| [2008/07/11 00:06:50] @ Quit: silk: | ||
| [2008/07/11 00:07:12] @ silk joined channel #puppet | ||
| [2008/07/11 00:11:11] @ Quit: silk: Client Quit | ||
| [2008/07/11 00:25:50] @ Quit: jvanzyl: | ||
| [2008/07/11 00:27:32] @ Quit: Maliuta: Remote closed the connection | ||
| [2008/07/11 00:32:03] @ Quit: elijahwright: Read error: 110 (Connection timed out) | ||
| [2008/07/11 00:36:47] @ Quit: catdude: | ||
| [2008/07/11 00:43:46] <randybias> lak: you around? | ||
| [2008/07/11 00:45:36] @ M- joined channel #puppet | ||
| [2008/07/11 01:03:10] @ shadoi joined channel #puppet | ||
| [2008/07/11 01:13:06] @ Quit: btm: kornbluth.freenode.net irc.freenode.net | ||
| [2008/07/11 01:13:08] @ Quit: bartc__: kornbluth.freenode.net irc.freenode.net | ||
| [2008/07/11 01:13:09] @ Quit: Omahn: kornbluth.freenode.net irc.freenode.net | ||
| [2008/07/11 01:13:10] @ Quit: thijso: kornbluth.freenode.net irc.freenode.net | ||
| [2008/07/11 01:14:46] @ Quit: hacim: Remote closed the connection | ||
| [2008/07/11 01:15:52] @ Quit: jason^: Remote closed the connection | ||
| [2008/07/11 01:16:09] @ jason^ joined channel #puppet | ||
| [2008/07/11 01:18:23] @ M-- joined channel #puppet | ||
| [2008/07/11 01:18:23] @ btm joined channel #puppet | ||
| [2008/07/11 01:18:24] @ bartc__ joined channel #puppet | ||
| [2008/07/11 01:18:24] @ thijso joined channel #puppet | ||
| [2008/07/11 01:18:25] @ Omahn joined channel #puppet | ||
| [2008/07/11 01:22:58] @ hacim joined channel #puppet | ||
| [2008/07/11 01:27:05] @ Quit: M-: Connection timed out | ||
| [2008/07/11 01:28:19] @ Maliuta joined channel #puppet | ||
| [2008/07/11 01:34:25] @ andrewcshafer joined channel #puppet | ||
| [2008/07/11 01:35:43] @ Quit: andrewcshafer: Client Quit | ||
| [2008/07/11 01:39:18] @ meandtheshell joined channel #puppet | ||
| [2008/07/11 01:40:47] @ sparanjape joined channel #puppet | ||
| [2008/07/11 01:41:16] @ andrewcshafer joined channel #puppet | ||
| [2008/07/11 01:45:45] @ andrewcshafer_ joined channel #puppet | ||
| [2008/07/11 01:47:14] @ Quit: andrewcshafer: Read error: 104 (Connection reset by peer) | ||
| [2008/07/11 01:48:39] @ Quit: asa_: | ||
| [2008/07/11 01:49:44] @ DavidS1 joined channel #puppet | ||
| [2008/07/11 01:50:11] @ DavidS1 is now known as DavidS | ||
| [2008/07/11 01:50:30] <DavidS> hi all! | ||
| [2008/07/11 01:51:39] <fujin> afternoon david | ||
| [2008/07/11 01:52:40] * DavidS loves TZ differences | ||
| [2008/07/11 01:53:04] <fujin> :D | ||
| [2008/07/11 01:53:19] <fujin> So do I, especially when I get my iPhone the first in the world. | ||
| [2008/07/11 01:53:19] <fujin> :D | ||
| [2008/07/11 01:54:58] <f3ew> lol | ||
| [2008/07/11 01:54:58] <DavidS> brb | ||
| [2008/07/11 01:55:05] @ Quit: DavidS: "Leaving." | ||
| [2008/07/11 01:55:12] * fujin strokes | ||
| [2008/07/11 01:55:49] @ DavidS joined channel #puppet | ||
| [2008/07/11 01:56:05] <DavidS> darned vpn software | ||
| [2008/07/11 01:56:21] <fujin> blast | ||
| [2008/07/11 01:56:32] <fujin> oooh that reminds me, I have to test the new VPN stuff on this 3g iphone | ||
| [2008/07/11 01:56:42] <fujin> apparently it supports cisco ipsec | ||
| [2008/07/11 01:56:48] @ Quit: andrewcshafer_: | ||
| [2008/07/11 01:58:57] @ Quit: randybias: | ||
| [2008/07/11 02:02:20] <DavidS> fujin: http://www.geekculture.com/joyoftech/joyarchives/1116.html | ||
| [2008/07/11 02:04:18] <bgupta> Oh.. Android, when will it be here? The temptation is hard to resist... :( | ||
| [2008/07/11 02:08:08] <shadoi> fujin: damn you. damn you to hell. | ||
| [2008/07/11 02:08:44] <DavidS> *lol* | ||
| [2008/07/11 02:09:17] <bda> Heh. Got iPhone 2.0 on my first-gen. App Store is pretty spiffy. | ||
| [2008/07/11 02:09:23] <bda> The Remote app is freakin' awesome. | ||
| [2008/07/11 02:09:40] <shadoi> bda: IR remote control stuff? | ||
| [2008/07/11 02:09:58] <bda> No, it's an application that runs on your iPhone that lets you control a (paired) iTunes instance over the wifi. | ||
| [2008/07/11 02:10:05] <shadoi> ah right | ||
| [2008/07/11 02:10:12] <bda> Or AppleTV, presumably. | ||
| [2008/07/11 02:10:19] <bda> Very nice. | ||
| [2008/07/11 02:10:22] <shadoi> I remember seeing the hackish one they had for jailbreaks | ||
| [2008/07/11 02:10:25] <bda> nod. | ||
| [2008/07/11 02:10:27] <bgupta> Hey I asked awhile ago. I'll ask again "Wondering if it's sane to monkey patch the gem provider to have --no-ri and --no-rdoc? Or is there a better way to specify that in the manifest?" (Basically generating the ruby docs on a bunch of servers is a big waste of time memory and cpu cycles. Those flags to the gem command tell it not to generate and install documentation.). | ||
| [2008/07/11 02:10:49] <shadoi> bgupta: I'd just wrap the gem binary | ||
| [2008/07/11 02:11:44] @ silent-ion joined channel #puppet | ||
| [2008/07/11 02:12:17] <shadoi> bgupta: or use a .gemrc maybe? | ||
| [2008/07/11 02:12:49] <bgupta> Hmmm.. I like the gemrc better... let me see if that can do it.. | ||
| [2008/07/11 02:13:20] <shadoi> http://docs.rubygems.org/read/chapter/2#page11 | ||
| [2008/07/11 02:16:49] <bgupta> Thanks! I think this seems the easiest thing to do, and is prolly what I will do short term... I'm wondering if adding an option to gem.rb provider is a patch Luke would want.. (IE: Am I the only person that would want this, and if not, would baking the option into the provider be desirable) | ||
| [2008/07/11 02:17:38] <shadoi> seems easy enough with the .gemrc, maybe just a sample recipe on the wiki | ||
| [2008/07/11 02:18:28] <DavidS> bgupta: provider-specific parameters are a long standing open issue, it's just really low priority | ||
| [2008/07/11 02:18:48] <DavidS> (especially if there is a easy local workaround) | ||
| [2008/07/11 02:19:54] <bgupta> Ok.. didn't know that that was the state, I am just starting to look under the hood.. | ||
| [2008/07/11 02:21:08] <bgupta> I'm testing the gemrc method now.. | ||
| [2008/07/11 02:21:30] <bgupta> sweet it worked. | ||
| [2008/07/11 02:21:35] @ lak joined channel #puppet | ||
| [2008/07/11 02:21:52] <DavidS> hi lak! | ||
| [2008/07/11 02:22:10] <lak> hi | ||
| [2008/07/11 02:22:11] <bgupta> more ~/.gemrc | ||
| [2008/07/11 02:22:11] <bgupta> update: -B 10 | ||
| [2008/07/11 02:22:12] <bgupta> install: -B 10 | ||
| [2008/07/11 02:22:12] <bgupta> gem: --no-ri --no-rdoc | ||
| [2008/07/11 02:22:41] <bgupta> gonna go answer my own mailing list email now.. | ||
| [2008/07/11 02:26:15] @ McBofh joined channel #puppet | ||
| [2008/07/11 02:28:36] <DavidS> lak: how was velocity? | ||
| [2008/07/11 02:28:48] <lak> it was good | ||
| [2008/07/11 02:28:50] <lak> but a while ago :) | ||
| [2008/07/11 02:29:00] <lak> i'm now at o'reilly's startupcamp then foocamp | ||
| [2008/07/11 02:29:23] <DavidS> quite a bit of "*camps" around :) | ||
| [2008/07/11 02:29:34] <lak> yeah | ||
| [2008/07/11 02:29:37] @ McBofh left channel #puppet ("Leaving") | ||
| [2008/07/11 02:29:51] <DavidS> aand I haven't seen you in a while and it's the last entry on your blog *hint* | ||
| [2008/07/11 02:35:37] <bgupta> shadoi: Thanks for your help... My name is Brian.. I plan to be hanging out here alot.. (I've got a grand vision of a mashup with EC2, DynDNS, iClassify, puppet, and erb templates... with dynamic monitoring of haproxy connection tables to spin up and down additional mongrel servers.) Not sure where reality meets the vision... Unfortunately it looks like I have to construct alot of the pieces myself as the recipes aren't out there. :( | ||
| [2008/07/11 02:36:21] <lak> DavidS: yeah, i know :/ | ||
| [2008/07/11 02:36:24] <lak> i need to blog more | ||
| [2008/07/11 02:36:34] <lak> i'm flagellating myself about it, but too damn much to do | ||
| [2008/07/11 02:36:43] <DavidS> bgupta: you might want to talk to randybias of cloudscale.net fame | ||
| [2008/07/11 02:36:45] <lak> hey, did you see there's a 0.24.5rc1 out? :) | ||
| [2008/07/11 02:37:32] <DavidS> lak: i didn't get 'round reading the lists for quite a while. (apropos speaking about flagellating and too much work ...) | ||
| [2008/07/11 02:37:37] <tim|imac> lak: twitter works out great... replace your blog with tweets :) | ||
| [2008/07/11 02:37:39] <lak> heh | ||
| [2008/07/11 02:37:46] <lak> i kinda have been | ||
| [2008/07/11 02:39:08] <bgupta> DavidS... What's the best way to reach Randy? I am unfortunately not up on the scene yet. | ||
| [2008/07/11 02:39:22] <DavidS> !seen randybias | ||
| [2008/07/11 02:39:22] <gepetto> DavidS: randybias was last seen 40 minutes and 25 seconds ago, quitting IRC () | ||
| [2008/07/11 02:39:47] <bgupta> Oh right here. doh.. | ||
| [2008/07/11 02:39:49] <bgupta> :) | ||
| [2008/07/11 02:39:49] @ Tempt joined channel #puppet | ||
| [2008/07/11 02:39:56] @ Quit: lak: | ||
| [2008/07/11 02:40:06] <Tempt> bgupta: Hey, someone just pasted me something you said - tell me about your grand vision, again? | ||
| [2008/07/11 02:40:29] @ McBofh joined channel #puppet | ||
| [2008/07/11 02:40:39] <bgupta> Did they paste it to you? | ||
| [2008/07/11 02:41:50] <Tempt> I just wanted to confirm it was true. | ||
| [2008/07/11 02:41:52] <bgupta> Do you need me to repeat it? I am confused... "g out here alot.. (I've got a grand vision of a mashup with EC2, DynDNS, iClassify, puppet, and erb templates... with dynamic monitoring of haproxy connection tables to spin up and down additional mongrel servers.) Not sure where reality meets the vision... Unfortunately it looks like I have to construct alot of the pieces myself as the recipes aren't out there." | ||
| [2008/07/11 02:42:09] <Tempt> Wow, I thought he was pulling my leg. | ||
| [2008/07/11 02:42:33] <bgupta> I dont think I am gonna accomplish it in short order though... | ||
| [2008/07/11 02:42:43] <McBofh> grand visions generally aren't short-term | ||
| [2008/07/11 02:42:47] <McBofh> that's my experience, anyway | ||
| [2008/07/11 02:42:51] <bgupta> lol | ||
| [2008/07/11 02:43:16] <DavidS> McBofh: wouldn't be "grand" if it's done in a day, isn't it? | ||
| [2008/07/11 02:43:28] <McBofh> DavidS: that's about the size of it | ||
| [2008/07/11 02:43:35] <bgupta> I'd say it was grand, if I got it done in a day. ;) | ||
| [2008/07/11 02:43:36] <Tempt> What are you going to call it? | ||
| [2008/07/11 02:44:05] <bgupta> nothing.. It won't be a product, just bits of glue | ||
| [2008/07/11 02:44:27] <Tempt> Aah, like last year's pony. | ||
| [2008/07/11 02:44:35] <bgupta> right | ||
| [2008/07/11 02:44:39] <f3ew> DavidS depends on what you call grand | ||
| [2008/07/11 02:45:06] <f3ew> My idea of grand is to be able to go from bare metal to a full fledged service in a single day | ||
| [2008/07/11 02:45:16] <bgupta> Right now, I have puppet managing my EC2 deployments.. in a completely non automated fashion.. | ||
| [2008/07/11 02:45:17] <f3ew> for any number of hosts | ||
| [2008/07/11 02:45:42] <Tempt> Isn't that called Custom Jumpstart with a few packages on top? | ||
| [2008/07/11 02:45:55] <DavidS> can someone explain "last year's pony" to us (me) non-natives? | ||
