[00:02] <savid> Is anyone here familiar with puppet?  Do you know if there's a ppa available for puppet 2.x?
[00:03] <grape> I think puppet has their own repository you can add. not sure about a ppa.
[00:04] <grape> savid: you can also just install the gem
[00:05] <savid> grape,  I tried installing gem,  it doesn't set up anything (config files,  /usr/local/bin symlinks, init scripts)
[00:06] <grape> Ah, ok. Best I could suggest is to try searching launchpad
[00:11] <grape> Can anyone tell me what the standard way of installing UbuntuCloud in 11.10 is going to be? Will it be an option on the Ubuntu Server Edition boot menu?
[00:59] <twb`> So: who's sick of rsnapshot?
[01:08] <twb`> Re. the vsftpd vs. lxc cgroup bitch-fight in lucid, I notice the kernel plumbers have asked for a proper fix for it within cgroups
[01:08] <twb`> "* fork throttling mechanism as basic cgroup functionality that is available in all hierarchies independent of the controllers used: This is important to implement race-free killing of all members of a cgroup, so that cgroup member processes cannot fork faster then a cgroup supervisor process could kill them. This needs to be recursive, so that not only a cgroup but all its subgroups are covered as well."
[02:47] <TDJACR> I updated to oneirc, and I've been having troubles with my mail server since.
[02:47] <TDJACR> I fixed the first two by installing missing packages
[02:47] <TDJACR> The third isn't as easy
[02:47] <TDJACR> I try to send mail via SMTP
[02:47] <TDJACR> And I'm getting authentication errors
[02:48] <TDJACR> In my logs
[02:48] <TDJACR> I get
[02:48] <TDJACR> SASL LOGIN authentication failed: no mechanism available
[02:48] <TDJACR> Even though mysql through pam is speficied
[02:48] <TDJACR> And worked flawlessly previously
[02:50] <twb`> SASL isn't PAM
[02:50] <TDJACR> I know
[02:50] <TDJACR> It's using the PAM authentication method
[02:50] <twb`> Ask etckeeper what configuration changes have occurred since the upgrade
[02:51] <TDJACR> etckeeper?
[02:51] <twb`> It version-controls your /etc dir, so you can answer such questions.
[02:51] <TDJACR> Whoops
[02:51] <TDJACR> Oh
[02:51] <twb`> Of course, I'm assuming you thought to install it BEFORE the upgrade
[02:51] <TDJACR> I declined any changes to config files during the upgrade
[02:51] <TDJACR> I hadn't heard about it
[02:52] <twb`> TDJACR: did you bother to understand the changes, or did you just click "no" every time?
[02:52] <TDJACR> twb`: I did bother to understand them, I diffed most of them
[02:53] <TDJACR> And edited the ones with potential issue
[02:53] <TDJACR> s
[02:55] <TDJACR> Is there a way to make sasld and postfix more verbose, so I can see the exact issue?
[02:57] <twb`> Probably
[02:57] <twb`> Are you using dovecot's sasl implementation?
[02:57]  * TDJACR checks
[02:58] <TDJACR> sasl2-bin
[02:59] <TDJACR> So cyrus
[03:01] <twb`> Don't do that.
[03:01] <TDJACR> (Originally from http://library.linode.com/email/postfix/dovecot-mysql-ubuntu-10.10-maverick#sph_configure-mail-aliases)
[03:01] <TDJACR> Ignore the link to the section, the whole page
[03:01] <twb`> I mean, unless you're using cyrus for imap or something, but AFAIK dovecot is the recommended IMAP/SASL backend.
[03:02] <TDJACR> I use dovecot for imap
[03:02] <TDJACR> See that link for my setu
[03:02] <TDJACR> p
[03:11] <TDJACR> twb`: What do you reccomend I do nwo
[03:11] <TDJACR> now*
[03:13] <twb`> I dunno, man, I don't really care
[03:14] <TDJACR> Ah alright
[03:14] <TDJACR> Well thank you anyway
[03:14] <TDJACR> I appreciate the etckeeper idea
[04:37] <twb`> https://packages.ubuntu.com times out instead of either failing immediately, or working.
[05:21] <bikcmp> is there a utility to figure out how fast the network traffic is flying on a server? (as in, 2 mbps down, 1 mbps up)
[05:22] <twb`> bikcmp: dd
[05:23] <twb`> Or mii-tool
[05:23]  * bikcmp looks
[05:26] <bikcmp> twb`: how would I use dd to do that?
[05:27] <twb`> ssh -f gw nc -l -p 12345; dd if=/dev/zero bs=1M count=1k | nc gw 12345 & while sleep 1m && kill -USR1 $!; do :; done
[05:27] <twb`> Or similar; salt to taste
[05:28] <bikcmp> that looks chinese to me.
[05:28] <bikcmp> god
[05:28] <bikcmp> haha
[05:28] <bikcmp> gw?
[05:28] <twb`> gw would be the host you're testing against
[05:28] <bikcmp> -p is port, i'm assuming. i want total traffic
[05:28] <twb`> bikcmp: then stop all other traffic
[05:28] <bikcmp> hm.
[05:30] <twb`> If you want to measure the actual traffic, rather than manufacturing a flow and measuring its overall speed, you could look at something like ifconfig's RX and TX lines
[05:31] <twb`> Obviously diff them over time, similar to the proof for differential calculus
[05:31] <bikcmp> yeah
[05:31] <bikcmp> i was thinking about writing a script to do that
[05:31] <twb`> There should be a file in /proc somewhere that has the TX and RX as raw fields, so you don't need to parse it out.
[05:31] <twb`> Probably someone has already solved that
[05:31] <bikcmp> still dirty as heck, haha
[07:09] <memoryleak> hi
[07:11] <jamespage> morning all
[09:01] <lynxman> jamespage: morning :)
[09:01] <RoyK> anyone that knows where I can find a jaunty repo?
[09:02] <RoyK> I have this machine still running jaunty, and I can't upgrade it yet - since if something goes wrong (such as grub), the machine is at Ny-Ålesund, Svalbard, so it'll take some time to get there :P
[09:04] <Kiall> RoyK: you're out of luck with the official mirrors ..
[09:05] <RoyK> I know
[09:05] <Kiall> Its unlikely you'll find a reliable+public 3rd party ,,
[09:09] <MuNk``> anyone know about installing & configuaring Emulex LP950 fibercards in Ubunut?
[09:11] <MuNk``> or if its possible ot install the Emulex lpfc utils into ubuntu 10.04?
[09:15] <TeTeT> RoyK: http://old-releases.ubuntu.com/releases/
[09:17] <koolhead17> hi all
[09:23] <RoyK> TeTeT: what would be the repo lines in /etc/apt/sources.list to use this?
[09:25] <TeTeT> RoyK: guess you need to replace archive.ubuntu.com with old-releases.ubuntu.com
[09:25] <TeTeT> RoyK: but keep in mind that of course noone will update any packages there with security fixes. But if you need to install some additional software or prepare the upgrade, that's the path
[09:26] <RoyK> TeTeT: I know - I just can't upgrade this box at the moment. I can't risk it hanging - as I said, it's at Svalbard...
[10:02] <soren> Daviey: Any chance this will be looked at for Oneiric? https://bugs.launchpad.net/bugs/844670
[10:07] <soren> Daviey: WEll, that and whatever other issue there might be with the dashboard. I kept my hands off because I was told there were fixes underway.
[10:13] <Kiall> soren: ah .. while you're on ;) I know you did the nova upstart scripts, should they not have respawn enabled?
[10:13] <Kiall> With an all in one setup, OS rearely starts after a reboot ..
[10:13] <Kiall> it either cant connect to libvirt/mysql (they havent started yet)
[10:16] <soren> Kiall: I would say that they should. They used to.
[10:17] <Kiall> Ah, none of the oneiric packages have it enabled...
