren> jayvee: LXC.
[11:05] <jayvee> bit like what some LTSP setups use, I spose
[11:05] <iLLiZT> Default behaviours have changed a bit
[11:05] <iLLiZT> Seems promising.
[11:05] <jayvee> soren: interesting — not heard of that
[11:05] <jayvee> and there's a libvirt driver
[11:05] <soren> Yes.
[11:06] <jayvee> speaking of which, I was trying to get libvirt to talk to virtualbox the other day
[11:06] <jayvee> had some error
[11:06] <jayvee> hang on
[11:06] <jayvee> hmm, wow, it works
[11:06] <jayvee> heh, never mind :)
[11:07] <jayvee> hmm, not as good as I hoped though. doesn't support the libvirt chrome, and doesn't run headless.
[11:07] <blinkiz> Currently setting up a new virtualization environment here. One storage server and two virtualization servers. Using iSCSI with cluster LVM. have it working! Pretty cool.
[11:08] <jayvee> cool, lxc looks awesome
[11:09] <jayvee> even less vm'ish than openvz
[11:09] <jayvee> does it use UML?
[11:11] <jayvee> actually, doesn't look like it uses separate kernels at all
[11:11] <jayvee> similar to openvz, from what I recall
[11:12] <blinkiz> jayvee, yeah, LXC is the future of openvz.
[11:14] <jayvee> I presume my kernel already supports it.
[11:15] <blinkiz> jayvee, yes. You have to ask someone else how mature LXC is. Ask soren
[11:21] <soren> Well, people are using it in production. It's been around for a while. I believe I heard its security model leaves a bit to be desired, though, but that may not matter for your use case.
[11:28] <jayvee> hmm, that was a bad idea
[11:29] <jayvee> ran lxc-create --name=larry, then lxc-start --name=larry
[11:29] <jayvee> and my whole system came to a grinding halt
[11:29] <jayvee> my display went all pretty colours, and then got dumped back to a tty
[11:29] <jayvee> I think I should have read the man page first. ;)
[11:48] <twb> Manpages are for wimps
[12:03] <garymc> Hi guys, anyone know where I look to see if bots are hitting my web server?
[12:03] <garymc> and what i look for?
[12:14] <oly> hi, can anyone tell me if there would be any issues with running openvz and kvm along side each other
[12:51] <pecisk> Hi people, I can't boot upgraded server system, it is known bug/issue? It was pure Karmic server install and then do-release-upgrade -d
[12:52] <pecisk> Hangs on kernel 'Checking 'hlt' instruction... OK.'
[12:57] <pmatulis> pecisk: what kind of system is it?
[12:58] <pecisk> well, I solved problem, sorry to bother :)
[12:58] <twb> pecisk: rather, it probably hangs immediately AFTER that.
[12:58] <pecisk> it was VirtualBox, needed acpi=off for some strange reason
[12:59] <pmatulis> pecisk: right, it's a VB issue
[12:59] <smoser> ttx, ping
[13:00] <ttx> smoser: pong
[13:00] <zul> morning
[13:01] <smoser> ttx, pm
[13:25] <kirkland> ttx: ping
[13:25] <kirkland> ttx: i'm going to a cherry pick of the 4 commits at https://code.edge.launchpad.net/~eucalyptus-maintainers/eucalyptus/1.6.2
[13:25] <kirkland> ttx: any objections?
[13:26] <kirkland> ttx: i'll build/upgrade my local packages and sniff test before uploading
[13:26] <ttx> kirkland: no objection
[13:26] <ttx> Note that I uploaded a new release
[13:26] <bogeyd6> mandriva instanton is a very neat concept
[13:26] <ttx> a couple hours ago
[13:30] <twb> Meh.
[13:31] <twb> I'd rather get Debian booting faster than adopt Mandrivia.
[13:32] <twb> http://en.wikipedia.org/wiki/Instant_on indicates almost all the vendors run Splashtop, a proprietary Linux distro.
[13:32] <kirkland> ttx: okay
[13:32] <kirkland> ttx: i'll get on it
[13:34] <bogeyd6> twb, agreed
[13:34] <twb> Seems to me that with live-helper you get pretty good boot times.
[13:35] <twb> I'm using it to boot a 200MB-ish Lucid system that just runs surf and mplayer, and that comes up in... let's time it...