| [2008/07/11 02:46:01] <DavidS> please? | ||
| [2008/07/11 02:46:05] <Tempt> last year's pony -> glue | ||
| [2008/07/11 02:46:16] <bgupta> they make glue out of horse parts. | ||
| [2008/07/11 02:46:19] <DavidS> ah | ||
| [2008/07/11 02:47:11] <bgupta> I'm really liking iClassify.. It's so damn simple... I love it.. | ||
| [2008/07/11 02:47:37] <bgupta> Not using it yet just playing with it.. | ||
| [2008/07/11 02:48:49] @ Quit: fbe: Read error: 60 (Operation timed out) | ||
| [2008/07/11 02:52:29] <shadoi> bgupta: that's similar to a part of what we're doing for CloudScale | ||
| [2008/07/11 02:52:41] @ McBofh left channel #puppet ("Leaving") | ||
| [2008/07/11 02:53:02] <shadoi> DavidS: that reminds me, I'm supposed to ping you again about the collectd stuff | ||
| [2008/07/11 02:53:06] <bgupta> Oh neat.. It's a big problem we should get together a team and split up the work.. | ||
| [2008/07/11 02:53:28] <shadoi> bgupta: we're pretty much done with most of it. | ||
| [2008/07/11 02:53:35] <shadoi> I'm not sure yet what parts we'll be opensourcing | ||
| [2008/07/11 02:53:43] <DavidS> shadoi: working on it ... I'm currently doing the threshold provider .. | ||
| [2008/07/11 02:53:50] <shadoi> DavidS: cool | ||
| [2008/07/11 02:53:57] <bgupta> Oh.. cool... I wish I could have used scalr... but I really like haproxy... | ||
| [2008/07/11 02:54:11] <shadoi> Why HA proxy? | ||
| [2008/07/11 02:54:15] <shadoi> it's a pain in the ass to setup | ||
| [2008/07/11 02:54:20] <shadoi> and it doesn't support SSL | ||
| [2008/07/11 02:54:29] <bgupta> No the instructions are just crap.. | ||
| [2008/07/11 02:54:35] <shadoi> I recommend pound | ||
| [2008/07/11 02:54:37] @ Tempt left channel #puppet () | ||
| [2008/07/11 02:54:49] <bgupta> Been running tons of load tests... | ||
| [2008/07/11 02:55:07] <DavidS> shadoi: i'll hope to send you something testable tonight (my tonight, that's in, like, 7-8h) | ||
| [2008/07/11 02:55:21] <shadoi> If you go beyond what pound can do, I highly recommend just getting an F5 BigIP | ||
| [2008/07/11 02:55:29] <shadoi> DavidS: sounds great | ||
| [2008/07/11 02:56:19] @ Quit: M--: "Leaving" | ||
| [2008/07/11 02:56:37] <shadoi> bgupta: software load-balancers are definitely not massively scalable solutions IMO. | ||
| [2008/07/11 02:56:40] <bgupta> There was a reason I didn't want to use pound.. it's been a little while, let me see if I can remember... As for F5 stuff, I wrote the first iRules wiki... That said F5 is lagging in the Ruby world.. | ||
| [2008/07/11 02:57:15] <bgupta> The F5 is really a software loadbalancer.. | ||
| [2008/07/11 02:57:18] @ emerose joined channel #puppet | ||
| [2008/07/11 02:57:22] <bgupta> for most of what people use it for | ||
| [2008/07/11 02:57:23] <shadoi> extremely optimized | ||
| [2008/07/11 02:57:42] <shadoi> compare the throughput specs with any software package | ||
| [2008/07/11 02:57:48] <shadoi> it's pretty dramatic | ||
| [2008/07/11 02:57:49] <bgupta> Yeah... and very powerful.. | ||
| [2008/07/11 02:58:22] <shadoi> I know Joyent uses BigIPs for most of their clients... a lot of which run rails | ||
| [2008/07/11 02:58:26] <shadoi> *shrug* | ||
| [2008/07/11 02:58:45] <bgupta> Um.. haproxy can push 10GB wirespeed.. it's the highest performing opensource software load balancer that I am aware of.. | ||
| [2008/07/11 02:58:59] <shadoi> yeah, it's fast. | ||
| [2008/07/11 02:59:30] <bgupta> It doesn't have the ability to do deep packet inspection like the F5 though, but for our needs, we don't need that. | ||
| [2008/07/11 02:59:41] <DavidS> darn, what h/w do you need to shovel 10GB around? | ||
| [2008/07/11 02:59:45] <shadoi> SSL support is required for a software package IMO, requiring a separate SSL server too costly | ||
| [2008/07/11 02:59:47] <f3ew> a switch | ||
| [2008/07/11 03:00:30] <bgupta> Right now, I am using nginx in front of haproxy for a variety of functions, including ssl offload.. | ||
| [2008/07/11 03:00:30] <shadoi> bgupta: I'd love to see a puppet module for HA Proxy though | ||
| [2008/07/11 03:01:19] <bgupta> We use it to intercept statics from hitting the mongrels, in addition, we plan to use it to serve the rails cache.. | ||
| [2008/07/11 03:01:30] <bgupta> (From memcached) | ||
| [2008/07/11 03:02:05] <bgupta> nginx has a memcached module builtin, in addition Igor will be releasing a full inline caching module for nginx in the next month or so | ||
| [2008/07/11 03:02:25] <shadoi> yeah, nginx is very nice as well. | ||
| [2008/07/11 03:02:47] <shadoi> The config syntax can get extremely hairy, but it's better than HA Proxy from what i've seen. | ||
| [2008/07/11 03:03:28] <bgupta> Well it's not competing against haproxy, it's competing against apache, and it that case, the config syntax wins by miles.. | ||
| [2008/07/11 03:03:59] <shadoi> I wouldn't put nginx in the same class as apache | ||
| [2008/07/11 03:04:00] <bgupta> I don't want to have to clean them to post them puplically, but I can show you my configs, and how elegent they can be.. | ||
| [2008/07/11 03:04:21] <shadoi> maybe mod_proxy_balancer but that's it, apache is like a 40 lb. swiss army knife these days. | ||
| [2008/07/11 03:05:15] @ [newbie] joined channel #puppet | ||
| [2008/07/11 03:05:37] <shadoi> I dunno, I'm probably just biased towards pound because it's very simple, quite fast, and supports SSL. | ||
| [2008/07/11 03:05:38] <DavidS> shadoi: yeah, who needs a scalpel, when he can have a jackhammer? | ||
| [2008/07/11 03:06:04] <bgupta> Well, you give up some features by going with Nginx, but if you don't need those features, you gain on performance and memory footprint.. For 90 of people's needs Nginx will meet them. | ||
| [2008/07/11 03:06:10] <bgupta> 90% I mean | ||
| [2008/07/11 03:06:26] <bgupta> The biggest problem with nginx is that the docs are written in russian | ||
| [2008/07/11 03:06:34] <shadoi> yeah | ||
| [2008/07/11 03:06:56] <shadoi> I like that the authors of pound reject features constantly | ||
| [2008/07/11 03:07:10] <shadoi> The biggest drawback to pound is that it doesn't do graceful restarts | ||
| [2008/07/11 03:07:13] <shadoi> very annoying. | ||
| [2008/07/11 03:07:57] <shadoi> I will probably write an nginx module soonish | ||
| [2008/07/11 03:08:02] <bgupta> Let me see if I can figure out why I shied away from pound.. (This was a few months ago, I made the call to go with haproxy)... | ||
| [2008/07/11 03:08:08] <bgupta> for puppet? | ||
| [2008/07/11 03:08:11] <shadoi> yeah | ||
| [2008/07/11 03:08:16] <bgupta> cool.. | ||
| [2008/07/11 03:08:21] @ emerose_ joined channel #puppet | ||
| [2008/07/11 03:08:42] <shadoi> I'd kind of like to write a simple Proxy native type, but that's probably a little too ambitious for my free time. :) | ||
| [2008/07/11 03:09:04] <bgupta> Right now, I am just pushing out a flat config... and since it's not doing the loadbalncing I can keep it that way for a while.. | ||
| [2008/07/11 03:09:49] @ pleemans joined channel #puppet | ||
| [2008/07/11 03:10:00] <bgupta> brb... | ||
| [2008/07/11 03:10:04] @ pleemans_ joined channel #puppet | ||
| [2008/07/11 03:10:20] @ Quit: pleemans_: Read error: 104 (Connection reset by peer) | ||
| [2008/07/11 03:14:02] <DavidS> shadoi: i had one question with collectd's thresholds, perhaps you can help me there | ||
| [2008/07/11 03:14:13] <shadoi> DavidS: sure | ||
| [2008/07/11 03:14:27] <DavidS> there is this "Instance" statement to restrict the threshold specification | ||
| [2008/07/11 03:14:49] @ tim|macbook joined channel #puppet | ||
| [2008/07/11 03:14:50] <DavidS> is it significant whether it is in the Type or the Plugin level ? | ||
| [2008/07/11 03:14:56] <DavidS> g'morning tim|macbook! | ||
| [2008/07/11 03:15:24] <tim|macbook> hi DavidS :) | ||
| [2008/07/11 03:16:19] <DavidS> shadoi: i'd expect it to be significant only insofar as which thresholds are constrained by it? | ||
| [2008/07/11 03:16:34] <DavidS> but the documentation is very spotty there | ||
| [2008/07/11 03:16:39] <shadoi> hrmm... | ||
| [2008/07/11 03:16:45] <shadoi> I haven't played with instances | ||
| [2008/07/11 03:16:51] @ Quit: emerose: Read error: 110 (Connection timed out) | ||
| [2008/07/11 03:16:57] <shadoi> I'll have to check it out once you've got the basics ready | ||
| [2008/07/11 03:17:22] <shadoi> I wouldn't worry too much about it at first | ||
| [2008/07/11 03:17:28] <DavidS> then I'll just propagate the value to the lowest level and we'll see how it works out ... | ||
| [2008/07/11 03:21:40] @ Quit: tim|macbook: "Leaving" | ||
| [2008/07/11 03:23:52] <bgupta> shadoi: I remember, haproxy was MUCH more CPU efficient than pound... | ||
| [2008/07/11 03:24:22] <shadoi> bgupta: I can imagine that's very true especially when pound is doing SSL. | ||
| [2008/07/11 03:24:54] @ tim|macbook joined channel #puppet | ||
| [2008/07/11 03:30:08] <shadoi> bgupta: I'd be curious what configurations people are having high CPU usage with pound | ||
| [2008/07/11 03:30:17] <bgupta> Here is a question. Can pound limit connections to each backend server to a single connection? | ||
| [2008/07/11 03:30:49] <bgupta> and have a frontend queue of incoming connections that drains to the firsta available backend server? | ||
| [2008/07/11 03:31:12] <bgupta> That was one of the main features that attracted us to haproxy.. | ||
| [2008/07/11 03:33:06] <shadoi> I don't think so, but that's not something you'd generally want to do for a front-facing load-balancer. If you're using nginx or pound in front of HA proxy I can see that being a good thing.... | ||
| [2008/07/11 03:35:29] <bgupta> Yeah that is what we are doing.. nginx in front.. | ||
| [2008/07/11 03:36:03] <bgupta> but assuming for a sec.. that it was a front facing load balancer, why wouldn't you want it to behave like htat. What am i missing? | ||
| [2008/07/11 03:36:20] <shadoi> But like I said, when you start throwing multiple tiers of machines at loadbalancing... you're getting diminishing returns, you should start looking at BigIPs, etc. | ||
| [2008/07/11 03:36:46] <bgupta> Oh no... nginx and haproxy run on the same machine... | ||
| [2008/07/11 03:37:14] <shadoi> bgupta: well, in that scenario you're not "load balancing" you're round-robining connections, and the load-balancer acts like a funnel, which is pretty nasty when the load gets past a certain point. | ||
| [2008/07/11 03:38:06] <shadoi> if you're certain that you'll never hit that point, then it's great I'm sure. | ||
| [2008/07/11 03:39:35] @ aymerick joined channel #puppet | ||
| [2008/07/11 03:40:33] <bgupta> You know I think we are working on different problem sets.. are you dealing with mostly rails apps? I am simply load balancing the rails.. almost all the statics are being served out of a CDN.... so the database is more of a bottleneck for us at this point.. so they have been doing tons of optimization on the backend... (As have the DBAs..) | ||
| [2008/07/11 03:41:17] <shadoi> *nod* | ||