[10:18] <soren> It got yanked out for some reason that escapes me.
[10:24] <Kiall> fair enough, I guess there was probably a good reason!
[10:24] <Kiall> Its not the sort of thing anything would disable on a whim!
[10:28] <koolhead17> hello soren
[10:30] <soren> o/
[10:30]  * soren heads to lunch
[10:36] <koolhead17> soren: am stuck with role based access control, can you point me whom should i ping for the same. am checking the DB tables to get some clue. :P
[10:40] <koolhead17> robbiew: came to know about the usb powered with ubuntu as giveaway at the summit :)
[10:46] <robbiew> koolhead17: ;)
[11:05] <Daviey> soren: That bug was not on my radar :(
[11:17] <soren> Daviey: I wonder if we should rip out the dashboard package? It's 100% useless in its current state.
[11:21] <Daviey> soren, is it really that bad?
[11:21] <Daviey> soren, note - that someone motivated to fix dashboard could have it done before releadse.
[11:21] <Daviey> It would be sad to just drop it IMO.
[11:24] <soren> Daviey: I'd rather have no package than one that is completely broken.
[11:24] <soren> Daviey: As I said: I kept my hands off because there were supposedly fixes underway.
[11:24] <soren> Daviey: Oh.
[11:24] <soren> "Could have it done" not "Could have done it". My bad.
[11:28] <Daviey> soren: I did feel that the dash was something that was quite importiant.
[11:28] <Daviey> It's part of the reason work went into making sure keystone worked.
[11:28] <Daviey> fml
[11:29] <soren> Daviey: I'd be nothing short of thrilled if it worked.
[11:29] <soren> Daviey: Is Keystone in good shape in Oneiric?
[11:29]  * soren is slightly out of touch
[11:31] <Daviey> soren, it's not the GA.. but seems to work.
[11:31] <Daviey> zul is thinking about having a look at updating t o release.
[11:31] <Daviey> The delta is pretty large, so we are not quirte sure.
[11:34] <rbasak> Daviey: bug 862129 seems to be breaking a lot of upgrades - only on desktop as far as I can see, but it's showing up on our bug lists seeing as samba is shared. Is there someone specific I should be referring this to? No response in #ubuntu-devel so far today.
[11:39] <iclebyte-work> i'm having an issue with running my own repo with some custom packages. the repo was built using apt-ftparchive utility but when adding it to the sources of a node I get an error about Error Occured while processing (package name) - problem with merge list - the list or status file could not be parsed or opened?
[11:44] <Daviey> rbasak: That is probably one for us to pick up.. :(
[11:49] <koolhead17> Daviey: i heard that oneiric will have diablo minus keystone. is that the plan? I am able to run dashborad minus keystone on oneiric :P
[11:51] <Daviey> koolhead17: from source?
[11:51] <Daviey> or packages?
[11:52] <Daviey> AIUI if you use the dash without keystone, you are using noauth?
[11:52] <koolhead17> Daviey: source. soren`s trunk on launchpad
[11:53] <koolhead17> Daviey: yes. we are not using keystone.
[11:53] <Daviey> ko	soren's trunk of the dash?
[11:53] <koolhead17> saperate db for nova and saperate for dashboard
[11:53] <koolhead17> Daviey: yay
[11:54] <koolhead17> :)
[11:54] <koolhead17> its very old with same bugs but works
[11:54] <Daviey> hah
[11:54] <koolhead17> now am checking on role based access control via dashboard
[11:54] <koolhead17> LO
[11:54] <koolhead17> :)
[11:55] <Daviey> lovely
[11:55] <koolhead17> had to show some love to django to authenticate users account  created via dashboard
[11:56] <koolhead17> that is why i was troubling soren  :P
[12:10] <jamespage> utlemming, smoser: when are we likely to have ec2 images for testing?
[12:21] <koolhead17> lynxman: hey
[12:21] <koolhead17> hello TeTeT
[12:21] <TeTeT> koolhead17: hi there
[12:21] <lynxman> koolhead17: ello there :)
[12:22] <lynxman> koolhead17: what revision of dashboard you got working without keystone?
[12:22] <koolhead17> lynxman: 65 :)
[12:22] <lynxman> koolhead17: cool vibes
[12:22] <koolhead17> TeTeT: seems all busy for oneiric release :)
[12:23] <koolhead17> lynxman: hehe. trying to show my love to django
[12:24] <lynxman> koolhead17: I have some experience in django if it can be of any help
[12:31] <koolhead17> lynxman: am runnning natty + diablo for nova and dashboard is on oneiric
[12:32] <sanderj_> Do anyone know what is using that much cpu here?: http://pastebin.com/ft5nGhCv
[12:34] <patdk-wk> sanderj_, no idea, those are completely incorrect stats to figure this out
[12:35] <sanderj_> patdk-wk, How can I see it?
[12:36] <patdk-wk> ps axl, doesn't show it?
[12:37] <patdk-wk> the first results from iostat are not useful, unless you want averages since boot
[12:38] <sanderj_> ps axl only displays time used by each process... As I can see.
[12:39] <patdk-wk> sorry, I can't help you, as I can't see your screen
[12:39] <sanderj_> patdk-wk, the paste from "top" displays "93% system time used"..
[12:41] <sanderj_> patdk-wk, http://pastebin.com/Pw0aH3A5
[12:42] <sanderj_> patdk-wk, is the %sys from top also counted since last reboot?
[12:46] <patdk-wk> sanderj_, check dmesg
[12:46] <patdk-wk> I want to say your disk is thrashing, but no waitio
[12:46] <patdk-wk> but your doing 4 mysqldumps at the same time
[12:50] <sanderj_> I stopped those mysqldumps now.
[12:51] <sanderj_> patdk-wk, how do you know my disks is bad?
[12:51] <patdk-wk> I don't
[12:52] <sanderj_> I'm using a san, trough vmware 3.. pretty old.
[12:53] <sanderj_> Thats what I dont understand myself.. why there is no iowait or %util used.
[13:12]  * koolhead17 finds RBAC totally broken in diablo :(
[13:17] <BrixSAt> hello, i just installed a psql server but i dont know wich is the default username\password
[13:17] <ahasenack> hi guys, will oneiric's openstack have ebs support? ("swift" iirc?)
[13:17] <ahasenack> BrixSAt: the default admin is the "postgres" linux user, just become it
[13:18] <ahasenack> BrixSAt: i.e., sudo su - postgres
[13:18] <BrixSAt> and the password?
[13:18] <ahasenack> BrixSAt: there isn't one, only root can become that user
[13:18] <BrixSAt> so if i install phppgadmin how do i login?
[13:19] <ahasenack> I don't know that program
[13:19] <BrixSAt> ahasenack:  it is the phpmyadmin for pgsql
[13:19] <ahasenack> you can always create another admin user and give it a password, see man createuser
[13:24] <koolhead17> soren: the sysadmin role assigned to a user by project admin via dashboard should allow him to run an instance. correct me if am wrong
[13:25] <BrixSAt> thks ahasenack
[13:28] <soren> koolhead17: No cule.
[13:29] <soren> clue, even.
[13:29] <koolhead17> soren: :)
[13:30] <koolhead17> nova-manage user modify   atul "" ""  T
[13:30] <koolhead17> this makes me a admin and then i can do whatever i want via dashboard
[13:56] <Kiall> ahasenack: yes, oneiric's openstack suppports attaching EBS volumes to instances .. I'm not entirely sure about booting from them though, there *seems* to be a suggestion that that is now implemented, but I might be wrong, and havent managed to get it working
[13:57] <ahasenack> Kiall: cool, thanks
[14:18] <tjaalton> so I installed ubuntu-orchestra-server on my machine hoping that it would help me maintain all my local and virtual machines, but I've no idea where to go next, and the documentation isn't helping either
[14:18] <tjaalton> i don't need juju, instead telling the system which hosts to manage & install
[14:43] <koolhead17> lynxman: around
[14:43] <lynxman> koolhead17: indeed
[14:44] <koolhead17> lynxman: seems like nova-manage command has changed in diablo.
