Tuesday, 2010-03-09

[2010/03/09 00:16:45] @ Log started by gepetto
[2010/03/09 00:16:45] @ rmiller4pi8 joined channel #puppet
[2010/03/09 00:18:06] @ Quit: rmiller4pi81: Ping timeout: 265 seconds
[2010/03/09 00:19:23] @ Quit: OpenMedia: Quit: Leaving.
[2010/03/09 00:25:10] @ Quit: cwebber: Quit: cwebber
[2010/03/09 00:29:46] @ Quit: pheezy: Remote host closed the connection
[2010/03/09 00:37:34] @ Quit: biertie: Ping timeout: 264 seconds
[2010/03/09 00:48:40] @ Quit: \ask: Remote host closed the connection
[2010/03/09 00:49:07] @ \ask joined channel #puppet
[2010/03/09 00:50:49] @ Quit: jfreeman: Remote host closed the connection
[2010/03/09 00:51:31] @ Quit: \ask: Client Quit
[2010/03/09 00:51:47] @ kc7zzv joined channel #puppet
[2010/03/09 00:55:00] @ Quit: bug: Quit: bug
[2010/03/09 00:59:39] @ nexx joined channel #puppet
[2010/03/09 01:09:22] @ bug joined channel #puppet
[2010/03/09 01:14:49] @ Cope joined channel #puppet
[2010/03/09 01:23:31] @ Quit: ahuman: Remote host closed the connection
[2010/03/09 01:28:06] @ Quit: bkero: Quit: ZNC - http://znc.sourceforge.net
[2010/03/09 01:28:59] @ bkero joined channel #puppet
[2010/03/09 01:29:10] @ Quit: bkero: Changing host
[2010/03/09 01:29:10] @ bkero joined channel #puppet
[2010/03/09 01:31:06] @ Quit: bkero: Excess Flood
[2010/03/09 01:31:17] @ vermeer joined channel #puppet
[2010/03/09 01:32:06] @ bkero joined channel #puppet
[2010/03/09 01:32:16] @ Quit: bkero: Excess Flood
[2010/03/09 01:32:37] @ bkero joined channel #puppet
[2010/03/09 01:32:38] @ Quit: bkero: Excess Flood
[2010/03/09 01:33:14] @ JJA joined channel #puppet
[2010/03/09 01:33:36] @ bkero joined channel #puppet
[2010/03/09 01:33:37] @ Quit: bkero: Excess Flood
[2010/03/09 01:34:40] @ bkero joined channel #puppet
[2010/03/09 01:34:44] @ Quit: bkero: Excess Flood
[2010/03/09 01:35:07] @ bkero joined channel #puppet
[2010/03/09 01:35:12] @ Quit: bkero: Excess Flood
[2010/03/09 01:37:08] @ bkero joined channel #puppet
[2010/03/09 01:37:15] @ Quit: bkero: Excess Flood
[2010/03/09 01:38:38] @ bkero joined channel #puppet
[2010/03/09 01:39:51] @ PaulWay[w] left channel #puppet ("I must go, there are too many tangrams on this channel.")
[2010/03/09 01:50:13] @ Quit: JJA: Ping timeout: 264 seconds
[2010/03/09 01:50:40] @ agentk joined channel #puppet
[2010/03/09 01:52:28] <ohadlevy> volcane: I've pushed extlookup queries into foreman stable branch
[2010/03/09 01:55:26] @ m1nish joined channel #puppet
[2010/03/09 01:58:05] @ Quit: kc7zzv: Quit: kc7zzv
[2010/03/09 01:58:55] @ kubicek joined channel #puppet
[2010/03/09 02:11:26] @ kc7zzv joined channel #puppet
[2010/03/09 02:17:25] @ Quit: alban2: Ping timeout: 268 seconds
[2010/03/09 02:17:29] @ agentk left channel #puppet ()
[2010/03/09 02:26:20] @ pmorillo joined channel #puppet
[2010/03/09 02:27:05] @ zipkid joined channel #puppet
[2010/03/09 02:30:31] @ Quit: bodepd: Quit: bodepd
[2010/03/09 02:31:27] @ ckauhaus joined channel #puppet
[2010/03/09 02:37:07] @ ohlolohlol joined channel #puppet
[2010/03/09 02:47:11] @ elementai joined channel #puppet
[2010/03/09 02:52:39] @ JJA joined channel #puppet
[2010/03/09 02:54:35] @ Quit: zipkid: Quit: leaving
[2010/03/09 02:55:35] @ zipkid joined channel #puppet
[2010/03/09 02:59:55] @ DavidS joined channel #puppet
[2010/03/09 03:10:08] @ erlingre joined channel #puppet
[2010/03/09 03:10:20] @ giskard joined channel #puppet
[2010/03/09 03:14:00] @ Quit: mizzy: Ping timeout: 252 seconds
[2010/03/09 03:16:58] <telmich> tcwofür das?
[2010/03/09 03:17:53] <SyTonnerre> ww? ;)
[2010/03/09 03:22:05] <telmich> hohe latenz
[2010/03/09 03:28:43] @ dim0 joined channel #puppet
[2010/03/09 03:31:07] @ Ramonster joined channel #puppet
[2010/03/09 03:32:09] @ Quit: cliff-hm: Ping timeout: 245 seconds
[2010/03/09 03:33:16] @ suchu joined channel #puppet
[2010/03/09 03:33:47] @ mizzy joined channel #puppet
[2010/03/09 03:33:59] @ mattwynne joined channel #puppet
[2010/03/09 03:36:22] @ Quit: vermeer: Ping timeout: 264 seconds
[2010/03/09 03:39:08] @ tim|macbook joined channel #puppet
[2010/03/09 03:39:15] @ verwilst joined channel #puppet
[2010/03/09 03:49:47] @ Groenleer is now known as Groenleer`Away
[2010/03/09 03:51:32] @ MattyM joined channel #puppet
[2010/03/09 03:51:40] @ gebi joined channel #puppet
[2010/03/09 03:54:49] @ Quit: giskard: Remote host closed the connection
[2010/03/09 04:00:25] @ alban2 joined channel #puppet
[2010/03/09 04:00:58] @ danielbln joined channel #puppet
[2010/03/09 04:03:54] @ Robbie_ joined channel #puppet
[2010/03/09 04:06:30] @ Quit: macbar: Remote host closed the connection
[2010/03/09 04:08:09] @ Quit: mizzy: Ping timeout: 265 seconds
[2010/03/09 04:09:58] @ Djelibeybi joined channel #puppet
[2010/03/09 04:12:36] @ biertie joined channel #puppet
[2010/03/09 04:12:41] @ vermeer joined channel #puppet
[2010/03/09 04:12:43] @ TREllis joined channel #puppet
[2010/03/09 04:19:30] <DavidS> hi, just changed all File%source arguments to "puppet:///modules/...", but I still get the depreciation notice. What now?
[2010/03/09 04:20:49] @ Quit: kvick: Ping timeout: 264 seconds
[2010/03/09 04:21:27] @ kolla joined channel #puppet
[2010/03/09 04:21:30] <Volcane> you must have missed some cos that fixes it for me
[2010/03/09 04:22:20] @ giskard joined channel #puppet
[2010/03/09 04:24:13] @ Quit: Djelibeybi: Quit: Leaving
[2010/03/09 04:25:34] * DavidS 's grep-fu was failing. Thanks, Volcane!
[2010/03/09 04:27:32] @ mizzy joined channel #puppet
[2010/03/09 04:27:34] <Volcane> kewl, its a pita :)
[2010/03/09 04:31:46] @ Quit: mizzy: Ping timeout: 248 seconds
[2010/03/09 04:33:13] * DavidS needs more caffeine
[2010/03/09 04:33:29] * odyi is on his 4th pot for the day
[2010/03/09 04:33:49] @ Quit: bobbyz: Ping timeout: 260 seconds
[2010/03/09 04:34:12] <ohadlevy> DavidS: if thats the only problem you had with upgrading to 0.25.x consider your self lucky ;)
[2010/03/09 04:34:22] <SyTonnerre> http://www.youtube.com/watch?v=osTW2fDsZHk
[2010/03/09 04:34:26] <SyTonnerre> About cups of coffee
[2010/03/09 04:34:45] <SyTonnerre> I really wonder how I can get 0.25 to install its bins to the right place
[2010/03/09 04:34:56] <SyTonnerre> Once that's solved I can upgrade
[2010/03/09 04:35:16] <tim|macbook> DavidS: !! you're still alive!! :D
[2010/03/09 04:35:24] <DavidS> ohadlevy: actually, there were not that many problems, this one was only annoying, because I missed one place, because i grepped too much
[2010/03/09 04:36:00] <DavidS> tim|macbook: yeah, although it was a close call (http://www.cheesy.at/en/2010/03/hoellenritt/) :-) :-)
[2010/03/09 04:36:23] <DavidS> SyTonnerre: what about using one of the available packages (gem/rpm/deb) ?
[2010/03/09 04:36:49] <SyTonnerre> DavidS, I'm the package maintainer P:
[2010/03/09 04:36:50] <SyTonnerre> :P
[2010/03/09 04:37:10] <ohadlevy> SyTonnerre: for which dist?
[2010/03/09 04:37:43] <SyTonnerre> http://bit.ly/cS4IOV
[2010/03/09 04:37:54] <DavidS> SyTonnerre: :-P
[2010/03/09 04:38:21] <tim|macbook> DavidS: meh... it's time for spring to start, I'm kinda done with all the snow and ice stuff
[2010/03/09 04:39:18] <SyTonnerre> ohadlevy, pkgsrc
[2010/03/09 04:39:54] <SyTonnerre> ohadlevy, I tried to upgrade but couldn't keep it from putting puppetd and puppetmasterd to some ruby dir
[2010/03/09 04:42:00] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 04:42:24] <DavidS> tim|macbook: +1
[2010/03/09 04:42:38] <DavidS> oh, ruby love: err: Report munin failed: Internal error: The :owner setting for munin report hack: basedir must be either 'root' or 'service', not 'service'
[2010/03/09 04:43:06] @ Quit: ckauhaus: Read error: Operation timed out
[2010/03/09 04:45:17] @ bobbyz joined channel #puppet
[2010/03/09 04:48:45] @ mizzy joined channel #puppet
[2010/03/09 04:52:10] @ kvick joined channel #puppet
[2010/03/09 04:53:59] @ mattwynne joined channel #puppet
[2010/03/09 04:54:12] <DavidS> -> #3354
[2010/03/09 04:54:13] <gepetto> DavidS: #3354 is http://projects.reductivelabs.com/issues/show/3354 "Puppet - Bug #3354: puppet/util/settings/file_setting.rb should internalize strings when checking for AllowedOwners and AllowedGroups - ReductiveLabs.com"
[2010/03/09 04:55:31] <robinbowes> Anyone point me at puppetrun docco?
[2010/03/09 04:55:54] <DavidS> first hit on google: http://linux.die.net/man/8/puppetrun
[2010/03/09 04:56:02] <robinbowes> I have the man page
[2010/03/09 04:56:12] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 04:56:15] <robinbowes> It doesn't mention what I want to know
[2010/03/09 04:56:17] <DavidS> See the configuration file documentation at http://reductivelabs.com/projects/puppet/reference/configref.html for the full list of acceptable parameters. 
[2010/03/09 04:56:23] <DavidS> (from the manpage)
[2010/03/09 04:56:36] <robinbowes> Specifically, do I need to open a firewall port on the clients?
[2010/03/09 04:56:40] <DavidS> if that doesn't help either, you might need to specify your problem more precisely
[2010/03/09 04:56:46] <robinbowes> ^^^^
[2010/03/09 04:57:10] <DavidS> robinbowes: race :-) yes, puppetrun connects to the clients to their listenport
[2010/03/09 04:58:33] <robinbowes> (the link in the manpage gives a 404, BTW)
[2010/03/09 04:58:45] <DavidS> http://docs.reductivelabs.com/references/latest/configuration.html
[2010/03/09 05:00:57] @ Quit: ohlolohlol: Quit: ohlolohlol
[2010/03/09 05:01:06] <robinbowes> listen
[2010/03/09 05:01:06] <robinbowes> Whether puppetd should listen for connections. If this is true, then by default only the runner server is started, which allows remote authorized and authenticated nodes to connect and trigger puppetd runs.
[2010/03/09 05:01:18] <robinbowes> *by default only the runner server is started*
[2010/03/09 05:01:40] <DavidS> see also --client
[2010/03/09 05:01:57] <DavidS> (i think)
[2010/03/09 05:02:31] <robinbowes> There is no client option on the latest configuration.html page
[2010/03/09 05:03:45] <kc7zzv> You thinking of --listen on puppet?
[2010/03/09 05:03:54] <kc7zzv> I think that's what it's called.
[2010/03/09 05:04:06] @ tjoe joined channel #puppet
[2010/03/09 05:04:56] <robinbowes> Dunno - DavidS mentioned --client
[2010/03/09 05:05:26] <robinbowes> I can see the "listen" config option (same as --listen)
[2010/03/09 05:05:35] @ mattwynne joined channel #puppet
[2010/03/09 05:05:48] <kc7zzv> I see that now. Somehow I missed your comment.
[2010/03/09 05:05:54] <robinbowes> But the docs seem to say that using it disables the normal puppet client
[2010/03/09 05:06:54] @ PsychoSid joined channel #puppet
[2010/03/09 05:10:18] <robinbowes> Hmm, I also just noticed:
[2010/03/09 05:10:26] <robinbowes> all: Connect to all available hosts. Requires LDAP support at this point.
[2010/03/09 05:10:38] <robinbowes> class: Specify a class of machines to which to connect. This only works if you have LDAP configured, at the moment.
[2010/03/09 05:11:06] <robinbowes> So, looks like puppetrun is not the tool I want at the present itme
[2010/03/09 05:11:10] <robinbowes> *time
[2010/03/09 05:11:24] @ nexx_ joined channel #puppet
[2010/03/09 05:11:26] * robinbowes thinks he will wait until mpdehaan re-works it
[2010/03/09 05:14:04] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 05:14:37] @ Quit: nexx_: Read error: No route to host
[2010/03/09 05:16:10] @ Quit: FiXion_: Ping timeout: 276 seconds
[2010/03/09 05:16:31] @ tyll joined channel #puppet
[2010/03/09 05:19:05] @ Quit: pmorillo: Quit: pmorillo
[2010/03/09 05:19:52] @ nexx_ joined channel #puppet
[2010/03/09 05:19:57] <nexx_> hi
[2010/03/09 05:20:30] <nexx_> usage of "user"-paramater "password" -> can i set a pwd in plaintext per password => "foo"?
[2010/03/09 05:29:37] @ Quit: tjoe: Quit: leaving
[2010/03/09 05:33:37] @ Groenleer`Away is now known as Groenleer
[2010/03/09 05:35:04] @ FiXion joined channel #puppet
[2010/03/09 05:50:56] @ Quit: xerxas: Quit: Leaving.
[2010/03/09 05:51:27] @ andrewh joined channel #puppet
[2010/03/09 05:53:55] @ mattwynne joined channel #puppet
[2010/03/09 05:54:55] @ pmorillo joined channel #puppet
[2010/03/09 05:55:03] @ Quit: robinbowes: Read error: Connection reset by peer
[2010/03/09 05:55:44] @ GioGio joined channel #puppet
[2010/03/09 05:56:39] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 05:58:21] @ rmiller4pi81 joined channel #puppet
[2010/03/09 06:00:17] @ Quit: rmiller4pi8: Ping timeout: 265 seconds
[2010/03/09 06:07:05] @ mattwynne joined channel #puppet
[2010/03/09 06:09:20] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 06:12:17] <idimmu_> HI
[2010/03/09 06:12:21] @ idimmu_ is now known as idimmu
[2010/03/09 06:13:15] <idimmu> is it possible to loop through all the mac addreses thrown out by facter ?
[2010/03/09 06:13:30] <idimmu> link eth0 eth-00:22:19:62:d4:27 - ether - -
[2010/03/09 06:13:30] <idimmu> link eth1 eth-00:22:19:62:d4:29 - ether - -
[2010/03/09 06:13:44] <idimmu> i need to template that, and the data is all there in factor, but not presented as an array afaik .. ?
[2010/03/09 06:14:47] <nasrat> idimmu: yeah facter currently has a flat structure
[2010/03/09 06:15:12] <nasrat> but what you can do is split interfaces fact on ',' to be an array
[2010/03/09 06:15:13] @ xerxas joined channel #puppet
[2010/03/09 06:16:06] <idimmu> ooh
[2010/03/09 06:16:25] * idimmu researches
[2010/03/09 06:16:30] <idimmu> cheers, youve been inspiring :D
[2010/03/09 06:17:57] <nasrat> if you're using a template you can do something like <%= ipaddress_${iface} %> in the loop on the split fact
[2010/03/09 06:18:00] <nasrat> http://blog.hokkertjes.nl/2010/02/15/puppet-tipstricks-variable-variables/
[2010/03/09 06:19:00] <nasrat> in the distant future (facter 2.0) we want to handle richer data structures so you could iterate over interfaces, get the interface and query it for mac
[2010/03/09 06:19:05] <nasrat> but that'll be quite intrusive
[2010/03/09 06:20:01] @ Quit: Maliuta: Read error: Connection reset by peer
[2010/03/09 06:21:25] @ Quit: jh1: Remote host closed the connection
[2010/03/09 06:22:52] @ gnosek joined channel #puppet
[2010/03/09 06:24:00] @ mattwynne joined channel #puppet
[2010/03/09 06:24:38] @ Maliuta joined channel #puppet
[2010/03/09 06:28:32] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 06:33:55] <madduck> wouldn't it make sense to be able to specify an APT pin in a package type stanza??
[2010/03/09 06:34:29] <DavidS> finally got around uploading my "recent" changes to puppet-modules. I'm surprised how painless the actual upload to github was, once I got the submodules sorted out
[2010/03/09 06:35:21] <madduck> DavidS: so is puppet-modules up-to-date again?
[2010/03/09 06:35:36] <madduck> DavidS: i just looked yesterday and it seemed like the project has died a bit. was i wrong?
[2010/03/09 06:36:05] @ Quit: Spads: Ping timeout: 260 seconds
[2010/03/09 06:36:51] <DavidS> madduck: I was drowning in other projects. I finally got tuits today to upload the changes from my migration to 0.25
[2010/03/09 06:37:20] <DavidS> now I'll take a look at Al's monitor:: meta module
[2010/03/09 06:37:38] <DavidS> and perhaps fix one or two outstanding issues with the currently uploaded modules
[2010/03/09 06:39:54] @ morphic joined channel #puppet
[2010/03/09 06:40:03] @ Quit: monachus: Remote host closed the connection
[2010/03/09 06:41:14] @ monachus joined channel #puppet
[2010/03/09 06:42:12] @ Quit: monachus: Remote host closed the connection
[2010/03/09 06:43:12] @ monachus joined channel #puppet
[2010/03/09 06:45:24] <madduck> DavidS: ok. I am trying to decide whether to work off your stuff or not. it's not an easy decision. ;)
[2010/03/09 06:45:37] <madduck> my last attempt failed, but I also never really learnt puppet
[2010/03/09 06:46:03] <madduck> and I looked at your APT module and didn't like it, so I was thinking about trying myself using augeas
[2010/03/09 06:46:26] @ Quit: monachus: Remote host closed the connection
[2010/03/09 06:46:31] <madduck> and then it occurred to me that I should really be able to specify pins on versions in package stanzas
[2010/03/09 06:46:35] <madduck> DavidS: do you agree with that?
[2010/03/09 06:47:19] <DavidS> madduck: like package { "foo": version => "unstable" } ?
[2010/03/09 06:47:26] @ monachus joined channel #puppet
[2010/03/09 06:47:40] <madduck> yeah
[2010/03/09 06:47:48] <ohadlevy> robinbowes: maybe this can help you http://theforeman.org/wiki/foreman/Puppetrun
[2010/03/09 06:47:50] @ monachus left channel #puppet ()
[2010/03/09 06:47:58] <DavidS> re augeas: for managing sources.list and the pinning stuff?
[2010/03/09 06:48:02] <madduck> yes
[2010/03/09 06:48:24] @ Spads joined channel #puppet
[2010/03/09 06:48:28] <madduck> DavidS: at least debian's augeas-lenses comes with trees for both\
[2010/03/09 06:48:28] <DavidS> madduck: for the first, you'll probably need to hack on the apt package provider
[2010/03/09 06:48:43] <madduck> DavidS: yeah, and that's ruby, so quite a showstopper for me
[2010/03/09 06:48:45] <DavidS> for the second I'm open for patches.
[2010/03/09 06:48:48] <madduck> but maybe i'll just try
[2010/03/09 06:49:15] <madduck> DavidS: i know you are open to patches, you even accepted from me previously, and from micah, but this feels like a completely different way to do things
[2010/03/09 06:49:23] <madduck> i am not sure how much gain there will be in reuse
[2010/03/09 06:50:02] <madduck> i am thinking along the lines of url variables for debian, backports, volatile, security
[2010/03/09 06:50:13] <DavidS> I'm thinking about a define apt::source, which takes a url and an array of components or so and creates a proper entry for sources.list
[2010/03/09 06:50:16] <madduck> and specifying them would basically enable the sources.list entries
[2010/03/09 06:50:20] <DavidS> that would meet your needs, no?
[2010/03/09 06:50:27] <madduck> DavidS: yes, absolutely
[2010/03/09 06:50:38] <madduck> though it should have key management included
[2010/03/09 06:50:50] <madduck> is this a definition?
[2010/03/09 06:50:56] <madduck> or a plugin?
[2010/03/09 06:51:42] <DavidS> it could BE a define :-)
[2010/03/09 06:51:52] <DavidS> I'm off to lunch. are you around later?
[2010/03/09 06:51:52] <madduck> it could, especially if it uses augeas
[2010/03/09 06:51:55] <madduck> yes
[2010/03/09 06:51:59] <DavidS> k
[2010/03/09 06:52:01] <madduck> all day set aside for puppet. \o/
[2010/03/09 06:52:03] <madduck> enjoy
[2010/03/09 06:52:07] <madduck> mahlzeit
[2010/03/09 06:53:08] <nasrat> madduck: http://projects.tryphon.eu/puppet-box/browser/manifests/defines/apt.pp?rev=36ba6f404aa9a128510a9e2c5df9db1f26c83c40
[2010/03/09 06:53:27] <nasrat> not used just came across but seems to be related to what DavidS suggested
[2010/03/09 06:53:49] @ Quit: tim|macbook: Ping timeout: 260 seconds
[2010/03/09 06:54:18] @ robinbowes joined channel #puppet
[2010/03/09 06:54:40] <madduck> nasrat: yeah, totally. do you know who the author is? is s/he on IRC?