| [2008/07/11 03:41:34] @ DerekW joined channel #puppet | ||
| [2008/07/11 03:41:39] <shadoi> unless you've built a sharded application it's always the DB that's the bottleneck | ||
| [2008/07/11 03:42:01] <shadoi> The last app I used pound for was sharded | ||
| [2008/07/11 03:42:10] <shadoi> So the DB was rarely a factor | ||
| [2008/07/11 03:42:27] @ thegcat joined channel #puppet | ||
| [2008/07/11 03:42:42] <bgupta> They are working on it... but it's about three steps away from being the low hanging fruit.. | ||
| [2008/07/11 03:42:57] <shadoi> there was a recent release for rails sharding | ||
| [2008/07/11 03:43:02] <shadoi> like.. 2 days ago | ||
| [2008/07/11 03:43:48] <bgupta> RIght now... they need to refactor most of the sight so that almost everything is cachable, and those bits that aren't are simply ajax calls. | ||
| [2008/07/11 03:44:33] <shadoi> http://www.railsinside.com/deployment/40-datafabric-rails-sharding-library.html | ||
| [2008/07/11 03:44:46] <bgupta> after that we are gonna move the rails cache into memcached, and serve it out of nginx directly, and after that maybe we can think about sharding.. | ||
| [2008/07/11 03:45:01] <bgupta> Cool.. | ||
| [2008/07/11 03:45:10] <shadoi> anyway, gonna go watch a movie | ||
| [2008/07/11 03:45:15] <shadoi> bgupta: good luck | ||
| [2008/07/11 03:45:19] <bgupta> you too.. | ||
| [2008/07/11 03:45:20] <DavidS> shadoi: have fun! | ||
| [2008/07/11 03:45:35] <shadoi> DavidS: hehe, you too | ||
| [2008/07/11 03:45:36] @ shadoi left channel #puppet () | ||
| [2008/07/11 03:47:08] <bgupta> Another possible option there is mysqlproxy.. Jury is still out though | ||
| [2008/07/11 03:55:23] <bgupta> nite guys.. | ||
| [2008/07/11 03:56:44] <bda> Why might a custom fact work via puppetd --test but not when puppetd is run standalone? | ||
| [2008/07/11 03:56:55] <bda> factsync is enabled. The facts exist on the system. | ||
| [2008/07/11 03:56:58] <tim|macbook> Volcane: i got the permission to switch to mod_proxy_ajp!! yay for me! | ||
| [2008/07/11 03:58:50] <DavidS> bda: something in your environment? test with "env" | ||
| [2008/07/11 03:58:56] <Volcane> tim|macbook: ah kewl, glad that worked out, its so much easier | ||
| [2008/07/11 03:59:08] <bda> DavidS: Oh. Yes. FACTORLIB is set. argh. | ||
| [2008/07/11 03:59:18] <DavidS> bda: glad to be of service :-D | ||
| [2008/07/11 03:59:31] <bda> Good call. :\ | ||
| [2008/07/11 04:02:38] @ fbe joined channel #puppet | ||
| [2008/07/11 04:02:50] @ Quit: jon___: "will you miss me when I'm gone?" | ||
| [2008/07/11 04:03:26] <bda> bah. puppetd says it's loading the facts (logs), but then fails to parse a template with them in it. | ||
| [2008/07/11 04:05:52] <bda> unset FACTERLIB, rerun --test, and it works. Daemon doesn't. argh. | ||
| [2008/07/11 04:14:34] @ jgm left channel #puppet () | ||
| [2008/07/11 04:37:53] @ Llama left channel #puppet ("ÏÐÁÎØËÉ...") | ||
| [2008/07/11 04:42:52] @ gileswork_ joined channel #puppet | ||
| [2008/07/11 04:42:58] @ gileswork_ is now known as gileswork | ||
| [2008/07/11 05:02:18] @ kolla joined channel #puppet | ||
| [2008/07/11 05:12:09] @ silk joined channel #puppet | ||
| [2008/07/11 05:25:42] @ Quit: silk: Remote closed the connection | ||
| [2008/07/11 05:26:04] @ silk joined channel #puppet | ||
| [2008/07/11 06:16:12] <duritong> jamesturnbull: about #1387 | ||
| [2008/07/11 06:16:14] <gepetto> duritong: jamesturnbull: #1387 is http://reductivelabs.com/redmine/issues/show/1387 | ||
| [2008/07/11 06:17:07] <jamesturnbull> duritong: yep | ||
| [2008/07/11 06:18:05] <duritong> i can not get more logs than the one I provided, I can't get a stacktrace or some debugs from puppet --parseonly | ||
| [2008/07/11 06:18:20] <duritong> or you like more to have an example of how to reproduce? | ||
| [2008/07/11 06:19:21] <jamesturnbull> duritong: reall it won't --trace? | ||
| [2008/07/11 06:19:27] <duritong> no | ||
| [2008/07/11 06:19:30] <jamesturnbull> duritong: oh | ||
| [2008/07/11 06:19:33] <duritong> puppet --parseonly --debug --trace manifests/munin.pp | ||
| [2008/07/11 06:19:33] <duritong> err: Could not parse for environment development: Unknown function mysql_password at /home/mh/immerda/git/puppet/modules/mysql/manifests/munin.pp:9 | ||
| [2008/07/11 06:20:10] @ rraphink is now known as raphink | ||
| [2008/07/11 06:22:16] <jamesturnbull> duritong: hmmm I wonder why --trace doesn't work - must fix that too | ||
| [2008/07/11 06:22:24] <jamesturnbull> duritong: okay I'll accept the ticket | ||
| [2008/07/11 06:24:48] <duritong> i provide a way of how to reproduce it | ||
| [2008/07/11 06:27:17] <jamesturnbull> duritong: weird - trace should work | ||
| [2008/07/11 06:33:13] @ keithlard joined channel #puppet | ||
| [2008/07/11 06:33:44] @ Quit: thegcat: | ||
| [2008/07/11 06:39:37] <msf> jamesturnbull: are you guys finally going to announce 1.5.0 ? | ||
| [2008/07/11 06:39:47] <msf> facter that is | ||
| [2008/07/11 06:41:00] <jamesturnbull> msf: we already did | ||
| [2008/07/11 06:41:13] <msf> k | ||
| [2008/07/11 06:41:25] <jamesturnbull> msf: it was on the mailing list | ||
| [2008/07/11 06:41:27] <msf> I just committed 1.5.0 to the openbsd ports tree | ||
| [2008/07/11 06:41:36] <jamesturnbull> msf: and it's in the /download directory and on Redmine | ||
| [2008/07/11 06:41:42] <msf> I haven't had time to really read any mail other than work this week | ||
| [2008/07/11 06:41:54] <msf> no I know you can download it | ||
| [2008/07/11 06:42:33] <msf> the next three weeks we are madly trying to finish developping everything we need to get our new market installed and up and running | ||
| [2008/07/11 06:42:44] <msf> it's kind of hectic | ||
| [2008/07/11 06:42:52] <jamesturnbull> msf: market? | ||
| [2008/07/11 06:43:01] <msf> yeah :-D | ||
| [2008/07/11 06:43:14] <msf> I build stock markets for a living :-P | ||
| [2008/07/11 06:43:25] <jamesturnbull> msf: oh cool - who for? | ||
| [2008/07/11 06:43:29] <msf> sbi japannext | ||
| [2008/07/11 06:43:51] <msf> we're what's reffered to as a PTS in japan | ||
| [2008/07/11 06:44:00] <msf> "proprietary trading system" | ||
| [2008/07/11 06:44:13] <msf> right now we have one market running at night | ||
| [2008/07/11 06:44:18] <msf> from 7pm to 2am | ||
| [2008/07/11 06:44:26] <msf> we're building a new one to service daytime trading | ||
| [2008/07/11 06:44:29] <jamesturnbull> msf: yeah I work in a bank and before that an institutional bank | ||
| [2008/07/11 06:49:14] <DavidS> i love ruby for the flexibility of the language, but I really hate it for the puts+trace debug neccessity | ||
| [2008/07/11 06:50:52] <jamesturnbull> DavidS: agreed | ||
| [2008/07/11 07:06:04] <msf> so yeah... right now we're running close to 10,000 lines of manifest just to install all the software for our market :-P | ||
| [2008/07/11 07:08:46] * Volcane thinks some improved logging is needed here: http://pastie.org/231914 | ||
| [2008/07/11 07:09:01] <Volcane> with "test" not in the path specified it just silently goes on without any warnings or anything | ||
| [2008/07/11 07:09:20] <jamesturnbull> msf: that's pretty cool | ||
| [2008/07/11 07:11:03] <msf> well.. yeah we;ve been working on it for months now | ||
| [2008/07/11 07:15:03] <DavidS> "find /srv/puppet/modules -type f | egrep -v '/.git|(.deb$)' | xargs wc -l" tells me 12k lines of manifests | ||
| [2008/07/11 07:16:11] <Volcane> wont that include cofig files and whatnots? | ||
| [2008/07/11 07:16:27] <msf> you should only find -name '*.pp' | ||
| [2008/07/11 07:16:47] <DavidS> 4.7k | ||
| [2008/07/11 07:17:05] <msf> we have 13K+ lines of pp for the new environment, doesn't count file or templates | ||
| [2008/07/11 07:17:09] <DavidS> I'm glad that IÄm not _that_ crazy ;) | ||
| [2008/07/11 07:17:13] <Volcane> heh | ||
| [2008/07/11 07:17:14] <DavidS> ä=' | ||
| [2008/07/11 07:17:27] <msf> and there's still about 30 modules we haven't ported over from the old source tree | ||
| [2008/07/11 07:17:33] <msf> that we will be doing later | ||
| [2008/07/11 07:19:52] * Volcane 's counts are massively inflated with comments though | ||
| [2008/07/11 07:20:53] <DavidS> counting's overrated anyways as the best stuff will reduce LOCs anyways, e.g. "define" | ||
| [2008/07/11 07:21:07] <Volcane> yeah | ||
| [2008/07/11 07:21:26] <Volcane> classic managers view on life | ||
| [2008/07/11 07:21:35] <Volcane> coder a is so much better cos he has 500 lines code more | ||
| [2008/07/11 07:21:42] <tim|macbook> 7940 total and still growing, here... not counting all the templates and stuff | ||
| [2008/07/11 07:23:28] <DavidS> 388 lines containing 'class|define' | ||
| [2008/07/11 07:24:18] <msf> honestly.. most of our stuff is working around the fact that the market software we use follows murphy's law of enterprise software | ||
| [2008/07/11 07:24:42] <DavidS> that's approximately 12 lines per class or define :) | ||
| [2008/07/11 07:24:44] <msf> ie. the quality of a software release is inversely proportional to the amount of money we paid for it | ||
| [2008/07/11 07:25:03] <DavidS> ouch | ||
| [2008/07/11 07:25:10] <msf> we paid ungodly amounts of money for this software | ||
| [2008/07/11 07:25:29] <msf> so it stands to reason that the vendor is incapable of providing us with releases | ||
| [2008/07/11 07:25:40] <msf> for example... when they send us bugfixes | ||
| [2008/07/11 07:25:52] <msf> very often they appear ui the form of single files at a time | ||
| [2008/07/11 07:25:59] <msf> that we must install manually | ||
| [2008/07/11 07:26:05] <msf> usually with typos in them | ||
| [2008/07/11 07:26:13] <msf> and no instructions :-P | ||
| [2008/07/11 07:26:20] <DavidS> wtf? | ||
| [2008/07/11 07:26:33] <msf> yeah it's depressing | ||
| [2008/07/11 07:26:39] <msf> anyways... it's miller time | ||
| [2008/07/11 07:26:45] <msf> see y'all later | ||
| [2008/07/11 07:26:48] <DavidS> have fun! | ||
| [2008/07/11 07:27:38] @ Quit: kenvandine: "Ex-Chat" | ||
| [2008/07/11 07:35:21] @ [newbie] is now known as f--z | ||
| [2008/07/11 07:41:54] @ Quit: DavidS: Read error: 104 (Connection reset by peer) | ||
| [2008/07/11 07:41:57] @ DavidS joined channel #puppet | ||
| [2008/07/11 07:48:37] <Volcane> what do ppl use for cacti packages in debian? | ||
| [2008/07/11 07:48:46] <Volcane> eth has an ancient and vulnerability ridden version | ||
| [2008/07/11 07:48:58] <Volcane> unstable is a bothed up broken distro without the cli bits | ||
| [2008/07/11 07:51:11] @ DavidS1 joined channel #puppet | ||
| [2008/07/11 07:51:11] <tim|macbook> we don't use cacti but munin | ||
| [2008/07/11 07:51:29] @ Quit: DavidS: Nick collision from services. | ||
| [2008/07/11 07:51:35] @ DavidS1 is now known as DavidS | ||
| [2008/07/11 07:51:52] <Volcane> munins so ugly though :( | ||
| [2008/07/11 07:52:23] <tim|macbook> cacti is way too complicated for what we want to achieve :) munin does the job and if you really think it's ugly, you can change the css ;-) | ||