[14:44] <lynxman> koolhead17: it has changed a lot
[14:44] <koolhead17> lynxman: any pointer would be great help :)
[14:45] <lynxman> koolhead17: let me get my notes, 1 sec
[14:45] <koolhead17> :)
[14:45] <koolhead17> even the nova.log not giving me intelligent errors which i could understand :P
[14:53] <lynxman> koolhead17: heh :)
[14:53] <lynxman> koolhead17: one sec, failing at multitasking
[14:53] <koolhead17> lynxman: np
[14:55] <lynxman> koolhead17: these are my notes http://pastebin.ubuntu.com/705430/
[14:55] <koolhead17> lynxman: thanks
[14:56] <koolhead17> lynxman: so after going through that command if i do
[14:56] <koolhead17> nova-manage role add --user=atul --role=netadmin --project=foo
[14:57] <koolhead17> this should work?
[14:58] <lynxman> koolhead17: it should
[14:58] <koolhead17> lynxman: but its not :(
[14:59] <lynxman> koolhead17: let me check my notes...
[14:59] <koolhead17> lynxman: it worked
[14:59] <koolhead17> :d
[14:59] <lynxman> koolhead17: ah
[15:00] <koolhead17> actully i allready assigned that user with netadmin role via dashboard
[15:03] <koolhead17> lynxman: am in interesting situation :D
[15:04] <lynxman> koolhead17: hehe good luck
[15:04] <koolhead17> lynxman: creating new user and mapping it to project and checking if few things work :)
[15:12] <boxybrown> hello all
[15:12] <boxybrown> anyone with experience doing Kerberos+LDAP client setup?
[15:13] <smoser> jamespage, the publish portion of the build failed last night.
[15:13] <smoser> i just kicked off a "republish"
[15:13] <jamespage> smoser: argh!
[15:13] <smoser> both utlemming and i are on holiday, so i dont know if he's around at all.
[15:13] <Daviey> smoser: erk.
[15:13] <jamespage> smoser: OK - I' just running hggdh through how to kick the testing off
[15:13] <smoser> but i'm willing to call 20111010 as "go"
[15:14] <smoser> unless there are known reasons not to
[15:14] <Daviey> smoser: can you document what happend, and what you did?
[15:14] <jamespage> so I'm not a bottle neck
[15:14] <utlemming> looks like we both kicked a republish
[15:14] <utlemming> smoser: nope, I'm around today
[15:15] <utlemming> smoser: there is one issue that I am hunting down today, and that is that the APT configuration is lacking the daily update setting
[15:15] <smoser> ugh..
[15:15] <smoser> i really would not think we should bother pulling that
[15:15] <smoser> its been like that forever (i think)
[15:15] <smoser> so i dont see any reason to fix it in 3 days before a release.
[15:15] <smoser> is there a bug on that ? or just Daviey complaining?
[15:15] <utlemming> yeah, I believe so as well, but it appears to have been fixed upstream
[15:16] <ttx> Daviey: in London already ?
[15:16] <Daviey> ttx: yes
[15:16] <koolhead17> lynxman: what is this concept of private image?
[15:16] <Daviey> utlemming: 'fixed upstream'... have a link?
[15:17] <utlemming> Daviey: no, I'm confirming now...eta 10 minutes.
[15:17] <Daviey> bug 870121
[15:21] <smoser> Daviey, really, if it has been this way in lucid -> oneiric beta, i would not change it now.
[15:22] <smoser> i have to run, but seriously, don't "fix" that
[15:22] <smoser> it could have potential fallout.
[15:22] <smoser> such as t1.micros going MIA for time periods when it runs.
[15:23] <Daviey> smoser: true
[15:24] <Daviey> smoser: are you sure it wasn't like that pre-oneiric?
[15:24] <smoser> i'm not certain, but you can check that.
[15:24] <Daviey> (the other issue is that Desktop seems to be seeing a change in behaviour)
[15:24] <Daviey> i think i checked natty
[15:25] <smoser> desktop is not a release
[15:25] <smoser> so it has nothing to do with this
[15:26] <smoser> i am fine to turn on desktop builds, but it is completely unrelated to thursdays release.
[15:26] <smoser> as there has never been a "release" version of desktop build.
[15:26] <smoser> i never enabled them during oneiric bring up, because i was going to wait until they stabalized.
[15:27] <Daviey> smoser: 'desktop' as in the iso release. not cloud img.
[15:27] <smoser> ah.
[15:28] <smoser> ok. i have to run.
[15:28] <smoser> utlemming, you can watch the republish at /tmp/republish-oneiric-20111010.log if you're interested
[15:28] <utlemming> smoser: my republish is just finishing
[15:29] <smoser> oh?
[15:29] <smoser> oh crap
[15:29] <utlemming> it's adding the launch ACL's now...which means a redo
[15:29] <smoser> hmm.. strange
[15:29] <utlemming> ap-northeast-1  ami-9eaa1e9f    i386    ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-i386-server-20111010
[15:29] <utlemming> ap-northeast-1  ami-a2aa1ea3    amd64   ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-amd64-server-20111010
[15:29] <utlemming> ap-southeast-1  ami-f20379a0    i386    ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-i386-server-20111010
[15:29] <utlemming> ap-southeast-1  ami-e00379b2    amd64   ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-amd64-server-20111010
[15:29] <utlemming> eu-west-1       ami-03c7f577    i386    ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-i386-server-20111010
[15:29] <utlemming> eu-west-1       ami-f7c7f583    amd64   ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-amd64-server-20111010
[15:29] <utlemming> us-east-1       ami-23d11d4a    i386    ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-i386-server-20111010
[15:29] <utlemming> us-east-1       ami-3dd11d54    amd64   ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-amd64-server-20111010
[15:29] <utlemming> us-east-1       ami-09d11d60    amd64   hvm     hvm/ubuntu-images-testing/ubuntu-oneiric-daily-amd64-server-20111010
[15:30] <utlemming> us-west-1       ami-977f23d2    i386    ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-i386-server-20111010
[15:30] <Kiall> -_-
[15:30] <utlemming> us-west-1       ami-a57f23e0    amd64   ebs     ebs/ubuntu-images-testing/ubuntu-oneiric-daily-amd64-server-20111010
[15:30] <smoser> hm... strange. my republish actually thought it should create a us-west-1 snapshot.
[15:30] <smoser> shouldnt have done that
[15:31] <RoyK> !pastebin
[15:32] <smoser> ok. i'm out for a bit. it looks like utlemming has things covered.
[15:32] <Daviey> smoser: o/
[15:46] <rbasak> Daviey: I've created a merge proposal for bug 862129 - this fixes a problem with samba upgrading to oneiric.
[15:47] <rbasak> https://code.launchpad.net/~racb/ubuntu/oneiric/samba/862129/+merge/78862
[15:48] <Daviey> rbasak: neat
[15:49] <rbasak> Daviey: please could you review/merge/upload, etc? Will it make oneiric?
[15:50] <Daviey> rbasak: looking, thanks
[16:13] <elb0w`> Why cant I get tftpd to work
[16:13] <elb0w`> I installed the packages
[16:14] <elb0w`> it just doesnt start
[16:28] <utlemming> Daviey: it looks like the update setting is still not getting set. I've confirmed that the server installs do not set it while the desktop settings do
[16:28] <utlemming> s/settings/install
[16:31] <RoyK> elb0w`: iirc you need to enable it in the inetd or similar. try atftpd to get a bit more control
[16:32] <jamespage> utlemming, smoser, Daviey: ping me and hggdh when ec2 images are ready for testing
[16:32] <jamespage> at set to go
[16:32] <jamespage> /at/all
[16:33] <utlemming> jamespage: they're hot and ready
[16:33] <utlemming> jampesage: nevermind...