[2010/03/09 06:56:45] <madduck> nasrat: except it uses /etc/apt/preferences.d, which is not available on lenny, and i'd prefer some sort of abstraction like augeas
[2010/03/09 06:57:08] <madduck> DavidS: re your apt::source idea — augeas really gives it all to you already, so it would really just be a simple define.
[2010/03/09 06:58:40] <madduck> DavidS: I think I am just going to give it a try, a new APT module
[2010/03/09 06:58:59] @ Quit: andrewh: Ping timeout: 246 seconds
[2010/03/09 06:59:03] <madduck> if you are interested in the pinning by the package type, maybe we could have a quick look at how I would accomplish that later?
[2010/03/09 06:59:49] @ tim|macbook joined channel #puppet
[2010/03/09 07:04:33] @ Quit: tim|macbook: Ping timeout: 260 seconds
[2010/03/09 07:10:01] @ fluxdude joined channel #puppet
[2010/03/09 07:11:57] @ Babbelaar joined channel #puppet
[2010/03/09 07:12:15] @ Quit: gnosek: Quit: bye
[2010/03/09 07:13:06] @ mattwynne joined channel #puppet
[2010/03/09 07:13:58] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 07:14:13] @ Quit: m1nish: Ping timeout: 264 seconds
[2010/03/09 07:20:18] @ choffee joined channel #puppet
[2010/03/09 07:21:48] @ Quit: Spads: Ping timeout: 276 seconds
[2010/03/09 07:23:39] @ PsychoSid1 joined channel #puppet
[2010/03/09 07:23:52] @ Quit: suchu: Read error: Connection reset by peer
[2010/03/09 07:24:05] @ suchu joined channel #puppet
[2010/03/09 07:24:15] @ Quit: PsychoSid: Ping timeout: 245 seconds
[2010/03/09 07:26:00] @ joe-mac left channel #puppet ()
[2010/03/09 07:27:56] @ francois1 joined channel #puppet
[2010/03/09 07:29:13] <madduck> Puppet::Provider::Package::APT defines the 'latest' method, but I cannot figure out where it's called at all.
[2010/03/09 07:30:16] <SyTonnerre> madduck, ensure => latest
[2010/03/09 07:30:32] <madduck> oh, and that actually causes the function to be called, eh?
[2010/03/09 07:32:41] <nasrat> madduck: dunno here is the project page http://projects.tryphon.eu/puppet-box
[2010/03/09 07:32:41] <matti> How hard is to replace Mongrel or Webrick in puppet with something like ngix or Apache?
[2010/03/09 07:33:12] <madduck> matti: you don't replace mongrel with apache, you proxy to mongrel with apache/nginx
[2010/03/09 07:33:15] <madduck> and it's pretty trivial
[2010/03/09 07:33:21] <nasrat> madduck: copyright says Alban Peignier, Florent Peyraud
[2010/03/09 07:34:05] <matti> madduck: Would that be able to handle 500+ server provisions, what do you reckon?
[2010/03/09 07:34:41] <madduck> matti: i don't know. in my experience, puppet doesn't scale well at all, but the queue manager might improve that
[2010/03/09 07:35:03] <matti> madduck: What do you mean?
[2010/03/09 07:35:20] <matti> madduck: I am not that familiar with puppet yet (I am just humble software guy).
[2010/03/09 07:35:48] <matti> madduck: We have custom system called "muppet" which my team wrote long time ago, and now looking to replace it with puppet in production :)
[2010/03/09 07:36:57] <madduck> matti: ask others not me, since I am also not that familiar
[2010/03/09 07:37:46] <matti> OK
[2010/03/09 07:37:51] @ tim|macbook joined channel #puppet
[2010/03/09 07:42:10] @ Quit: morphic: Ping timeout: 248 seconds
[2010/03/09 07:43:16] @ ahasenack joined channel #puppet
[2010/03/09 07:45:33] @ a55 joined channel #puppet
[2010/03/09 07:46:04] @ tim|macbook_ joined channel #puppet
[2010/03/09 07:47:52] <a55> hi... is there any "nice and tight" bacula modules ?
[2010/03/09 07:48:06] @ Quit: tim|macbook: Ping timeout: 245 seconds
[2010/03/09 07:48:06] @ tim|macbook_ is now known as tim|macbook
[2010/03/09 07:49:48] @ a55 left channel #puppet ()
[2010/03/09 07:50:57] @ a55 joined channel #puppet
[2010/03/09 07:51:33] <a55> just /part unexpectedly! what a shame ;(
[2010/03/09 07:52:30] @ inveratulo left channel #puppet ()
[2010/03/09 07:59:55] <madduck> is there still not a better way to do define-unless-defined à la http://slexy.org/view/s20PMdjarh?
[2010/03/09 08:05:30] @ Quit: Maliuta: Ping timeout: 245 seconds
[2010/03/09 08:05:49] @ rmiller4pi8 joined channel #puppet
[2010/03/09 08:06:00] @ Quit: kc7zzv: Quit: kc7zzv
[2010/03/09 08:06:21] @ Quit: tim|macbook: Ping timeout: 256 seconds
[2010/03/09 08:07:16] @ Quit: rmiller4pi81: Ping timeout: 245 seconds
[2010/03/09 08:08:25] @ tim|macbook joined channel #puppet
[2010/03/09 08:10:12] <chrisg> have any of you guys written a function to get user passwords from local shadow resource, to then use in the password field when defining a user?
[2010/03/09 08:10:38] <chrisg> i'd like to basically manage passwords for my users defined in puppet, as local unix users on the system puppetmaster is running on, and then propagate it out to the puppet clients
[2010/03/09 08:11:02] <chrisg> i wrote a little ruby function to grab it from shadow using rubys shadow lib, but i think it's got a bit confused and trying to maybe run on the client and looping
[2010/03/09 08:11:13] <chrisg> 'cause whenever i use it the client just sticks on tryinf to run through the catalogue :)
[2010/03/09 08:11:27] <Volcane> puppetmaster doesnt run as root
[2010/03/09 08:11:28] <madduck> http://reductivelabs.com/trac/puppet/wiki/Recipes/PasswordManagement
[2010/03/09 08:11:32] <Volcane> it wouldnt be able to read shadow file
[2010/03/09 08:11:34] <ashp> chrisg: i just stuck the hash into the .pp!
[2010/03/09 08:11:49] <ashp> chrisg: i actually made a /etc/puppet/secrets/, put them in there, then include them in the .pp so they are hidden away
[2010/03/09 08:12:28] <chrisg> ah do you automagically parse from the shadow source into /etc/puppet/secrets then?
[2010/03/09 08:12:56] <chrisg> just say a crontab'd run of a loop doing @getent shadow $user@
[2010/03/09 08:13:30] @ Spads joined channel #puppet
[2010/03/09 08:13:36] <ashp> I don't, but it would be really easy to do
[2010/03/09 08:13:54] <ashp> That's how I'd do it if I had to, just do a quick loop and shovel it into files, then include them later
[2010/03/09 08:14:22] <ashp> it's kind of dumb and ugly, but I just have a list at the top of the .pp
[2010/03/09 08:14:25] <ashp> $apenney = file("/etc/puppet/secrets/users/apenney")
[2010/03/09 08:14:34] <ashp> again, that would be easy to autogenerate
[2010/03/09 08:21:28] @ Quit: kubicek: Quit: kubicek
[2010/03/09 08:24:04] @ mpdehaan joined channel #puppet
[2010/03/09 08:24:19] @ Quit: mpdehaan: Client Quit
[2010/03/09 08:25:05] @ spawnyd joined channel #puppet
[2010/03/09 08:26:09] @ kubicek joined channel #puppet
[2010/03/09 08:28:06] @ cwebber joined channel #puppet
[2010/03/09 08:29:49] @ Quit: cwebber: Client Quit
[2010/03/09 08:32:11] @ ckauhaus joined channel #puppet
[2010/03/09 08:33:52] @ Quit: kubicek: Quit: kubicek
[2010/03/09 08:34:44] @ doubleukay joined channel #puppet
[2010/03/09 08:35:29] <doubleukay> hi guys. let's say I wanna declare an array of resources, how do I use the title of the resource as a variable?
[2010/03/09 08:35:30] <nico> hmm, this monitor module looks like a nice idea
[2010/03/09 08:35:52] <doubleukay> e.g. file { [ "/foo", "/bar" ]: contents => $name }
[2010/03/09 08:38:47] <tim|macbook> doubleukay: exactly like that
[2010/03/09 08:39:05] <Volcane> tim|macbook: nope
[2010/03/09 08:39:14] @ Quit: erlingre: Ping timeout: 248 seconds
[2010/03/09 08:39:15] <tim|macbook> Volcane: are you sure?
[2010/03/09 08:39:20] <doubleukay> yes
[2010/03/09 08:39:25] <tim|macbook> or wait... that only works for defined types...
[2010/03/09 08:39:31] @ cliff-hm joined channel #puppet
[2010/03/09 08:39:39] <doubleukay> I tried that, but got the name of the module
[2010/03/09 08:40:04] @ Quit: yure: Quit: Odhajam
[2010/03/09 08:40:11] <tim|macbook> doubleukay: then I don't know... but you can create a workaround for it by making a wrapper around file
[2010/03/09 08:40:43] <tim|macbook> defined file_with_content_of_the_name { file { $name: content => $name } }
[2010/03/09 08:41:04] <tim|macbook> file_with_content_of_the_name { ["/foo","/bar"]:; }
[2010/03/09 08:41:10] <doubleukay> thanks, I think something like that will suit what I'm looking for
[2010/03/09 08:44:42] @ uphillian joined channel #puppet
[2010/03/09 08:46:42] @ Quit: alban2: Quit: Leaving.
[2010/03/09 08:50:02] <DavidS> chrisg: take a look at ralsh, it might be able to do what ya need
[2010/03/09 08:50:25] <DavidS> madduck: pong
[2010/03/09 08:52:19] <DavidS> doubleukay: define loop() { file { $name: content => $name } loop { ["/foo", "/bar"]: }
[2010/03/09 08:52:21] @ Quit: danny: Ping timeout: 260 seconds
[2010/03/09 08:52:51] <DavidS> i.e. what tim|macbook said O:-)
[2010/03/09 08:53:07] <doubleukay> noted
[2010/03/09 08:53:07] <DavidS> not to self: read first, post later
[2010/03/09 08:53:13] <DavidS> +e
[2010/03/09 08:53:53] <madduck> DavidS: i have started work on APT and I am getting closer, let me quickly fix this one issue and then push it to Git
[2010/03/09 08:54:06] <doubleukay> perhaps there will one day be syntactic sugar to handle this
[2010/03/09 08:54:48] <madduck> DavidS: i also found out that I have to go see a client soon, so little hacking today. :(
[2010/03/09 08:54:53] @ Quit: Av: Remote host closed the connection
[2010/03/09 08:54:54] <madduck> DavidS: but tomorrow…?
[2010/03/09 08:56:45] @ kubicek joined channel #puppet
[2010/03/09 08:57:19] <tim|macbook> madduck: an apt module?
[2010/03/09 08:57:22] * tim|macbook would love to see that
[2010/03/09 08:57:30] <tim|macbook> I'm in the middle of a rewrite of our entire codebase myself
[2010/03/09 08:58:04] <madduck> sec, about to push
[2010/03/09 08:58:06] <madduck> tim|macbook: ^
[2010/03/09 08:58:16] <tim|macbook> ok :D
[2010/03/09 08:58:40] <madduck> count down from 50 ;)
[2010/03/09 08:58:57] <tim|macbook> 0!
[2010/03/09 08:59:01] <madduck> pffffffff
[2010/03/09 08:59:06] * tim|macbook is a fast counter
[2010/03/09 08:59:10] <madduck> 50 was given to you in base 256
[2010/03/09 08:59:13] <kjetilho> uh, please refresh my memory: can you do multiple values on the LHS of a selector?
[2010/03/09 08:59:30] <kjetilho> e.g., $foo ? { "val1", "val2" => "blah" }
[2010/03/09 08:59:39] <tim|macbook> kjetilho: yes
[2010/03/09 08:59:42] <kjetilho> (the above is obviously not valid syntax)
[2010/03/09 08:59:50] <tim|macbook> i think so
[2010/03/09 08:59:53] <tim|macbook> using an array?
[2010/03/09 09:00:04] <kjetilho> I get an error for that, too
[2010/03/09 09:00:27] <kjetilho> Syntax error at '['; expected ']'
[2010/03/09 09:00:30] <kjetilho> (!)
[2010/03/09 09:00:46] <madduck> tim|macbook, DavidS: http://git.madduck.net/v/puppet/modules/apt.git/blob/HEAD:/manifests/init.pp
[2010/03/09 09:01:14] <tim|macbook> kjetilho: $owner = $operatingsystem ? { /(redhat|debian)/ => "bin", < from the language tutorial
[2010/03/09 09:01:27] <madduck> oh, funky. i dropped the main mirror ;)
[2010/03/09 09:01:41] <kjetilho> hrm. yeah, I could use a regexp
[2010/03/09 09:01:55] <kjetilho> (need to use /i, though)
[2010/03/09 09:02:22] <kjetilho> or at least you'll need it if lak accepts masterzen's patch for case-sensitivity and regexps
[2010/03/09 09:02:49] <tim|macbook> madduck: hm... our module is like... totally different...
[2010/03/09 09:02:54] <tim|macbook> wanna see?
[2010/03/09 09:02:56] <madduck> of course
[2010/03/09 09:03:22] <masterzen> kjetilho: you can workaround: /(?i)blahblah/
[2010/03/09 09:03:36] <madduck> man puppet isn't easy. If I include a .pp file with a definition from inside a class, the definition does not seem to be available as class::definition
[2010/03/09 09:03:48] <kjetilho> masterzen: oh, isn't /i supported?
[2010/03/09 09:03:53] <masterzen> kjetilho: no
[2010/03/09 09:04:26] <masterzen> madduck: it should work. I'm doing this all the time
[2010/03/09 09:04:43] <kjetilho> madduck: you mean import?
[2010/03/09 09:04:52] <madduck> masterzen: well, see http://git.madduck.net/v/puppet/modules/apt.git/blob/HEAD:/manifests/init.pp and No file(s) found for import of 'definitions/repository.pp' at /etc/puppet/modules/apt/manifests/init.pp:15 on node swan.madduck.net
[2010/03/09 09:04:55] <kjetilho> include doesn't affect class hierarchy
[2010/03/09 09:05:11] <tim|macbook> madduck: http://pastie.org/861330
[2010/03/09 09:05:19] <madduck> kjetilho: that was your error above ^, not for masterzen
[2010/03/09 09:05:23] <madduck> masterzen: "Could not find resource type apt::repository at /etc/puppet/manifests/site.pp:51 on node swan.madduck.net"
[2010/03/09 09:05:47] @ mikepea joined channel #puppet
[2010/03/09 09:06:05] <madduck> tim|macbook: i suggest you include a fingerprint check on the key!
[2010/03/09 09:06:08] <masterzen> madduck: you're confusing include and import.
[2010/03/09 09:06:49] <masterzen> madduck: include is to be followed by a classname, import by a manifest file path
[2010/03/09 09:07:18] <madduck> masterzen: okay, and import takes a relative filename?
[2010/03/09 09:07:37] <masterzen> it does, and even globs
[2010/03/09 09:07:43] <madduck> i know about globs
[2010/03/09 09:07:48] <madduck> anyway, with
[2010/03/09 09:07:48] <madduck> import "variables.pp"
[2010/03/09 09:07:48] <madduck> import "definitions/repository.pp"
[2010/03/09 09:08:00] <tim|macbook> madduck: hm indeed... hadn't thought about that yet... and the cachesize is a bit... arbitrary, of course... needs to be a bit more fine grained and maybe controlled from a higher level... but okay
[2010/03/09 09:08:13] <madduck> i still cannot use the definition apt::repository (which I import into class apt) from outside the class
[2010/03/09 09:08:17] @ jmccune joined channel #puppet
[2010/03/09 09:08:31] <madduck> tim|macbook: some interesting ideas there.
[2010/03/09 09:08:39] <madduck> we aren't *too* different, you know.
[2010/03/09 09:08:56] <masterzen> madduck: put the import outside the class. You're actually importing apt::apt:repository :-)
[2010/03/09 09:08:57] <madduck> but i'd prefer to add pinning to the Package type, rather than to create my own stuff.
[2010/03/09 09:09:17] <madduck> masterzen: why? the definition is not inside a class.
[2010/03/09 09:09:23] * Volcane wonders why sane people still use import
[2010/03/09 09:09:39] * masterzen also
[2010/03/09 09:09:40] <madduck> Volcane: how else can I split the manifest into smaller chunks?
[2010/03/09 09:09:44] <Volcane> modules
[2010/03/09 09:09:50] <masterzen> auto-load
[2010/03/09 09:10:01] <madduck> oh, of course, but I cannot have modules inside modules, can I?
[2010/03/09 09:10:14] <Volcane> foo::bar::baz::blah::meh
[2010/03/09 09:10:16] <Volcane> all valid
[2010/03/09 09:10:29] <Volcane> all automagically auto loading
[2010/03/09 09:10:30] @ alban2 joined channel #puppet
[2010/03/09 09:10:32] <Volcane> no imports needed
[2010/03/09 09:10:40] <madduck> and this is a definition I require for my module, so I kinda prefer to provide it alongside, rather than to declare a dependency
[2010/03/09 09:10:54] <Volcane> foo::bar bar being a define stores in foo/manifests/bar.pp
[2010/03/09 09:10:55] <madduck> or do i simply create ./modules/foo inside my module?
[2010/03/09 09:10:57] <tim|macbook> madduck: is repository selection available for the package{} resource?
[2010/03/09 09:10:57] <Volcane> all automagically auto loading
[2010/03/09 09:11:14] <madduck> tim|macbook: no, that would be counter-debian, wouldn't it?
[2010/03/09 09:11:18] <madduck> Volcane: okay, will try
[2010/03/09 09:11:38] <tim|macbook> madduck: then I don't get your comment about adding it to package instead of preferences
[2010/03/09 09:11:54] @ fzzzt joined channel #puppet
[2010/03/09 09:12:33] <madduck> tim|macbook: package { [ puppet, puppet-common ]: ensure => "unstable" }
[2010/03/09 09:12:35] <madduck> or the like
[2010/03/09 09:12:42] * masterzen <3 autoloading
[2010/03/09 09:12:53] <madduck> it could use apt-get -t unstable install
[2010/03/09 09:12:59] <madduck> or it could maintain a file in /etc/apt/preferences.d
[2010/03/09 09:13:29] <tim|macbook> ah you're talking about actually doing some development in puppet :) not something that's already available?
[2010/03/09 09:13:47] <tim|macbook> and yeah, /etc/apt/preferences.d is prefered, but not yet available in lenny
[2010/03/09 09:13:56] <madduck> exactly
[2010/03/09 09:14:16] @ Quit: a55: Ping timeout: 276 seconds
[2010/03/09 09:14:21] <madduck> so we also need some sort of exec to run-parts+cat all files therein into /etc/apt/preferences on <=lenny
[2010/03/09 09:14:30] <henk> /etc/sudoers.d/ neither :(
[2010/03/09 09:14:46] <madduck> henk: is there a bug report? does sudoers.d exist?
[2010/03/09 09:14:56] <henk> madduck: it will after lenny afaik
[2010/03/09 09:15:02] <tim|macbook> I already have provisions in place for it, though :) as soon as I get an apt version installed that supports /etc/apt/preferences.d, it'll add files to that directory instead of concatenated_file
[2010/03/09 09:15:19] <madduck> so I have modules/apt/manifests.repository.pp with a definition of 'repository', but: "Could not find resource type apt::repository at /etc/puppet/manifests/site.pp:51 on node swan.madduck.net" from the outside. :(
[2010/03/09 09:15:30] <joe-mac1> i love .d directories
[2010/03/09 09:15:38] <henk> madduck: oh, no, sorry. "includedir" will be added to sudoers... but that's almost the same ;)
[2010/03/09 09:15:40] <joe-mac1> they make mass management so easy. and they make me feel funny inside.
[2010/03/09 09:15:49] <madduck> henk: right
[2010/03/09 09:16:08] @ Quit: afletcher: Quit: afletcher
[2010/03/09 09:17:00] @ a55 joined channel #puppet
[2010/03/09 09:17:17] <fzzzt> Is there a way to have two module affect the same file? I have two Tomcat apps that both need to write entries into tomcat-users.xml but can't think of how to do it
[2010/03/09 09:17:29] <barn> virtual aliases?
[2010/03/09 09:17:52] <barn> virtual resources, sorry
[2010/03/09 09:18:03] <Volcane> fzzzt: http://www.devco.net/archives/2010/02/19/building_files_from_fragments_with_puppet.php
[2010/03/09 09:18:32] @ unxfrek joined channel #puppet
[2010/03/09 09:19:41] <tim|macbook> fzzzt: or using a line{} resource
[2010/03/09 09:19:41] <madduck> any idea about the auto-import?
[2010/03/09 09:20:04] <madduck> tim|macbook: i don't think there is a line{} resource
[2010/03/09 09:20:17] <tim|macbook> madduck: no, but there are a lot of examples for it :)
[2010/03/09 09:20:23] <madduck> yeah yeah
[2010/03/09 09:20:29] @ pheezy joined channel #puppet
[2010/03/09 09:21:14] @ rgsteele joined channel #puppet
[2010/03/09 09:21:30] <Volcane> line{} is evil and broken by design
[2010/03/09 09:21:31] <tim|macbook> madduck: i think auto import only works with classes?
[2010/03/09 09:21:45] <tim|macbook> Volcane: I like line{} and use it a lot :)
[2010/03/09 09:21:45] <Volcane> tim|macbook: defines too
[2010/03/09 09:21:51] <tim|macbook> ah hm
[2010/03/09 09:22:36] <madduck> what does configfile do, tim|macbook?