| [2008/07/11 07:52:24] <Volcane> and while its nice to make aggregated graphs easily, munin on debian at least has some issues with that, maybe its just another classic debian out of date package | ||
| [2008/07/11 07:52:38] <tim|macbook> we use munin from debian, no troubles there | ||
| [2008/07/11 07:52:38] <Volcane> tim|macbook: not the html, the graphs | ||
| [2008/07/11 07:52:59] <tim|macbook> the graphs are readable and do the trick... that's all i need :) | ||
| [2008/07/11 07:53:14] <tim|macbook> although you could change some rrdtool settings if you really want to | ||
| [2008/07/11 07:53:45] <Volcane> tim|macbook: i made this aggregate graph: http://nephilim.ml.org/~rip/Totals-web_bandwidth-day.png | ||
| [2008/07/11 07:54:06] <Volcane> tim|macbook: so the screenshot was made a bit too late, but the totals in cur never goes over 0 | ||
| [2008/07/11 07:54:13] <Volcane> even when the graphs does | ||
| [2008/07/11 07:54:38] <tim|macbook> that sounds like a bug indeed... our graphs are behind 401s, I'm afraid | ||
| [2008/07/11 07:54:48] <tim|macbook> i'll check on some aggregated graphs | ||
| [2008/07/11 07:55:30] <Volcane> yeah so going back to cacti, only to discover that once again debian is the achilles heel of everything | ||
| [2008/07/11 07:56:44] <tim|macbook> well, go fix it and submit bugreports :) that's what I'm doing with the samba packages :) | ||
| [2008/07/11 07:57:06] <Volcane> if i had to submit bug reports for everyting i find wrong with debian I'll never get any work done | ||
| [2008/07/11 07:57:30] <tim|macbook> then switch to a better distro :P | ||
| [2008/07/11 07:57:38] * tim|macbook has no troubles at all with debian | ||
| [2008/07/11 07:57:40] <Volcane> wasnt my decision :( | ||
| [2008/07/11 07:57:41] <tim|macbook> everything works for us | ||
| [2008/07/11 07:57:50] <thom> us likewise | ||
| [2008/07/11 07:58:05] <DavidS> debian is like democracy: it sucks, but all others in the field are even worse | ||
| [2008/07/11 07:58:07] <tim|macbook> and if it doesn't, we either build our own custom packages (debian.kumina.nl) or use backports or testing | ||
| [2008/07/11 07:58:10] <Volcane> tim|macbook: i think your expectations level has gone down from too much debian use so it just seems like alls well :P | ||
| [2008/07/11 07:58:50] <tim|macbook> Volcane: actually, my expectations went up... had to install Wordpress on a Suse machine for a friend a few weeks ago... i was baffled by the fact that suse didn't have support for easy installation of multiple blogs using the same codebase | ||
| [2008/07/11 07:59:58] <Volcane> tim|macbook: nod, a lot of the customisations debian make to packages are well intended but often the quality of those customisations are poor | ||
| [2008/07/11 08:00:58] <Volcane> tim|macbook: so probably that ability on debian is via debian home grown mods to how things work | ||
| [2008/07/11 08:01:03] <Volcane> i dont know, dont use wordpress | ||
| [2008/07/11 08:01:07] <tim|macbook> maybe, but it usually works, as long as you don't try to bend the rules too much... if you want something completely different, you always have the choice not to use the customizations | ||
| [2008/07/11 08:01:25] <Volcane> how can you not use the customisations? | ||
| [2008/07/11 08:01:37] @ andrewcshafer joined channel #puppet | ||
| [2008/07/11 08:01:48] <Volcane> taking cacti for example, they modified the actual cacti php source to inlude some other php that deb-conf generates to do configs | ||
| [2008/07/11 08:01:49] <DavidS> Volcane: install from upstream, like everywhere else | ||
| [2008/07/11 08:02:17] <tim|macbook> that's always an option | ||
| [2008/07/11 08:02:48] <Volcane> DavidS: its *horrid* that you need to resort to installing by source onto a package based system just cos the maintainers have munged it beyond recognition | ||
| [2008/07/11 08:02:57] <tim|macbook> the only problem we encounter is when we either need a newer version for some dependency or if there is no package for debian at all | ||
| [2008/07/11 08:03:47] <tim|macbook> i don't think i ever encountered that kind of customization... only stuff i know doesn't go beyond enhanced init scripts or custom creation scripts (for wordpress, eg.) | ||
| [2008/07/11 08:04:32] <tim|macbook> and another thing we encounter is that sometimes the dependencies are a bit too much... like a forced install of X when you are installing a curses app or something | ||
| [2008/07/11 08:05:10] <Volcane> tim|macbook: i get hte impression when debian ppl package something they find great joy in customizing it and making it debian specific so they go overboard and do things that are completely unneeded | ||
| [2008/07/11 08:05:20] <Volcane> example, sendmail is totally capable of figuring out its hostname | ||
| [2008/07/11 08:05:22] <Volcane> no problems at all | ||
| [2008/07/11 08:05:34] <Volcane> debian sendmail package however, tries to hardocde the hostname into the cf | ||
| [2008/07/11 08:05:46] <Volcane> eventhough the comments say: # ... define this only if sendmail cannot automatically determine your domain | ||
| [2008/07/11 08:06:08] <tim|macbook> honestly, i'd want that too... since we seldom use the hostname from a domU as the mailname... | ||
| [2008/07/11 08:06:11] <Volcane> so they do dnslookups during deb-conf and then fill that in there | ||
| [2008/07/11 08:06:30] <DavidS> Volcane: from exim4 i know that they (only) try to hardcode the hostname in "dialup" mode to suppress the DNS request | ||
| [2008/07/11 08:06:30] <Volcane> but the quality of the lookup script is terrible | ||
| [2008/07/11 08:06:33] <tim|macbook> but we don't use sendmail :) | ||
| [2008/07/11 08:06:46] <tim|macbook> we use postfix, mainly | ||
| [2008/07/11 08:07:16] <Volcane> so that if for whatever reason during install time your nameservers isnt there, or isnt setup correctly you end up iwth a mailname of ';; connection timed out; no servers could be reached' | ||
| [2008/07/11 08:07:49] <Volcane> so you have to agree, in this case that customisation was completely unneeded and done really poorly | ||
| [2008/07/11 08:08:15] <Volcane> and it increases the complexity of the system, cos now *everyone* who wants to change their hostname has to go grepping in etc and everywhere to find all mentions of old hostname | ||
| [2008/07/11 08:08:40] <Volcane> where if they left it to default, the ones who wish to have a different mailname from hostname can do so and know its been done | ||
| [2008/07/11 08:09:13] <tim|macbook> yeah probably... but i can't believe they have no good reason for doing it anyway... | ||
| [2008/07/11 08:09:22] <DavidS> Volcane: only if *everyone* is restricted to people with gratis DNS lookups | ||
| [2008/07/11 08:10:37] <Volcane> DavidS: good point, it also though introduces uncertainty into the quality of your builds, you just cannot rely that 2 x debian machines installed look the same even if you do it the same, behind the scenes deb-conf can be very fragile i found | ||
| [2008/07/11 08:11:11] <Volcane> DavidS: thats prolly my biggest gripe of all with the debian way, the uncertainty and unpredictability - cos like witht he sendmail example, half my machines had that error in hostnames | ||
| [2008/07/11 08:11:54] <tim|macbook> Volcane: that's why we use preseeding... at least you can be sure that the debconf values are correct from the start :) | ||
| [2008/07/11 08:12:09] <tim|macbook> preseeding + puppet = tha bomb | ||
| [2008/07/11 08:12:23] <Volcane> only if the package is good quality | ||
| [2008/07/11 08:12:33] <Volcane> cacti package, it asks if you you want to db install | ||
| [2008/07/11 08:12:46] <Volcane> you say no, it then asks for your root password on your mysql and bombs if it cant connect | ||
| [2008/07/11 08:13:02] <Volcane> so only way round is to preseed it to ignore the error, it should just not try when you tell it not too! | ||
| [2008/07/11 08:13:14] <Volcane> even when it *asks* if it shouldnt | ||
| [2008/07/11 08:13:33] <DavidS> I got the impression that using puppet makes most of debian's debconf unneccessary, also becaus i couldn't change values in config files with preseeding | ||
| [2008/07/11 08:13:41] <Volcane> it just strikes me as if all these debconf code is of lower quality than upstreams code and have lower quality testing associated | ||
| [2008/07/11 08:13:47] <DavidS> therefore I generally ignore what debian creates and push my own configs | ||
| [2008/07/11 08:14:02] <DavidS> though I tend to use debian's config as template | ||
| [2008/07/11 08:14:10] <Volcane> DavidS: yeah, then when you try to uninstall a package it winges that it wasnt deb-conf'd properly so do you want to do it then, sigh! | ||
| [2008/07/11 08:14:43] <tim|macbook> DavidS: that too, but preseeding works nicely for a new install... we have fully automated installation of dom0 and domU (xen) and puppet only kicks in for host-specific alterations (like, what does the server actually *do*) | ||
| [2008/07/11 08:16:03] <DavidS> tim|macbook: so you re-install the dom0 if something in your preseeding files changes? | ||
| [2008/07/11 08:16:52] <tim|macbook> no, changes are made from puppet, but preseeding is nice for new installations | ||
| [2008/07/11 08:17:22] <tim|macbook> we tell the hardware guys to just plug the server in and pxe boot and 20 minutes later we have a working machine with users for us, of which the ssh keys are already set :) | ||
| [2008/07/11 08:17:36] <Volcane> preseeding is conceptually against what you're trying to achieve with puppet imho | ||
| [2008/07/11 08:18:18] <Volcane> which is why it really grates me that just to get something to kind of work on debian you need to preceed stuff | ||
| [2008/07/11 08:18:22] @ Quit: andrewcshafer: Read error: 110 (Connection timed out) | ||
| [2008/07/11 08:18:23] <tim|macbook> no, it just changes the defaults, so the values are correct from the start... puppet changes them if needed or checks if they're still the correct values | ||
| [2008/07/11 08:20:14] <Volcane> from my experience the answers you give during deb-conf can significantly alter the final result on your drive, even pretty major changes | ||
| [2008/07/11 08:20:20] <Volcane> so if you preseed, then want to change your mind | ||
| [2008/07/11 08:20:29] <Volcane> you'll have pretty hairy puppet configs to rectify that | ||
| [2008/07/11 08:21:13] <f3ew> tim|imac can you share the preseeding bits? | ||
| [2008/07/11 08:21:24] * f3ew is going to have to write those otherwise :| | ||
| [2008/07/11 08:21:40] <tim|macbook> the automated install or the actual preseeding-to-debconf? | ||
| [2008/07/11 08:21:48] <f3ew> automated install | ||
| [2008/07/11 08:22:17] <f3ew> We have some crappily written internal code to do it, which is going to need a rewrite anyway | ||
| [2008/07/11 08:22:26] <tim|macbook> it's actually a modified version from hands.com/d-i... all the customization are for our setup, i guess | ||
| [2008/07/11 08:22:44] <f3ew> Mostly because the guy doing xen had no clue of how things needed to be setup | ||