[16:34] <utlemming> jamespage: okay, now they're ready http://uec-images.ubuntu.com/oneiric/20111010/
[16:34] <koolhead17> lynxman: around
[16:35] <Daviey> jamespage: hold fire until tomorrow i think
[16:36] <lynxman> koolhead17: around
[16:37] <koolhead17> lynxman: cool. got some answers from here https://answers.launchpad.net/nova/+question/154272 will ping you if stuck :)
[16:37] <lynxman> koolhead17: aah cool :)
[16:48] <elb0w`> no what I need is fucking app armor
[16:48] <elb0w`> to die in a fire
[16:49] <ikonia> done the language down please.
[16:49] <ikonia> tone
[16:51] <elb0w`> who likes apparmor though seriously
[16:52] <elb0w`> its like if you forget its there you sit for an hour debugging something
[16:56] <RoyK> elb0w`: I think ikonia is referring to the ubuntu practice, that language should be family friendly instead of admin friendly :P
[17:01] <elb0w`> sorry
[17:01] <elb0w`> I just forget about it all the time
[17:01] <elb0w`> and sit there
[17:01] <elb0w`> then after literally nothing works
[17:01] <elb0w`> its like oh hey is app armor here?
[17:02] <RoyK> elb0w`: now, what was it that stopped working?
[17:02] <elb0w`> Just wasnt able to connect to the tftp server
[17:03] <elb0w`> soon as I removed apparmor it worked like magic
[17:05]  * RoyK wonders how apparmor could have broken that....
[17:05] <jdstrand> yeah, there isn't a profile for that...
[17:12] <axisys> how do I find out if mod_php is vulnerable with this apache2?
[17:12] <axisys>  Apache/2.2.14 (Ubuntu) PHP/5.3.2-1ubuntu4.9 with Suhosin-Patch mod_ssl/2.2.14 OpenSSL/0.9.8k configured
[17:13] <elb0w`> Dont ask me
[17:13] <elb0w`> all I know is it worked after I removed it
[17:14] <koolhead17> elb0w`: use selinux then
[17:34] <koolhead17> lynxman: help help!!
[17:34] <lynxman> koolhead17: :)
[17:35] <koolhead17> now what i have done is i added a user to a project
[17:35] <koolhead17> when i created this user i got secert key and api key right
[17:35] <lynxman> koolhead17: correct
[17:36] <koolhead17> also i gave this user sysadmin privileges for my project
[17:37] <koolhead17> i also changed the novarc file and modified its access key as zipfile by defualt assigns access key as "username:project"
[17:39] <koolhead17> lynxman: am getthing nice error http://paste.ubuntu.com/705516/
[17:39] <koolhead17> am i missing something
[17:41] <lynxman> koolhead17: looks like your key is definitely borked
[17:41] <lynxman> koolhead17: do you have any logs for that operation?
[17:41] <lynxman> koolhead17: /var/log/nova/*log
[17:42] <koolhead17> indeed!! :)
[17:43] <koolhead17> http://paste.ubuntu.com/705518/
[17:43] <lynxman> koolhead17: hmm how did you add this user to the managers group? It really looks like you're missing permissions
[17:45] <koolhead17> lynxman: nova-manage role add --user=asysadmin --role=sysadmin --project=proj
[17:46] <lynxman> koolhead17: what shows up if you list users and roles
[17:48] <koolhead17> lynxman: nova-manage role has command you meant ?
[17:48] <lynxman> koolhead17: yeah let me check (don't have openstack installed here, darn)
[17:49] <koolhead17> its giving me False
[17:51] <lynxman> koolhead17: ova-manage user list
[17:51] <lynxman> koolhead17: nova-manage user list
[17:51] <koolhead17> k
[17:52] <koolhead17> i can see the user there
[17:54] <lynxman> koolhead17: try nova-manage role has asysadmin sysadmin projectname
[17:54] <lynxman> koolhead17: see if that returns true
[17:54] <koolhead17> k
[17:56] <koolhead17> False :(
[17:56] <lynxman> koolhead17: there you go, missing creds :)
[17:56] <lynxman> koolhead17: also roles have changed
[17:56] <lynxman> koolhead17: admin, itsec, projectmanager, netadmin, developer
[17:57] <lynxman> koolhead17: role add asysadmin admin projectname
[17:57] <koolhead17> lynxman: but that includes global plus local role
[17:58] <lynxman> koolhead17: will do for the project yeah
[17:58] <koolhead17> lynxman: lemme try :D
[18:27] <Takyoji> So how would someone connect this with just a few SATA harddrives directly? http://www.newegg.com/Product/Product.aspx?Item=N82E16816115100 xP
[18:28] <Takyoji> Being that it has two SAS connectors, which I'm unfamiliar with
[18:29] <patdk-wk> with sff-8087 to sata breakout cables
[18:29] <patdk-wk> that isn't an sas connector
[18:29] <patdk-wk> that is a quad sas/sata connector
[18:30] <patdk-wk> http://www.newegg.com/Product/Product.aspx?Item=N82E16816116097
[18:31] <Takyoji> ahh; as for price, would that be sane for just a 2-4 disk RAID 1?
[18:32] <patdk-wk> depends
[18:33] <patdk-wk> highpoint hasn't had a good history if their cards working with bios's very nicely
[18:33] <Takyoji> any personal brand recommendations?
[18:33] <patdk-wk> these days, I've been on lsi
[18:33] <patdk-wk> but I normally always get battery backed raid cards
[18:34] <patdk-wk> a 4port lsi is like $180 I think on there
[18:35] <patdk-wk> almost every one uses lsi, intel, supermicro, ibm, hp, ...
[18:35] <patdk-wk> they are all just rebranded
[18:36] <Takyoji> well, isn't it moreso the ICs produced by LSI and assembled by Intel, IBM, and so on? :P
[18:37] <patdk-wk> exactly how much customization can one do and still have a working card?
[18:38] <patdk-wk> in the case of ibm, disable the raid5 support in firmware
[18:38] <patdk-wk> but the real issue is, lsi won't help you if you don't have a real lsi card
[18:38] <patdk-wk> and lsi is normally good about support
[18:53] <Takyoji> Gah, it's inconvenient how NewEgg just has an asston of obscure brands for controllers and bays
[18:54] <Takyoji> also, there are backplanes with controllers, yes? Not just FakeRAID or anything?
[18:55] <RoyK> Takyoji: software raid is usually faster anyway :P
[18:57] <Takyoji> in what way? :P
[18:57] <Takyoji> Are the controllers usually even a bottleneck sometimes, or?
[18:58] <RoyK> the little processor on the board is generally slower than the one you have as your cpu
[18:58] <RoyK> so as long as you have sufficient bandwidth between disk and cpu and memory, software raid is likely to be better
[18:59] <RoyK> with SATA ports on PCIex, that shouldn't be an issue
[19:00] <RoyK> also, software raid is usually far more flexible than those controller-based RAID systems unless they are rather expensive
[19:02] <Takyoji> So then in what case would a controller be sufficient over software RAID?
[19:02] <RoyK> any sata controller will do
[19:02] <Takyoji> or hardware RAID I should say
[19:02] <RoyK> Takyoji: why hardware raid?
[19:03] <Takyoji> So that it's transparent from the kernel?
[19:03] <RoyK> what is the benefit in that?
[19:03] <RoyK> software raid is faster and easier to troubleshoot once the shit hits the fan
[19:03] <RoyK> and more flexible
[19:04] <dkn> i've also had software raid work on multiple machines by just moving the disks between them in the event of a motherboard, or controller failure
[19:05] <RoyK> with hardware raid, you'll need the same raid board if you need to move the drives - sometimes the same firmware version
[19:05] <RoyK> and even if the board says LSI something, it could be anything - old 3ware or some other stuff
[19:06] <dkn> particularly troublesome if you use the fake raid built into the motherboard.... and then your 3yr old motherboard bites the dust....