[13:38] <twb> Not too good -- 32 seconds for a 1.5GHz Via C7
[13:38] <twb> My laptop's around 15s for a 630MHz Celeron, but I don't start X.
[13:41] <ttx> kirkland: saw my pm ?
[13:41] <kirkland> ttx: no
[13:41] <ttx> kirkland: and now ?
[13:41] <kirkland> ttx: yup
[14:11] <Jeeves_> Bah. Live migration still doens't work :(
[14:15] <Daviey> zul: Have you seen that php 5.3.2 bug and "Security Enhancements" release has been made?
[14:15] <zul> Daviey: yep I have
[14:16] <Daviey> zul: what do you think about bumping to it?
[14:16] <zul> Daviey: umm....maybe :)
[14:17] <zul> Daviey: ill take a look at it this afternon
[14:19] <Daviey> zul: neat
[14:25] <Jeeves_> IRC-REVOKE. Live migration does work
[14:43] <_ruben> Jeeves_: in what kind of environment? plain kvm? "free" live migration sure make me wanna play around with kvm (and possibly xen). as our current choice of virtualization product, vmware esxi, surely has a hefty pricetag on its live migration feature
[14:44] <Jeeves_> _ruben: Libvirt, KVM, Ubuntu Lucid
[14:45] <smoser> kirkland, ttx, bug 532682
[14:45] <_ruben> Jeeves_: nice, nfs storage i guess?
[14:46] <Jeeves_> _ruben: No, iscsi
[14:46] <_ruben> Jeeves_: ah, interesting
[14:46] <_ruben> Jeeves_: where do you handle the iscsi? in the vm or the host?
[14:47] <Jeeves_> Sun Unified Storage
[14:48] <_ruben> you and your suns ;)
[14:49] <Jeeves_> :)
[14:52] <ogra> arent they oracles nowadays ?
[14:53] <Jeeves_> Yes
[14:53] <Jeeves_> So we switched to Fujitsu servers
[14:53] <ogra> so what do they predict ? :)
[14:53] <Jeeves_> But the Sun storage kicks ass
[14:57] <_ruben> Jeeves_: what fs do you for your shared storage?
[14:57] <_ruben> +use
[14:58] <henkjan> _ruben: shared storage to host the vms?
[14:58] <henkjan> _ruben: eacht vm has its own iscsi lun
[15:02] <Jeeves_> _ruben: What henkjan says
[15:03] <_ruben> ah ok
[16:03] <zul> ivoks: ping
[16:07] <ttx> zul: he's not around
[16:07] <zul> bah
[16:51] <dvrcoder> hi. question: is there any kind of problem with .ssh/authorized_keys when /home/USERNAME is a symlink to /bla/home/USERNAME?
[16:56] <jbroome> The tinyurl in the /topic is broken.
[17:09] <Andy-at-home> guys, see to access my webmin on my vps, is that not just http://xxx.xxx.xxx.xxx:10000 to log into it?
[17:19] <jbroome> IIRC, that's the default port for webmin
[17:19] <jbroome> may be https, i don't remember
[17:20] <_ruben> !webmin
[17:22] <Andy-at-home> still supported in 9.0.4 by any chance?
[17:24] <fullstop> Has anyone seen a segfault generating grub.cfg when updating the linux-server-image?
[17:24] <fullstop> http://www.pastebin.org/101841
[17:27] <_ruben> Andy-at-home: no, iirc webmin was dropped years ago
and its 9.04, not 9.0.4</nitpick>
[17:28] <fullstop> actually update-grub segfaults on 9.10 server.. odd.
[17:34] <qman__> Andy-at-home, webmin was dropped before 8.04 LTS, and is not in any supported versions
[17:53] <zul> kirkland: ping what happens when qemu-kvm userland and kernel is mismatched?
[17:54] <kirkland> zul: how mismatched?
[17:54] <kirkland> zul: there are "recommended combinations", I'd say
[17:54] <kirkland> zul: but modern qemu-kvm should work with most kvm's from the last, say year+ ?
[17:54] <zul> so should the qemu-kvm use dkms?
[17:55] <zul> kirkland: okies then
[17:55] <kirkland> zul: oh, the dailies?