| [2008/07/11 08:22:56] <tim|macbook> ah ok | ||
| [2008/07/11 08:23:09] <tim|macbook> just a sec, checking if there's any setup-specific (or secret) stuff in there | ||
| [2008/07/11 08:24:26] <f3ew> (setup currently involves a dd of a LV from one place to another, followed by a bunch of sed scripts) | ||
| [2008/07/11 08:25:06] <DavidS> f3ew: I always wondered whether that isn'T better than debootstrapping .. | ||
| [2008/07/11 08:25:16] <DavidS> minus sed plus puppet of course | ||
| [2008/07/11 08:25:34] <f3ew> DavidS, I prefer kickstart | ||
| [2008/07/11 08:29:40] <tim|macbook> f3ew: http://www.kumina.nl/preseed/example-preseed.iso <-- iso with all our non-specific installation stuff... you can easily copy it over for tftpboot/pxe installation... be warned, booting from the iso will totally screw up your computer, bank account and love life, so be sure to check the code before running it | ||
| [2008/07/11 08:29:53] <f3ew> heh | ||
| [2008/07/11 08:29:56] <f3ew> ty | ||
| [2008/07/11 08:30:05] <tim|macbook> it wasn't really created for general distribution (we use that CD when we need to install a machine without pxe capabilities) | ||
| [2008/07/11 08:30:11] @ Quit: pjw: | ||
| [2008/07/11 08:30:44] * f3ew is just going to have to figure out this stuff | ||
| [2008/07/11 08:31:34] <tim|macbook> ow nvm, the iso is for netinstall :P | ||
| [2008/07/11 08:31:39] <tim|macbook> i'll tar the rest for you | ||
| [2008/07/11 08:33:15] <f3ew> ty | ||
| [2008/07/11 08:36:24] @ kenvandine joined channel #puppet | ||
| [2008/07/11 08:37:44] @ mdray joined channel #puppet | ||
| [2008/07/11 08:59:43] @ elshaa joined channel #puppet | ||
| [2008/07/11 08:59:44] <elshaa> hi | ||
| [2008/07/11 09:00:24] <f3ew> lo | ||
| [2008/07/11 09:00:30] <elshaa> According to the doc, the if statements seems limited, but though, is it a way to negate a variable ? | ||
| [2008/07/11 09:00:53] <elshaa> In a .erb , I would like to delete next line if a variable is undef. | ||
| [2008/07/11 09:01:36] <elshaa> a kind of <% !variable -%> ... | ||
| [2008/07/11 09:01:52] <fsweetser> elshaa: erb files don't use puppet manifest syntax, they use full ruby syntax | ||
| [2008/07/11 09:01:55] <elshaa> <% if !variable -%> | ||
| [2008/07/11 09:02:07] <elshaa> fsweetser: woooo. nice ! | ||
| [2008/07/11 09:02:15] <fujin> you probably want unless variable.nil? | ||
| [2008/07/11 09:02:27] <fujin> Do vars evaluate to nil in the bind passed to ERB? | ||
| [2008/07/11 09:02:30] * fujin forgets | ||
| [2008/07/11 09:02:36] <Volcane> no i doubt it | ||
| [2008/07/11 09:02:40] <Volcane> there were some chatter here about that | ||
| [2008/07/11 09:02:45] <elshaa> fujin: I don't know anything about ruby, but I'll check it | ||
| [2008/07/11 09:02:59] <Volcane> and i thihnk you made a ticket and a fix in a ticket to add a function :P | ||
| [2008/07/11 09:03:00] <fsweetser> IIRC, the vars are injected into erb via method_missing | ||
| [2008/07/11 09:03:03] <mdray> .empty? or is that rails-only? | ||
| [2008/07/11 09:03:10] <fsweetser> so I don't think that the nil? method worked | ||
| [2008/07/11 09:03:33] <DavidS> check for == "" | ||
| [2008/07/11 09:03:48] <Volcane> that wont catch undefineds | ||
| [2008/07/11 09:04:13] <Volcane> since next version you can do has_variable?(foo) | ||
| [2008/07/11 09:05:02] <Volcane> holoway was saing something like: @scope.lookupvar("variable"), false) will work, but i couldnt get that going | ||
| [2008/07/11 09:05:25] <elshaa> well, maybe I can deal with an empty variable, so == "" atm. | ||
| [2008/07/11 09:06:38] <Volcane> yeah i just do case statements and set it to known values when i dont want things to happen and check for that in if's | ||
| [2008/07/11 09:16:20] @ Quit: sparanjape: "Easy as 3.14159265358979323846..." | ||
| [2008/07/11 09:20:06] <mdray> when searching $modulepath does puppet pick the first or last class is finds? if i have modulepath=/foo:/bar and a myclass in each, which is used when i "import myclass" | ||
| [2008/07/11 09:21:50] <Volcane> first | ||
| [2008/07/11 09:22:17] <Volcane> i use that to keep bulk of my modules in common/modules and then be able to override it per environment with environment/modules | ||
| [2008/07/11 09:22:31] <mdray> good good | ||
| [2008/07/11 09:22:45] <mdray> (snap! (not surprise as i copied you)) | ||
| [2008/07/11 09:23:10] <Volcane> :) | ||
| [2008/07/11 09:23:29] @ Quit: meandtheshell: Read error: 110 (Connection timed out) | ||
| [2008/07/11 09:27:46] <elshaa> And in a .pp file, can I use if statement negation ? | ||
| [2008/07/11 09:27:55] <fujin> no | ||
| [2008/07/11 09:27:58] <fujin> well, kind of. | ||
| [2008/07/11 09:28:02] <fujin> What are you trying to do? | ||
| [2008/07/11 09:28:10] <fujin> if the value is empty? unset? | ||
| [2008/07/11 09:29:40] <elshaa> unset again :) | ||
| [2008/07/11 09:30:12] <elshaa> basically, since I can't check if a variable is undef in the .erb, I was thinking to have something like that in the .pp : | ||
| [2008/07/11 09:30:18] <fujin> use a case | ||
| [2008/07/11 09:30:34] <elshaa> if $var { | ||
| [2008/07/11 09:30:35] <elshaa> } | ||
| [2008/07/11 09:30:40] <elshaa> else { | ||
| [2008/07/11 09:30:47] <elshaa> my stuff | ||
| [2008/07/11 09:30:48] <elshaa> } | ||
| [2008/07/11 09:30:50] <elshaa> but this is hugly | ||
| [2008/07/11 09:31:06] <elshaa> fujin: ok | ||
| [2008/07/11 09:31:37] <fujin> seen them on the wiki:LanguageTutorial ? | ||
| [2008/07/11 09:31:39] <gepetto> fujin: wiki:LanguageTutorial is http://reductivelabs.com/trac/puppet/wiki/LanguageTutorial | ||
| [2008/07/11 09:31:47] @ Quit: silk: | ||
| [2008/07/11 09:31:59] <fujin> something like http://pastie.org/231969 | ||
| [2008/07/11 09:32:04] <fujin> I use it as a semi bootstrap | ||
| [2008/07/11 09:32:07] <elshaa> fujin: I'm on it, at the case section | ||
| [2008/07/11 09:33:24] @ andrewcshafer joined channel #puppet | ||
| [2008/07/11 09:33:39] <elshaa> fujin: I don't understand how the case can help to know if my variable is unset. | ||
| [2008/07/11 09:33:46] <elshaa> if it's unset, it won't even pass in the case | ||
| [2008/07/11 09:33:49] <elshaa> (no ?) | ||
| [2008/07/11 09:34:52] @ Quit: andrewcshafer: Client Quit | ||
| [2008/07/11 09:35:21] <elshaa> I think it | ||
| [2008/07/11 09:35:28] <elshaa> 's simpler to go with empty variables | ||
| [2008/07/11 09:42:44] <fujin> elshaa: no, | ||
| [2008/07/11 09:42:50] <fujin> if it's unset, it'll evaluate to "". | ||
| [2008/07/11 09:43:02] <fujin> wait | ||
| [2008/07/11 09:43:08] <elshaa> hum ok. You mean in the .pp | ||
| [2008/07/11 09:43:30] <fujin> yeah. | ||
| [2008/07/11 09:43:33] <elshaa> k | ||
| [2008/07/11 09:43:36] <fujin> If the variable is unset, it'll evaluate in a case to "" | ||
| [2008/07/11 09:44:42] <elshaa> hum, ok in a case. | ||
| [2008/07/11 09:44:44] <elshaa> Thx | ||
| [2008/07/11 09:48:30] @ kambiz joined channel #puppet | ||
| [2008/07/11 09:48:42] @ pjw joined channel #puppet | ||
| [2008/07/11 09:50:58] @ jvanzyl joined channel #puppet | ||
| [2008/07/11 10:01:24] @ yure joined channel #puppet | ||
| [2008/07/11 10:03:06] @ jbrothers joined channel #puppet | ||
| [2008/07/11 10:06:37] <mdray> any ideas on this: http://pastie.org/231987 | ||
| [2008/07/11 10:06:53] @ ski98033 joined channel #puppet | ||
| [2008/07/11 10:07:11] <mdray> i can't figure out how to ensure virtual users are realized in time | ||
| [2008/07/11 10:09:22] <tim|macbook> how about require => User["admin"] ? | ||
| [2008/07/11 10:09:34] <tim|macbook> also, does "class user" work? doesn't it conflict with the type? | ||
| [2008/07/11 10:09:41] <tim|macbook> not that it should, just curious | ||
| [2008/07/11 10:10:36] @ lazzurs joined channel #puppet | ||
| [2008/07/11 10:11:04] <lazzurs> Hello, for storedconfigs do the clients also have to have ruby-$dbofchoice support? | ||
| [2008/07/11 10:12:02] <mdray> no | ||
| [2008/07/11 10:13:20] <mdray> tim|macbook: warning: Configuration could not be instantiated: Could not find dependency User[admin] for Cron[timesync] at /etc/puppet/modules/development/ntp/manifests/init.pp:13 | ||
| [2008/07/11 10:13:37] <mdray> tim|macbook: will try users, not user | ||
| [2008/07/11 10:13:46] <tim|macbook> should be user... no idea | ||
| [2008/07/11 10:30:45] @ catdude joined channel #puppet | ||
| [2008/07/11 10:37:37] @ zenTourist joined channel #puppet | ||
| [2008/07/11 10:41:14] <jbooth> Volcane: Turns out I only have 5 places where I 'file { require=> }', so I think I'm just going to set it globally in a File {} for now and hand-fix those 5 cases. I'll deal with a "proper" fix later. :-/ | ||
| [2008/07/11 10:41:57] <lazzurs> ah yes, this was the problem I hit last time I went to play with puppetshow, hobo is not found by gem install hobo | ||
| [2008/07/11 10:42:06] <lazzurs> is there a way I can install it manually? | ||
| [2008/07/11 10:44:05] <jbooth> Volcane: I don't think a wrapper is the way to fix it anyway -- it means that everything needs to support that wrapper, so sharing puppet configs gets a lot less useful. | ||
| [2008/07/11 10:47:22] <DavidS> tim|macbook: realize(User[admin]) | ||
| [2008/07/11 10:47:30] <DavidS> then the require will work | ||
| [2008/07/11 10:52:32] @ oxtail joined channel #puppet | ||
| [2008/07/11 10:56:33] @ Quit: f--z: Read error: 110 (Connection timed out) | ||
| [2008/07/11 10:57:13] <Disconnect> anyone want to read/review an rsyslog module? | ||
| [2008/07/11 10:59:17] @ brscott joined channel #puppet | ||
| [2008/07/11 11:01:34] <Volcane> jbooth: i agree, puppet needs to be able to append to requires | ||
| [2008/07/11 11:04:24] @ Quit: pjw: Read error: 110 (Connection timed out) | ||
| [2008/07/11 11:12:05] <duritong> Disconnect: i could | ||
| [2008/07/11 11:13:43] @ andrewcshafer joined channel #puppet | ||
| [2008/07/11 11:13:48] @ Quit: catdude: | ||
| [2008/07/11 11:14:32] @ Quit: oxtail: "Ex-Chat" | ||
| [2008/07/11 11:18:04] <sigmonsays> RHEL5.2 users beware of "yum clean" bug! | ||
| [2008/07/11 11:18:12] <Volcane> yeah? | ||
| [2008/07/11 11:18:14] <mdray> mm? | ||
| [2008/07/11 11:18:59] <lazzurs> sigmonsays: what is the bug? | ||
| [2008/07/11 11:19:25] <sigmonsays> If you 'yum clean' with 3rd party repos, rhn-plugins adds the repo twice so the clean isn'tpossible | ||
| [2008/07/11 11:19:32] <sigmonsays> gives you "$REPO is already in configuration" | ||
| [2008/07/11 11:19:39] <sigmonsays> was killing my puppets | ||
| [2008/07/11 11:20:03] <Volcane> sheesh lame | ||
| [2008/07/11 11:20:09] <sigmonsays> https://bugzilla.redhat.com/show_bug.cgi?id=448012 | ||
| [2008/07/11 11:20:13] <Volcane> not seen it cos i'm a centosser | ||
| [2008/07/11 11:20:33] <lazzurs> Volcane: I have to day I have been unimpressed with rhel5 in general | ||
| [2008/07/11 11:20:33] <sigmonsays> I will be soon too | ||
| [2008/07/11 11:20:51] <lazzurs> the automounter appears to have issues as well due to moving to autofs5 | ||
| [2008/07/11 11:21:02] <Volcane> lazzurs: I'm really happy with centos 5.x apart from the sillyness with redhat-release file but thats fixed now too | ||