[19:06] <dkn> beem there....
[19:06] <dkn> software raid from here on out unless you're running nodes and you can afford to have then entire machine down or service
[19:06] <RoyK> most cheap NAS things use linux software raid
[19:07]  * dkn is full of typo's today
[19:07]  * RoyK uses ZFS for important stuff
[19:08] <dkn> well some of the hardware raid stuff is really well integrated with the chasis so you get the nice red light next to the failed drive so the datacenter guys can swap it out right away without needing access to the system... but that might be a little bigger scope then some of us here...
[19:09] <RoyK> dkn: there are tools that can do that for even software raid/zfs solutions
[19:09] <RoyK> most controllers come with such tools
[19:10] <RoyK> and then, when, not if, the shit hits the fan, you can easily reassembly the RAID on another box, with other controllers etc
[19:11] <dkn> i don't know if my motherboard came with the cable to attach to the backplane of the hot swap bays though.... i'd be interested in getting it working if you knew how to pass them message from mdadm to the supermicro LSI chip if i had the port =
[19:12] <RoyK> dkn: sas2ircu
[19:13] <RoyK> dkn: that works for the SAS6 controllers, such as 92xx
[19:13] <RoyK> or SAS2, perhaps, 6Gbps SAS
[19:13] <dkn> royk: thank's i'll give it a go later on :)
[19:14] <RoyK> mdadm won't flash automatically, but a small shell script might do :P
[19:14] <RoyK> dkn: using those with openindiana for some high-volume servers
[19:16] <RobinBAwesome> i'm trying to make an automated server install cd, and the installer can't see my preseed. i've got preseed.cfg at the top level of the cd and i'm trying to get it using preseed/url=file:///cdrom/preseed.cfg.
[19:17] <RobinBAwesome> does anybody know what it should be?
[19:18] <RobinBAwesome> so far the other stuff that doesn't work is url=file:///preseed.cfg, url=file:preseed.cfg, url=file:///cdrom/./preseed.cfg, file=/cdrom/preseed.cfg, and file=preseed.cfg
[19:29] <Takyoji> How do you check for harddrive failures in a software RAID (using mdadm) then?
[19:30] <RoyK> smartctl?
[19:34] <koolhead11> lynxman: around :P
[19:34] <koolhead11> i dont see much of difference in nova-manage commands :D
[19:35] <RoyK> Takyoji: most RAID controllers don't check data integrity - they just rely on the harddisk's CRC, meaning if the disk says it's an error, they mark it an error. The only filesystems doing actual data checks are (AFAIK) ZFS and btrfs. the former only works on solaris and BSD and under FUSE on linux, the latter isn't stable yet
[19:41] <Takyoji> ahh
[19:41] <Takyoji> but what I mean is for the whole entire drive failing
[19:41] <Takyoji> is there a way to write a script to just occasionally check?
[19:42] <Takyoji> Or is it just a matter of trying to read a file, and if there's an IO error, the harddrive has failed? :P
[19:45] <RoyK> Takyoji: if a drive fails, md will see that quite easily
[19:45] <RoyK> so if you have a spare, that will be set into service
[19:45] <RoyK> if not, well, no harm done unless you find bad data elsewhere
[19:46] <Takyoji> So in software RAID, if there's an error on one of them (but not complete failure at all); it wouldn't be able to autocorrect it, or move it to a different sector, or? (if using EXT3 or EXT4)
[19:46] <RoyK> Takyoji: I use zfs on openindiana for important stuff
[19:47] <RoyK> it's not related to the filesystem on top
[19:47] <RoyK> md will find the data if it's available
[19:48] <RoyK> and reconstruct it from parity if a sector goes bad
[19:50] <RoyK> Takyoji: generally, software raid is like hardware raid - the only difference is that the software is running on linux and not on some chip
[19:50] <RoyK> "hardware" raid is just software raid offloaded to another cpu
[19:51] <RoyK> some of the expensive boards have hardware accellerators for that, but I somewhat doubt that makes a big difference
[19:51] <RoyK> it's just XOR after all
[19:53] <Takyoji> ahh
[19:53] <Takyoji> also, is this really the form factor of 2 CD-ROM drives? http://www.newegg.com/Product/Product.aspx?Item=N82E16817994075
[19:54] <Takyoji> 5.25"
[19:54] <RoyK> I have something like that
[19:54] <RoyK> works well
[19:54] <RoyK> not that model, but something from Lian Li
[19:55] <RoyK> without hotswap, though
[19:55] <RoyK> keep in mind that the controller and driver must support hotswap for it to be usable
[19:55] <Takyoji> ahh
[19:55] <Takyoji> but anyway, I measured  5.75" x 3.5" on my system, which doesn't seem to match the proportions specified
[19:56] <Takyoji> unless if it's just the faceplate that is that size.
[19:56] <RoyK> 5.25"
[19:56] <RoyK> that's the size of a CDROM
[19:57] <RoyK> or a normal DVD drive
[19:57] <Takyoji> alright
[19:59] <cce> For the purposes of a virtual appliance, I'm wondering how small the "minimal" unbuntu server distribution is, using vmbuilder it seems to be making 350MB images.  Is there a way to get a much smaller one?
[19:59] <RoyK> Takyoji: if you don't need the flexibility of md, as in adding drives to an existing raid5 to expand it, I'd suggest you should try zfs on openindiana or fbsd - it's far better when it comes to data safety
[20:00] <RoyK> cce: with current USB sticks at 16GB, is it really necessary to make it that small?
[20:00] <cce> RoyK: well, if you're downloading your app as an appliance?
[20:00] <RoyK> I get your point
[20:00] <cce> 350M is ~145M compressed, that's a good 10 min download
[20:00] <RoyK> cce: perhaps try debian - it's a bit slimmer
[20:01] <RoyK> ubuntu includes a bunch of stuff you normally don't need
[20:01]  * cce can get FreeBSD down to 30-40M, but it's increasingly difficult to fight linux dominance 
[20:02] <RoyK> did you try debian?
[20:02] <cce> not yet... linux noob I am
[20:02] <RoyK> it should be down to 100 megs on a standard install
[20:02] <cce> RoyK does it use libvirt and stuff?
[20:02] <cce> RoyK: that's useful
[20:02] <cce> Thank you.
[20:03] <RoyK> ubuntu came from the land of debian, so it's the same tools, same packaging system etc
[20:04] <Takyoji> It's just RAID 1. :P
[20:04] <Takyoji> 3 harddrives
[20:04] <RoyK> what is?
[20:04] <Takyoji> of what we'll have.
[20:05] <RoyK> three-side mirrors?
[20:05] <Takyoji> yes
[20:05] <RoyK> is that what you use?
[20:06] <Takyoji> a 3-disk RAID 1
[20:06] <RoyK> then what's the problem?
[20:07] <Takyoji> just referring to "as in adding drives to an existing raid5"
[20:07] <RoyK> Takyoji: md supports that
[20:07] <RoyK> or converting between raid5 and raid6
[20:07] <RoyK> linux md is far more than just mirrors
[20:07] <RoyK> lvm only supports mirrors
[20:08] <RoyK> but that's another thing
[20:08] <RoyK> if you only need mirroring, anything will do
[20:09] <RoyK> md, lvs, zfs, even btrfs can do that
[20:10] <SpamapS> LVM has striping too
[20:11] <Takyoji> and this will have the OS install on it and everything
[20:11] <Takyoji> as an aside
[20:11] <Omni_Lynk> what would be a good small gui for Ubuntu Server? Something that will not be used much past initial setup?