[17:56] <kirkland> zul: that would be phenomenal for the dailies ...  is the kvm kernel source in the qemu-kvm tarball?  /me checks
[18:31] <hink> How do i reprocess all messages in the mailq
[18:32] <ivoks> postqueue -f
[18:33] <zul> smoser: dude that was quick...the ebs stuff congrats
[18:33] <zul> ivoks: or you could it the traditional way and telnet localhost 25 and do a ETRN ;0
[18:34] <smoser> thanks zul.
[18:34] <smoser> have a nice weekend all
[18:34] <smoser> i'nm out now
[18:47] <skettles> ciao a tutti
[18:47] <skettles> chi mi aiuta?
[18:47] <skettles> hi all anyone help me?
[18:48] <skettles> c'è qualcuno?
[18:49] <guntbert> !it | skettles
[18:49] <skettles> sorry me
[18:52] <ivoks> !hr | ivoks
[18:52] <ivoks> lol, nice
[18:53] <acalvo_> hi
[18:53] <acalvo_> is there any way to monitor the memory usage and output it to a file or so?
[18:54] <acalvo_> a server started crashing once a day for the last week
[18:54] <acalvo_> and I want to know which process takes it down
[18:54] <acalvo_> since it gets hung
[18:54] <acalvo_> I'm trying to figure out how to detect the process that is eating all the memory
[18:55] <ivoks> well, there's that thing called oom-killer
[18:55] <acalvo_> yes, I've seen it
[18:55] <ivoks> it randomly kills process when there's no memory
[18:55] <acalvo_> but the server dies always
[18:56] <ivoks> it dies in the sense you can't connect to it or it freezes?
[18:57] <acalvo_> I can't connect to it
[18:57] <acalvo_> but it respond to pings
[18:57] <ivoks> so it kills ssh :)
[18:57] <acalvo_> it's a VM
[18:57] <acalvo_> and I can't log in neither
[18:57] <ivoks> huh
[18:57] <acalvo_> http://ubuntuforums.org/showthread.php?t=1421823
[18:57] <acalvo_> I've posted a question
[18:58] <acalvo_> and I'd like to know which process is killing it
[18:58] <ivoks> top -b > logfile
[18:58] <acalvo_> ok, fine!
[18:59] <ivoks> funny name fot RealtimeKit user
[18:59] <ivoks> rtkit
[18:59] <ivoks> running rtkit-daemon
[19:00] <ivoks> makes your heart skip a beat
[19:01] <acalvo_> btw
[19:01] <acalvo_> ivoks: I guess you're developers of ubuntu, aren't you?
[19:01] <acalvo_> I've spend the last months around here asking (basic) questions
[19:01] <acalvo_> and some of you are always here
[19:02] <Andy-at-home> so webmin isnt going to work on version 9.04?
[19:02] <Andy-at-home> im logged in and can view the navigation sidebar but no links seem to load
[19:02] <ivoks> acalvo_: yes
[19:02] <ivoks> Andy-at-home: webmin ins't something we care about
[19:02] <ivoks> Andy-at-home: isn't
[19:03] <acalvo_> working in cloud computing for ubuntu (eukalyptus) i guess
[19:03] <ivoks> Andy-at-home: but i don't see how non-working web page is caused by OS
[19:03] <jbroome> wow, the webmin thing is still going on
[19:03] <Andy-at-home> what would you recommend?
[19:04] <jbroome> ssh and the CLI like linus intended
[19:04] <qman__>  /agree
[19:04] <Andy-at-home> mmm
[19:05] <_ruben> !ebox
[19:05] <ivoks> ebox is cool
[19:06] <acalvo_> thanks!
[19:06] <RoAkSoAx> ivoks, applying for UDS sponsorship i guess?
[19:06] <ivoks> RoAkSoAx: maybe
[19:07] <RoAkSoAx> ivoks, why not? not going?
[19:07] <ivoks> i'm planing on moving to redhat
[19:07] <james_w> hi. I'd like to try the cloud-init/puppet integration, but I can't see how to bootstrap it to tell it to do the puppet stuff. Is it done via user-data?
[19:07] <RoAkSoAx> ivoks, lol
[19:08] <ivoks> RoAkSoAx: :D
[19:08] <RoAkSoAx> seriosly?