| [2008/07/11 11:21:32] * Volcane doesnt generally automount, didnt notice that one | ||
| [2008/07/11 11:22:21] * lazzurs is beginning to not like it | ||
| [2008/07/11 11:27:24] @ meandtheshell joined channel #puppet | ||
| [2008/07/11 11:34:57] @ lak joined channel #puppet | ||
| [2008/07/11 11:46:04] <DavidS> g'morning lak :) | ||
| [2008/07/11 11:46:17] <lak> morning | ||
| [2008/07/11 11:48:19] <Disconnect> duritong: ok lemme post it up (sorry, was @ review board explaining - again - how this helps the enterprise..) | ||
| [2008/07/11 11:49:15] <benp-> sillyness with redhat-release? | ||
| [2008/07/11 11:49:24] <DavidS> Disconnect: don't be so negative .. there are few activities with equal return on investmant like marketing ... | ||
| [2008/07/11 11:49:55] <Volcane> benp-: centos redhat-release worked different from upstream | ||
| [2008/07/11 11:50:01] <Volcane> benp-: fixed now in 5.2 | ||
| [2008/07/11 11:51:07] <Disconnect> we're a public-serving internet .com (think geocities) .. I can't point to these servers (esp since i'm stalled on prod deployments) and say "look, its {faster/more reliable}" or "hey, i brought in {x} paid users"... so i have to explain that periodically. thats not negative, its just life. | ||
| [2008/07/11 11:52:18] @ brscott left channel #puppet () | ||
| [2008/07/11 11:53:06] <Disconnect> duritong: http://pastie.org/232063 | ||
| [2008/07/11 11:57:35] @ Quit: lak: | ||
| [2008/07/11 11:59:22] @ Quit: tim|macbook: "This computer has gone to sleep" | ||
| [2008/07/11 12:01:56] @ Quit: andrewcshafer: | ||
| [2008/07/11 12:07:47] @ andrewcshafer_ joined channel #puppet | ||
| [2008/07/11 12:07:50] @ Quit: andrewcshafer_: Client Quit | ||
| [2008/07/11 12:08:15] @ lak joined channel #puppet | ||
| [2008/07/11 12:10:03] @ andrewcshafer_ joined channel #puppet | ||
| [2008/07/11 12:13:38] <duritong> Disconnect: looks fine | ||
| [2008/07/11 12:20:42] @ Quit: DerekW: "Leaving" | ||
| [2008/07/11 12:26:03] @ Quit: jvanzyl: | ||
| [2008/07/11 12:33:52] @ Quit: fbe: Read error: 110 (Connection timed out) | ||
| [2008/07/11 12:34:58] @ Quit: mdray: "Leaving" | ||
| [2008/07/11 12:40:49] <Disconnect> seems to work but i've been known to do stupid stuff in the past :) | ||
| [2008/07/11 12:42:59] <sigmonsays> I have 300+ nodes that I need to deploy code to | ||
| [2008/07/11 12:43:15] <Disconnect> capistrano | ||
| [2008/07/11 12:43:41] <sigmonsays> I was thinking of writing a system that allows nodes to copy code to eachother (once a new reiceves the update, it can copy to another node) allowing faster deployment | ||
| [2008/07/11 12:43:54] <sigmonsays> spread like a super virus | ||
| [2008/07/11 12:44:12] <sigmonsays> I am not a ruby shop | ||
| [2008/07/11 12:44:32] <sigmonsays> ruby or rails* | ||
| [2008/07/11 12:44:40] <thom> sigmonsays: packages! | ||
| [2008/07/11 12:44:54] <sigmonsays> packages? | ||
| [2008/07/11 12:46:42] <thom> sigmonsays: we build debian packages for all our code releases | ||
| [2008/07/11 12:47:10] <sigmonsays> i'm a bit more XP than that | ||
| [2008/07/11 12:47:20] <Disconnect> sigmonsays: i maintained that system at my last job. used rsync. its a disaster. | ||
| [2008/07/11 12:47:33] <sigmonsays> i'm using rsync now. I want it to change | ||
| [2008/07/11 12:47:37] <Disconnect> sigmonsays: you really want to speed it up, just use multiple deployment servers. (or rsync with batching) | ||
| [2008/07/11 12:47:55] <sigmonsays> that is a better idea | ||
| [2008/07/11 12:49:02] <Disconnect> we were using it for client files (video processing and bulk web serving) .. inherited it.. it broke -all- the time. and as soon as you get system time drift (and you will, eventually - oops, oom killed ntp. or ntp crashed. or ntp changed timezones randomly to piss you off. or it didn't start @ boot cuz the clock was too far out of sync...) it starts fighting itself. and it goes downhill from there. | ||
| [2008/07/11 12:49:28] <sigmonsays> hehehe | ||
| [2008/07/11 12:49:42] <sigmonsays> I have a +2 second drift nagios alert on my servers | ||
| [2008/07/11 12:49:59] <sigmonsays> but none the less, understandable | ||
| [2008/07/11 12:50:24] <Disconnect> had a couple of clients that dumped backups to us. imagine how long it takes rsync to scan a tree of about 5k 10-200 meg files (checksumming all of them - see time discussion above..) and compare it to another similar tree? .. now you can do that once and deploy teh batch, or you can do that 'n' times where 'n' is something roughly equivalent to 50x the number of actual servers... | ||
| [2008/07/11 12:50:44] <Disconnect> 2 seconds is -more- than enough to get rsync fighting itself | ||
| [2008/07/11 12:50:58] <sigmonsays> that's specific to batch mode of rsync? i'm unfamiliar w/ batch mode | ||
| [2008/07/11 12:51:01] @ shadoi joined channel #puppet | ||
| [2008/07/11 12:51:22] <Disconnect> no thats specific to your p2p distribution method | ||
| [2008/07/11 12:51:54] <Disconnect> batch mode is just "run rsync between host 1 and 2, save the changes to file Z." .. then you distribute file Z to host 3-n (should start identical to 2) | ||
| [2008/07/11 12:52:05] <Disconnect> no compares, no nothing. just a bindiff basically. | ||
| [2008/07/11 12:52:44] <Disconnect> so you need to track serials and such but thats all real easy. (I threw together a system to do that in about 3 days, using bash with a tiny bit of python for reporting and db access.) | ||
| [2008/07/11 12:53:25] <sigmonsays> nice. i'll scope that | ||
| [2008/07/11 12:57:09] @ Quit: kolla: Remote closed the connection | ||
| [2008/07/11 12:59:40] <thom> sigmonsays: belatedly, our continuous integration stuff builds packages, then when devs nominate a build for qa it goes to our staging env | ||
| [2008/07/11 13:03:22] <Disconnect> if its actual code and such, i'd do something like that.. the rsync batch is great for moving static content between a master and a bunch of slaves.. (which works, but you lose things like rollback and such from a proper "deploy code" app) | ||
| [2008/07/11 13:07:41] <sigmonsays> what's qa? | ||
| [2008/07/11 13:07:42] <sigmonsays> (jk) | ||
| [2008/07/11 13:09:51] * Disconnect is in an agile shop, so 'whats qa' has more meaning than not.. :( | ||
| [2008/07/11 13:14:12] @ Quit: zenTourist: Read error: 110 (Connection timed out) | ||
| [2008/07/11 13:15:40] <thom> heh | ||
| [2008/07/11 13:33:25] <sigmonsays> How do you tell if something is jacked in production? I know unit tests should cover that but c'mon. it's not perfect | ||
| [2008/07/11 13:33:35] <sigmonsays> real time metrics or something? | ||
| [2008/07/11 13:33:57] <sigmonsays> brb. gotta step out | ||
| [2008/07/11 13:34:02] @ sigmonsays is now known as sigmonsays-brb | ||
| [2008/07/11 13:41:18] @ punkcut joined channel #puppet | ||
| [2008/07/11 13:42:34] <punkcut> on RHEL 5 I have the following config for iptables ( http://pastebin.ca/1069342 ) . A simple one to ensure its stopped. However, puppet does not stop it when it is running... any advice? | ||
| [2008/07/11 13:43:36] <sigmonsays-brb> does /etc/init.d/iptables stop | ||
| [2008/07/11 13:43:39] <sigmonsays-brb> ? | ||
| [2008/07/11 13:44:40] <Volcane> punkcut: you need to tell puppet how to figure out if its running | ||
| [2008/07/11 13:44:45] <Volcane> punkcut: try hasstatus | ||
| [2008/07/11 13:45:16] <Volcane> punkcut: else it does a ps looking for a process called iptables | ||
| [2008/07/11 13:45:33] <sigmonsays-brb> Ahhh! genius :) | ||
| [2008/07/11 13:45:53] @ sigmonsays-brb is now known as sigmonsays | ||
| [2008/07/11 13:49:03] @ Quit: euclid: Connection timed out | ||
| [2008/07/11 13:52:00] <benp-> ruh roh. anyone know what this error might mean? | ||
| [2008/07/11 13:52:02] <benp-> err: Could not retrieve catalog: Puppet::Parser::Compiler failed with error NameError: uninitialized constant ParamValue on node | ||
| [2008/07/11 13:53:22] <Volcane> cant even guess | ||
| [2008/07/11 13:53:37] <Volcane> unless you have very munge node files thats lke node { } | ||
| [2008/07/11 13:53:41] <Volcane> instead of giving a name | ||
| [2008/07/11 13:53:52] <Volcane> but thats just wild stabbing in the dark | ||
| [2008/07/11 13:54:42] <benp-> it actually said the name of the node at the end, i cut it off | ||
| [2008/07/11 13:54:47] <Volcane> ah :) | ||
| [2008/07/11 13:55:13] @ Quit: lak: | ||
| [2008/07/11 13:56:02] <Volcane> you're doing collect/export? | ||
| [2008/07/11 13:56:15] <benp-> yes, very much so on that node in particular | ||
| [2008/07/11 13:57:04] <benp-> it's the one that takes 2000 seconds to run | ||
| [2008/07/11 13:57:06] <benp-> :) | ||
| [2008/07/11 13:57:15] <punkcut> thanks Volcane | ||
| [2008/07/11 13:57:44] <Volcane> benp-: hmmm, well as much as i can figure thats what its related too, something is not completely exporting correctly | ||
| [2008/07/11 13:59:06] @ pasha joined channel #puppet | ||
| [2008/07/11 13:59:07] @ \ask_ joined channel #puppet | ||
| [2008/07/11 14:00:11] <benp-> it's a new puppetmaster, but with an old database that i dumped from mysql running on the other puppetmaster box. | ||
| [2008/07/11 14:00:35] <pasha> i know i can copy file form the master to client using file resource. Can i copy more then one at at time? | ||
| [2008/07/11 14:00:43] <pasha> that the syntax for that? | ||
| [2008/07/11 14:00:59] <Volcane> if you make a define, then you can pass arrays of file names to the define | ||
| [2008/07/11 14:01:34] <pasha> ic | ||
| [2008/07/11 14:03:43] <benp-> totally off topic but, do people line up to buy iphones because supplies are limited or because they want to be the first to get an iphone? | ||
| [2008/07/11 14:03:47] <Volcane> pasha: http://pastie.org/232176 | ||
| [2008/07/11 14:04:23] <Volcane> benp-: because they need to go seek help but havnt realised it yet:P | ||
| [2008/07/11 14:04:52] <benp-> makes sense | ||
| [2008/07/11 14:04:57] <Volcane> pasha: something like that | ||
| [2008/07/11 14:12:37] <Disconnect> benp-: because they haven't realized there's gonna be another price cut | ||
| [2008/07/11 14:14:34] <Volcane> or that v 2.1 wont have yellow casts on the displays and issues activing like the current bunch :P | ||
| [2008/07/11 14:15:04] @ Quit: keithlard: | ||
| [2008/07/11 14:17:25] @ lak joined channel #puppet | ||
| [2008/07/11 14:24:18] <shadoi> ugh... I can't remember how use github anymore. bleh | ||
| [2008/07/11 14:24:54] <shadoi> need coffeeeeeeee | ||
| [2008/07/11 14:25:58] @ Quit: shadoi: | ||
| [2008/07/11 14:28:07] @ asa_ joined channel #puppet | ||
| [2008/07/11 14:28:40] @ jvanzyl joined channel #puppet | ||
| [2008/07/11 14:41:27] @ Quit: DavidS: Read error: 110 (Connection timed out) | ||
| [2008/07/11 14:58:15] @ Quit: emerose_: | ||
| [2008/07/11 15:10:33] @ Quit: kambiz: "Leaving" | ||
| [2008/07/11 15:12:24] @ josb joined channel #puppet | ||