[20:11] <RoyK> Takyoji: last I checked, you can't boot from RAID[56]
[20:11] <Takyoji> I know mdadm is in the alternate installer for Ubuntu; but not sure if there's lvm, zfs, etc
[20:11] <RoyK> SpamapS: striping doesn't count
[20:12] <Takyoji> As stated, RAID 1. :P
[20:12] <RoyK> as for raid-1, md is fine, so is lvs
[20:12] <Takyoji> I just don't know how I'd go at setting up anything other than mdadm
[20:13] <RoyK> well, mdadm is safe
[20:13] <RoyK> and very stable
[20:13] <Takyoji> Or would it be as simple as: installing to one harddrive, then in the OS environment to install and setup the RAID?
[20:13] <RoyK> just setup a three-drive mirror
[20:14] <RoyK> the installer will install grub on all of them and it should be safe as heaven
[20:14] <Takyoji> So just go with mdadm then? :P
[20:14] <RoyK> Takyoji: yes
[20:14] <Takyoji> alright
[20:15] <RoyK> mdadm has been stable for what, 10 years? 15?
[20:19] <Omni_Lynk> what would be a good small gui for Ubuntu Server? Something that will not be used much past initial setup?
[20:20] <RoyK> apt-get install ubuntu-desktop :P
[20:20] <RoyK> Omni_Lynk: I'd suggest you learn administrating by the commandline, though
[20:21] <Omni_Lynk> i'm planning on it, however i'm under the gun and i do not know all of the commands yet to actually swim in the Command Line.  best i could do is stay aflot for 5 minutes before drowning.
[20:22] <RoyK> Omni_Lynk: there are few tools available on the gui - the gui is meant for users, not admins
[20:23] <RoyK> Omni_Lynk: do it the hard way :P
[20:24] <Omni_Lynk> yeah, trial by fire. That is what i asked, had a good 2 months to play with it, then server's motherboard blew. now i have everything running on a desktop machine and need to get it moved & shares done as fast as possible. :(
[20:27] <jason_II> So I'm on an Ubuntu 10.04 box with no network manager... static IP, subnet, and gateway set in /interfaces and DNS servers along with search and domain are in /etc/resolv.conf. I can ping IPs externally, but I cannot ping hostnames externally (www.weather.com, etc(. What am I missing??
[20:31] <Patrickdk> a working resolv.conf file
[20:31] <jason_II> nevermind
[20:32] <jason_II> I didnt realize I had to have "nameserver" before each DNS
[20:32] <jason_II> I thought I could just have the IP of the DNS *facepalm*
[20:32] <Patrickdk> heh
[20:32] <Patrickdk> I personally would remove the domain and search
[20:32] <Patrickdk> unless you love to use just hostnames
[20:32] <jason_II> from past experience I may need it. This is an LTSP thin client server.
[20:32] <Patrickdk> I always find it annoying, and abuses the crap out of my dns servers
[20:32] <Patrickdk> ok :)
[20:32] <jason_II> I thought I remember clients having an issue and putting in search/domain helped...
[20:32] <jason_II> I'll try without and see what happens. :P
[20:33] <Patrickdk> I always find it annoying that every single lookup failure causes it to trigger a search
[20:33] <Patrickdk> mygod.example.com doesn't exist? well does mygod.example.com.example.org exist?
[20:33] <jason_II> yeahhh
[20:33] <jason_II> gotcha
[20:34] <jason_II> Im just happy I got these darn powerpc systems running as clients...
[21:24] <dkn> http://www.norcotek.com/product_images/flyer/rpc4164ff.jpg
[21:26] <Dulcin> Hi, is it possible to have an ordered list start at 1.1 instead of 1?
[21:26] <Dulcin> oh crap wrong channel
[21:26] <Dulcin> my bad
[21:34] <koolhead17> Daviey: around?
[21:35] <soren> smoser: I'm trying to use the vmdk from http://cloud-images.ubuntu.com/releases/natty/release/ in kvm... That's not going to work is it?
[21:35] <soren> smoser: That's the VMDK+QEmu problem you mentioned at some point, right?
[21:35] <utlemming> soren: nope
[21:36] <soren> utlemming: Because the vmdk flavour isn't support by QEmu?
[21:36] <utlemming> soren: correct
[21:36] <soren> utlemming: It'll happily convert it to RAW, but the resulting image looks like nonsense.
[21:36] <utlemming> soren: those are compressed VMDK's
[21:36] <soren> utlemming: Oh.
[21:37] <utlemming> soren: I thought smoser converted those to qcow2's, so I'm surprised to see they are vmdk's
[21:38] <utlemming> your other option is to convert to raw using VBoxManage and then convert to qcow2 using qemu-img convert
[21:38] <utlemming> it sucks I know
[21:38] <soren> utlemming: I tried to figure out how to do it with VBoxManage, but failed.
[21:39] <soren> utlemming: Note: These are the natty images.
[22:03] <utlemming> soren: sorry...missing your chat...still around?
[22:04] <utlemming> soren: 1) vboxmanage clonehd --format raw natty.vmdk natty.raw
[22:05] <utlemming> soren: 2) qemu-img convert -c -O qcow2 natty.raw natty.qcow2
[22:05] <utlemming> soren: 3) profit....or something like that
[22:06] <utlemming> soren: drop the "-c" if you want uncompressed qcow2. I would go compressed because it is faster on the upload, and save you ~400MB
[22:06] <lynxman> koolhead17: hey, back now :
[22:06] <koolhead17> lynxman: :)
[22:07] <koolhead17> i was having conversation with vishy. he suggested i need to modify my nova.conf with deprecated  option
[22:08] <koolhead17> and also some change with older configuration has to be done in paste.ini or something, i need to check tomorrow as am home finally :)
[22:09] <koolhead17> lynxman: where you able to get toggle to True with opetion
[22:09] <koolhead17> nova-manage role has option?
[22:09] <lynxman> koolhead17: I normally just use the admin user straight
[22:10] <koolhead17> lynxman: i was testing RBAC control so i had to assign those options :)
[22:11] <koolhead17> one thing which i have to test tomorrow is nova-manage role modify "" "" True option
[22:12] <koolhead17> cos after assigning this i was able to add/manage an instance from dashboard even. :)
[22:14] <lynxman> koolhead17: cool :)
[22:15] <koolhead17> lynxman: i will paste exact wat vishy told me tomorrow once am in office as in about noauth and modifying paste.ini or sumthing :P
[22:16] <koolhead17> gosh i hate these moron economists and media channels screaming laud about recession :(
[22:16] <lynxman> koolhead17: lol
[22:17] <koolhead17> lynxman: http://www.bloomberg.com/video/77239564/
[22:17] <koolhead17> :D
[22:17] <koolhead17> i will move to agriculture then.
[22:18] <lynxman> koolhead17: bloomberg? jeez man :)
[22:18] <koolhead17> lynxman: i just hate all these big business channels :D
[22:19] <lynxman> koolhead17: heh
[22:19] <koolhead17> after watching them i ind economists are the folks who make most out of everything
[22:20] <nigelb> lynxman: hey got a minute?
[22:21] <lynxman> nigelb: sure!
[22:21] <nigelb> lynxman: what's the difference between orchestra and ensemble?
[22:21] <nigelb> (friend
[22:21] <nigelb> (friend's doing a pressentation)
[22:21] <lynxman> nigelb: orchestra and juju are two different tools for two very different purposes
[22:22] <nigelb> juju is create, use, destroy?
[22:22] <dkn> i think my rsync is doing something with the soft links too, http:/paste.ubuntu.com/705635/ i'm copying the serverbackup to backup..... 656gb to 1.1TB????