[2010/03/09 09:23:20] <fzzzt> barn: Close but I don't think that will work. Even if I set up a define for tomcatuser, I need to have it be realized only once (IIUC) and have all the tomcatusers merged into one file.
[2010/03/09 09:23:57] <tim|macbook> madduck: ah that's just our wrapper around file, adds some conveniences for us (like keeping track in a local file which files are modified/sourced/etc. and the like)
[2010/03/09 09:24:26] <madduck> Volcane: tim|macbook oh, interesting.
[2010/03/09 09:24:43] <tim|macbook> and allows me to do rmtsrc => "modules/bla/file" instead of source => "puppet://puppet/modules/bla/file"
[2010/03/09 09:24:52] <madduck> Volcane: http://git.madduck.net/v/puppet/modules/apt.git/blob/HEAD:/manifests/repository.pp doesn't work. :(
[2010/03/09 09:25:23] <tim|macbook> madduck: maybe you need to name the defined type apt::repository? (just guessing here)
[2010/03/09 09:25:36] <Volcane> madduck: perhaps read wiki:ModuleOrganisation again :)
[2010/03/09 09:25:36] <gepetto> Volcane: madduck: wiki:ModuleOrganisation is http://reductivelabs.com/trac/puppet/wiki/ModuleOrganisation
[2010/03/09 09:26:54] @ Robbie__ joined channel #puppet
[2010/03/09 09:27:45] @ danny joined channel #puppet
[2010/03/09 09:28:19] <fzzzt> Volcane: clever (concat::fragment)
[2010/03/09 09:28:27] <madduck> tim|macbook: that, and import "*", which I found thanks to Volcane's gentle RFTM
[2010/03/09 09:28:28] <madduck> ;)
[2010/03/09 09:28:38] <Volcane> (you dont need the import)
[2010/03/09 09:28:51] <Volcane> isnt this the entire point we're trying to avoid by telling you to use modules properly? :)
[2010/03/09 09:29:04] <Volcane> import was needed like 0.24.early only
[2010/03/09 09:30:16] @ Quit: ckauhaus: Quit: Leaving.
[2010/03/09 09:30:30] @ ckauhaus joined channel #puppet
[2010/03/09 09:30:32] @ Quit: Robbie_: Ping timeout: 246 seconds
[2010/03/09 09:30:53] <madduck> oh, i got confused
[2010/03/09 09:30:58] @ cwebber joined channel #puppet
[2010/03/09 09:31:14] <madduck> i still need import if I want to factor out the variable definitions, but that's not working. :(
[2010/03/09 09:31:22] <madduck> i guess i just won't factor them out.
[2010/03/09 09:31:42] <madduck> i wish there was a better way to do define-unless-defined than http://slexy.org/view/s20PMdjarh
[2010/03/09 09:33:05] <madduck> and i wish puppet would let me deal with variables a bit more sensibly in the first place. ;)
[2010/03/09 09:33:22] <Volcane> madduck: did you look at extlookup/
[2010/03/09 09:33:40] <madduck> you mean what we talked about months ago? not yet.
[2010/03/09 09:33:50] <madduck> but it sounds like a good idea
[2010/03/09 09:33:53] <madduck> in this context
[2010/03/09 09:34:34] @ yui joined channel #puppet
[2010/03/09 09:37:32] @ Robbie_ joined channel #puppet
[2010/03/09 09:37:35] @ Quit: Robbie__: Read error: No route to host
[2010/03/09 09:38:40] @ monachus joined channel #puppet
[2010/03/09 09:41:11] @ notbrien joined channel #puppet
[2010/03/09 09:41:14] @ Quit: yui: Ping timeout: 252 seconds
[2010/03/09 09:44:44] @ afletcher joined channel #puppet
[2010/03/09 09:45:22] @ Quit: Robbie_: Ping timeout: 248 seconds
[2010/03/09 09:46:16] <joe-mac1> guys, there are syntax errors slipping by puppet --parseonly
[2010/03/09 09:46:31] <joe-mac1> such as commas after resource parameters
[2010/03/09 09:46:47] <joe-mac1> so, my precommit is letting me commit stupid errors that i don't find until the node complains
[2010/03/09 09:47:07] @ Quit: rmiller4pi8: Quit: Leaving.
[2010/03/09 09:47:14] <tim|macbook> madduck: I have a feature request for the define-unless-defined thingy running
[2010/03/09 09:47:42] <fzzzt> Unfortunately concat won't work either (at least not without modifications)
[2010/03/09 09:47:43] <fzzzt> hmm
[2010/03/09 09:47:48] <Volcane> why?
[2010/03/09 09:47:50] <tim|macbook> madduck: #3181
[2010/03/09 09:47:50] <gepetto> tim|macbook: madduck: #3181 is http://projects.reductivelabs.com/issues/show/3181 "Puppet - Feature #3181: Create resource if it doesn't already exist (with array support) - ReductiveLabs.com"
[2010/03/09 09:48:07] <fzzzt> I have to assemble a line and put it in, along with other lines...
[2010/03/09 09:48:24] <Volcane> fzzzt: so the line is made up of various bits?
[2010/03/09 09:48:49] <joe-mac1> well that makes no sense
[2010/03/09 09:48:52] <joe-mac1> not you, me
[2010/03/09 09:48:55] @ Robbie_ joined channel #puppet
[2010/03/09 09:49:29] <joe-mac1> nevermind, my precommit is somehow not always checking things
[2010/03/09 09:49:35] @ lordb joined channel #puppet
[2010/03/09 09:49:38] <joe-mac1> i hate troubleshooting svn hooks too, omfg
[2010/03/09 09:49:41] @ mpdehaan_ joined channel #puppet
[2010/03/09 09:49:55] <fzzzt> Yeah... Basically, I have 2 tomcat apps that need shared auth. Each app has it's own roles <role .../> in the file, and the users have multiple roles <user roles="..."/> in the file. I need to put the roles into the file from each app, and alter the user lines to include both roles.
[2010/03/09 09:49:56] @ Quit: Robbie_: Read error: Connection reset by peer
[2010/03/09 09:50:02] @ kaptk2 joined channel #puppet
[2010/03/09 09:50:03] @ Robbie_ joined channel #puppet
[2010/03/09 09:50:04] <fzzzt> I'm looking to see if I can use multiple files in Tomcat right now
[2010/03/09 09:50:11] @ Quit: nexx_: Quit: leaving
[2010/03/09 09:50:29] <Volcane> fzzzt: you can probably also safely split those over lines - xml is very forgiving wrt that
[2010/03/09 09:51:01] <fzzzt> XML wise yeah, but I don't know if Tomcat would work with multiple user lines with the same name and different role attributes
[2010/03/09 09:51:29] @ swygue joined channel #puppet
[2010/03/09 09:51:39] @ Robbie__ joined channel #puppet
[2010/03/09 09:52:36] @ Quit: Robbie__: Read error: Connection reset by peer
[2010/03/09 09:52:44] @ Quit: Robbie_: Read error: Connection reset by peer
[2010/03/09 09:52:48] @ Robbie_ joined channel #puppet
[2010/03/09 09:53:57] @ Quit: Robbie_: Read error: Connection reset by peer
[2010/03/09 09:53:58] @ Robbie_ joined channel #puppet
[2010/03/09 09:59:55] @ yure joined channel #puppet
[2010/03/09 10:01:56] <DavidS> fzzzt: you can try chaining concats. but this will become very heinous very fast. Better implement a proper type and ParsedFile provider
[2010/03/09 10:02:25] <DavidS> (i.e. use the first concat to build the roles="" string and the second to actually create the complete xml)
[2010/03/09 10:02:40] <DavidS> "brittle" being the understatement of the year
[2010/03/09 10:04:14] <tim|macbook> or a augeas lens
[2010/03/09 10:05:05] <DavidS> tim|macbook: indeed
[2010/03/09 10:05:28] @ Quit: fsweetser: Read error: Connection reset by peer
[2010/03/09 10:08:35] <fzzzt> I have an iptables type that takes all iptables{} resources and (somehow) assembles them all in order at one point... I wonder if I could adapt that for a generic file.
[2010/03/09 10:09:09] @ rmiller4pi8 joined channel #puppet
[2010/03/09 10:09:13] <fzzzt> For now, I think I can work around it using Tomcat config
[2010/03/09 10:09:46] <DavidS> git(hub) question: I'd like to manage this network graph: http://github.com/puppet-modules/puppet-common/network
[2010/03/09 10:10:17] @ Quit: rmiller4pi8: Client Quit
[2010/03/09 10:10:18] <DavidS> simpsonjulian has non-merged changes. I could just merge -s ourse from him and the branch would vanish in this graph
[2010/03/09 10:10:31] @ Quit: tim|macbook: Quit: tim|macbook
[2010/03/09 10:10:46] <DavidS> but I presume the next time julian would pull from the master, his local changes would vanish.
[2010/03/09 10:11:15] <DavidS> is there another possibility to "clean" this graph without screwing users over?
[2010/03/09 10:17:16] @ Quit: biertie: Read error: Connection reset by peer
[2010/03/09 10:18:05] @ labrown joined channel #puppet
[2010/03/09 10:18:23] @ Quit: giskard: Quit: Leaving...
[2010/03/09 10:19:04] @ rcrowley joined channel #puppet
[2010/03/09 10:19:34] @ malraid joined channel #puppet
[2010/03/09 10:21:57] @ mattwynne joined channel #puppet
[2010/03/09 10:22:21] @ Quit: Robbie_: Remote host closed the connection
[2010/03/09 10:22:24] @ jab_doa joined channel #puppet
[2010/03/09 10:22:29] @ shenson joined channel #puppet
[2010/03/09 10:22:42] @ Robbie_ joined channel #puppet
[2010/03/09 10:23:22] @ fsweetser joined channel #puppet
[2010/03/09 10:23:26] @ Quit: mattwynne: Remote host closed the connection
[2010/03/09 10:25:24] @ rmiller4pi8 joined channel #puppet
[2010/03/09 10:27:56] @ Quit: verwilst: Ping timeout: 246 seconds
[2010/03/09 10:30:13] <DavidS> Volcane: may I publish your concat stuff in my repos?
[2010/03/09 10:31:36] @ Av joined channel #puppet
[2010/03/09 10:32:08] @ Quit: Dyresen: Quit: leaving
[2010/03/09 10:32:18] <DavidS> (and, is "2010/02/19" the most recent version?)
[2010/03/09 10:32:47] <Volcane> DavidS: yeah - have a new version about to come out (next week when i have time again)
[2010/03/09 10:33:29] <Volcane> changes will be:
[2010/03/09 10:33:40] <Volcane> - fix bug in printing the warning header
[2010/03/09 10:33:51] <Volcane> - fix some true != "true" confusion
[2010/03/09 10:34:10] <Volcane> - fix a 0.24.x compatilbity bug with path to empty dir
[2010/03/09 10:34:47] <Volcane> paul elliott reported those to me and sent patches
[2010/03/09 10:34:47] @ biertie joined channel #puppet
[2010/03/09 10:34:58] <DavidS> is any of those a showstopper? the first sounds bad...
[2010/03/09 10:35:12] <Volcane> nope - the first only apply if you use that feature
[2010/03/09 10:35:18] <DavidS> ok
[2010/03/09 10:35:18] <Volcane> and all that will happen is the header wont print :P
[2010/03/09 10:35:22] <DavidS> hehe
[2010/03/09 10:35:59] @ giskard joined channel #puppet
[2010/03/09 10:36:35] <DavidS> I'll upload as is to http://github.com/puppet-modules and will start using it in my other modules. If you don't come around to integrating paul's patches, forward them to david@dasz.at and I'll take apply them there
[2010/03/09 10:36:48] <Volcane> I'll def integrate them
[2010/03/09 10:37:55] <Volcane> those one liners of yours gave me nightmares, so opted on the shell script helper approach :P
[2010/03/09 10:37:59] @ Quit: kubicek: Quit: kubicek
[2010/03/09 10:38:32] <DavidS> O:-) yeah, much nicer
[2010/03/09 10:42:17] @ Quit: gebi: Ping timeout: 246 seconds
[2010/03/09 10:45:39] @ mpdehaan__ joined channel #puppet
[2010/03/09 10:48:30] @ Quit: JJA: Ping timeout: 276 seconds
[2010/03/09 10:50:22] @ Quit: JD_: Disconnected by services
[2010/03/09 10:50:25] <DavidS> mpdehaan__: hi!
[2010/03/09 10:50:30] <mpdehaan__> hi
[2010/03/09 10:50:42] @ kubicek joined channel #puppet
[2010/03/09 10:52:34] <madduck> DavidS: no client. I am back
[2010/03/09 10:52:48] @ bodepd joined channel #puppet
[2010/03/09 10:52:53] <DavidS> hi dan!
[2010/03/09 10:53:28] <DavidS> madduck: I really like what you do in your apt module
[2010/03/09 10:53:50] <madduck> DavidS: awesome; this means a lot to me.\
[2010/03/09 10:54:04] <madduck> i am now going to add key management
[2010/03/09 10:54:09] <madduck> and then look at pinning in the package type again
[2010/03/09 10:55:07] <DavidS> I'll probably rework my module to be - at least - provide apt::repository
[2010/03/09 10:56:05] <pheezy> is it really the official puppet-modules repository?
[2010/03/09 10:56:07] <madduck> DavidS: maybe once mine is done, we can find a way to combine them in a backwards-compatible way? i know micah would like to use it too
[2010/03/09 10:56:40] <DavidS> I'll have to take a look what then remains of my module :-)
[2010/03/09 10:57:06] <DavidS> perhaps I'll just switch wholesale to yours. one thing less I've got to worry about :-P
[2010/03/09 10:57:33] <madduck> well, thus far, it's not working yet. I cannot figure out a good way to let hosts set the configurables individually
[2010/03/09 10:57:40] <madduck> i need to read the docs, I think
[2010/03/09 10:57:49] <madduck> or maybe even start with lak's book, which I have
[2010/03/09 10:58:15] <DavidS> that should just be "node myhost { $debian_... = ... ; include apt; }
[2010/03/09 10:58:18] <madduck> wait, that's jamesturnbull, not lak ;)
[2010/03/09 10:58:35] * DavidS hates working with different github accounts
[2010/03/09 10:59:02] <madduck> DavidS: … hates … github ;)
[2010/03/09 10:59:43] @ dyresen joined channel #puppet
[2010/03/09 10:59:50] <fzzzt> I wish there was a way to have puppetrun not return until the run is finished
[2010/03/09 10:59:53] <DavidS> least resistance, and there are rumors that reductivelabs will provide a module-oriented github replacement soon. I'm totally looking forward to it
[2010/03/09 11:00:01] <madduck> no, the easy way doesn't seem to work, unfortunately.
[2010/03/09 11:00:09] <DavidS> fzzzt: is there no foreground or --no-daemonize?
[2010/03/09 11:00:10] <madduck> the easy way with variables i mean
[2010/03/09 11:00:26] <fzzzt> non-interactive
[2010/03/09 11:00:45] <DavidS> madduck: I think the problem is that you do not rename the variables when setting defaults
[2010/03/09 11:01:01] <DavidS> that might make troubles too
[2010/03/09 11:01:06] @ Quit: xerxas: Ping timeout: 248 seconds
[2010/03/09 11:01:17] <madduck> well, the real problem is "Cannot assign to variables in other namespaces at /etc/puppet/manifests/site.pp:91"
[2010/03/09 11:01:29] <madduck> because otherwise i could just define the variables in apt{} and then override them
[2010/03/09 11:02:16] <DavidS> with the contents of http://git.madduck.net/v/puppet/modules/apt.git/blob/HEAD:/manifests/init.pp , "node host { $debian_suite = "lenny"\n include apt\n }" should work
[2010/03/09 11:03:19] @ Quit: vermeer: Remote host closed the connection
[2010/03/09 11:03:35] @ Quit: PsychoSid1: Quit: Leaving.
[2010/03/09 11:03:39] <madduck> indeed, debian_suite works, but debian_mirror_base does not. :(
[2010/03/09 11:04:01] <DavidS> madduck: mirrorbase vs. mirror_base!
[2010/03/09 11:04:11] <madduck> YIKES
[2010/03/09 11:04:41] @ vermeer joined channel #puppet
[2010/03/09 11:05:06] @ francois1 left channel #puppet ()
[2010/03/09 11:05:53] @ xerxas joined channel #puppet
[2010/03/09 11:06:07] <madduck> okay, that's fixed. now i need to make $debian_volatile_mirror and $debian_backports_mirror work
[2010/03/09 11:06:37] @ verwilst joined channel #puppet
[2010/03/09 11:06:55] <DavidS> error?
[2010/03/09 11:07:02] <madduck> jsut empty when run
[2010/03/09 11:07:20] <DavidS> always use ${...} when referencing vars ins strings
[2010/03/09 11:08:14] <DavidS> I have no idea whether mirror     => "$debian_volatile_mirror-sloppy" references $debian_volatile_mirror or $debian_volatile_mirror-sloppy
[2010/03/09 11:09:42] @ Quit: vermeer: Read error: Connection reset by peer
[2010/03/09 11:13:37] * Volcane 's having doubts about file serving in 0.25.4
[2010/03/09 11:13:52] <Volcane> anyone here who have a setup like:
[2010/03/09 11:14:20] <Volcane> multiple environments, environment setup like: modulepath = /etc/puppet/manifests/development/modules:/etc/puppet/manifests/common/modules
[2010/03/09 11:14:30] <jb> i do..
[2010/03/09 11:14:50] <jb> but, im still 0.25.3
[2010/03/09 11:14:53] <Volcane> so if i have a mdule foo in both development/modules/foo/files/test.txt
[2010/03/09 11:15:02] <Volcane> and common/modules/foo/files/test.txt
[2010/03/09 11:15:11] <Volcane> with some manifests to go in each one
[2010/03/09 11:15:20] <Volcane> I am finding its sourcing the classes from the development one
[2010/03/09 11:15:24] <Volcane> but files from the common one
[2010/03/09 11:15:26] @ bodepd left channel #puppet ()
[2010/03/09 11:15:28] <jb> hm
[2010/03/09 11:16:24] <madduck> DavidS: you seem to have been spot on. danke!
[2010/03/09 11:16:31] <madduck> a few more things to fix though
[2010/03/09 11:16:45] <madduck> DavidS: anyway, if you wanted plain git hosting, I could happily set you up anytime
[2010/03/09 11:17:41] @ alfism joined channel #puppet
[2010/03/09 11:17:53] <DavidS> madduck: http://git.black.co.at is mine; doesn't change the fact that github is more visible and less overhead for me. also lak should have a password for the puppet-modules account on github, but not on my host
[2010/03/09 11:18:09] @ sbrock_ joined channel #puppet
[2010/03/09 11:18:17] <madduck> i use gitolite for hosting, so it's trivial to give out granular permissions. anyway, just an offer. ;)
[2010/03/09 11:18:51] * Volcane figures it out
[2010/03/09 11:18:57] <Volcane> its broken with setting environments as facts
[2010/03/09 11:18:58] <Volcane> ffs.
[2010/03/09 11:19:31] <Volcane> nigelk: do you still set environments in facts?
[2010/03/09 11:20:28] <nigelk> yes with caveats
[2010/03/09 11:20:32] @ nieko joined channel #puppet
[2010/03/09 11:20:36] <mpdehaan__> I'm thinking about how to define groups for puppetrun
[2010/03/09 11:20:47] <mpdehaan__> my first idea is just a INI style file in etcdir like hostgroups.txt
[2010/03/09 11:20:58] <Volcane> nigelk: sourcing files is broken if i do that - it fetches from the main modulepath for me :(
[2010/03/09 11:21:02] <mpdehaan__> so you should be able to do "puppetrun --host "!webservers" or something to dereference a group
[2010/03/09 11:21:05] <Volcane> nigelk: finding the class to serve though works
[2010/03/09 11:21:10] @ nexx_ joined channel #puppet
[2010/03/09 11:21:10] <mpdehaan__> preferences?
[2010/03/09 11:21:15] <nexx_> hiho
[2010/03/09 11:21:19] <nigelk> Volcane: so we had to change the way we did it
[2010/03/09 11:21:21] <madduck> if I want to define the debian_mirror_base in a class, i.e. have classes for machines-hosted-at-foo and machines-hosted-at-bar, and simply include those, how would I need to do it? It seems that include does not merge namespaces, but also $foo = $klass::foo does not work ("Could not look up qualified variable 'hoster_bytemark::debian_mirror_base'; class hoster_bytemark has not been evaluated")
[2010/03/09 11:21:39] @ Quit: sbrock: Ping timeout: 276 seconds
[2010/03/09 11:21:48] <nigelk> in 0.24x we could just set it as a fact, and you got the wrong modulepath (main) on first run, once your info was cached it was ok
[2010/03/09 11:22:02] <nigelk> so our puppet runs run against a specific tag in noop mode just to cache it on the server, then do a proper run
[2010/03/09 11:22:08] <Volcane> yes, ditto
[2010/03/09 11:22:10] <nexx_> is it possible to set a pwd for a single user without working with password-files? can i use password=>"foo" in the user-type?
[2010/03/09 11:22:14] <Volcane> in 0.25 that doesnt even work though
[2010/03/09 11:22:21] <nigelk> in 0.25.x we had to extend our puppet.conf.erb to include the environment
[2010/03/09 11:22:26] <Volcane> ffs.
[2010/03/09 11:22:27] <nigelk> (from the fact value)
[2010/03/09 11:22:39] <DavidS> madduck: use an external nodes classifier or volcane's extlookup (also contained in my common module)
[2010/03/09 11:22:49] <Volcane> nigelk: thanks, that fucking sux.
[2010/03/09 11:22:54] <nigelk> Volcane: this is part of a bigger problem lak etc are all trying to sort out
[2010/03/09 11:23:02] <nigelk> who is authoritative for environment info?
[2010/03/09 11:23:04] <nigelk> the server?
[2010/03/09 11:23:04] <Volcane> yeah i recall the conversations no :(
[2010/03/09 11:23:05] <nigelk> the client?