[19:08] <ivoks> of course not :D
[19:08] <RoAkSoAx> hahaha
[19:08] <RoAkSoAx> lol
[19:08] <RoAkSoAx> I will apply but prolly wont get it since I didn't contribute as I expected and as much as I did previous release cycle :(
[19:08] <RoAkSoAx> cause of my surgeries :(
[19:09] <ivoks> well, i don't know how they decide
[19:09] <ivoks> but since it's close, i might even go on a road trip :)
[19:10] <RoAkSoAx> yeah that's cool
[19:10] <RoAkSoAx> if i don't get sponsorship i'll probably go back to peru for a few months
[19:10] <ivoks> sounds like ultimatum :D
[19:11] <RoAkSoAx> ivoks, yeah from my dad hahaha
[19:11] <ivoks> 'sponsor me, or i'll go home'
[19:11] <ivoks> :D
[19:11] <RoAkSoAx> he doesn't want me to stay in the US through the summer
[19:12] <ivoks> take care
[19:12] <ruben23> hi i have lvextend a logical volume, but still i cant see its size what is the next step i should do..?
[19:12] <ivoks> extend partition
[19:12] <ivoks> is it ext3?
[19:12] <ruben23> yes
[19:12] <ivoks> bah, filesystem, not partition
[19:13] <ruben23> belogn to vg
[19:13] <ivoks> resize2fs
[19:14] <ivoks> it can expand mounted filesystem
[19:14] <ivoks> so, just fire it ;)
[19:14] <ivoks> resize2fs /dev/vg/lv
[19:15] <ruben23> online resizing now
[19:15] <ivoks> ruben23: and remember, df shows size of filesystem, not a partition
[19:15] <ivoks> your partition can be 1TB, but filesystem can be 1GB
[19:16] <ruben23>  ivoks: yes thank you, you save me form doom..
[19:16] <ivoks> good night everybody
[19:17] <ruben23> my var LVM is 98 percent on usage-in minute it will be full
[19:17] <ruben23> goodnight
[19:23] <james_w> aha, doc/examples/cloud-config-puppet.txt, score
[19:31] <Jeeves_Moss> are there any good open source managed services platforms that run on Ubuntu? I'm looking to manage Windows/Ubuntu/Mac systems, and I would love to keep it open source
[19:32] <|Mike|> "huh"?
[19:35] <Pirate_Hunter> does 8.04 come with lmv2 by default?
[19:41] <Jeeves_Moss> |Mike| managed services.  google man
[19:44] <|Mike|> Jeeves_Moss: then i still don't get your question. You're looking for *something* what offers Ubuntu virtual machines to run "services" on?
[19:46] <Jeeves_Moss> |Mike| nope.  I'm looking for a managed services platform that the main server runs on a Ubuntu box.  Something like N-Able, keyesa, etc
[19:48] <|Mike|> so basicly you're looking for a managed solution to deploy your (opensource, or licensed) application on?
[19:49] <|Mike|> (which runs Ubuntu)
[20:00] <ruben23> hi deleting a directory with a very large size--> what is the best syntaxs should i have..?
[20:00] <|Mike|> rm -rf bla
[20:01] <|Mike|> how big is big? :)
[20:01] <ruben23> |Mike|: hmmm strange i have run like that but the directory was not delete after i have check the next morning.. it 90 GB
[20:02] <|Mike|> it does contain a lot of files ruben23 ?
[20:03] <ruben23> yes large number of files, audio recordings
[20:04] <ruben23> |Mike|: i would like to create a simple script for it to be move to a cifs windows shared directory..but unsuccesful.
[20:04] <|Mike|> for i in `ls -l a*`; do rm -rf $1; done ?
[20:05] <|Mike|> (what is a cifs windows share?)
[20:06] <ruben23> yes
[20:07] <ruben23> how do i create the script for it..? someone suggested before about Xargs
[20:07] <|Mike|> they all have the same extension?
[20:12] <ruben23> no, they have different..coz it an audio files
[20:12] <ruben23> for every different person recorded
[20:13] <ruben23> mike sorry for a late reply..
[20:14] <ruben23> |Mike|:you there..?
[20:15] <|Mike|> yep
[20:16] <ruben23>  |Mike|:can i ask help on it, for a simple script
[20:18] <|Mike|> sure
[20:18] <|Mike|> pastebin it :)
[20:20] <ruben23> |Mike|: what should i pastebin.?
[20:21] <|Mike|> the script where you where working on?
[20:22] <|Mike|> but what are you trying to achieve with the script? Just remove a directory which contains a shitload of data?