[22:22] <lynxman> nigelb: juju is a high level orchestration service, giving you a 10,000 feet view of your system, being able to deploy charms (full apps) and all that they need around them
[22:22] <lynxman> nigelb: orchestra is a hardware provisioning service, to be able to preseed machines and for example, let them get configured by juju :)
[22:22] <nigelb> ah
[22:22] <koolhead17> lynxman: lol!! 10,000 feet
[22:23] <lynxman> nigelb: I'd say that in comparison for example with chef, orchestra would be below it and juju above
[22:23] <dkn> i did rsync -a serverbackup/ backup/
[22:23] <lynxman> nigelb: hehe :)
[22:23] <koolhead17> nigelb: your friend doing presentation where?
[22:23] <nigelb> lynxman: Thanks!
[22:23] <nigelb> koolhead17: Philadelphia LUG
[22:24] <koolhead17> nigelb: Woahh!! i i thought some meetup in blr :)
[22:24] <koolhead17> cool
[22:24] <lynxman> nigelb: there's some presentations you can tap some info from as well
[22:24] <lynxman> nigelb: I did one at puppetconf, let me get the URL...
[22:24] <nigelb> lynxman: ooh, that'd rock :)
[22:26] <lynxman> nigelb: http://www.slideshare.net/derleiermann/juju-puppetconf
[22:26] <nigelb> lynxman: *whee* thanks!
[22:26] <koolhead17> lynxman: i feel orchestra is daddy of cobbler!! :D
[22:26] <koolhead17> *big
[22:26] <lynxman> koolhead17: cobbler is a big part of it, but there's more :)
[22:26] <lynxman> koolhead17: like a python written ec2 metadata-like service (made by yours truly)
[22:27] <lynxman> koolhead17: nagios integration, rsyslog integration, squid integration, etc
[22:27] <koolhead17> lynxman: have to finish diablo integration on oneiric first then i will check orchestra :D
[22:27] <koolhead17> lynxman: that sounds cool though. :P
[22:27] <lynxman> koolhead17: sounds fun :)
[22:28] <koolhead17> lynxman: what is your current setup, your running everything from oneiric repo
[22:28] <koolhead17> am still using nova on natty and else on oneiric
[22:29] <koolhead17> and i am really curious since we have like 2 days before oneiric release are we shipping with keystone :D
[22:29] <Daviey> koolhead17: yo'll
[22:29] <koolhead17> hi Daviey
[22:29] <lynxman> koolhead17: yeah we need to work on the latest and greatest, so oneiric it is ;)
[22:29] <lynxman> Daviey: hey sir o/
[22:30] <koolhead17> lynxman: last time i tried oneiric i was not able to use mysql database so i left and moved back to natty
[22:30] <koolhead17> :D
[22:30] <koolhead17> Daviey: sir. spent whole day on role based access control, still no luck :(
[22:32] <koolhead17> zul: hey :)
[22:36] <Daviey> koolhead17: :(
[22:37] <Daviey> hey lynxman
[22:37] <koolhead17> Daviey: am sure i will have it done tomorrow for sure :D
[22:37] <koolhead17> just need to understand the noauth thing and what old change has to be done in paste.ini thing
[22:38] <Daviey> koolhead17: well i need to go to sleep, but i'd love to hear how you got on.
[22:38] <lynxman> Daviey: sleep tight and see you tomorrow ;)
[22:38] <Daviey> you don't want noauth.. that really does mean NO AUTH :)
[22:38] <Daviey> lynxman: o/
[22:38] <koolhead17> Daviey: i troubled vishy or a while and he replied with some suggestion
[22:38] <koolhead17> :D
[22:38] <utlemming> Daviey: when you get up, I tracked down the APT thing too...its a meta problem affecting ubuntu-server
[22:39] <koolhead17> *for
[22:39] <Daviey> koolhead17: oh rocking, fancy forwarding that - i'd love to read it
[22:39] <Daviey> utlemming: oh?
[22:39] <utlemming> Daviey: you can read the bug, but the fix is to mark "update-notifier-common" into the server tasks
[22:39] <utlemming> Daviey: the change happened between Lucid and Natty, so Natty and Oneiric are both affected
[22:40] <koolhead17> Daviey: tomorrow 1st thing will share it with you. i need to understand what he meant by change to older configuration in paste.ini
[22:40] <Daviey> utlemming: smoser was worried about introducing that on m1.tiny...
[22:40] <Daviey> koolhead17: thanks
[22:40] <koolhead17> if i crack this i think i can get dashboard working with all powers!! :D
[22:40] <soren> utlemming: "the APT thing"?
[22:40] <utlemming> Daviey: I have some reservations about that too
[22:40] <utlemming> soren: yup
[22:40] <Daviey> utlemming: wait, Lucid was smoser's first release, right?
[22:41] <soren> Daviey: No.
[22:41] <utlemming> Daviey: not sure
[22:41] <soren> Daviey: Karmic.
[22:41] <Daviey> bug 870121
[22:41] <Daviey> ie, you can't see pending updates - and the apt-cache is always stale.
[22:43] <Daviey> soren: So this is something we can blame on smoser?
[22:43] <soren> Daviey: We can blame lots of stuff on smoser. I don't know what we're talking about right now, though.
[22:43] <soren> Daviey: Oh, bug 870121?
[22:44] <utlemming> soren: that's the bug we're talking about
[22:44]  * Daviey takes comfort in attributing the blame to smoser
[22:46] <soren> utlemming: Do we know how it gets set on regular installs?
[22:46] <Daviey> soren: keep up at the back
[22:47] <Daviey> 23:39 < utlemming> Daviey: you can read the bug, but the fix is to mark "update-notifier-common" into the server tasks
[22:47] <Daviey> which implies that is done on the desktop
[22:48] <soren> Ah.
[22:48] <utlemming> Daviey, soren: the short story is that "ubuntu-desktop" pulls in "update-notifier-common" as a depedency. The server task list does not include update-notifier.
[22:48] <soren> So..
[22:48] <soren> The bug says:
[22:48] <utlemming> update-notifier-common provides the apt config that we want
[22:48] <soren> "In Desktop and Server ISO we set APT::Periodic::Update-Package-Lists:
[22:48] <soren> "
[22:49] <soren> So how/why does it happen on server installs?
[22:49] <utlemming> soren: I _installed_ a new minimial server on Friday and it did not have it
[22:49] <soren> utlemming: Ah.
[22:49] <utlemming> My beta1 ISO install for server had it set
[22:49] <soren> So the bug lies.
[22:49] <Daviey> soren: I think my initial anlaysis was incorrect.
[22:49] <soren> I agree with smoser. The cloud images should do what the server install does.
[22:50] <Daviey> and server is doing it wrong.
[22:50] <soren> So if the server isntall doesn't do this, I'd say the bug is invalid.
[22:50] <soren> In its current form, at least.
[22:50] <Daviey> the bug isn't invalid.
[22:50] <soren> We may want both server and cloud to have this enabled, but that's a different story.
[22:50] <soren> Ok.
[22:51] <Daviey> having an out of date caceh and not able to see pending updates (esp. with byobu enabled), isn't cool.
[22:51] <soren> I either disagree or am missing something.
[22:51] <Daviey> soren: the thought you you disagreeing with me astounds me :)
[22:54] <soren> Daviey: Yeah. Preposterous.
[22:55] <Daviey> it really comes down to target audiences.
[22:55] <Daviey> :)
[22:55] <utlemming> Daviey: how hard is it to change the meta-data? I may not be the biggest fan of byubo, but showing no updates is a bad user experience and will result in bugs. I'm inclined to agree that this is a bug.
[22:56] <Daviey> utlemming: Changing it is easy, the concern is, rightly so, is will m1.tiny choke on this?
[22:57] <utlemming> Daviey: I have concerns that it will affect the m1.tiny/t1.micros, but I also don't think that the t1.micros are usable for anything other than a glorified shell account.
[22:57] <Daviey> it's just and automated suod apt-get update... so should be ok... but very late.