[2010/03/09 11:23:08] <DavidS> madduck: and include $klass first
[2010/03/09 11:23:09] <Volcane> no/now
[2010/03/09 11:23:17] <madduck> DavidS: right, it works with include $klass
[2010/03/09 11:23:29] <nigelk> I'm about to try throwing external node providers into the mix
[2010/03/09 11:23:31] <madduck> but I would really like to avoid external dependencies
[2010/03/09 11:23:37] <nigelk> and see if not specifying environment there at all helps
[2010/03/09 11:23:54] <Volcane> nigelk: yeah, arse oh well guess i need to make templates
[2010/03/09 11:24:13] <madduck> of course, your common module is so common that it should just become part of puppet!
[2010/03/09 11:24:19] <nigelk> This is where I get frustrated by external node providers
[2010/03/09 11:24:30] <mpdehaan__> what now?
[2010/03/09 11:24:30] <nigelk> you should have *more* than just the node name passed as an argument
[2010/03/09 11:24:38] <nigelk> it's not enough info to perform logic with
[2010/03/09 11:24:43] <mpdehaan__> hmm. what else?
[2010/03/09 11:24:47] <nigelk> all the facts
[2010/03/09 11:24:49] <Volcane> nigelk: totally
[2010/03/09 11:25:06] <mpdehaan__> You could look it up if storeconfigs is on, but yeah, I see what you mean
[2010/03/09 11:25:14] <nigelk> You have an annoying catch 22 here, given facts come from plugins in modules which are in environments.....
[2010/03/09 11:25:15] <mpdehaan__> you can't implement smart groups and such without that data
[2010/03/09 11:25:18] <Volcane> sotreconfigs works well </sarcasm>
[2010/03/09 11:25:19] <nigelk> exactly
[2010/03/09 11:25:23] <mpdehaan__> *nod*
[2010/03/09 11:25:31] <nigelk> Volcane: it's a lot better when it's "thin" and queued :)
[2010/03/09 11:25:31] @ JJA joined channel #puppet
[2010/03/09 11:25:47] <Volcane> still sux if u have many masters in regions
[2010/03/09 11:25:52] <nieko> I am currently doing a Project at the University of Amsterdam on Puppet. I am especially interested in problems that has occured during the implementation of Puppet in large, heterogeneous environments. Maybe somebody that has experience with Puppet in these kind of environments wants to have a talk? Feel free to query me.
[2010/03/09 11:25:56] <nigelk> you don't need storeconfigs really
[2010/03/09 11:26:04] <nigelk> you could always peer into the cache on the server :)
[2010/03/09 11:26:15] <mpdehaan__> nieko, what sort of problems?
[2010/03/09 11:26:24] <mpdehaan__> I am incidentally looking for GSOC students...
[2010/03/09 11:26:29] <nigelk> All the problems come down to one
[2010/03/09 11:26:32] <nigelk> "Computers are hard"
[2010/03/09 11:26:36] <nigelk> Let's go shopping!
[2010/03/09 11:26:38] <mpdehaan__> let's go shopping
[2010/03/09 11:26:39] <mpdehaan__> doh
[2010/03/09 11:26:41] <nigelk> :)
[2010/03/09 11:26:58] <mpdehaan__> yes all computers must be destroyed
[2010/03/09 11:27:16] <mpdehaan__> if they were nicely manageable no one would have jobs here :)
[2010/03/09 11:27:22] <mpdehaan__> think of it that way! it's a good thing
[2010/03/09 11:27:48] @ Quit: verwilst: Quit: Ex-Chat
[2010/03/09 11:27:49] <nieko> mpdehaan__: none particular, for me it is difficult to create ane environment which resembles for example hundreds or even thousands of servers.. a part of my project is about the usability in large heterogenous environments.
[2010/03/09 11:28:13] <Volcane> DavidS: ok well I've pulled in all the new fixes and stuff for hte concat, turns out we've actually added quite a few features in the next itteration so will blog etc
[2010/03/09 11:28:55] <mpdehaan__> nieko, so it's more of a report/survey than trying to build software solutions for something?
[2010/03/09 11:28:56] <DavidS> I'm actually having problems to get it to actually create the file ...
[2010/03/09 11:29:18] <nieko> mpdehaan__: yes
[2010/03/09 11:29:19] <mpdehaan__> I'd suggest you figure out what you really want to know and post to the list
[2010/03/09 11:29:27] <mpdehaan__> i.e. small survey or something
[2010/03/09 11:29:55] <nieko> mpdehaan__: idd, I thought of that, but I was not sure if the mailing list would be the right place to do that..
[2010/03/09 11:30:00] <nexx_> is it possible to set a pwd for a single user without working with password-files? can i use password=>"foo" in the user-type?
[2010/03/09 11:30:08] @ Quit: erimar77: Remote host closed the connection
[2010/03/09 11:30:41] <nigelk> neek: you mean specify the clear text password rather than the hash?
[2010/03/09 11:30:53] <nigelk> woops. nexx_ was who that was for
[2010/03/09 11:31:11] <nexx_> yes
[2010/03/09 11:31:14] <nigelk> no
[2010/03/09 11:31:14] <mpdehaan__> nieko, replying to /msg
[2010/03/09 11:31:20] <nexx_> simplest (and maybe not the smartest) way
[2010/03/09 11:31:34] <nigelk> if you *really* need that, set up an exec to set the password
[2010/03/09 11:31:50] <DavidS> Volcane: what is the fragments.concat file for?
[2010/03/09 11:32:12] <Volcane> DavidS: a temp file that it uses in future runs to compare against etc so it only replace files when it needs to
[2010/03/09 11:32:33] <nexx_> nigelk: yes, but i dont know how to use passwd with exec, cause it needs user-input after the call
[2010/03/09 11:32:52] <DavidS> isn't it much easier to just concat to the tmpdir and then use file { $name: source => $tmpfile } ?
[2010/03/09 11:33:41] <Volcane> DavidS: ordering pains etc, this way i can just have a test mode on the shell script to prevent execs from notifying etc
[2010/03/09 11:34:10] <nigelk> nexx_: which OS?
[2010/03/09 11:34:22] <nexx_> nigelk: ubuntu
[2010/03/09 11:34:24] <nigelk> ah
[2010/03/09 11:34:35] <nigelk> yeah, I don't think that supports passwd --stdin
[2010/03/09 11:34:45] <nexx_> mh
[2010/03/09 11:34:48] <nigelk> you might have to build an expect script that takes an argument
[2010/03/09 11:34:52] <nigelk> and use that to set the password
[2010/03/09 11:35:03] <nexx_> thought of that, yeah
[2010/03/09 11:35:26] @ gaveen joined channel #puppet
[2010/03/09 11:35:27] <nexx_> but before i do such stuff, i rather add the user manually without puppet
[2010/03/09 11:35:49] <Volcane> DavidS: so it concats on each run - test mode or not - but if you're asking it to test for use in the unless bits in the exec, then it just cmp's against the temp file but doesnt copy, else it copies
[2010/03/09 11:35:54] <DavidS> Volcane: puppet only notifies downstream if the file source actually changes and is copied. I don't see the problem there
[2010/03/09 11:36:05] <nigelk> nexx_: http://azimyasin.wordpress.com/2007/09/25/adding-users-on-ubuntu-box-via-file/
[2010/03/09 11:36:06] <DavidS> ah
[2010/03/09 11:36:19] <Volcane> DavidS: oh, you're thinking the concat script is only useful/used in puppet, wrong :)
[2010/03/09 11:36:42] <DavidS> still, I'd use it so: "unless => "/usr/local/bin/concatfragments.sh -o ${fragdir}/fragments.concat -d ${fragdir} -t","
[2010/03/09 11:36:43] <nexx_> nigelk: nice, thanks!
[2010/03/09 11:36:51] <nigelk> nexx_: or ;) "man chpasswd"
[2010/03/09 11:37:09] <nigelk> note that it bypasses PAM, so that may not be good for you
[2010/03/09 11:37:11] <nexx_> nigelk: oh, didnt know that this exists
[2010/03/09 11:37:14] <nigelk> me neither :)
[2010/03/09 11:37:27] <DavidS> Volcane: and file { $name: source => "${fragdir}/fragments.concat", ... }
[2010/03/09 11:37:52] <nexx_> time to quit work, thanks nigelk & bye
[2010/03/09 11:37:55] <nigelk> np
[2010/03/09 11:38:08] <DavidS> also, this shows nice diffs when something changes
[2010/03/09 11:38:09] <madduck> okay, APT module works
[2010/03/09 11:38:12] @ Quit: nexx_: Quit: leaving
[2010/03/09 11:38:24] <Volcane> DavidS: ah that would be useful yeah
[2010/03/09 11:38:24] <nigelk> madduck: what are you working on?
[2010/03/09 11:38:25] @ Quit: kubicek: Read error: Connection reset by peer
[2010/03/09 11:38:25] <madduck> now to integrate some of the ideas of tim|imac ;)
[2010/03/09 11:38:36] <madduck> nigelk: http://git.madduck.net/v/puppet/modules/apt.git/blob/HEAD:/manifests/init.pp
[2010/03/09 11:38:36] <Volcane> DavidS: I'll have a play
[2010/03/09 11:39:06] @ kubicek joined channel #puppet
[2010/03/09 11:39:13] <nigelk> madduck: interesting
[2010/03/09 11:39:15] <Volcane> DavidS: but i use the concat other places too thats not under puppet
[2010/03/09 11:39:46] @ Quit: MattyM: Quit: ta ta
[2010/03/09 11:40:24] <madduck> is there an easy way to tell the ruby template to split a string into lines of width 70 at word boundaries and prepend each line with a '# '?
[2010/03/09 11:40:34] <nigelk> madduck: we've also set up /etc/apt/preferences.d/ that we use here
[2010/03/09 11:40:57] <madduck> nigelk: yeah, except i kinda want pinning to become part of the Package type
[2010/03/09 11:41:05] <madduck> nigelk: package { [ puppet, puppet-common ]: ensure => "unstable" }
[2010/03/09 11:41:11] <nigelk> yep
[2010/03/09 11:41:14] <madduck> or somesuch
[2010/03/09 11:41:18] <nigelk> I have to finish off my hold support patch too
[2010/03/09 11:41:30] <madduck> oh, i don't believe in hold
[2010/03/09 11:41:39] <nigelk> it's all working even, so you can do package { "puppet": ensure => held }
[2010/03/09 11:41:41] @ Quit: unxfrek: Quit: Leaving
[2010/03/09 11:41:52] <nigelk> me neither particularly, but the rest of the ubuntu/debian folks here at google do :)
[2010/03/09 11:41:57] <madduck> package { [ puppet, puppet-common ]: ensure => "0.25.4-2" }
[2010/03/09 11:41:57] <nigelk> never works the way I expect it to
[2010/03/09 11:42:03] <madduck> just pin a version!
[2010/03/09 11:42:15] <madduck> ensure with a version should automatically set a pin
[2010/03/09 11:42:17] <nigelk> you want ensure a specific version to also pin that version?
[2010/03/09 11:42:25] <madduck> of course!
[2010/03/09 11:42:27] <nigelk> that's a big behaviour change....
[2010/03/09 11:42:36] <madduck> hold is only respected by dpkg and aptitude
[2010/03/09 11:42:40] <madduck> not by apt-get or APT
[2010/03/09 11:42:57] <nigelk> how are you going to know what value to pin at?
[2010/03/09 11:43:13] <madduck> you pass the version number, so you pin it to that. ;)
[2010/03/09 11:43:14] @ Quit: yure: Quit: Odhajam
[2010/03/09 11:43:41] <madduck> and you have a $apt_pin_priority_version=1001 or so
[2010/03/09 11:43:46] <nigelk> no I mean the pin-priority
[2010/03/09 11:43:49] <nigelk> hrm
[2010/03/09 11:43:52] <nigelk> ew
[2010/03/09 11:43:52] <nigelk> :)
[2010/03/09 11:43:54] <madduck> vs. $apt_pin_priority_release=600
[2010/03/09 11:44:08] <nigelk> I kind of feel like you're overloading ensure too much
[2010/03/09 11:44:12] <madduck> or 'ensure => "0.25.4-2/1001"'
[2010/03/09 11:44:13] <nigelk> and this deserves another attribute
[2010/03/09 11:44:16] <madduck> could be
[2010/03/09 11:44:22] <madduck> i am just running ideas by you. ;)
[2010/03/09 11:44:26] <nigelk> I like the idea of being able to do this :)
[2010/03/09 11:44:32] <madduck> in the hope that you'll implement it. ;)
[2010/03/09 11:44:35] <nigelk> but we have lots of pins above 1000
[2010/03/09 11:44:36] <nigelk> hahahaha
[2010/03/09 11:44:41] <madduck> (/me does not know ruby at all (yet))
[2010/03/09 11:44:57] <madduck> if you want a specific version installed, a pin of 1001 is the right onw
[2010/03/09 11:44:57] <nigelk> http://angryruby.blogspot.com/
[2010/03/09 11:45:07] <nigelk> madduck: what if you have release pinning above 1001 :)
[2010/03/09 11:45:24] <madduck> then you are nuts
[2010/03/09 11:45:34] <madduck> so pin versions at 99999
[2010/03/09 11:45:45] <madduck> or 65535, depending on implementation. ;)
[2010/03/09 11:45:47] <nigelk> heh
[2010/03/09 11:45:57] <nigelk> so if we were to come up with a new attribute
[2010/03/09 11:46:05] <nigelk> what would be a reasonable cross-packaging-system name for it?
[2010/03/09 11:46:13] <nigelk> does redhat have anything like pinning?
[2010/03/09 11:46:16] @ DavidS left channel #puppet ()
[2010/03/09 11:46:17] <madduck> pin ? ;)
[2010/03/09 11:46:23] @ DavidS joined channel #puppet
[2010/03/09 11:46:28] <nigelk> biab
[2010/03/09 11:46:29] <madduck> i don't know any other system that has something like pinning
[2010/03/09 11:47:12] <DavidS> volcane: http://gist.github.com/326780 , /var/lib/puppet/concat/_etc_apache2_ports.conf/fragments.concat is created, but empty. /etc/apache2/ports.conf is also empty. what am I doing wrong?
[2010/03/09 11:47:41] @ Quit: biertie: Ping timeout: 245 seconds
[2010/03/09 11:48:27] <DavidS> the fragments have the right contents
[2010/03/09 11:51:17] @ lilmatt joined channel #puppet
[2010/03/09 11:52:09] @ heali joined channel #puppet
[2010/03/09 11:52:59] <madduck> hacim: ^^
[2010/03/09 11:53:23] <heali> could somebody explain to me why this does not work: http://pastebin.org/107536
[2010/03/09 11:53:52] <heali> [this code is placed inside a template]
[2010/03/09 11:54:04] <kjetilho> madduck: in RedHat, you'd add include/exclude directives as appropriate to the various repositories
[2010/03/09 11:54:08] <DavidS> heali: because @sep is not defined outside the loop
[2010/03/09 11:54:27] <heali> DavidS: how do i get it outside?
[2010/03/09 11:54:37] <heali> i thought the @ made it global
[2010/03/09 11:54:51] <madduck> kjetilho: i don't know what that means. could you give an example?
[2010/03/09 11:54:52] <DavidS> @sep = "" before the loop?
[2010/03/09 11:55:08] <madduck> kjetilho: (a) track puppet from unstable/development, (b) keep mutt at 1.5.20
[2010/03/09 11:55:09] @ Quit: Ramonster: Quit: So long, thanx for all the fish
[2010/03/09 11:55:12] <DavidS> heali: but I'm only guessing, I just noticed that you haven't posted the error message
[2010/03/09 11:55:12] <kjetilho> [epel-updates] excludepkgs = *
[2010/03/09 11:55:24] <kjetilho> includepkgs = *puppet*
[2010/03/09 11:55:49] <kjetilho> keep mutt at 1.5.20 is ensure => "1.5.20"
[2010/03/09 11:55:58] <kjetilho> no need for new syntax there
[2010/03/09 11:56:05] <madduck> kjetilho: what if someone upgrades mutt by hand?
[2010/03/09 11:56:10] @ phantez joined channel #puppet
[2010/03/09 11:56:24] <madduck> on debian right now, puppet would *not* downgrade
[2010/03/09 11:56:25] <heali> DavidS: Failed to parse template /etc/puppet...: undefined method 'join' for "item1": string at /etc/puppet/... on node test
[2010/03/09 11:56:28] <DavidS> madduck: puppet should downgrade and report. Then apply LART to admin
[2010/03/09 11:56:35] <kjetilho> madduck: what if someone types rm -rf / ?
[2010/03/09 11:56:35] <heali> DavidS: did not seem to fix my problem but thanks for teaching me that :)
[2010/03/09 11:56:39] <madduck> kjetilho: ;)
[2010/03/09 11:56:46] <DavidS> heali: do'h use [] instead of ""
[2010/03/09 11:56:56] <DavidS> heali: it should be an array, not a stirng
[2010/03/09 11:57:06] <madduck> DavidS: true, it does downgrade
[2010/03/09 11:57:31] <madduck> so we don't need to add pins for versions, but it would be consistent
[2010/03/09 11:58:18] <madduck> kjetilho: so if we did something like 'package { [ puppet, puppet-common ]: pin_to_release => "epel-updates" }', that would make sense in RedHat?
[2010/03/09 11:58:39] <kjetilho> s/pin_to_release/pin_to_repository/
[2010/03/09 11:59:23] <kjetilho> wouldn't just "repository" be a good keyword for this?
[2010/03/09 11:59:26] <madduck> ok, how about just "pin"?
[2010/03/09 11:59:30] <madduck> ;)
[2010/03/09 11:59:36] <madduck> repository would work for me too
[2010/03/09 11:59:43] <hacim> madduck: can you provide me with an executive summary?
[2010/03/09 12:00:06] @ Quit: danielbln: Quit: Leaving.
[2010/03/09 12:00:25] <madduck> hacim: 'package { [ puppet, puppet-common ]: repository => "lenny-backports" }' would install puppet and automatically maintain pins
[2010/03/09 12:00:39] <DavidS> Volcane: aaahhhh: /bin/sort: No such file or directory
[2010/03/09 12:01:08] <madduck> hacim: also, 'package { mutt: ensure => "1.5.20" }' should really add a versioned pin >1000 for consistency
[2010/03/09 12:01:21] <madduck> hacim: basically, pins should be managed by package types
[2010/03/09 12:01:29] <heali> DavidS: http://pastebin.org/107537
[2010/03/09 12:02:21] <heali> DavidS: i want to add a last item to each row
[2010/03/09 12:02:34] <heali> but its harder then i thought, perhaps i should ask this in #ruby
[2010/03/09 12:03:30] @ Quit: Robbie_: Remote host closed the connection
[2010/03/09 12:04:57] @ Quit: kubicek: Quit: kubicek
[2010/03/09 12:06:19] @ DavidS1 joined channel #puppet
[2010/03/09 12:06:31] @ Quit: DavidS1: Client Quit
[2010/03/09 12:06:39] @ Quit: DavidS: Ping timeout: 265 seconds
[2010/03/09 12:09:14] @ Quit: alban2: Quit: Leaving.
[2010/03/09 12:09:28] <hacim> madduck: that would be nice: http://projects.reductivelabs.com/issues/2406
[2010/03/09 12:09:43] <hacim> also http://projects.reductivelabs.com/issues/2836
[2010/03/09 12:11:28] <madduck> hacim: i don't think we'll need 'held' if pins work fine. i never trust holds anyway
[2010/03/09 12:11:50] <hacim> some people use them, even if they aren't you :)
[2010/03/09 12:12:11] <madduck> tim|imac: http://git.madduck.net/v/puppet/modules/apt.git/commitdiff/c789e4e434e5d22f09fee35773082580c7b97288 — but I just found out that it's no longer needed on linux kernels.
[2010/03/09 12:12:16] <madduck> hacim: NO!
[2010/03/09 12:12:51] <madduck> hacim: point being, I think we could gently nudge those people towards sanity with pins, or do you think holds are something with worth of their own?
[2010/03/09 12:12:52] <hacim> hehe
[2010/03/09 12:13:03] <hacim> i think pins are insanity
[2010/03/09 12:13:09] <hacim> i also think that about holds
[2010/03/09 12:13:12] <madduck> i think *you* are insanity
[2010/03/09 12:13:12] <hacim> but for different reasons
[2010/03/09 12:13:21] <madduck> ;)
[2010/03/09 12:13:26] <madduck> hacim: ♥ sorry!
[2010/03/09 12:17:02] <tim|imac> madduck: yeah, I should do that too... even if it's no longer needed with newer kernels...
[2010/03/09 12:17:09] <tim|imac> btw... define newer?
[2010/03/09 12:17:17] <madduck> tim|imac: no idea, i didn't research it
[2010/03/09 12:17:24] <tim|imac> oh ok
[2010/03/09 12:17:28] <madduck> just hadn't set the value on any lenny system
[2010/03/09 12:19:18] @ alexine_dsouza joined channel #puppet
[2010/03/09 12:19:26] @ eshamow joined channel #puppet
[2010/03/09 12:19:48] @ ch joined channel #puppet
[2010/03/09 12:20:44] @ Groenleer is now known as Groenleer`Away
[2010/03/09 12:20:56] @ Groenleer`Away is now known as Groenleer
[2010/03/09 12:23:15] @ Quit: jaredrhine: Read error: Operation timed out
[2010/03/09 12:24:16] @ EdRandom joined channel #puppet
[2010/03/09 12:26:04] @ Quit: fxpester: Read error: Connection reset by peer
[2010/03/09 12:28:30] @ Quit: Spads: Ping timeout: 260 seconds
[2010/03/09 12:35:04] @ gaveen_ joined channel #puppet
[2010/03/09 12:38:27] @ Quit: gaveen: Ping timeout: 258 seconds
[2010/03/09 12:42:09] @ Quit: TREllis: Quit: ciao
[2010/03/09 12:43:46] @ Quit: jab_doa: Remote host closed the connection
[2010/03/09 12:47:15] @ dyresen is now known as Dyresen
[2010/03/09 12:49:41] @ unxfrek joined channel #puppet
[2010/03/09 12:49:46] @ Quit: giskard: Remote host closed the connection
[2010/03/09 12:50:20] @ giskard joined channel #puppet
[2010/03/09 12:50:30] @ WALoeIII joined channel #puppet
[2010/03/09 12:50:43] @ gaveen_ is now known as gaveen
[2010/03/09 12:54:56] @ Quit: giskard: Ping timeout: 258 seconds
[2010/03/09 12:59:44] <dan__t> Howdy.