| [2008/07/11 15:15:33] <josb> Hi. The k5login type/provider expects the .k5login file to exist and won't create it if it doesn't, but produce an error instead. Is that expected behavior? | ||
| [2008/07/11 15:18:07] <pasha> Volcane does this make sense http://pastie.org/232234 | ||
| [2008/07/11 15:18:42] @ ptrmcrthr joined channel #puppet | ||
| [2008/07/11 15:20:41] <refuseresisted> pasha, you have to use an array | ||
| [2008/07/11 15:21:14] <refuseresisted> http://reductivelabs.com/trac/puppet/wiki/LanguageTutorial#arrays | ||
| [2008/07/11 15:21:59] <pasha> yeah i have a book | ||
| [2008/07/11 15:22:00] <pasha> :) | ||
| [2008/07/11 15:22:22] <refuseresisted> ok | ||
| [2008/07/11 15:25:28] <punkcut> how do I determine how often the puppetd service checks in with the puppetmaster for new or changed configs? | ||
| [2008/07/11 15:25:44] @ Quit: meandtheshell: "Leaving." | ||
| [2008/07/11 15:30:30] <punkcut> nvm. found it. | ||
| [2008/07/11 15:37:36] @ emerose joined channel #puppet | ||
| [2008/07/11 15:42:39] @ Quit: lak: | ||
| [2008/07/11 15:50:49] @ Quit: andrewcshafer_: | ||
| [2008/07/11 16:03:41] @ shadoi joined channel #puppet | ||
| [2008/07/11 16:16:51] @ jprall joined channel #puppet | ||
| [2008/07/11 16:22:14] * Volcane 's added a way to view facts for hosts to his little puppet report viewer thingy | ||
| [2008/07/11 16:25:46] @ pjw joined channel #puppet | ||
| [2008/07/11 16:28:31] @ pjw_ joined channel #puppet | ||
| [2008/07/11 16:29:07] <shadoi> Volcane: is yours the php thinger? | ||
| [2008/07/11 16:30:37] <Volcane> nods | ||
| [2008/07/11 16:31:57] <shadoi> it's a shame it's PHP, but very cool nonetheless | ||
| [2008/07/11 16:32:04] @ Quit: punkcut: "Leaving." | ||
| [2008/07/11 16:32:06] <Volcane> lol | ||
| [2008/07/11 16:32:17] @ andrewcshafer joined channel #puppet | ||
| [2008/07/11 16:32:24] <Volcane> so much easier to get php things running than rail | ||
| [2008/07/11 16:32:28] <Volcane> rails | ||
| [2008/07/11 16:32:34] @ kambiz joined channel #puppet | ||
| [2008/07/11 16:32:40] @ Quit: pleemans: "Ex-Chat" | ||
| [2008/07/11 16:32:42] <shadoi> eh.. pretty easy with mod_rails | ||
| [2008/07/11 16:32:54] <shadoi> but yeah, distributing rails apps has traditionally been a huge pain | ||
| [2008/07/11 16:32:58] <Volcane> plus having been coding php since 99ish helps | ||
| [2008/07/11 16:33:05] <shadoi> yeah | ||
| [2008/07/11 16:33:19] <shadoi> I swore never to write another line. :) | ||
| [2008/07/11 16:33:26] <Volcane> hehe | ||
| [2008/07/11 16:33:59] <Volcane> i have a framework with single signon etc that i have running at a few clients and for my own intranet, so the objects i make to run this easily plugs into any of them | ||
| [2008/07/11 16:34:56] @ lak joined channel #puppet | ||
| [2008/07/11 16:39:10] <Volcane> shadoi: also been kind of tinkering with a external node tool that i might combine with this, still deciding | ||
| [2008/07/11 16:39:38] @ pjw__ joined channel #puppet | ||
| [2008/07/11 16:39:39] <shadoi> Volcane: yeah, I had big plans for puppetshow, merging it with iClassify, adding reporting, etc. | ||
| [2008/07/11 16:39:53] <Volcane> web development just sux though | ||
| [2008/07/11 16:39:56] <shadoi> Probably won't have time to do any of it for a while though | ||
| [2008/07/11 16:40:00] <shadoi> hah.. tell me about it. | ||
| [2008/07/11 16:40:02] <Volcane> such a drag to even do simple interactive things | ||
| [2008/07/11 16:40:19] <shadoi> I've been writing an extremely javascript heavy app for the last few months | ||
| [2008/07/11 16:40:25] <shadoi> it's very tedious | ||
| [2008/07/11 16:40:35] <Volcane> yeah i dont have the patience | ||
| [2008/07/11 16:40:48] <refuseresisted> thats why you steal others javascript | ||
| [2008/07/11 16:41:59] @ Quit: aymerick: | ||
| [2008/07/11 16:42:03] <Volcane> at least i am pretty good at seperating design from code, so i might in theory pay someone off guru.com to make it pretty | ||
| [2008/07/11 16:42:23] <Volcane> or call in favours with several little web devs in poland I've help setup xen servers @ hetzner | ||
| [2008/07/11 16:45:18] @ euclid joined channel #puppet | ||
| [2008/07/11 16:48:01] @ jon__ joined channel #puppet | ||
| [2008/07/11 16:48:51] <shadoi> refuseresisted: oh trust me, I try to make use of as much as I can. the worst part is forms though, there's no really great way to do them. | ||
| [2008/07/11 16:49:44] <Volcane> shadoi: god yeah, get form data, something out of place, regenerate same form with little warnings of whats missing/wrong, repeat etc, nightmarish tedium | ||
| [2008/07/11 16:49:50] <Volcane> shadoi: i miss VB at times like that | ||
| [2008/07/11 16:50:02] @ Quit: pjw_: Read error: 110 (Connection timed out) | ||
| [2008/07/11 16:50:43] @ Quit: pjw: Read error: 110 (Connection timed out) | ||
| [2008/07/11 16:50:47] <refuseresisted> Volcane, don't know if i've ever had to worry about that with rails | ||
| [2008/07/11 16:51:02] <refuseresisted> validates rul | ||
| [2008/07/11 16:51:16] <refuseresisted> gah, rule | ||
| [2008/07/11 16:51:19] <shadoi> refuseresisted: when you're using something like ExtJS with rails it's a huge pain. | ||
| [2008/07/11 16:51:39] <shadoi> not validation, but the data modeling, etc. | ||
| [2008/07/11 16:51:40] <refuseresisted> shadoi, never heard of it. admittedly, my web dev is very limited as a sysadmin | ||
| [2008/07/11 16:52:04] <shadoi> essentially it's like you're writing two separate applications | ||
| [2008/07/11 16:53:34] <Volcane> last time i looked at rails it was a monumental pain to do anyting that doesnt fit into their idea of a data model | ||
| [2008/07/11 16:53:49] <Volcane> not sure if thats still the way | ||
| [2008/07/11 16:54:08] <shadoi> Volcane: what do you mean? | ||
| [2008/07/11 16:54:33] <shadoi> MVC / Active Record is not really rails specific at all. | ||
| [2008/07/11 16:55:12] <Volcane> back then to really take advantage of all the automation and code generation you had to have, say, table called animals, instances of animal, forms called animal | ||
| [2008/07/11 16:55:23] <Volcane> or somesuch, its ages ago that i looked atit | ||
| [2008/07/11 16:55:49] <Volcane> and if say you have old data and want to get all the advantage of the automation and stuff you had to either massage it or make views | ||
| [2008/07/11 16:55:51] <shadoi> ah well, you get quite a bit of benefit from following the convention | ||
| [2008/07/11 16:56:25] <Volcane> which is great for starting fresh yes | ||
| [2008/07/11 16:56:48] <Volcane> but for me at the time my biggest itch was my existing intranet app that was getting a bit shitty | ||
| [2008/07/11 16:56:55] <shadoi> You can make your model handle aliasing, and table names, etc. But yeah, it's not totally pretty to add legacy data to a rails app. | ||
| [2008/07/11 16:57:02] <Volcane> and it drives a whole lot of stuff, and my data model didnt fit well with their conventions | ||
| [2008/07/11 16:58:02] <shadoi> Volcane: here's what I was talking about the Active Record pattern: http://en.wikipedia.org/wiki/Active_record_pattern | ||
| [2008/07/11 16:58:13] <Volcane> i know the pattern yeah :) | ||
| [2008/07/11 16:58:42] <shadoi> *nod* | ||
| [2008/07/11 16:58:52] <shadoi> Just bugs me when people think Rails invented all this. :) | ||
| [2008/07/11 16:58:58] <Volcane> yeah | ||
| [2008/07/11 16:59:50] @ Quit: pjw__: Read error: 110 (Connection timed out) | ||
| [2008/07/11 17:01:22] <nDuff> shadoi, looked into GWT? Having the compiler generate javascript for you is a pretty good way of avoiding it. | ||
| [2008/07/11 17:02:00] <shadoi> nDuff: yeah, it's hideous to write in my opinion... I hate Java, but THIS makes me drool for that same reason: http://mysterycoder.blogspot.com/2008/07/announcing-rubyjsonrails-arax-sans.html | ||
| [2008/07/11 17:02:32] <nDuff> shadoi, eh; my experience differs, but *shrug*. | ||
| [2008/07/11 17:03:11] <shadoi> nDuff: I'll freely admit that I'm a completely lame coder. I absolutely hated programming until ruby, so take everything I say with that large lump of salt. | ||
| [2008/07/11 17:03:52] <nDuff> shadoi, I'm not interested unless it has serious debugging toolage; if I didn't care about that, I'd be using Hop (http://hop.inria.fr/) for webapps. | ||
| [2008/07/11 17:06:12] * nDuff isn't particularly fond of Ruby, but doesn't want to get into a flamewar here. The language borrows some good ideas, to be sure. | ||
| [2008/07/11 17:06:12] <shadoi> nDuff: hmm, hadn't heard of Hop, looks interesting | ||
| [2008/07/11 17:07:02] <nDuff> shadoi, it *is* interesting, but there aren't suitable debugging tools for the compiled client-side code. | ||
| [2008/07/11 17:07:05] @ Quit: kenvandine: "Ex-Chat" | ||
| [2008/07/11 17:07:30] <shadoi> man their default CSS is hideous | ||
| [2008/07/11 17:08:46] <shadoi> nDuff: have you tried ExtGWT? I'm using ExtJS, it's not bad for the standard stuff. | ||
| [2008/07/11 17:09:10] <nDuff> shadoi, not only have I tried it, it's what I use; my employer has an ExtJS license. | ||
| [2008/07/11 17:09:34] <shadoi> *nod* | ||
| [2008/07/11 17:09:35] <nDuff> ...when I'm doing webapps with significant client-side code, which I don't when it can be at all avoided. | ||
| [2008/07/11 17:09:49] * nDuff does backends. | ||
| [2008/07/11 17:09:56] <shadoi> yeah | ||
| [2008/07/11 17:10:08] <shadoi> I would prefer the same, but someone's gotta do it. :) | ||
| [2008/07/11 17:10:48] <nDuff> yes, but there are other someones on staff who are paid for it and willing... so that someone's damned well not going to be me. | ||
| [2008/07/11 17:11:19] <shadoi> hopefully we can hire someone who just loves doing front-end crap | ||
| [2008/07/11 17:11:25] <shadoi> sooooon | ||
| [2008/07/11 17:14:29] @ pjw joined channel #puppet | ||
| [2008/07/11 17:17:59] <btm> that's the nice thing about UX people, they get off on coding frontends. | ||
| [2008/07/11 17:18:46] <shadoi> I admit I enjoy it when everything comes together, but it's not worth the tedium. | ||
| [2008/07/11 17:18:46] <Volcane> UX? | ||
| [2008/07/11 17:18:59] <shadoi> User eXperience? | ||
| [2008/07/11 17:19:04] <shadoi> I'm guessing | ||
| [2008/07/11 17:19:05] <Volcane> all i enjoy is writing objects that work together etc, hate frontend junk | ||
| [2008/07/11 17:22:45] <Volcane> so my frontends all nicely seperated from the logic, but its all tables and ancient style html heh | ||
| [2008/07/11 17:25:11] <pasha> Volcane i am getting error on line 13 on the script you gave http://pastie.org/232176 | ||
| [2008/07/11 17:25:27] <pasha> can you tell me why please | ||
| [2008/07/11 17:29:42] <Volcane> because it was a guide you need to look at it, learn from it and make it work in your environemnt | ||
| [2008/07/11 17:31:55] <Volcane> also "an error" doesnt help much, I think I must have seen 100s of completely different errros mentioned on this channel | ||
| [2008/07/11 17:32:46] @ Quit: markl_: "Lost terminal" | ||
| [2008/07/11 17:35:17] @ mgarfias joined channel #puppet | ||