[20:23] <Pirate_Hunter> wonders if he should be adventurers and use alpha3 for a live server, butterflies in my gut but my mind says yes
[20:24] <ruben23> im planing to move it to a CIFS shared drive...so my directory would be empty
[20:24] <ruben23> need script to make it automatic..that i would missed that might my server drive will get full
[20:29] <|Mike|> (what is a CIFS share ?)
[20:31] <ruben23> windows shared drive/directory to my ubuntu-server..
[20:34] <RoAkSoAx> kirkland, is it possible to convert a vmware image into a kvm image?
[20:36] <|Mike|> ruben23: is it like drag and drop or.. ?
[20:40] <jbernard> kirkland: byobu_multi.svg seems to have not made it into the deb
[20:43] <ruben23> yes...mapped to a windows drive directory
[20:58] <histo> Alright I want to be able to use php's mail function just to send emails on a contact page.
[20:58] <histo> So I guess I have to install postfix?
[21:00] <histo> Or someone is suggesting to use the mail function to send thej mail with googles smtp
[21:11] <histo> Anyone around?
[21:40] <resno> question about running a mail server. whats involved in config, etc?
[21:41] <histo> yes like I just need postfix to work withe mail fuction of php. I don't need it for anythign elese
[21:41] <kirkland> jbernard: well, it's kinda big right now
[21:41] <histo> Like users aren't going to be connecting etc... And I want it secure so that someone can't start sending spam with it.
[21:41] <kirkland> jbernard: need to scale it way down
[21:41] <kirkland> jbernard: takes up too much space
[21:56] <zroysch> what steps should be taken to verify that a partitions data is completely unrecoverable
[21:56] <zroysch> before starting over
[22:08] <histo> I'm going to use ssmtp
[22:08] <histo> Can you do a server install from the mini.iso?
[22:10] <histo> man this channel is a hoping place.
[23:07] <james_w> anyone tried to boot the EBS-based AMIs that Scott posted? I'm apparently not authorized, so I wonder if they need to be made public or something?
[23:15] <veebull> can someone explain how 'internal locked-down USB ports' work?
[23:15] <veebull> i.e. how are they different from a regular USB port, and why would you use one?
[23:29] <histo> veebull: internal locked-down usb port?
[23:49] <AnRkey> I cant find any package for linux-headers-2.6.24-23-server in the 8.04 repos, am I missing something?
[23:51] <AnRkey> i do sudo apt-get install linux-headers-`uname -r` and I get E: Package linux-headers-2.6.24-23-server has no installation candidate
[23:54] <qman__> AnRkey, I only see linux-headers-lbm-2.6.24-23-server and linux-headers-lum-2.6.24-23-server
[23:54] <qman__> though that's an old kernel, you probably want to update to the latest first
[23:55] <AnRkey> i have been using apt-get upgrade so far
[23:55] <qman__> upgrade doesn't update your kernel
[23:55] <AnRkey> never had to install kernel myself
[23:55] <qman__> you need to use dist-upgrade to do that
[23:55] <AnRkey> oh ok
[23:55] <AnRkey> shweet
[23:55] <AnRkey> thanks, will give it a try quick
[23:56] <AnRkey> yeah, there it goes
[23:56] <AnRkey> thanks so much
[23:56] <qman__> no problem
[23:56] <AnRkey> i knew i was being stupid :)
[23:57] <AnRkey> the mirrors are ultra fast at 2am :)
[23:57] <qman__> I actually cache the packages in squid because I've got so many ubuntu machines here
[23:58] <qman__> cut overall update time in half
[23:58] <AnRkey> mine is going through my apt-proxy
[23:58] <AnRkey> but have not done a 8.04 box in a year so it's got some stuff to get
[23:59] <qman__> I'm about half and half, 8.04 and 9.10
[23:59] <AnRkey> 8.04.4 is soooo stable now, i'm a happy camper :)
[23:59] <AnRkey> 9.10 has been a bit crappy for me
[23:59] <qman__> looking forward to lucid so I can get all back on one version again
[23:59] <AnRkey> had to rebuild my mail server after the upgrade
[23:59] <AnRkey> yeah, lucid is gonna be shweet
[23:59] <AnRkey> gonna wait till 10.04.1 though
[23:59] <qman__> yeah