[2010/03/09 13:00:00] <dan__t> I can't figure out why solr_master_url is not being expanded in the following template snippet: http://pastebin.com/wXQciQTM
[2010/03/09 13:00:08] @ Quit: ckauhaus: Quit: Leaving.
[2010/03/09 13:00:12] <dan__t> Is it because its embedded inside an if statement?
[2010/03/09 13:00:34] @ Quit: JJA: Ping timeout: 240 seconds
[2010/03/09 13:01:12] <dan__t> solr_master_url is defined via extlookup, its actually there
[2010/03/09 13:04:00] @ Quit: heali: Ping timeout: 252 seconds
[2010/03/09 13:05:26] @ raz left channel #puppet ()
[2010/03/09 13:06:06] <nigelk> madduck: hacim I'm not convinced pinning should happen at the package resource level
[2010/03/09 13:06:30] <nigelk> as often we want to implement pins independently of actual package resource definition
[2010/03/09 13:06:42] <nigelk> we may have no packages automatically installed that would be affected by a given pin
[2010/03/09 13:07:00] <dan__t> I smell something wrong with scope... I did the scope.lookupvar trick that the Template docs suggest, but I don't think that's going to get me what I want anyway.
[2010/03/09 13:07:07] <madduck> nigelk: example?
[2010/03/09 13:07:33] <nigelk> madduck: we pin a certain repository/release that users can and will install software from, but we don't enforce the state of any of those packages
[2010/03/09 13:07:56] <nigelk> it's making me think we could do with a whole "apt" type of some kind around repos and pinning
[2010/03/09 13:08:15] <nigelk> like the module you posted above but a bit different
[2010/03/09 13:08:26] <madduck> nigelk: well, I think there's a repository type, for sure, and I am quite happy with mine so far, although the key management does not yet work
[2010/03/09 13:08:34] <madduck> how about adding a pin there??
[2010/03/09 13:08:44] <nigelk> yeah, I'm trying to think how we'd do this
[2010/03/09 13:08:56] <nigelk> currently we have /etc/apt/preferences.d we drop fragments into and construct apt preferences with
[2010/03/09 13:09:33] @ Quit: bgupta: Quit: bgupta
[2010/03/09 13:09:44] <nigelk> I'll have a poke through that list of fragments and think about how we'd want to model them instead
[2010/03/09 13:10:11] <nigelk> madduck: apt-get does understand holds?
[2010/03/09 13:10:18] <madduck> i don't think it does
[2010/03/09 13:10:20] <nigelk> for dist-upgrade doesn't it?
[2010/03/09 13:10:31] <madduck> i'd be surprised
[2010/03/09 13:10:34] <nigelk> i'm not sure why apt-get would have --ignore-hold then :)
[2010/03/09 13:10:35] <madduck> but not impossible
[2010/03/09 13:10:47] <madduck> convincing argument you have there ;)
[2010/03/09 13:10:50] * madduck retreats to his shell
[2010/03/09 13:11:24] @ lak joined channel #puppet
[2010/03/09 13:11:26] <nigelk> when I looked at this... the problem was aptitude implements holds differently
[2010/03/09 13:11:38] <nigelk> that may be resolved in very new versions that it respects dpkg holds
[2010/03/09 13:11:41] <nigelk> apt-get certainly does
[2010/03/09 13:12:00] <hacim> madduck: your apt module is nice! i still don't understand why you didn't just add this to an existing history
[2010/03/09 13:12:09] <madduck> hacim: i might still
[2010/03/09 13:12:24] <hacim> huzzah!
[2010/03/09 13:12:24] <madduck> grrrr, why can't I use $vardir from puppet.conf within a manifest… or how would I?
[2010/03/09 13:12:54] <madduck> hacim: mostly because there seems to be some not-quite-perpendicular stuff in module-apt, and I didn't want to worry about interoperation.
[2010/03/09 13:13:33] <nigelk> madduck: you shouldn't need to peer inside the client config like that? how come you need to do it?
[2010/03/09 13:14:02] <madduck> nigelk: i need to store temporary files from my module, and I thought $vardir/modules/$mymodulename/ seems like the place
[2010/03/09 13:14:23] <nigelk> oh you mean the server's $vardir ?
[2010/03/09 13:14:25] @ jaredrhine joined channel #puppet
[2010/03/09 13:14:29] <nigelk> I thought that worked in a manifest
[2010/03/09 13:14:35] <madduck> no, this is on the client.
[2010/03/09 13:15:32] <madduck> DavidS uses a whole define for it, module_dir { ["common", "common/dir1", "common/dir2" ]: }
[2010/03/09 13:16:01] <hacim> madduck: if you class out things so that different pieces can be chosen, then those things that you don't use you dont 'include'
[2010/03/09 13:16:27] <madduck> hacim: yeah well, i first wanted to gain some experience anyway
[2010/03/09 13:16:33] <hacim> its good!
[2010/03/09 13:16:44] <hacim> madduck levels up!
[2010/03/09 13:17:08] @ bgupta joined channel #puppet
[2010/03/09 13:17:26] <dan__t> Definitely a scoping problem.
[2010/03/09 13:17:34] <madduck> hacim: ;)
[2010/03/09 13:17:42] <madduck> hacim: wait for the key management solution. ;)
[2010/03/09 13:17:53] <madduck> it's nifty in theory, but has yet to work. ;)
[2010/03/09 13:18:53] <joe-mac1> ssh keys? i plan on writing an ssh key rotator that uses fetchmail and whatnot to get the new keys. gotta work out all the security issues.
[2010/03/09 13:19:12] <madduck> no, apt repo keys
[2010/03/09 13:19:34] <joe-mac1> ah, i c. i serve that out to some dir and notify a refreshonly exec to import it.
[2010/03/09 13:19:37] @ Quit: xerxas: Ping timeout: 264 seconds
[2010/03/09 13:20:03] <Clint> fetchmail as key distribution?
[2010/03/09 13:20:27] <dan__t> heh
[2010/03/09 13:20:34] <dan__t> that just sounds funny
[2010/03/09 13:20:42] <madduck> i am rewriting puppet to work off imap4rev1
[2010/03/09 13:21:09] @ giskard joined channel #puppet
[2010/03/09 13:21:16] <hacim> make it a dovecot extension
[2010/03/09 13:21:21] <joe-mac1> no, see my ssh users are only developers. so i want to parse the users class every quarter, send an e-mail out telling them to include their newly generated key and some shared secret that i have tio distribute some other way i haven't quite figured out so no one impersonates someone.
[2010/03/09 13:21:40] <dan__t> Anyone mind schooling me on a little variable scope please? Why doesn't solr_master_url expand in this example: http://pastebin.com/wXQciQTM
[2010/03/09 13:21:40] <joe-mac1> then fetchmail goes and gets newsshkeys@mysite.com or something, edits the manifest etc
[2010/03/09 13:21:59] <joe-mac1> this is just an idea i've been toying with, have had no time to start working on it yet since i talk in here too much :-P
[2010/03/09 13:22:37] <dan__t> fact
[2010/03/09 13:22:38] <dan__t> :)
[2010/03/09 13:22:38] <Volcane> joe-mac1: i just give mine access to svn
[2010/03/09 13:23:11] <joe-mac1> easier cause you stuck with files, like i should have done. i moved to the authorized keys type. to be all puppety and stuff.
[2010/03/09 13:23:15] <dan__t> wordup, Volcane
[2010/03/09 13:23:46] <Volcane> joe-mac1: you dont think your devs are clever enough to put their key in a manifest?
[2010/03/09 13:23:53] <dan__t> i maintain an authorized_keys/%u archive on each machine, puppet svn co's everyone's key and maintains it that way
[2010/03/09 13:24:03] <dan__t> That works out real well for us.
[2010/03/09 13:24:11] <joe-mac1> o they are. but they are also clever enough to edit the manifest since right now it's a class that contains all the users
[2010/03/09 13:24:18] <Volcane> yup, so what?
[2010/03/09 13:24:39] <joe-mac1> in which case they can make themselves root, or even change my key so i can't access my boxes
[2010/03/09 13:24:52] @ eshamow is now known as eshamow|away
[2010/03/09 13:24:57] <Volcane> you are running code they write on your servers, if you dont trust them you've lost already
[2010/03/09 13:25:12] <joe-mac1> interesting point
[2010/03/09 13:25:16] <Volcane> plus, since the code is the reason the business exist, you'll find its their boxes
[2010/03/09 13:25:23] <Volcane> you're just preparing it for them
[2010/03/09 13:25:36] @ Spads joined channel #puppet
[2010/03/09 13:25:58] <dan__t> Can I call other variables like that from within if statements? Am I really out of scope?
[2010/03/09 13:26:00] <joe-mac1> how do you manage svn authorization Volcane? i use plain unix perms right now cause most of my use cases are super simple... do uou have some define for authz files?
[2010/03/09 13:26:34] <Volcane> joe-mac1: have a user define that copies out keys and a profile - devs can provide their own profiles for their user accounts to set up svn and shells and whatever
[2010/03/09 13:26:49] @ Quit: giskard: Ping timeout: 264 seconds
[2010/03/09 13:27:05] <joe-mac1> but i mean authorization for the repo itself
[2010/03/09 13:27:11] @ giskard joined channel #puppet
[2010/03/09 13:27:14] <joe-mac1> is this one of your few manual tasks?
[2010/03/09 13:27:35] <Volcane> your assuming i restrict who can commit where :)
[2010/03/09 13:27:58] <Volcane> business knows development is ultimately respnsible for the app, its support, deploys and maintenance
[2010/03/09 13:28:06] @ Quit: pmorillo: Quit: pmorillo
[2010/03/09 13:28:11] <Volcane> if it fucks out, the CEO doesnt even ask me what happened, goes str8 to the lead dev
[2010/03/09 13:28:18] <Volcane> = they dont mess around
[2010/03/09 13:28:30] @ paxos joined channel #puppet
[2010/03/09 13:28:47] <Volcane> if alerts go off at 2am, i dont get them, dev does
[2010/03/09 13:29:01] <Volcane> u can bet that if unstable code got rolled out today, it gets fixed morrow
[2010/03/09 13:29:59] <hggh> uh. that's cool, if dev gets notified.
[2010/03/09 13:30:27] <dan__t> Volcane, can I borrow your brain on this for a minute please?
[2010/03/09 13:30:32] <Volcane> they can deploy, roll out rpms, build vhosts, restart service - everything is logged and audited though via sudo and whatevers
[2010/03/09 13:30:42] <dan__t> sudo su -
[2010/03/09 13:31:02] <Volcane> and they're ultimately responsible, leaving me to do things i care for - build better monitoring, backups, networks and such
[2010/03/09 13:31:19] <Volcane> dan__t: wassup?
[2010/03/09 13:31:27] <hggh> yeah, if you have good devs, you can make it.
[2010/03/09 13:31:40] <dan__t> http://pastebin.com/wXQciQTM - Is solr_master_url out of scope? It doesn't expand where it is right now.
[2010/03/09 13:32:23] <Volcane> dan__t: depends on your puppet manifest - which isnt shown
[2010/03/09 13:32:34] <dan__t> Sorry, one second.
[2010/03/09 13:33:02] <dan__t> http://pastebin.com/USxcT23C
[2010/03/09 13:33:45] @ biertie joined channel #puppet
[2010/03/09 13:33:57] <dan__t> I'm still in love with your extlookup heh
[2010/03/09 13:34:10] <Volcane> dan__t: should be in scope, add notify{"solr master url: ${solr_master_url}": } and see what --test says
[2010/03/09 13:36:14] <dan__t> Good call, I need to use that more often.
[2010/03/09 13:36:18] <dan__t> Yes, it does expand properly.
[2010/03/09 13:37:18] <Volcane> oh
[2010/03/09 13:37:23] <Volcane> <%= instead of <%
[2010/03/09 13:38:51] @ Quit: p3rror: Read error: Operation timed out
[2010/03/09 13:39:42] <dan__t> Nosir.
[2010/03/09 13:39:49] @ \ask joined channel #puppet
[2010/03/09 13:40:24] <Volcane> line 3 in your template? should have <%= solr_master_url %>
[2010/03/09 13:41:01] <dan__t> Yeah that's what I changed it to.
[2010/03/09 13:41:33] <Volcane> then i am guessing the problem lies in solr_slave
[2010/03/09 13:42:03] <Volcane> or does it run the block just with nothing where the master url should be?
[2010/03/09 13:42:25] <dan__t> BAH
[2010/03/09 13:42:52] <dan__t> Yeah you were right, I was changing some shit and used $solr_master_url
[2010/03/09 13:43:02] <Volcane> heh
[2010/03/09 13:43:28] <dan__t> Thanks man. Again.
[2010/03/09 13:43:37] <Volcane> np
[2010/03/09 13:44:00] <Volcane> you should learn about Rubber duck debugging, it solves most of this stuff :P
[2010/03/09 13:44:15] <dan__t> fact.
[2010/03/09 13:44:16] <dan__t> thanks man :)
[2010/03/09 13:44:18] <dan__t> Back in a few.
[2010/03/09 13:45:27] <Babbelaar> I'm trying to build a redhat rpm for 0.24.6 and I keep running into a dependency issue regarding /sbin/runscript
[2010/03/09 13:46:08] <Babbelaar> I imagine this is explained somewhere, but I can't find anything, is there an option I can put in the spec file to tell it I'm not using gentoo?
[2010/03/09 13:47:11] <Volcane> Babbelaar: why dont you use the rpm spec file that came in the tarfile?
[2010/03/09 13:47:41] <Babbelaar> I'm trying to build it from the gem
[2010/03/09 13:47:53] <Volcane> get the tar file, there's a rpm spec file in it
[2010/03/09 13:47:54] @ mvn071 joined channel #puppet
[2010/03/09 13:47:54] <Volcane> use that
[2010/03/09 13:47:56] <Babbelaar> I can build the one from the tarfile with no problem
[2010/03/09 13:48:01] <Babbelaar> ok
[2010/03/09 13:48:16] <Volcane> gem -> rpm when not needed is just GIGO
[2010/03/09 13:51:07] @ p3rror joined channel #puppet
[2010/03/09 13:56:01] <madduck> okay, APT repository key management done: http://git.madduck.net/v/puppet/modules/apt.git/commitdiff/578cc415ab3d0c8b4f9176d28485f4da62e845cd
[2010/03/09 13:56:04] <madduck> hacim: ^
[2010/03/09 13:56:37] <madduck> this is the way it's used: http://slexy.org/view/s2vgTz9mrM
[2010/03/09 13:56:39] @ weaselkeeper joined channel #puppet
[2010/03/09 13:57:09] <weaselkeeper> anyone using puppet 0.25 with passenger ? (2.2.11 if that matters)
[2010/03/09 13:57:45] <nigelk> weaselkeeper: yes but with 2.2.5
[2010/03/09 13:57:52] @ gebi joined channel #puppet
[2010/03/09 13:58:02] <madduck> wtf, a file is lost… :(
[2010/03/09 13:58:03] <nigelk> i've been wary of upgrading as the passenger folks seem to be quite happy to break compatibility in point releases
[2010/03/09 13:58:31] <tim|imac> madduck: aren't you getting in an infinite loop that way? you need gnupg to check the key, but you need to add the key before you can install gnupg...
[2010/03/09 13:58:55] <weaselkeeper> nigelk: Is that gem installed passenger? or centos/etc?
[2010/03/09 13:59:10] <weaselkeeper> yeah, I have been noticing that on another issue
[2010/03/09 13:59:26] @ Quit: suchu: Quit: ChatZilla 0.9.86 [Firefox 3.5.8/20100202165920]
[2010/03/09 13:59:33] <weaselkeeper> think I will roll it back to 2.2.5 and see if my problem goes away before I try to bug someone else with it. :P
[2010/03/09 13:59:37] <nigelk> weaselkeeper: I built my own debs for ubuntu based upon the brightbox PPAs
[2010/03/09 13:59:48] <nigelk> as their debs are 32bit and we're all 64bit here
[2010/03/09 13:59:52] <weaselkeeper> nigelk: ok, thanks
[2010/03/09 13:59:56] <madduck> tim|imac: same problem with debian-archive-keyring
[2010/03/09 14:00:10] <nigelk> weaselkeeper: yeah, I'd try 2.2.5 at least you know that can work :)
[2010/03/09 14:00:11] <madduck> tim|imac: also, gnupg is installed on debian by default, I think
[2010/03/09 14:00:46] @ ezmobius joined channel #puppet
[2010/03/09 14:00:47] @ Quit: ezmobius: Excess Flood
[2010/03/09 14:01:11] @ ezmobius joined channel #puppet
[2010/03/09 14:01:25] <tim|imac> madduck: hm ok
[2010/03/09 14:01:55] @ pting joined channel #puppet
[2010/03/09 14:02:07] @ alban2 joined channel #puppet
[2010/03/09 14:02:36] * tim|imac likes the key class
[2010/03/09 14:02:42] <tim|imac> i might steal it from you :P
[2010/03/09 14:03:29] <madduck> tim|imac: maybe we can fuse out two apt modules?
[2010/03/09 14:03:35] <madduck> i took a bit from you
[2010/03/09 14:04:05] <tim|imac> aye :)
[2010/03/09 14:04:25] <madduck> anyway, if you checked out git, go a fetch and force-update to origin/master
[2010/03/09 14:04:26] <madduck> ;)
[2010/03/09 14:04:46] <madduck> http://git.madduck.net/v/puppet/modules/apt.git/commitdiff/416874b39faa30c3ff20aad5b554fd58ea73d2cd is the proper commit
[2010/03/09 14:05:59] @ Quit: mpdehaan_: Quit: mpdehaan_
[2010/03/09 14:06:05] <tim|imac> although I am developing in git, I'm fairly new to it all... you kinda lost me after "checked out git"
[2010/03/09 14:06:13] <madduck> did you clone my repo?
[2010/03/09 14:06:23] <tim|imac> no, i was already working in my own
[2010/03/09 14:06:30] <madduck> okay, then you're fine
[2010/03/09 14:06:33] <tim|imac> but i don't have a repo per module, either
[2010/03/09 14:06:39] <madduck> i did something bad to my repo (rewrote public history)
[2010/03/09 14:06:49] <madduck> that fucks with everyone who tracks my repo
[2010/03/09 14:06:54] <tim|imac> lol
[2010/03/09 14:06:58] <madduck> and thus i am trying to make sure you don't get screwed ;)
[2010/03/09 14:07:07] <madduck> yeah, bad madduck
[2010/03/09 14:07:43] @ kubicek joined channel #puppet
[2010/03/09 14:07:47] <tim|imac> hm... does an undefined variable make it equal ""?
[2010/03/09 14:07:51] <weaselkeeper> <sigh> still getting the same problem with passenger 2.2.5 as well. So it's my config. I have puppet working fine with the built in server, no worries. Running via passenger returns an error to the error_log for apache that bitches about /var/puppet and /etc/puppet/ssl but those dirs aren't defined anywhere. I have /var/lib/puppet and /var/lib/puppet/ssl, all deined in configs.
[2010/03/09 14:08:02] <madduck> tim|imac: i think so, and to undefine, you assign undef
[2010/03/09 14:08:13] <weaselkeeper> I Have no idea why puppet via passenger is looking for /var/puppet and /etc/puppet/ssl
[2010/03/09 14:08:40] <gebi> has anyone already tried puppet with unicorn?
[2010/03/09 14:08:43] <tim|imac> madduck: instead of those selectors, I'm using a "if $var { $real_var = $var } else { $real_var = "default_value" }" everywhere
[2010/03/09 14:09:03] <madduck> tim|imac: i don't like either way. :(
[2010/03/09 14:09:03] <weaselkeeper> any ideas where those dirs would be configured? I tried symlinking to /var/lib/puppet and /var/lib/puppet/ssl, but that breaks on owner=root of the symlink, and passenger doesn't like that.
[2010/03/09 14:09:17] <madduck> thanks for #3181
[2010/03/09 14:09:18] <gepetto> madduck: #3181 is http://projects.reductivelabs.com/issues/show/3181 "Puppet - Feature #3181: Create resource if it doesn't already exist (with array support) - ReductiveLabs.com"
[2010/03/09 14:09:23] <Volcane> weaselkeeper: does the user that puppet run as - look in ps for puppetmaster - have access to /etc/puppet/puppet.conf?
[2010/03/09 14:10:11] <tim|imac> madduck: vote it up or expand :) it's been quiet for a few days and I don't want to force the issue ;-)
[2010/03/09 14:10:16] <weaselkeeper> nope, he sure doesn't. That was it.
[2010/03/09 14:10:18] <weaselkeeper> Thank you!
[2010/03/09 14:10:28] <weaselkeeper> so simple.... doh! retarded admin ftl.
[2010/03/09 14:10:29] <weaselkeeper> Thanks
[2010/03/09 14:10:34] <Volcane> weaselkeeper: make it run as puppet - by setting the rack up file to the right user
[2010/03/09 14:10:39] @ Quit: unxfrek: Ping timeout: 276 seconds
[2010/03/09 14:11:49] <weaselkeeper> Volcane: it is running as puppet, but the file was owned by root and 440 perms
[2010/03/09 14:11:51] @ JuddJudd joined channel #puppet
[2010/03/09 14:12:13] <weaselkeeper> changing /etc/puppet/puppet.conf to be owned by puppet works.
[2010/03/09 14:12:22] <Volcane> ah
[2010/03/09 14:12:38] @ Quit: Babbelaar: Remote host closed the connection
[2010/03/09 14:12:55] <JuddJudd> Heya folks. I've got a silly templating question.
[2010/03/09 14:14:12] <JuddJudd> I've read the http://reductivelabs.com/trac/puppet/wiki/PuppetTemplating
[2010/03/09 14:14:19] @ unxfrek joined channel #puppet
[2010/03/09 14:14:54] <JuddJudd> but I'm left wondering how I can loop over multiple values at the same time, perhaps lists of hostnames & ports.
[2010/03/09 14:15:10] <JuddJudd> The example shows looping over only one value.