| [2008/07/11 17:36:34] <shadoi> pasha: your usage would lead to this source url being used: puppet://puppet/files//bar.hostname | ||
| [2008/07/11 17:36:40] <shadoi> which probably isn't want you want | ||
| [2008/07/11 17:36:57] <pasha> i know | ||
| [2008/07/11 17:37:08] <pasha> Syntax error at ',' at /etc/puppet/manifests/classes/munin.pp:13 | ||
| [2008/07/11 17:37:33] <shadoi> yeah it has to be an array | ||
| [2008/07/11 17:37:52] <shadoi> better to define it like "$filenames = [ "/foo", "/bar" ] | ||
| [2008/07/11 17:37:54] <pasha> something wrong with syntax, maybe you can put multiple values when you call the definition | ||
| [2008/07/11 17:38:12] <shadoi> remotefile { $filenames: ... } | ||
| [2008/07/11 17:38:27] <pasha> ic | ||
| [2008/07/11 17:38:39] <refuseresisted> pasha, i told you earlier to use an array hahaha | ||
| [2008/07/11 17:38:40] <pasha> put the value in array | ||
| [2008/07/11 17:38:50] <refuseresisted> and linked to the wiki docs on arrays | ||
| [2008/07/11 17:39:06] <pasha> i am learning scripting sorry guys | ||
| [2008/07/11 17:39:09] <refuseresisted> ;) | ||
| [2008/07/11 17:39:23] <pasha> i am suffering here | ||
| [2008/07/11 17:39:23] <shadoi> also, I think using remotefile defines is sort of a bad idea | ||
| [2008/07/11 17:39:34] <shadoi> you should use a resource default for the scope you're in instead | ||
| [2008/07/11 17:39:55] <pasha> well i want to copy a bunch of file from master to client | ||
| [2008/07/11 17:40:03] <Volcane> shadoi: why? | ||
| [2008/07/11 17:40:16] <hMz> Failed to retrieve current state of resource: index 23401 out of string | ||
| [2008/07/11 17:40:25] <hMz> huh? | ||
| [2008/07/11 17:40:26] <shadoi> Volcane: because it breaks a few things by hiding the actual File resources inside the define | ||
| [2008/07/11 17:41:09] <Volcane> which things? | ||
| [2008/07/11 17:41:15] <shadoi> other resource defaults | ||
| [2008/07/11 17:41:23] <shadoi> some inheritance bits | ||
| [2008/07/11 17:41:30] <Volcane> (i dont use them since moving to modules, but didnt have problems with them yonks ago) | ||
| [2008/07/11 17:41:33] <shadoi> I forget exactly, I just remember it was bad juju. | ||
| [2008/07/11 17:41:40] <Volcane> ah :) | ||
| [2008/07/11 17:41:55] <Volcane> well I'd be keen to see a sample of a problem it causes | ||
| [2008/07/11 17:41:59] <shadoi> resource defaults in general are the best way to do most things I think. | ||
| [2008/07/11 17:42:30] <Volcane> shadoi: well, i had to make 100 cronjobs, only difference betwen them was a number | ||
| [2008/07/11 17:42:54] <shadoi> *nod* | ||
| [2008/07/11 17:42:58] <Volcane> shadoi: a define to createcron{["1", "5".....]: } was much easier than 100 cron{}s with defaults set | ||
| [2008/07/11 17:43:18] <shadoi> indeed | ||
| [2008/07/11 17:43:55] <Volcane> same for files really, if there was an easy way to figure out what current module i was in I'd have made a new remotefile to work with mods i guess | ||
| [2008/07/11 17:44:20] @ Quit: pjw: | ||
| [2008/07/11 17:44:31] <shadoi> Ah, one big badness was that File resources inside a define don't use filebuckets by default and aren't backed up. | ||
| [2008/07/11 17:44:36] <shadoi> that bit rather hard. | ||
| [2008/07/11 17:45:23] <pasha> can i call a definition name in a node as a class? | ||
| [2008/07/11 17:45:27] <Volcane> interesting, though i never rely on them buckets | ||
| [2008/07/11 17:45:32] <pasha> like i class | ||
| [2008/07/11 17:45:32] <hMz> any clue about my error? | ||
| [2008/07/11 17:45:51] <Volcane> hMz: no bells | ||
| [2008/07/11 17:45:58] <lak> hMz: what error? | ||
| [2008/07/11 17:45:58] <hMz> :( | ||
| [2008/07/11 17:46:11] <shadoi> hMz: what is that error from? | ||
| [2008/07/11 17:46:12] <hMz> Failed to retrieve current state of resource: index 23401 out of string | ||
| [2008/07/11 17:46:25] <hMz> let me pastebin i guess | ||
| [2008/07/11 17:46:49] <lak> hMz: i'd need a stack trace for that; it's a programming bug | ||
| [2008/07/11 17:46:57] <lak> trying to get an index on a string when you shouldn't | ||
| [2008/07/11 17:47:28] <lak> s/you/i/ | ||
| [2008/07/11 17:47:38] <shadoi> bad lak, bad. | ||
| [2008/07/11 17:47:53] <lak> all about love | ||
| [2008/07/11 17:48:01] <hMz> i think this is my doing | ||
| [2008/07/11 17:48:02] <shadoi> hehe | ||
| [2008/07/11 17:48:06] <hMz> i'm writing a custom package provider | ||
| [2008/07/11 17:48:11] <hMz> so i probably did something dumb | ||
| [2008/07/11 17:48:14] <shadoi> bad hMz, bad. | ||
| [2008/07/11 17:48:52] <shadoi> lak: I wrote a patch for that environment column in the hosts table for storeconfigs | ||
| [2008/07/11 17:49:02] <hMz> how do i acquire said trace? | ||
| [2008/07/11 17:49:06] <shadoi> lak: if I can just remember how to use git for a minute I'll point you to it. | ||
| [2008/07/11 17:49:17] <shadoi> hMz: run with --trace | ||
| [2008/07/11 17:49:42] <lak> shadoi: didn't someone else do that already? | ||
| [2008/07/11 17:49:52] <lak> i thought that patch was already integrated and the ticket was closed... | ||
| [2008/07/11 17:49:58] <shadoi> lak: oh probably, I'm terribly slow these days. | ||
| [2008/07/11 17:50:01] * lak is only half-paying attention, at a conference | ||
| [2008/07/11 17:50:40] <shadoi> yup they did, and it's identical to mine, so... meh | ||
| [2008/07/11 17:50:57] <shadoi> except I wrote a migration instead of just modifying the schema | ||
| [2008/07/11 17:51:16] <hMz> ahh yup, my stupidity strikes again :) | ||
| [2008/07/11 17:53:03] @ Jonny joined channel #puppet | ||
| [2008/07/11 17:53:58] <lak> shadoi: err, it's not a migration? can you, um, add your patch and reopen the bug? | ||
| [2008/07/11 17:54:08] <shadoi> sure | ||
| [2008/07/11 17:54:30] <lak> thanks | ||
| [2008/07/11 17:54:59] <shadoi> ah, fsweetser did | ||
| [2008/07/11 17:55:06] <shadoi> all good | ||
| [2008/07/11 17:55:20] <shadoi> or whoever | ||
| [2008/07/11 17:55:30] <shadoi> gah, need more coffee again, brain not work. | ||
| [2008/07/11 18:00:45] @ Quit: jon__: "going, going, gone" | ||
| [2008/07/11 18:07:26] @ randybias joined channel #puppet | ||
| [2008/07/11 18:07:28] @ marti_martinez joined channel #puppet | ||
| [2008/07/11 18:07:29] @ Quit: marti_martinez: Client Quit | ||
| [2008/07/11 18:07:55] @ marti_martinez joined channel #puppet | ||
| [2008/07/11 18:10:08] @ Quit: marti_martinez: Client Quit | ||
| [2008/07/11 18:10:29] @ marti_martinez joined channel #puppet | ||
| [2008/07/11 18:13:24] @ jbrothers left channel #puppet () | ||
| [2008/07/11 18:13:50] @ punkcut joined channel #puppet | ||
| [2008/07/11 18:22:59] <punkcut> On RHEL5, using the puppetd client, I keep getting error messages about not being able to create the PID file: I think this may be a bug in the initscript... im on 24.4 | ||
| [2008/07/11 18:25:42] <marti_martinez> Is there a reference/strategy for using puppet with dhcp clients (specifically, managing certificates on machines that will be changing ips on a regular basis) | ||
| [2008/07/11 18:26:09] <punkcut> wouldnt it be better to use dyndns with DHCP and have the certs signed to hostname? | ||
| [2008/07/11 18:26:53] <Volcane> punkcut: so you have puppetd runing already and then run puppetd on the command line and get the error right? | ||
| [2008/07/11 18:29:24] <marti_martinez> punkcut: yes, it probably would be, but I don't have that option right now | ||
| [2008/07/11 18:40:11] @ Quit: jvanzyl: | ||
| [2008/07/11 18:47:15] <punkcut> Volcane : no. Im just running puppetd and then when it hits its checkin interval it spits out a the could not create PID file error | ||
| [2008/07/11 18:47:34] <punkcut> Volcane: it actually seems to be stopping puppetd from doing its job | ||
| [2008/07/11 18:48:51] <punkcut> Volcane: I have --trace enabled but it doesnt seem to spit out a trace for that... thinking about keeping it in the foreground and running strace... | ||
| [2008/07/11 18:49:47] <Volcane> only seen it do that when there were 2 running at once | ||
| [2008/07/11 18:50:28] <Volcane> whats in /var/run/puppet/puppetd.pid ? | ||
| [2008/07/11 18:50:43] <shadoi> punkcut: make sure you're running it as root? | ||
| [2008/07/11 18:51:17] <punkcut> Volcane: the correct pid | ||
| [2008/07/11 18:51:31] <punkcut> shadoi: yup | ||
| [2008/07/11 18:51:38] <Volcane> punkcut: odd | ||
| [2008/07/11 18:51:51] <shadoi> punkcut: what are the perms on that file and dir? | ||
| [2008/07/11 18:52:14] <shadoi> what what does: puppetd --configprint rundir | ||
| [2008/07/11 18:52:14] <shadoi> say | ||
| [2008/07/11 18:52:25] <shadoi> can I get a what what | ||
| [2008/07/11 18:52:53] <punkcut> http://pastebin.ca/1069585 | ||
| [2008/07/11 18:53:13] <shadoi> wrong perms, should be owned by root | ||
| [2008/07/11 18:53:18] <punkcut> shadoi: /var/run/puppet | ||
| [2008/07/11 18:53:46] <punkcut> shadoi : changed the perms. still getting the error. | ||
| [2008/07/11 18:54:09] <punkcut> its now root.root - same octal | ||
| [2008/07/11 18:54:20] <shadoi> with +t? | ||
| [2008/07/11 18:54:36] <shadoi> restricted deletion flag | ||
| [2008/07/11 18:54:41] <shadoi> hrmm | ||
| [2008/07/11 18:55:07] <shadoi> try chmod -t | ||
| [2008/07/11 18:55:09] <shadoi> curious | ||
| [2008/07/11 18:55:30] <Volcane> drwxrwxrwt 2 puppet puppet 4096 Jul 10 22:07 /var/run/puppet | ||
| [2008/07/11 18:55:43] <punkcut> k so it should be puppet.puppet 777 | ||
| [2008/07/11 18:55:59] <punkcut> Volcane: is that 777 recursive? | ||
| [2008/07/11 18:56:27] <Volcane> -rw-r--r-- 1 root root 5 Jul 10 22:07 /var/run/puppet/puppetd.pid | ||
| [2008/07/11 18:56:48] <punkcut> k so i match your perms now Volcane | ||
| [2008/07/11 18:56:53] <punkcut> waiting for the next check in... | ||
| [2008/07/11 18:57:25] <shadoi> punkcut: probably not the issue, the dir on my puppetmaster was like yours and it works fine | ||
| [2008/07/11 18:57:34] <shadoi> puppetmasterd requires puppet.puppet | ||
| [2008/07/11 18:57:45] <punkcut> you guys want me to setup an strace? | ||
| [2008/07/11 18:58:02] <shadoi> probably not a bad idea | ||
| [2008/07/11 18:58:08] <punkcut> nothing like killing a fly with a shotgun... | ||
| [2008/07/11 18:58:23] <punkcut> gimme a second to get puppetd in the foreground | ||
| [2008/07/11 18:58:34] <punkcut> Volcane: FYI: just got a checkin. same error with new perms. | ||
| [2008/07/11 18:58:46] <punkcut> TIA for all the help shadoi, Volcane | ||
| [2008/07/11 18:58:56] <Volcane> rpm -V puppet | ||
| [2008/07/11 18:58:56] <Volcane> ? | ||
| [2008/07/11 18:59:29] <punkcut> /var/run/puppet M | ||
| [2008/07/11 18:59:29] <punkcut> /etc/sysconfig/puppet S.5....T c | ||
| [2008/07/11 19:00:01] <punkcut> would you guys mind saving me a trip to the uber slow puppet docs and give me the option to disable daemonizing? | ||
| [2008/07/11 19:00:31] <refuseresisted> --no-daemonize | ||
| [2008/07/11 19:00:32] <Volcane> --no-daemonize | ||
| [2008/07/11 19:00:36] <refuseresisted> I WIN! | ||
| [2008/07/11 19:01:17] <Volcane> :) | ||
| [2008/07/11 19:01:38] <Volcane> being midnight means i get |