[22:57] <utlemming> Daviey: I have argued that we should limit support on them just because they are so resource constrained
[22:58] <Daviey> well it did take too much support to try and gtet java working on them.
[22:58] <Daviey> So i would back that claim, i think.
[22:59] <soren> Sorry, what? Are the m1.tiny or m1.micro sizes unable to apt-get update?!?
[22:59] <utlemming> Java is not working on them yet
[23:00] <utlemming> soren: no, the concern is that having cronjobs could affect the performance of a t1.micro/m1.tiny.
[23:00] <utlemming> soren: the t1.micro is very starved and cronjobs and the like can push them into a hyper-visor enforced coma
[23:02] <soren> utlemming: The overhead of being a cronjob should be miniscule.
[23:02] <utlemming> soren: at least with the t1.micro, the machine just drops off the face of the planet until the load characteristics match what Amazon wants.
[23:03] <soren> utlemming: So it's really about the ability to run "apt-get update". Right?
[23:03]  * utlemming looks the EC2 doc...
[23:03] <Daviey> bed
[23:03] <Daviey> fiish this tomrrow
[23:04] <Daviey> nn
[23:04] <utlemming> 'night Daviey
[23:04] <smw> utlemming,  the micro does not drop off completely
[23:05] <smw> utlemming, it does go very slow though. When doing a large number of updates at once, it finishes slowly. But it is not too bad
[23:06] <utlemming> smw: I have to respectfully disagree. If you have a mostly idle micro, then occasional spikes are okay. If you one under sustained load, it will become unresponsive until the backlog clears. Try running LTP on a t1.micro and watch it become unusable for the next 21 hours.
[23:06] <koolhead17> nyt Daviey :D
[23:06] <smw> utlemming, 21 hours?!
[23:06] <utlemming> smw: yup
[23:06] <smw> utlemming, I have never had even close to that type of problems... what is LTP?
[23:07] <utlemming> smw: Linux Test Suite
[23:07] <smw> utlemming, but then again, the most computationally intensive thing I do on a micro is large updates or compiling
[23:08] <utlemming> smw: which fits the performance expectations for a t1.micro -- occasional spikes in CPU usage.
[23:08] <smw> utlemming, yeah
[23:08] <utlemming> so the fear here is that a t1.micro performing near the baseline will get pushed over the top and then die
[23:08] <smw> utlemming, my company is not stingy and ups the instance type whenever we do anything with computation or we put something into prod
[23:09] <utlemming> smw: glad to hear :)
[23:11] <utlemming> smw: the t1.micro concerns me for a couple reasons. The first is that most people don't understand it and the second is that when it behaves like it is supposed to, they blame Ubuntu.
[23:12] <utlemming> with all said, I am not sure that turning the job on is a bad thing per se -- having updates available is a good thing on every other instance type -- but the perception is what bothers me
[23:12] <smw> utlemming, nice :-)
[23:12] <smw> utlemming, we don't get many people confused with a slow micro on ##aws
[23:23] <dkn> in lvm..... a vg can span multiple pv, can lv span multiple vg?
[23:27] <Jasonn> Tohuw:
[23:27] <Tohuw> Hello. This will be better, so as not to further fill up #ubuntu. Anyway... you moved the lock file, and did it spawn another one?
[23:27] <Tohuw> Run ls -a /var/lib/dpkg to find out.
[23:28] <Jasonn> .  ..  alternatives  available  available-old  cmethopt  diversions  diversions-old  format  info  lock  lock.OLD  parts  statoverride  statoverride-old  status  status-old  triggers  updates
[23:28] <Jasonn>  
[23:28] <Jasonn> It appears so :o
[23:28] <Jasonn> and I also dont want to restart the server :3
[23:30] <Tohuw> Understood. Try removing any processes employing the file: sudo fuser -cuk /var/lib/dpkg/lock
[23:30] <Tohuw> Then try removing it
[23:30] <Tohuw> or moving it; either is fine. It's just a lock file and we can see your system is capable of using it.
[23:30] <Tohuw> Ruh roh?
[23:32] <Tohuw> dkn: no
[23:32] <Jasonn> Tohuw: well, you coulda given me a heads up
[23:33] <Jasonn> :3
[23:33] <dkn> kinda odd.. it's supposed to be this gloriously flexible system...
[23:33] <Tohuw> Jasonn: :\ your network connection died when you fuser'd the file, didn't it?
[23:33] <dkn> i guess the vg does all the flexing..
[23:33] <Tohuw> dkn: Don't confuse flexibility with "uses the logic I expect it to use"
[23:34] <dkn> lol
[23:34] <Tohuw> Jasonn: If so, that was an unforseen consequence. I'm sorry to hear that happend.
[23:38] <Tohuw> hrm. I'll be semi-around Jasonn, should you need additional assistance.
[23:46] <Jasonn> Tohuw: yeah, it fixed it, thanks :D - Actually, it killed most of my processes
[23:48] <Tohuw> Jasonn: eek. That may be because dpkg was dealing with multiple dependencies or touching a lot of areas to get mysql installed... dpkg follows instructions to stop services and what not, and you just told it to go get bent. Hopefully, you should be up and running now. You may want to run some checks though. man dpkg for information; it's important to learn about this.
[23:48] <z3r0c0d3> hi alll
[23:49] <Tohuw> !hello | z3r0c0d3
[23:49] <z3r0c0d3> anybody could help me with SAMBA
[23:49] <Tohuw> !ask | z3r0c0d3
[23:50] <z3r0c0d3> i cant put it to work
[23:50] <Tohuw> !doesntwork | z3r0c0d3
[23:50] <Tohuw> (ubottu gives more support on the ubuntu channels than everyone else combined.)
[23:51] <patdk-lap> ya, but the postfix bot puts it to shame :)
[23:51] <patdk-lap> it's triggers a few times a second :)
[23:51] <z3r0c0d3> i cant access ir from ubuntu desk
[23:51] <twb`> patdk-lap: dpkg's is much better, too
[23:51] <z3r0c0d3> only for windows
[23:51] <twb`> patdk-lap: I don't know why ubottu wasn't seeded with dpkg bots info entries
[23:52] <z3r0c0d3> how can i get this working ?
[23:52] <patdk-lap> z3r0c0d3, what is the error?
[23:52] <patdk-lap> you aren't giving any details
[23:52] <z3r0c0d3> sorry
[23:53] <z3r0c0d3> when i access from windows it ask me for a user name and password
[23:53] <patdk-lap> I mean, ive been running samba for a good 15years now, haven't had an issue
[23:53] <Tohuw> z3r0c0d3: "can't access it" is not enough information. What is the error? Where is the share? What version of Ubuntu Server/what version of SAMBA (run smbstatus). Do you have a login saved for the share (check gnome-keyring)?
[23:53] <z3r0c0d3> but from linux it says that cant retrive de server list
[23:54] <Tohuw> Ah. Now you're providing something. So, what happens if you type smb://yourserver/yoursharename?
[23:54] <Tohuw> (into nautilus)
[23:55] <z3r0c0d3> error to mount windows share
[23:55] <patdk-lap> it didn't ask for username/pass?
[23:55] <z3r0c0d3> error: failed to mount the windows share
[23:55] <z3r0c0d3> no
[23:55] <z3r0c0d3> didnt
[23:57] <Tohuw> Ok, from terminal on the Ubuntu client: smbclient -L servername
[23:57] <patdk-lap> must be saved or you have guest access on
[23:57] <patdk-lap> it asks me atleast
[23:58] <Tohuw> patdk-lap: there are other possibilities.
[23:58] <patdk-lap> ya, I know, hundreds :)
[23:59] <z3r0c0d3> connection to server failed (Error NT_STATUS_UNSUCCESFUL)
[23:59] <Tohuw> However, I need to run. z3r0c0d3, I will be back in ~30 minutes, if no one has helped you by then.
[23:59] <patdk-lap> what version of samba is on the server?