[2010/03/09 14:15:16] @ tonyskapunk joined channel #puppet
[2010/03/09 14:15:28] <JuddJudd> s/value/list/
[2010/03/09 14:15:37] @ Quit: Desdic: Remote host closed the connection
[2010/03/09 14:15:51] <tim|imac> JuddJudd: like, you have multiple arrays?
[2010/03/09 14:15:54] @ gwar9999 joined channel #puppet
[2010/03/09 14:16:03] <JuddJudd> yup.
[2010/03/09 14:16:11] <tim|imac> you can simply add a loop within a loop
[2010/03/09 14:16:28] <JuddJudd> not exactly.
[2010/03/09 14:17:07] <JuddJudd> $hostnames = [ host1, host2, host3 ]
[2010/03/09 14:17:21] <JuddJudd> $ports = [ 8080, 8180, 8280 ]
[2010/03/09 14:17:30] <tim|imac> <% values.each do |val| %> <% values2.each do |val2| %> Yay, I got <%= val %> with <%= val2 %>!! <% end %> <% end %>
[2010/03/09 14:18:11] <tim|imac> you want to add some newlines in there, though ;-)
[2010/03/09 14:18:21] <JuddJudd> then I'd have :
[2010/03/09 14:18:53] <JuddJudd> I'd have each host represented with each of the ports.. not just the port in the same array position as the hostname.
[2010/03/09 14:19:01] <tim|imac> ah
[2010/03/09 14:19:08] <JuddJudd> How can I increment two arrays at the same time, or just use a different data structure?
[2010/03/09 14:19:09] <tim|imac> i got you
[2010/03/09 14:19:18] <tim|imac> use a different data structure, i'd say
[2010/03/09 14:19:24] <Volcane> JuddJudd: hostnames.each_with_index do |host, idx|
[2010/03/09 14:19:38] <tim|imac> or do what Volcane says, which is probably smarter :)
[2010/03/09 14:19:38] <Volcane> JuddJudd: ports[idx] wil then be port matching the host
[2010/03/09 14:19:46] <JuddJudd> Volcane - THANKS!
[2010/03/09 14:20:03] <JuddJudd> or should I really be using Augeus for this?
[2010/03/09 14:20:06] * tim|imac really should learn ruby one of these days (I think I've been saying that for the last two years now)
[2010/03/09 14:20:20] <tim|imac> JuddJudd: depends on which file you're trying to recreate
[2010/03/09 14:20:24] @ verwilst joined channel #puppet
[2010/03/09 14:20:38] <JuddJudd> time|imac : mod_jk workers.properties file.
[2010/03/09 14:21:19] <JuddJudd> time|imac : I should really learn Ruby too, and not keep assuming it's "cool perl."
[2010/03/09 14:22:08] <JuddJudd> Is there a wiki page I can update with this case?
[2010/03/09 14:22:44] <tim|imac> JuddJudd: ah, we've switched to mod_proxy_ajp, because you can ditch the silly workers.properties file :)
[2010/03/09 14:22:45] <JuddJudd> I'd rather others not face the same fate as me.
[2010/03/09 14:23:09] <tim|imac> if you up to making a lens for it, augeas is probably easiest... otherwise I'd use concatenated_file resources
[2010/03/09 14:24:07] <JuddJudd> time|imac : I wish I could ditch it. Turns out the app developers have already created a workaround for the failings of mod_jk (whatever they are), so it's chicken-and-egg issue.
[2010/03/09 14:24:16] * Volcane had unimaginable hatefull experiences with mod_proxy_ajp under heavy load
[2010/03/09 14:25:01] <tim|imac> Volcane: i don't think our load is very heavy, but we've reached maxclients a few times without mod_proxy_ajp problems
[2010/03/09 14:25:02] <Volcane> it owuld literally mix up responses - send response that left tomcat for user a to user b's browser
[2010/03/09 14:25:14] <tim|imac> haven't seen that
[2010/03/09 14:25:15] <nigelk> Volcane: I've been there! It sucked
[2010/03/09 14:25:23] <nigelk> that was a while ago though
[2010/03/09 14:25:33] <Volcane> refused to believe it sucked that bad, got the devs to spend AGES debugging their stuff
[2010/03/09 14:25:33] <nasrat> yeah I thought it'd all moved to straight mod_proxy as trwtdi
[2010/03/09 14:25:42] * tim|imac thinks mod_jk is fairly draconian
[2010/03/09 14:25:45] <Volcane> eventually i rolled back to the old way and it all worked :(
[2010/03/09 14:26:02] <Volcane> ajp is nice that it keeps original requestors ips and stuff
[2010/03/09 14:26:12] <Volcane> but really, mod_proxy_ajp is rubbish
[2010/03/09 14:26:14] <JuddJudd> what about ditching ajp altogether and using http?
[2010/03/09 14:26:40] <tim|imac> Volcane: when was your experience with it? we've switched fairly recent... last summer
[2010/03/09 14:26:50] <Volcane> yeah back then my apache's infront were serving lots of other stuff - can do now since my tomcat machines does just that
[2010/03/09 14:26:58] <JuddJudd> Volcane: you can get original IP from http headers.
[2010/03/09 14:27:04] <Volcane> tim|imac: late last year - we ran latest versions etc
[2010/03/09 14:27:20] <tim|imac> hm... we run debian packages
[2010/03/09 14:27:28] <tim|imac> haven't seen what you described, really
[2010/03/09 14:27:41] <Volcane> JuddJudd: you'd need to interpret the forwarded for headers and stuff
[2010/03/09 14:27:46] <nasrat> we were having fun with trying to backport mod_remoteip last week from 2.3 to 2.2.x
[2010/03/09 14:28:01] <JuddJudd> Volcane: tcp headers or http?
[2010/03/09 14:28:26] <Volcane> nasrat: mod_extract_forwarded is nice too
[2010/03/09 14:28:30] <nasrat> JuddJudd: you need to parse X-Forwarded-For, which is spoofable to some degree
[2010/03/09 14:28:37] <Volcane> JuddJudd: mod_proxy (non ajp one) would hide the source ip
[2010/03/09 14:28:37] * JuddJudd thinks this is turning into an apache/tomcat channel.
[2010/03/09 14:28:47] @ Quit: himanshu: Quit: Leaving
[2010/03/09 14:29:01] <nasrat> more appropriate for ##infra-talk
[2010/03/09 14:29:04] <tim|imac> JuddJudd: fairly common, lots of sysadmins here, all with their own opinions and experiences :) makes the channel so cool
[2010/03/09 14:29:05] <nasrat> war stories, etc
[2010/03/09 14:29:05] <Volcane> nasrat: ah mod_extract_forwarded has some security options of who to trust and such
[2010/03/09 14:29:16] <nasrat> Volcane: as does mod_remoteip
[2010/03/09 14:29:19] <Volcane> kewl
[2010/03/09 14:29:34] <JuddJudd> we trust akamai. :)
[2010/03/09 14:29:36] <Volcane> tim|imac, JuddJudd: should join ##infra-talk though :)
[2010/03/09 14:29:54] <tim|imac> I can hardly keep up with #puppet is it is :S
[2010/03/09 14:30:04] @ Quit: lak: Quit: lak
[2010/03/09 14:30:13] <fzzzt> Speaking of... I have a tomcat app proxied by apache and want to turn on basic auth...can I just do that in tomcat, or so i need to set it up in apache also/instead? :P
[2010/03/09 14:30:15] <nasrat> tim|imac: but just imagine keeping it On topic ;()
[2010/03/09 14:30:30] @ Quit: gebi: Ping timeout: 246 seconds
[2010/03/09 14:30:44] <fzzzt> unfortunately #tomcat is dead :(
[2010/03/09 14:30:59] <JuddJudd> fzzzt try the jbossers?
[2010/03/09 14:31:13] <JuddJudd> but apache basic auth is super easy.
[2010/03/09 14:31:23] <fzzzt> i know
[2010/03/09 14:35:11] @ Quit: giskard: Ping timeout: 245 seconds
[2010/03/09 14:38:37] @ Quit: russell_h: Ping timeout: 260 seconds
[2010/03/09 14:40:14] @ giskard joined channel #puppet
[2010/03/09 14:42:12] @ mikepea_ joined channel #puppet
[2010/03/09 14:42:31] @ Quit: mikepea_: Client Quit
[2010/03/09 14:43:22] @ Quit: jsman: Remote host closed the connection
[2010/03/09 14:47:24] <madduck> alright, this concludes my APT module work for today: http://git.madduck.net/v/puppet/modules/apt.git
[2010/03/09 14:48:43] <tim|imac> madduck: in you init.pp, line 66-69... why?
[2010/03/09 14:49:15] <madduck> well, because I don't have that in my site.pp yet, nor any other place to instantiate
[2010/03/09 14:49:19] @ jes5 joined channel #puppet
[2010/03/09 14:49:42] <madduck> tim|imac: did I misunderstand virtual resources? Isn't this saying "if noone else has caused wget|gnupg to be installed yet, do it now?"
[2010/03/09 14:50:07] <tim|imac> uh kinda
[2010/03/09 14:50:21] <tim|imac> it more like, if you create the virtual resource in one place, you can realize it as often as you like
[2010/03/09 14:50:30] <tim|imac> you still can't create the virtual resource multiple times
[2010/03/09 14:50:40] <madduck> you can't create multiple virtual resources of the same name?
[2010/03/09 14:50:55] <tim|imac> so that's yet another place where I'd do "if ! defined(Package["wget"]) { etc."
[2010/03/09 14:51:06] <tim|imac> not the last time I tried that, no
[2010/03/09 14:51:06] <madduck> sure
[2010/03/09 14:52:04] @ Maliuta joined channel #puppet
[2010/03/09 14:52:04] @ jsman joined channel #puppet
[2010/03/09 14:52:05] @ Quit: jsman: Changing host
[2010/03/09 14:52:05] @ jsman joined channel #puppet
[2010/03/09 14:52:24] <madduck> tim|imac: might well be, I'll fix it one of these days.
[2010/03/09 14:52:27] <madduck> ok. no.
[2010/03/09 14:52:30] <madduck> NOW
[2010/03/09 14:52:46] <tim|imac> hehehe
[2010/03/09 14:53:12] @ cbrinke1 joined channel #puppet
[2010/03/09 14:53:24] <madduck> done
[2010/03/09 14:53:26] <madduck> and now good night
[2010/03/09 14:54:06] <cbrinke1> Hello, I have been reading through the docs/looking for how to do the following: How do you have a "variable" that ends up in a template, that can be overridden in the node/class definition?
[2010/03/09 14:54:46] <cbrinke1> Example: class php { $mem = 16M} node foo { $php::mem = 32M }
[2010/03/09 14:54:47] <tim|imac> cbrinke1: define "overridden"
[2010/03/09 14:54:53] <tim|imac> ah
[2010/03/09 14:54:55] <tim|imac> you cant
[2010/03/09 14:55:06] <tim|imac> you'd do something like:
[2010/03/09 14:55:51] <fzzzt> you could do $mymem = $php::mem, or inherit a node and override it that way
[2010/03/09 14:55:52] <tim|imac> class php { if $php_mem { $mem = $php_mem } else { $mem = 16M } } node foo { $php_mem = 32M include php
[2010/03/09 14:55:53] <tim|imac> }
[2010/03/09 14:56:07] <tim|imac> or rather, that's how I'd deal with it
[2010/03/09 14:57:01] <Volcane> extlookup("php_mem", $php_mem) :)
[2010/03/09 14:57:04] <cbrinke1> so it seems that you have to manually define the "override" procedure? Is that a good summary of how it is done?
[2010/03/09 14:57:31] <fzzzt> Well, generally there is no override procedure
[2010/03/09 14:57:42] <fzzzt> You can override an inherited variable
[2010/03/09 14:57:54] @ plathrop-away is now known as plathrop
[2010/03/09 14:58:58] <cbrinke1> Perhaps there is a better way to skin this cat: I am interested in having n sets of app servers, depending on environment/pool/hw determines the amount of memory set in the php.ini Is the above method, the best way to handle this?
[2010/03/09 14:59:40] <Volcane> cbrinke1: http://www.devco.net/archives/2009/08/31/complex_data_and_puppet.php
[2010/03/09 15:00:03] <fzzzt> I would use directives in the vhost config for that
[2010/03/09 15:00:04] @ kc7zzv joined channel #puppet
[2010/03/09 15:00:37] <fzzzt> php_admin_value
[2010/03/09 15:00:44] <fzzzt> but that's just me :)
[2010/03/09 15:04:37] @ Chiku joined channel #puppet
[2010/03/09 15:04:49] <cbrinke1> Volcane: so it seems that puppet itself was not made for this sort of issue?
[2010/03/09 15:05:19] <Volcane> cbrinke1: well u can use the pattern suggested above, extlookup just makes that easier for u
[2010/03/09 15:05:42] <cbrinke1> (I mean we have an existing m4 system that handles this well, and will dovetail nicely with the extlookup function you have described) but we were hoping puppet would do this more smoothly
[2010/03/09 15:05:53] <Volcane> cbrinke1: and sometime soon puppet will have the same or matching functionality
[2010/03/09 15:06:00] <plathrop> cbrinke1: Yeah, you need to use generate() or extlookup, etc. Or you can switch to an external nodes util
[2010/03/09 15:06:20] @ Quit: rcrowley: Quit: rcrowley
[2010/03/09 15:06:34] <Volcane> cbrinke1: but the big thing about puppet is that its easy to extend either through functions like this or through node tools - it cant be everything to everyone but it can provide copious hooks
[2010/03/09 15:06:49] <cbrinke1> I had considered external nodes however, in a node definition you can't seem to influence the templates/variables.
[2010/03/09 15:06:49] <fzzzt> code monkey like fritos!
[2010/03/09 15:06:52] @ allsystemsarego joined channel #puppet
[2010/03/09 15:07:03] <fzzzt> big warm fuzzy secret heart...code monkey like yooouuuu
[2010/03/09 15:07:19] @ Quit: bgupta: Read error: Connection timed out
[2010/03/09 15:07:25] <Volcane> cbrinke1: in my nodes i can set $nrpe_mailq_warn = 20
[2010/03/09 15:07:36] <fzzzt> (http://www.youtube.com/watch?v=7s8S7QxpjeY)
[2010/03/09 15:07:48] <Volcane> cbrinke1: in my nrpe classes i do if $nrpe_mailq_warn { $mailq_warn = $nrpe_mailq_warn } else {$mailq_warn = 10}
[2010/03/09 15:07:55] <Volcane> cbrinke1: and then u just use $mailq_warn in your template
[2010/03/09 15:08:12] <Volcane> cbrinke1: then u can set it in your node, and have defaults and have it effect eveything
[2010/03/09 15:08:36] <cbrinke1> Volcane: that seems like what I have been wanting.
[2010/03/09 15:08:42] @ bgupta joined channel #puppet
[2010/03/09 15:08:44] <gwar9999> I have a "define" definition that creates a package and exec clause. I want to be able to require either this definition by name or the internal package/exec clauses. Can I do that?
[2010/03/09 15:09:06] <Volcane> cbrinke1: thats what tim|imac said about :)
[2010/03/09 15:09:10] <Volcane> s/about/above
[2010/03/09 15:09:24] @ plathrop is now known as plathrop-away
[2010/03/09 15:10:15] <cbrinke1> will the "inherits" work as expected? or is that variable only defined for that node?
[2010/03/09 15:10:59] <Volcane> in the above case I'd define $nrpe_mailq_warn in the node or the external node tool
[2010/03/09 15:11:07] <Volcane> if undefined, it defaults to 10
[2010/03/09 15:12:15] <cbrinke1> Right, just saying node hwtype1 {$foo_var = 10} ; node serv01 inherits hwtype1 {include foo} ... will serv01's template have a $foo_var = 10?
[2010/03/09 15:14:49] <fzzzt> yes
[2010/03/09 15:15:04] <Volcane> nope
[2010/03/09 15:15:09] <Volcane> errr, sorry
[2010/03/09 15:15:16] <Volcane> i mean you couldnt then set it specifically in the node
[2010/03/09 15:15:45] <fzzzt> you sure? I think I've done that
[2010/03/09 15:15:54] <Volcane> wiki:CommonMisconceptions
[2010/03/09 15:16:18] <gepetto> Volcane: wiki: wiki:CommonMisconceptions is http://reductivelabs.com/trac/puppet/wiki/CommonMisconceptions
[2010/03/09 15:16:45] <Volcane> the fact that u're changing a var thats also in the base node might change something slightly
[2010/03/09 15:16:50] <Volcane> but mostly node inheritance is a bug
[2010/03/09 15:17:40] <Volcane> jamesturnbull: bugs with target versions in the past is unloved, not even loved enough to get the attention during your usual near release bug spamathons :)
[2010/03/09 15:17:42] <fzzzt> That's not quite the same
[2010/03/09 15:17:53] <fzzzt> I believe what he wants to do will work...
[2010/03/09 15:18:13] <Volcane> fzzzt: yeah i think due to it being also in the base node it might change things
[2010/03/09 15:18:16] <fzzzt> In the FAQ it has an included class that gets inherited and shows that you can't override a variable after you inherit it
[2010/03/09 15:18:39] <fzzzt> In his example, he includes the class in the inherited node, after it's inherited and overridden
[2010/03/09 15:18:55] <Volcane> fzzzt: no its not the class that gets inherited
[2010/03/09 15:18:58] <Volcane> fzzzt: in the faq
[2010/03/09 15:19:08] <fzzzt> This is how I did package versions for a while
[2010/03/09 15:19:11] <fzzzt> it works
[2010/03/09 15:19:20] <fzzzt> The class is in the base node, which gets inherited
[2010/03/09 15:19:23] <fzzzt> in the faq:)
[2010/03/09 15:19:39] <jamesturnbull> Volcane: but this isn't in the past - it is set for 0.5 - it didn't make .4 I'll admit but we ran out of time there with Ubuntu
[2010/03/09 15:19:55] <jamesturnbull> Volcane: I try :(
[2010/03/09 15:20:10] <Volcane> jamesturnbull: hehe i know, i think its related to the other exec bugs though
[2010/03/09 15:20:19] <jamesturnbull> Volcane: didn't all that code get reverted?
[2010/03/09 15:20:45] <jamesturnbull> Volcane: we dropped it from .4 because it all k-fucked or do you mean the underlying issue?
[2010/03/09 15:21:04] <jamesturnbull> Volcane: also do you get a report each run with nothign specified in config_version?
[2010/03/09 15:21:19] * jamesturnbull is brb transit to work
[2010/03/09 15:22:08] <Volcane> fzzzt: http://pastie.org/861977 thats what i mean wont work
[2010/03/09 15:22:27] <Volcane> jamesturnbull: i dont really pay that close attention to my reports :)
[2010/03/09 15:23:06] <fzzzt> Volcane: I agree, that's what the FAQ says. What cbrinke1 said is different though
[2010/03/09 15:23:23] <Volcane> fzzzt: i know, and i said what he did will work *but* he couldnt then do node specific variables
[2010/03/09 15:24:00] <fzzzt> I guess I don't know what you mean by node-specific variables
[2010/03/09 15:24:08] <Volcane> jamesturnbull: my store reports dont have that
[2010/03/09 15:24:46] <Volcane> fzzzt: as per my pastie, he couldnt set $foo_var = 20 in a node he could only control it as fine grained as his base nodes
[2010/03/09 15:25:41] <fzzzt> You mean he couldn't create new variables in the inheriting nodes?
[2010/03/09 15:25:57] <Volcane> he said: 20:12 <cbrinke1> Right, just saying node hwtype1 {$foo_var = 10} ; node serv01 inherits hwtype1 {include foo} ... will serv01's template have a $foo_var = 10?
[2010/03/09 15:26:09] <fzzzt> yeah
[2010/03/09 15:26:17] <Volcane> in that example, he could set $foo_var only in hwtype1, if for any reason some node had a special foo_var, he couldnt do it
[2010/03/09 15:26:38] <fzzzt> he could override it
[2010/03/09 15:26:55] <fzzzt> just like your pastie does with $bar
[2010/03/09 15:26:56] <Volcane> pastie a sample
[2010/03/09 15:27:02] <Volcane> but it doesnt work in my pastie
[2010/03/09 15:27:03] <fzzzt> 10 -> 20
[2010/03/09 15:27:19] <fzzzt> because you include foo in the base node, not the inheriting node
[2010/03/09 15:27:55] <fzzzt> it doesn't make much sense to set it in the base if it's not included, but it works that way
[2010/03/09 15:28:17] <Volcane> right, yup, that works, makes for quite heavy node blocks since each box would then get all the includes
[2010/03/09 15:28:30] <fzzzt> yep
[2010/03/09 15:28:46] <himynameisthor> is it possible to get the date in puppet? I'm trying to make a puppet define to ensure a directory exists, whose name is based on the current date. for example: /data/2010/03/09/
[2010/03/09 15:28:49] <Volcane> not particularly flexible either way
[2010/03/09 15:28:57] <fzzzt> but you could have a php node include base_node, and web nodes include that...heh
[2010/03/09 15:28:58] @ gwar9999 left channel #puppet ()
[2010/03/09 15:29:11] <fzzzt> this is exactly what i did when i started out :)
[2010/03/09 15:29:25] * Volcane sticks to extlookup :P
[2010/03/09 15:29:38] <jb> foreman has extlookupish support now :)
[2010/03/09 15:29:51] <tim|imac> heh
[2010/03/09 15:29:59] <Volcane> i know, he's waiting for me and co-conspirator to finish our pluggable extlookup :(
[2010/03/09 15:30:16] <Volcane> so my extlookup would talk to foreman for example, or csv, or ldap or whatever
[2010/03/09 15:30:27] * tim|imac wouldn't want to set such things from a web interface... rather do it from files
[2010/03/09 15:30:41] <Volcane> all the same :)
[2010/03/09 15:30:47] <tim|imac> I'm spartan like that
[2010/03/09 15:30:59] <Volcane> web -> mysql -> file :P
[2010/03/09 15:31:34] @ Quit: swygue: Ping timeout: 265 seconds
[2010/03/09 15:31:35] <tim|imac> 322Tbit/s?!?! damn...
[2010/03/09 15:31:39] <tim|imac> i want one
[2010/03/09 15:31:50] <joe-mac1> that's a lot of prawn
[2010/03/09 15:32:03] <tim|imac> exactly
[2010/03/09 15:32:10] <kjetilho> hardly revolutionising the Internet, though
[2010/03/09 15:33:21] @ Quit: tyler79: Quit: tyler79
[2010/03/09 15:38:56] @ stahnma joined channel #puppet
[2010/03/09 15:40:14] <fzzzt> 4k webcams+pron will revolutionize the internet though
[2010/03/09 15:40:21] <fzzzt> rather, 8k
[2010/03/09 15:40:29] <fzzzt> mmm 8k resolution
[2010/03/09 15:45:21] @ Quit: Bass10: Quit: Leaving
[2010/03/09 15:46:24] @ Quit: giskard: Remote host closed the connection
[2010/03/09 15:46:33] @ miah joined channel #puppet
[2010/03/09 15:46:53] <miah> hi
[2010/03/09 15:46:55] @ giskard joined channel #puppet
[2010/03/09 15:47:44] <jbooth> himynameisthor: probably best done in an inline template
[2010/03/09 15:48:37] @ Quit: jmeeuwen: Ping timeout: 260 seconds
[2010/03/09 15:51:14] @ Quit: unxfrek: Ping timeout: 240 seconds
[2010/03/09 15:51:57] @ Quit: allsystemsarego: Quit: Leaving
[2010/03/09 15:54:33] @ gwar9999 joined channel #puppet
[2010/03/09 15:55:29] @ flakrat joined channel #puppet
[2010/03/09 15:55:31] <gwar9999> Is it possible (and if so, how) to provide a list of items to a definition and then iterate over them by calling another definition w/ each item?
[2010/03/09 15:55:50] @ c4rg joined channel #puppet
[2010/03/09 15:56:56] <jbooth> If you can pull the list into the title/name of the resource sure.
[2010/03/09 15:57:25] @ fzzzt left channel #puppet ()
[2010/03/09 15:57:25] <whack> gwar9999: defingin resouces like foo { ["one", "two"] : ... ; } doesn't invoke 'foo' resource with an array
[2010/03/09 15:57:26] <nasrat> mpdehaan__: have you read the stuff I've posted on facter roadmap?
[2010/03/09 15:57:40] <whack> so the foo define won't even know you gave a list of resources
[2010/03/09 15:57:50] <nasrat> mpdehaan__: I know and have discussed the namespace stuff
[2010/03/09 15:58:19] <gwar9999> whack: so I guess it's not possible then. Bummer.
[2010/03/09 15:58:25] @ plathrop-away is now known as plathrop
[2010/03/09 15:58:40] <whack> gwar9999: well, your 'foo' define could conditionally invoke other defines based on the value of $name
[2010/03/09 15:58:43] <whack> if that's really what you wanted
[2010/03/09 16:00:58] @ Quit: verwilst: Quit: Ex-Chat
[2010/03/09 16:01:52] <gwar9999> whack: well, I want to pass 5 values into the def'n and then for each value call another. Not sure if there is even a way to iterate over things in a recipe.
[2010/03/09 16:02:14] @ jmeeuwen joined channel #puppet
[2010/03/09 16:02:18] <whack> you mean like foo { "bar": value1 => "blah", ... ; } ?
[2010/03/09 16:03:08] <gwar9999> well, the number varies, which is why I wanted a list func { values => ['foo', 'bar', 'baz'] }
[2010/03/09 16:03:13] <whack> oh I see
[2010/03/09 16:03:33] <mpdehaan__> nasrat, I have, I'll review it again shortly.
[2010/03/09 16:03:40] <nasrat> mpdehaan__: cool
[2010/03/09 16:03:51] <c4rg> |Mike|: .
[2010/03/09 16:04:08] @ unxfrek joined channel #puppet
[2010/03/09 16:04:25] <jbooth> What's the difference between that and def func(){...do stuff...} // func { ['foo', ...]: ensure => present; } ?
[2010/03/09 16:04:39] <whack> gwar9999: check this
[2010/03/09 16:05:05] <whack> http://pastebin.com/Gkh6Swvx
[2010/03/09 16:05:12] <whack> I didn't really expect that to work
[2010/03/09 16:06:07] <jbooth> huh, I wouldn't have expected it to either.
[2010/03/09 16:06:56] <nasrat> mpdehaan__: basically, I think we're in violent agreement. But I want to start seeing small patches come in from you before we start on the big stuff
[2010/03/09 16:07:11] <mpdehaan__> jbooth, puppet deals in terms of resources more so than functions, if that's what you meant by 'func'
[2010/03/09 16:07:12] <nasrat> as atm, facter is v. brittle
[2010/03/09 16:07:13] <gwar9999> whack: cool. thanks
[2010/03/09 16:07:21] <mpdehaan__> jbooth, I can explain more about 'ensure' if you like
[2010/03/09 16:07:31] <mpdehaan__> nasrat, good, I like violent agreement :)
[2010/03/09 16:07:48] <jbooth> mpdehaan__: I just pulled gwar's name of his def. I realize func is a really bad choice for a defined type name.
[2010/03/09 16:08:28] <mpdehaan__> you mean 'define' IIRC
[2010/03/09 16:08:47] <jbooth> Yeah, just being shorthanded where I shouldn't
[2010/03/09 16:08:52] <nasrat> mpdehaan__: lets spend the next two weeks getting 1.5.8 out and revisit the *future*
[2010/03/09 16:08:55] <mpdehaan__> ok just checking
[2010/03/09 16:09:08] <jbooth> Actually I think I've spent too much time in the firewall type in ruby recently. Thus def.
[2010/03/09 16:09:15] <mpdehaan__> nasrat, I should have some time for that... kinda multitasking, but yeah
[2010/03/09 16:09:20] <mpdehaan__> more so next
[2010/03/09 16:09:54] <mpdehaan__> jbooth, have you read about defines at http://docs.reductivelabs.com/guides/introduction.html ?
[2010/03/09 16:10:09] <mpdehaan__> http://docs.reductivelabs.com/guides/introduction.html#definitions
[2010/03/09 16:10:58] @ Quit: meister_: Ping timeout: 248 seconds
[2010/03/09 16:11:22] <Volcane> 'defined type' is the correct term though
[2010/03/09 16:12:06] @ meister joined channel #puppet
[2010/03/09 16:12:42] <mpdehaan__> jbooth, sorry, didn't grok the question ... Volcane *nod* but "function" wasn't :)
[2010/03/09 16:12:53] <mpdehaan__> so the question was really about ensure?
[2010/03/09 16:13:49] <jbooth> I think the question was can you $var = ['array', 'here'] then file { $var: ... ; } and apparently the answer is yes.
[2010/03/09 16:14:18] @ mattwynne joined channel #puppet
[2010/03/09 16:20:10] @ Quit: jmccune: Ping timeout: 264 seconds
[2010/03/09 16:20:33] * mpdehaan__ ponders how to best use Puppet's parser for his new /etc/puppet/hostgroups.txt idea
[2010/03/09 16:21:05] <mpdehaan__> ideally it would be in INI-format (by Parser, I mean the config file one)
[2010/03/09 16:23:31] @ Quit: cliff-hm: Ping timeout: 245 seconds
[2010/03/09 16:23:54] @ Quit: ezmobius: Remote host closed the connection
[2010/03/09 16:27:18] <dan__t> Hey Volcane, thanks again
[2010/03/09 16:32:45] @ Quit: nexx: Quit: quit
[2010/03/09 16:34:40] @ Quit: mvn071: Quit: Leaving
[2010/03/09 16:42:37] @ Quit: ahasenack: Ping timeout: 265 seconds
[2010/03/09 16:42:59] @ Quit: bgupta: Quit: bgupta
[2010/03/09 16:49:02] @ Quit: EdRandom: Ping timeout: 264 seconds
[2010/03/09 16:50:18] @ Quit: tyll: Ping timeout: 260 seconds
[2010/03/09 16:50:20] @ rcrowley joined channel #puppet
[2010/03/09 16:59:38] @ Quit: rgsteele: Quit: ChatZilla 0.9.86 [Firefox 3.0.7/2009030423]
[2010/03/09 17:00:40] @ morphic joined channel #puppet
[2010/03/09 17:00:42] @ Quit: paxos: Quit: paxos
[2010/03/09 17:05:14] <cbrinke1> Volcane: I have gotten extlookup function working, I think this will suit my needs for now. Thanks for the pointers, as I certainly would not have found it without your feedback.
[2010/03/09 17:12:36] @ Quit: uphillian: Quit: Leaving.
[2010/03/09 17:16:07] <plathrop> SSL makes my brain bleed
[2010/03/09 17:23:44] <Tonnerre> SSL is simple
[2010/03/09 17:24:36] <jb> no, its really not..
[2010/03/09 17:24:50] <jb> "_
[2010/03/09 17:24:51] <jb> :)
[2010/03/09 17:24:57] <Tonnerre> What's supposed to be so hard about SSL?
[2010/03/09 17:25:03] @ Quit: labrown: Quit: Leaving
[2010/03/09 17:25:22] @ Quit: afletcher: Quit: afletcher
[2010/03/09 17:28:00] <zirpu> Tonnerre: the "Secure" part. it's definately layered like an onion. :-)
[2010/03/09 17:29:58] @ bgupta joined channel #puppet
[2010/03/09 17:30:26] @ Quit: spawnyd: Ping timeout: 260 seconds
[2010/03/09 17:31:08] @ ashleyb joined channel #puppet
[2010/03/09 17:31:40] * jamesturnbull is a security guy so SSL is in his blood
[2010/03/09 17:32:04] <Volcane> oh, that explains it :P
[2010/03/09 17:32:21] <Tonnerre> jamesturnbull, yeah, it's all pretty straightforward ;)
[2010/03/09 17:32:28] <jamesturnbull> Volcane: yeah it certainly explains the weeping sores and the cough
[2010/03/09 17:32:34] <nasrat> jamesturnbull: STARTTLS
[2010/03/09 17:32:52] * Tonnerre shakes nasrats hand
[2010/03/09 17:33:00] @ verwilst joined channel #puppet
[2010/03/09 17:33:11] <jamesturnbull> nasrat: 220 Go Ahead
[2010/03/09 17:33:12] @ Quit: ashleyb: Client Quit
[2010/03/09 17:34:24] * nasrat recalls someone at HAL 2001 mentioning they reseeded blowfish IIRC so protocol negotiation would fail if you weren't using counterpanes SSL
[2010/03/09 17:34:36] <zirpu> 220.1 Go Ahead, this is not the command you wanted.
[2010/03/09 17:35:04] * nasrat wants a protocol with response code Make my Day
[2010/03/09 17:35:15] @ Bass10 joined channel #puppet
[2010/03/09 17:35:19] <Tonnerre> nasrat, you can use any text response you like in SMTP
[2010/03/09 17:35:27] <Tonnerre> nasrat, just like in FTP
[2010/03/09 17:35:59] <jamesturnbull> nasrat: hell you could write a simple protocol and have whatever error messages you want :)
[2010/03/09 17:36:32] <nasrat> like bofh excuse of the day server
[2010/03/09 17:37:05] <Tonnerre> ftp> ls qdbm*
[2010/03/09 17:37:11] @ Quit: Bass10: Max SendQ exceeded
[2010/03/09 17:37:11] <Tonnerre> 150 Have a Gorilla.
[2010/03/09 17:37:18] <Tonnerre> [...]
[2010/03/09 17:37:20] <Tonnerre> 226 There, everyone likes a Gorilla.
[2010/03/09 17:37:25] <Tonnerre> That's valid ftp
[2010/03/09 17:37:38] @ ch left channel #puppet (">")
[2010/03/09 17:37:40] @ cbrinke1 left channel #puppet ()
[2010/03/09 17:37:41] @ Bass10 joined channel #puppet
[2010/03/09 17:37:49] <nasrat> RFCs are a wonderful wonderful thing
[2010/03/09 17:38:24] <nasrat> wibbling, time for bed ...
[2010/03/09 17:44:06] <jbooth> oh-bother. 0.25 client talking to 0.24 server. Snowball's chance in heck?
[2010/03/09 17:47:11] <plathrop> can you use puppet's functions (like "defined") in a template?
[2010/03/09 17:48:47] <jbooth> I don't see why not?
[2010/03/09 17:49:20] <Volcane> jbooth: no can do
[2010/03/09 17:50:16] <jbooth> Volcane: bleh!
[2010/03/09 17:56:29] <cwebber> is there any reason that it would be a bad idea to add a view to the stored config database?
[2010/03/09 17:57:53] @ Quit: blahdeblah: Ping timeout: 260 seconds
[2010/03/09 17:58:33] <mpdehaan__> cwebber, I don't see how a view could hurt anything
[2010/03/09 17:58:45] <cwebber> mpdehaan__: that was my thought
[2010/03/09 17:58:45] <mpdehaan__> i.e. for your own usage
[2010/03/09 17:59:20] <cwebber> mpdehaan__: exactly... I need to be able to query against info that isnt well represented unless i create a view.
[2010/03/09 18:00:30] @ Quit: unxfrek: Remote host closed the connection
[2010/03/09 18:02:45] @ Quit: morphic: Ping timeout: 258 seconds
[2010/03/09 18:02:51] @ Quit: kolla: Remote host closed the connection
[2010/03/09 18:03:03] @ Quit: nasrat: Ping timeout: 256 seconds
[2010/03/09 18:04:50] @ Quit: malraid: Quit: malraid
[2010/03/09 18:08:15] @ Quit: verwilst: Quit: Ex-Chat
[2010/03/09 18:08:25] @ Quit: kubicek: Quit: kubicek
[2010/03/09 18:08:46] @ Quit: p3rror: Quit: أستودعكم الله
[2010/03/09 18:10:26] @ Quit: jamesturnbull: Ping timeout: 248 seconds
[2010/03/09 18:10:35] @ Quit: kc7zzv: Quit: kc7zzv
[2010/03/09 18:11:26] @ Quit: pheezy: Remote host closed the connection
[2010/03/09 18:18:09] @ jamesturnbull joined channel #puppet
[2010/03/09 18:22:48] @ blahdeblah joined channel #puppet
[2010/03/09 18:26:58] @ jab_doa joined channel #puppet
[2010/03/09 18:27:09] @ Quit: kaptk2: Quit: Leaving.
[2010/03/09 18:28:57] @ Quit: rmiller4pi8: Ping timeout: 265 seconds
[2010/03/09 18:44:58] <gwar9999> is there a noop value that I could use as a default for require => ? That is, in a definition I want to pass in require and use it if it's there and have it default to requiring nothing if it's not passed. I'd rather avoid having an if/else conditional where one defines a resource with require and the other specifies it without.
[2010/03/09 18:45:20] @ OpenMedia joined channel #puppet
[2010/03/09 18:45:23] <himynameisthor> does anyone have a recipe for keeping the puppet package up to date? ensure => latest isn't upgrading it, i think it might have to do that the current version is x86_64 and the new version is noarch
[2010/03/09 18:45:38] <himynameisthor> ensure => '0.25.4-1' isn't working either
[2010/03/09 18:48:01] <whack> gwar9999: try 'undef' but otherwise you'll have to conditional the whole resource
[2010/03/09 18:48:25] <gwar9999> whack: thanks, I'll give that a shot.
[2010/03/09 18:50:27] <whack> I had the same problem, but didn't put much time into solving it. I basically did define foo($blah) { if ($blah) { someresource } else { someresource_without_requires } }
[2010/03/09 18:51:49] <Volcane> whack: you can use the setting defaults thing too
[2010/03/09 18:51:51] <gwar9999> whack: yeah, that's what I did initially but thought there could be a better way.
[2010/03/09 18:51:57] @ Quit: cwebber: Quit: cwebber
[2010/03/09 18:52:54] @ diranged joined channel #puppet
[2010/03/09 18:52:55] <Volcane> if $content { File{ content => $content } } else { File { source => $source } } and then your file block without specifying either source or content
[2010/03/09 18:53:07] <diranged> hmm ok.. something broke on my system between puppet 0.25.1 and 0.25.4..
[2010/03/09 18:53:24] <diranged> all kinds of errors like this: Failed to retrieve current state of resource: Could not intern from pson: source '"#<Puppet::FileServi' not in PSON! Could not retrieve file metadata fo
[2010/03/09 18:53:30] <whack> Volcane: Ahh, using defaults. Good call.
[2010/03/09 18:54:34] <gwar9999> Volcane: not sure I follow... is there an example of that defaults logic somewhere?
[2010/03/09 18:54:53] @ shenson is now known as shenson_away
[2010/03/09 18:56:23] <Volcane> gwar9999: look in the language toturial
[2010/03/09 18:56:36] <gwar9999> Volcane: thanks, I'll take a peek.
[2010/03/09 18:56:40] <Volcane> tutorial even
[2010/03/09 18:56:59] <diranged> in .25.4, do i need to specify a file-type or something with my files that get passed down?
[2010/03/09 18:58:11] @ Quit: mpdehaan__: Remote host closed the connection
[2010/03/09 19:01:48] <diranged> oh, thats an ugly error when the puppet server is older than the puppet client
[2010/03/09 19:01:48] <diranged> geeze
[2010/03/09 19:05:20] @ morphic joined channel #puppet
[2010/03/09 19:05:24] <diranged> working once i upgraded the server..
[2010/03/09 19:05:26] @ Quit: diranged: Quit: Leaving
[2010/03/09 19:05:26] @ lilmatt_ joined channel #puppet
[2010/03/09 19:05:46] @ Quit: lilmatt: Ping timeout: 264 seconds
[2010/03/09 19:05:50] @ bodepd joined channel #puppet
[2010/03/09 19:07:12] @ snerd joined channel #puppet
[2010/03/09 19:07:39] @ Quit: notbrien: Ping timeout: 256 seconds
[2010/03/09 19:09:03] <plathrop> Volcane: any chance you are around?
[2010/03/09 19:09:51] <plathrop> Well, anyone who uses external nodes might be able to help me
[2010/03/09 19:10:30] * Volcane doesnt use ext nodes :)
[2010/03/09 19:10:40] <whack> plathrop: what's up?
[2010/03/09 19:11:22] <plathrop> Prior to switching to external nodes, we tested manifests before pushing them to production by doing a noop run on a production server with --environment=testing
[2010/03/09 19:11:36] <plathrop> But it seems with external nodes the client can no longer set the environment?
[2010/03/09 19:11:36] <whack> you can override the node classifier if you're doing testing
[2010/03/09 19:11:43] <Volcane> plathrop: version of puppet?
[2010/03/09 19:11:49] <plathrop> 0.25.4
[2010/03/09 19:11:54] <plathrop> whack: how?
[2010/03/09 19:11:55] <Volcane> its a mess
[2010/03/09 19:12:16] <whack> plathrop: node_terminus=plain
[2010/03/09 19:12:16] <plathrop> oh dear, that doesn't sound good. Care to elaborate, Volcane ?
[2010/03/09 19:12:27] <plathrop> whack: I set that client-side?
[2010/03/09 19:12:41] <Volcane> plathrop: there ws a thread on the list about facts/environments/and where the preferences live
[2010/03/09 19:12:42] <whack> yeah, like: puppetd --test --node_terminus=plain etc
[2010/03/09 19:13:09] <plathrop> Volcane: yeah, haven't been following the list, it got too noisy. *sigh*
[2010/03/09 19:13:18] <plathrop> whack: I'll give that a try...
[2010/03/09 19:13:35] <plathrop> whack: not sure it will help, though, because that doesn't test the real config.
[2010/03/09 19:13:41] <whack> plathrop: true
[2010/03/09 19:13:57] <whack> converting external nodes yaml to a puppet node manifest wouldn't be too hard
[2010/03/09 19:14:06] <whack> though perhasp would defeat the point and wouldn't really help you get true testing
[2010/03/09 19:14:13] <whack> I don't use environments, so I can't help you here.
[2010/03/09 19:14:36] <plathrop> whack: oh? How do you do testing then?
[2010/03/09 19:14:39] <Volcane> plathrop: i set environment in a fact - worked in 24, totally screwed up in 25
[2010/03/09 19:15:02] <Volcane> plathrop: classes come from one environment while files come from the default :(
[2010/03/09 19:15:20] <plathrop> Volcane: switching to external nodes has changed so much behavior it is like learning puppet all over again :-(
[2010/03/09 19:15:44] <Volcane> nods
[2010/03/09 19:15:59] @ cwebber joined channel #puppet
[2010/03/09 19:16:26] <plathrop> This is really going to put a crimp in my work. My entire methodology of testing changes before they go out is broken and I can't see an easy way to fix it.
[2010/03/09 19:19:37] @ captkirk joined channel #puppet
[2010/03/09 19:19:50] @ sebas891 joined channel #puppet
[2010/03/09 19:21:58] @ Quit: snerd: Remote host closed the connection
[2010/03/09 19:23:38] <zirpu> you can't test with just a static checkout of the puppet config hierarchy? and a temp node.conf ?
[2010/03/09 19:24:15] <plathrop> zirpu: node.conf? and no, because we're pulling in a LOT of info from the external node tool.
[2010/03/09 19:25:01] <plathrop> we want to see what actually happens when that data is present. It's all about being able to tell the puppetmaster "use these other manifests than the ones you usually use on this machine"
[2010/03/09 19:26:21] <plathrop> I guess we'll just have to a) stop puppet b) move the node around in our external db c) do the noop run d) move the node back e) start puppet
[2010/03/09 19:26:59] @ lak joined channel #puppet
[2010/03/09 19:27:54] <zirpu> so environment is totally broken w/ external nodes?
[2010/03/09 19:28:24] <plathrop> No, you just have to set it from the opposite direction
[2010/03/09 19:28:30] <plathrop> arguably, that's the way it should be
[2010/03/09 19:30:39] <Volcane> its the subtle changes between versions that gets you
[2010/03/09 19:30:43] <Volcane> like this environment thing
[2010/03/09 19:30:54] <Volcane> and where in the past if $foo behaved like if defined($foo)
[2010/03/09 19:31:04] <Volcane> now it doesnt == broken missbehaving code
[2010/03/09 19:33:47] @ Quit: cwebber: Quit: cwebber
[2010/03/09 19:34:20] <plathrop> Volcane: oh crap. Really? *runs to audit his code, where he is sure he depends on that behavior*
[2010/03/09 19:35:53] <Volcane> plathrop: well, its a bit more subtle, $foo = false \n if ! $foo { notice("its not set") } will print the notice :(
[2010/03/09 19:36:09] <Volcane> leaving foo unset wil have the same result
[2010/03/09 19:36:28] <Volcane> setting it to "false" doesnt print the notice
[2010/03/09 19:38:31] <Volcane> dont have a 24 box now but i think in the past setting it to false would have not printed the notice
[2010/03/09 19:40:51] @ Quit: jaredrhine: Ping timeout: 268 seconds
[2010/03/09 19:42:59] <Volcane> guess though thats a change that came in with proper if structures late 24 though
[2010/03/09 19:44:39] @ Quit: morphic: Ping timeout: 240 seconds
[2010/03/09 19:46:13] @ Quit: jab_doa: Quit: Verlassend
[2010/03/09 19:46:14] <captkirk> Does anyone have experience installing software through puppet using the dpkg provider?
[2010/03/09 19:47:12] @ Quit: lak: Quit: lak
[2010/03/09 19:47:33] <pastamancer> captkirk: why dpkg instead of apt?
[2010/03/09 19:48:44] <captkirk> pastamancer: apt repos are often a few versions out of date, also, it's possible to package software like MATLAB in .debs to deploy in your domain.
[2010/03/09 19:49:47] <pastamancer> captkirk: I have a local repository for custom-build packages
[2010/03/09 19:50:36] <pastamancer> captkirk: take a look at reprepro, it was pretty easy to setup
[2010/03/09 19:52:08] @ morphic joined channel #puppet
[2010/03/09 19:53:02] @ tjoe joined channel #puppet
[2010/03/09 19:53:10] @ tessier joined channel #puppet
[2010/03/09 19:53:42] <captkirk> pastamancer: cool, thanks... don't suppose you also know of a way to pull off packaging on Solaris in a nice way?
[2010/03/09 19:54:00] <pastamancer> captkirk: I've managed to avoid solaris for the most part
[2010/03/09 19:55:29] <captkirk> pastamacer: haha, it's really not that bad for servers, just sucks for workstations, and since the sun package provider is not versionable I'm worried about how tricky pushing packages could get
[2010/03/09 19:57:54] <tessier> Hello all. I am trying to get apache+passenger+puppet working on RHEL5.4 following the instructions at http://reductivelabs.com/trac/puppet/wiki/UsingPassenger
[2010/03/09 19:58:21] <tessier> I don't understand what is supposed to create mod_passenger.so for Apache to link to
[2010/03/09 19:58:21] <tessier> I have a /usr/lib/ruby/gems/1.8/gems/passenger-2.2.11/ext/apache2/mod_passenger.c
[2010/03/09 19:59:07] @ Quit: tonyskapunk: Quit: Leaving
[2010/03/09 20:04:21] <tessier> passenger-install-apache2-module ah-hah
[2010/03/09 20:04:27] <tessier> That wasn't mentioned on the wiki but fortunately google found it
[2010/03/09 20:04:51] @ M- joined channel #puppet
[2010/03/09 20:10:29] @ pheezy joined channel #puppet
[2010/03/09 20:10:55] <JuddJudd> Hey folks. I just converted my code to a module.. and my /var/lib/puppet/module/hsww/manifests/init.pp has an import "classes/*" which only works when init.pp is touched. If I make changes to any of the imported classes and save them, they fail to load until init.pp is touched. Anything I can do?
[2010/03/09 20:15:35] @ jheiss joined channel #puppet
[2010/03/09 20:18:14] @ Edgan joined channel #puppet
[2010/03/09 20:18:57] @ diranged joined channel #puppet
[2010/03/09 20:19:23] <diranged> ok i am using the sysctl modules that someone else wrote.. and on 95% of my centos hosts it works fine. but on 4 hosts ive cloned recently it fails .. and only fails when run automatically. if i run by hand, it works
[2010/03/09 20:19:57] <diranged> fails with this error: http://www.pastie.org/862472
[2010/03/09 20:21:42] <diranged> looking at the log:
[2010/03/09 20:21:44] <diranged> http://www.pastie.org/862474
[2010/03/09 20:22:00] <diranged> so what does "ndefined method `provide' for nil:NilClass" mean? why does it work on most of my hosts.. but not all?
[2010/03/09 20:23:27] <diranged> and for good measure, sysctl.rb:
[2010/03/09 20:23:28] <diranged> http://www.pastie.org/862476
[2010/03/09 20:24:21] <tjoe> hi, I am runnin puppet localy and I have two questions: firs is when I try to get puppet output in a log file like 'puppet -l /tmp/manifest.log manifest.pp' I get a 'Unknown destination type' error
[2010/03/09 20:26:08] <tessier> Yeay. puppet+apache+passenger are all working.
[2010/03/09 20:26:27] <tessier> Hopefully the load on my server will go down now.
[2010/03/09 20:27:11] <tjoe> second is I have a receipe where it should check for a file but that file does not exist and I don't get any error or warning message.
[2010/03/09 20:28:15] @ lilmatt joined channel #puppet
[2010/03/09 20:28:25] @ cliff-hm joined channel #puppet
[2010/03/09 20:29:51] @ lak joined channel #puppet
[2010/03/09 20:30:15] @ cwebber joined channel #puppet
[2010/03/09 20:31:30] @ Quit: tessier: Ping timeout: 260 seconds
[2010/03/09 20:31:33] @ Quit: lilmatt_: Ping timeout: 240 seconds
[2010/03/09 20:31:41] @ Quit: pheezy: Remote host closed the connection
[2010/03/09 20:32:15] @ jaredrhine joined channel #puppet
[2010/03/09 20:40:31] @ swygue joined channel #puppet
[2010/03/09 20:48:46] @ Quit: miah: Quit: meh
[2010/03/09 20:49:11] @ Quit: lilmatt: Read error: Connection reset by peer
[2010/03/09 20:49:15] @ pheezy joined channel #puppet
[2010/03/09 20:49:29] @ lilmatt joined channel #puppet
[2010/03/09 20:52:53] @ notbrien joined channel #puppet
[2010/03/09 20:53:15] @ pipegeek left channel #puppet ("Now, oh, now, I needs must /part")
[2010/03/09 20:54:22] @ Quit: jes5: Quit: Leaving.
[2010/03/09 20:54:49] @ bwood joined channel #puppet
[2010/03/09 20:55:46] <bwood> tidy { "/home/bwood/tmp/backups": age => '0s' } fails to remove anything. Any clues??
[2010/03/09 20:58:27] @ Djelibeybi joined channel #puppet
[2010/03/09 20:59:58] @ Quit: jheiss: Ping timeout: 260 seconds
[2010/03/09 21:01:12] @ LinuxCode joined channel #puppet
[2010/03/09 21:01:17] @ Quit: diranged: Ping timeout: 260 seconds
[2010/03/09 21:06:54] @ Quit: cwebber: Quit: cwebber
[2010/03/09 21:11:31] @ Quit: alexine_dsouza: Read error: Connection reset by peer
[2010/03/09 21:11:34] @ Quit: WALoeIII: Quit: WALoeIII
[2010/03/09 21:11:37] @ Quit: lak: Quit: lak
[2010/03/09 21:13:26] @ rmiller4pi8 joined channel #puppet
[2010/03/09 21:14:13] @ Quit: biertie: Ping timeout: 264 seconds
[2010/03/09 21:15:13] @ biertie joined channel #puppet
[2010/03/09 21:20:21] <ericlee> hi, do I have to set up the hostname on a client first then puppet can manage it? Could I use puppet to set up a hostname?
[2010/03/09 21:21:18] <odyi> Facter uses the host name to match it to a node definition.
[2010/03/09 21:21:35] <odyi> Plus it is needed for proper SSL cert generation.
[2010/03/09 21:21:46] <ericlee> so I shall set up the hostname first.
[2010/03/09 21:22:02] <odyi> Will probably give you less headaches, yes.
[2010/03/09 21:22:51] <odyi> Are you loading boxes with a generic hostname and then changing them?
[2010/03/09 21:22:53] <ericlee> Then I don't think I need /etc/hosts file any more, because puppet will make kit properly according to nodes information.
[2010/03/09 21:23:30] <ericlee> actually the clients just have ip addresses I hope set up hostnames for them.
[2010/03/09 21:24:35] <ericlee> Am I right?
[2010/03/09 21:25:04] <ohadlevy> ericlee: i use my provisioning system to setup the host names, dns records etc
[2010/03/09 21:25:14] @ afletcher joined channel #puppet
[2010/03/09 21:25:54] <ericlee> Okay, so I don't need /etc/hosts file for clients, right?
[2010/03/09 21:26:11] <odyi> How many hosts are you talking about?
[2010/03/09 21:26:18] <ohadlevy> ericlee: that depends if the machine can resolve itself in another way or not
[2010/03/09 21:26:25] <ericlee> 20~30
[2010/03/09 21:26:38] <odyi> o.O set up bind if you can and manage DNS
[2010/03/09 21:26:54] @ \ask_ joined channel #puppet
[2010/03/09 21:26:59] <ericlee> bind and dhcp, right?
[2010/03/09 21:27:17] <odyi> yeah
[2010/03/09 21:28:26] @ Quit: eshamow|away: Quit: leaving
[2010/03/09 21:29:13] <ericlee> What minimal requests before I install puppet on the client?
[2010/03/09 21:30:46] @ Quit: \ask: Ping timeout: 260 seconds
[2010/03/09 21:32:14] @ jes5 joined channel #puppet
[2010/03/09 21:32:51] @ Quit: \ask_: Remote host closed the connection
[2010/03/09 21:36:55] <ericlee> could puppet set up DNS and other networking dtuff?
[2010/03/09 21:36:59] <ericlee> could puppet set up DNS and other networking stuff?
[2010/03/09 21:38:22] <odyi> Yes it could
[2010/03/09 21:39:03] <LinuxCode> correction
[2010/03/09 21:39:07] <LinuxCode> Yes it can and does
[2010/03/09 21:39:33] * LinuxCode handles both in puppet
[2010/03/09 21:39:43] <ericlee> Could you show me some example?
[2010/03/09 21:39:45] <ericlee> In a doc?
[2010/03/09 21:39:55] <LinuxCode> in a doc, no
[2010/03/09 21:40:07] <LinuxCode> ericlee, best advice, just start hacking something
[2010/03/09 21:40:14] <LinuxCode> and if you get stuck come and ask here
[2010/03/09 21:41:25] <ericlee> Okay, well like I wanna set up DNS ip addr for clients, how?
[2010/03/09 21:42:10] <ohadlevy> ericlee: how would you normally do it/
[2010/03/09 21:42:11] <ohadlevy> ?
[2010/03/09 21:42:22] <odyi> You would use it to manage the files and services associated with bind or some other dns implementation
[2010/03/09 21:42:30] <ericlee> Got it.
[2010/03/09 21:42:37] <ericlee> I understand.
[2010/03/09 21:42:51] <ericlee> just sync /etc.resolv.conf
[2010/03/09 21:42:57] <ericlee> just sync /etc/resolv.conf
[2010/03/09 21:43:32] <odyi> uhm...that points a host at a specific DNS server it doesn't manage DNS
[2010/03/09 21:44:47] <LinuxCode> odyi, lets be more specific, a DNS resolver
[2010/03/09 21:44:58] <LinuxCode> ericlee, what exactly do you want to achieve ?
[2010/03/09 21:45:12] <LinuxCode> handling resolvcers is trivial
[2010/03/09 21:49:03] @ cwebber joined channel #puppet
[2010/03/09 21:50:38] @ Quit: notbrien: Quit: notbrien
[2010/03/09 21:55:02] <ericlee> Well, I need to set up resolver, yum repo, users, off/on services.
[2010/03/09 21:55:26] <ohadlevy> ericlee: just start one at a time, you'll get there :)
[2010/03/09 21:55:39] <ericlee> I see.
[2010/03/09 21:55:48] <ericlee> So they are basic usages?
[2010/03/09 21:56:22] <LinuxCode> ericlee, basic and more complex yes
[2010/03/09 21:56:37] <LinuxCode> ericlee, I suggest you just start by readong a how to on the reductive labs site
[2010/03/09 21:56:44] <LinuxCode> or even buy James Puppet book
[2010/03/09 21:56:51] <LinuxCode> I found that quite useful
[2010/03/09 21:57:20] @ Quit: tjoe: Quit: leaving
[2010/03/09 21:58:53] <ericlee> it can also set up NFS mount?
[2010/03/09 22:00:25] <odyi> Sure...
[2010/03/09 22:00:51] @ prekitt_ joined channel #puppet
[2010/03/09 22:01:05] <LinuxCode> ericlee, JUST START CODING
[2010/03/09 22:01:07] <LinuxCode> ;-D
[2010/03/09 22:01:11] <LinuxCode> trust me!
[2010/03/09 22:01:22] <Djelibeybi> LinuxCode: it's not code. :)
[2010/03/09 22:01:27] <ericlee> yep
[2010/03/09 22:01:31] <ericlee> Thanks.
[2010/03/09 22:01:33] <LinuxCode> Djelibeybi, what is it then ?
[2010/03/09 22:01:37] <LinuxCode> gummi bears ?
[2010/03/09 22:01:50] <Djelibeybi> LinuxCode: configuration, perhaps. declaration.
[2010/03/09 22:02:11] <Djelibeybi> An interesting question though. I don't consider writing Puppet manifests as "coding"
[2010/03/09 22:02:20] <LinuxCode> Djelibeybi, you call it whatever you like
[2010/03/09 22:02:28] <ohadlevy> I prefer the gummi bears
[2010/03/09 22:02:34] <LinuxCode> ohadlevy, haha
[2010/03/09 22:02:47] <Djelibeybi> ohadlevy: yes, me too. ;-)
[2010/03/09 22:04:19] @ Quit: elementai: Ping timeout: 276 seconds
[2010/03/09 22:04:20] @ Quit: prekitt: Ping timeout: 276 seconds
[2010/03/09 22:06:11] @ Quit: pheezy: Remote host closed the connection
[2010/03/09 22:06:31] @ pheezy joined channel #puppet
[2010/03/09 22:07:14] @ plathrop is now known as plathrop-away
[2010/03/09 22:09:11] @ Quit: pheezy: Read error: Operation timed out
[2010/03/09 22:11:31] @ elementai joined channel #puppet
[2010/03/09 22:11:38] @ rgsteele joined channel #puppet
[2010/03/09 22:16:58] @ jmeeuwen_ joined channel #puppet
[2010/03/09 22:19:42] @ bodepd left channel #puppet ()
[2010/03/09 22:20:37] @ colyte_ joined channel #puppet
[2010/03/09 22:20:40] @ anakha_ joined channel #puppet
[2010/03/09 22:20:50] @ danny_ joined channel #puppet
[2010/03/09 22:20:55] @ agaffney_ joined channel #puppet
[2010/03/09 22:20:57] @ MrHeavy_ joined channel #puppet
[2010/03/09 22:21:01] @ rodjek_ joined channel #puppet
[2010/03/09 22:21:28] @ Dyresen_ joined channel #puppet
[2010/03/09 22:21:38] @ zeroXten_ joined channel #puppet
[2010/03/09 22:21:45] @ jbarratt_ joined channel #puppet
[2010/03/09 22:21:58] @ Quit: rodjek_: Changing host
[2010/03/09 22:21:59] @ rodjek_ joined channel #puppet
[2010/03/09 22:22:11] @ Quit: cwebber: Quit: cwebber
[2010/03/09 22:23:27] @ LinuxCode_aw joined channel #puppet
[2010/03/09 22:23:32] @ xmslack_ joined channel #puppet
[2010/03/09 22:25:37] @ Quit: LinuxCode: *.net *.split
[2010/03/09 22:25:38] @ Quit: Djelibeybi: *.net *.split
[2010/03/09 22:25:39] @ Quit: jaredrhine: *.net *.split
[2010/03/09 22:25:39] @ Quit: jmeeuwen: *.net *.split
[2010/03/09 22:25:40] @ Quit: Dyresen: *.net *.split
[2010/03/09 22:25:40] @ Quit: danny: *.net *.split
[2010/03/09 22:25:42] @ Quit: MrHeavy: *.net *.split
[2010/03/09 22:25:42] @ Quit: f3ew: *.net *.split
[2010/03/09 22:25:43] @ Quit: tuf: *.net *.split
[2010/03/09 22:25:43] @ Quit: |shad0w|: *.net *.split
[2010/03/09 22:25:44] @ Quit: zeroXten: *.net *.split
[2010/03/09 22:25:44] @ Quit: colyte: *.net *.split
[2010/03/09 22:25:45] @ Quit: jbarratt: *.net *.split
[2010/03/09 22:25:45] @ Quit: xmslack: *.net *.split
[2010/03/09 22:25:46] @ Quit: anakha: *.net *.split
[2010/03/09 22:25:47] @ Quit: dan__t: *.net *.split
[2010/03/09 22:25:47] @ Quit: rodjek: *.net *.split
[2010/03/09 22:25:48] @ Quit: agaffney: *.net *.split
[2010/03/09 22:25:48] @ Quit: Insyte: *.net *.split
[2010/03/09 22:25:49] @ tuf_ is now known as tuf
[2010/03/09 22:26:05] @ jmeeuwen_ is now known as jmeeuwen
[2010/03/09 22:26:40] @ Djelibeybi joined channel #puppet
[2010/03/09 22:28:01] @ dan__t joined channel #puppet
[2010/03/09 22:28:01] @ jaredrhine joined channel #puppet
[2010/03/09 22:28:02] @ danny joined channel #puppet
[2010/03/09 22:28:02] @ f3ew joined channel #puppet
[2010/03/09 22:28:02] @ 16SAAH7BO joined channel #puppet
[2010/03/09 22:28:02] @ |shad0w| joined channel #puppet
[2010/03/09 22:28:02] @ rodjek joined channel #puppet
[2010/03/09 22:28:02] @ jbarratt joined channel #puppet
[2010/03/09 22:28:02] @ xmslack joined channel #puppet
[2010/03/09 22:28:02] @ agaffney joined channel #puppet
[2010/03/09 22:28:02] @ Insyte joined channel #puppet
[2010/03/09 22:28:47] @ Quit: danny: Ping timeout: 277 seconds
[2010/03/09 22:28:48] @ Quit: jbarratt: Ping timeout: 277 seconds
[2010/03/09 22:28:48] @ Quit: xmslack: Ping timeout: 277 seconds
[2010/03/09 22:28:49] @ Quit: rodjek: Ping timeout: 277 seconds
[2010/03/09 22:28:49] @ Quit: agaffney: Ping timeout: 277 seconds
[2010/03/09 22:28:50] @ Quit: Insyte: Ping timeout: 277 seconds
[2010/03/09 22:28:59] @ rodjek_ is now known as rodjek
[2010/03/09 22:29:34] @ Insyte joined channel #puppet
[2010/03/09 22:29:53] @ p3rror joined channel #puppet
[2010/03/09 22:33:42] @ agaffney_ is now known as agaffney
[2010/03/09 22:34:39] @ eshamow joined channel #puppet
[2010/03/09 22:35:06] @ gaveen_ joined channel #puppet
[2010/03/09 22:35:09] @ Quit: gaveen: Read error: No route to host
[2010/03/09 22:39:16] @ cwebber joined channel #puppet
[2010/03/09 22:39:36] @ Quit: jes5: Quit: Leaving.
[2010/03/09 22:39:53] @ eshamow_ joined channel #puppet
[2010/03/09 22:41:33] @ Quit: eshamow: Ping timeout: 268 seconds
[2010/03/09 22:44:04] @ Quit: JuddJudd: Quit: Leaving
[2010/03/09 22:45:40] @ malraid joined channel #puppet
[2010/03/09 22:46:31] @ Quit: rgsteele: Remote host closed the connection
[2010/03/09 22:47:59] @ Quit: cliff-hm: Ping timeout: 245 seconds
[2010/03/09 22:53:34] @ joe-mac joined channel #puppet
[2010/03/09 22:55:13] @ Quit: alfism: Quit: http://opensolaris.com/
[2010/03/09 22:58:54] @ Quit: p3rror: Read error: Operation timed out
[2010/03/09 22:59:20] @ Quit: eshamow_: Quit: leaving
[2010/03/09 22:59:40] @ Quit: bug: Quit: bug
[2010/03/09 23:00:06] @ Quit: malraid: Quit: malraid
[2010/03/09 23:03:36] @ sebas891 left channel #puppet ()
[2010/03/09 23:05:45] @ bug joined channel #puppet
[2010/03/09 23:09:38] @ malraid joined channel #puppet
[2010/03/09 23:09:47] @ Quit: malraid: Client Quit
[2010/03/09 23:14:10] @ p3rror joined channel #puppet
[2010/03/09 23:15:15] @ Quit: cwebber: Quit: cwebber
[2010/03/09 23:17:46] @ cwebber joined channel #puppet
[2010/03/09 23:18:47] @ Quit: pting: Quit: Ex-Chat
[2010/03/09 23:25:14] @ jes5 joined channel #puppet
[2010/03/09 23:33:27] @ Quit: p3rror: Read error: Operation timed out
[2010/03/09 23:40:14] @ bodepd joined channel #puppet
[2010/03/09 23:44:09] @ bodepd_ joined channel #puppet
[2010/03/09 23:44:10] @ Quit: bodepd: Read error: Connection reset by peer
[2010/03/09 23:44:10] @ bodepd_ is now known as bodepd
[2010/03/09 23:44:29] @ nasrat joined channel #puppet
[2010/03/09 23:44:32] @ Quit: alban2: Ping timeout: 246 seconds
[2010/03/09 23:44:38] @ Quit: bwood: Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org
[2010/03/09 23:47:19] @ p3rror joined channel #puppet
[2010/03/09 23:48:37] @ Quit: Djelibeybi: Quit: Leaving
[2010/03/09 23:50:13] @ Quit: LinuxCode_aw: Changing host
[2010/03/09 23:50:13] @ LinuxCode_aw joined channel #puppet
[2010/03/09 23:50:24] @ LinuxCode_aw is now known as LinuxCode
[2010/03/09 23:50:35] @ Quit: p3rror: Read error: Operation timed out
[2010/03/09 23:53:42] @ Quit: lilmatt: Ping timeout: 252 seconds
[2010/03/09 23:56:53] @ Quit: M-: Quit: Leaving

Generated by irclog2html.py 2.8 by Marius Gedminas - find it at mg.pov.lt!