[04:57] grrr [04:57] freajking raining [06:10] :D [06:11] ;D [06:19] what you up to being up so early mate [06:30] XD [06:30] being one damaged individual [06:31] just finished a film [06:31] http://www.imdb.com/title/tt0338013/ [06:36] hm let me see [06:37] not sure ive seen thst [06:37] gonna watch the cube tomorrow..never seen it:) [06:47] ah yeah [06:47] think there are a few to that too? [06:55] yea 3 [06:55] u seen 1st? [06:56] yip think so, many moons ago === Lcawte|Away is now known as Lcawte === zmoylan-1i is now known as zmoylan-pi === DanChapman_ is now known as DanChapman [10:11] morning boys and girls. [15:21] afternoon [15:28] can anyone explain how this works? http://pastebin.com/mVFLutDg [15:28] according to du, I've got like 40GB used, not 250. [15:33] * penguin42 looks [15:34] Azelphur: under where? [15:34] what do you mean under where? [15:34] oh I see [15:34] Azelphur: So your /media is a separate 2TB mount? [15:34] yea [15:35] Azelphur: OK, so there are a few reasons that can happen [15:35] i too do not understand the question [15:35] Azelphur: One reason is that you might have some stuff in /media on the main drive that's hidden by the mount [15:35] ali1234: I'm trying to find out where all my drive space has gone, du is being less than helpful [15:35] Azelphur: So try umounting /media and duing it [15:36] use ncdu not du [15:36] du sucks [15:36] Azelphur: Similarly for /tmp or the like that's a shmfs mount [15:36] penguin42: that was the correct answer, looks like it hasn't mounted my second drive for some reason and is instead writing backups onto the tiny 250GB drive. [15:37] there we go, now things are much happier. [15:39] thanks :) [15:40] ali1234: btw, did you see my silly lights video? :) [15:40] yes [15:40] hehe [15:40] looks like a good way to get rid of unwanted house guests [15:40] that's true, it'd probably be remarkably effective [15:40] especially if I played country music instead [15:41] why do you have windows that go to the floor? [15:42] ali1234: the building is circa 1700s and is protected [15:48] hmm [15:48] my raspberrypi stopped working again..yet was plugged in and net up..anyopne got any ideas what could be causing it?! [15:53] mapps: I'd try adding some warm custard [15:56] :D [16:00] \o/ New drive installed & setup, just copying data off the old drive now [16:06] mapps: So without knowing the Pi specifically; 1) Check your SD card, they're made of cheese, 2) Check how you're powering it [16:09] pah:P i cant im not there..so just check sd and cables afaik all fine [16:09] where would it log any errors to that could cause it to die dmesg? [16:11] mapps: Well yes but that depends why it died, if it couldn't get to the SD for example then they won;t have been able to get recorded onto the said SD [16:13] mapps: If it does it regularly then try and setup network or serial logging to another machine [16:17] its done it after a week of 24/7 [16:20] my dad unplugged it and plugged back in it boots up and all works fine agao [16:21] if it runs out of memory..too much running would that d oit? [16:21] mapps: The other thing then to look for is how dead it is - does it still ping for example? Also look if there is support for any watchdog timer on it so it can reset itself [16:21] mapps: If you've got a leak somewhere then yes [16:21] couldnt ssh to it/reach apache or vpn didnt try pinging [16:21] mapps: ping first :-) [16:22] watchdog timer?! surely there wouldnt be anythong to reset after x days if thats what you mean?! [16:22] mapps: The idea with a watchdog is that it's a timer that you have to prod every few seconds, if you don't it sends a reset [16:23] ah [16:23] mapps: Then you get something that just checks everything is ok, and prods the timer; if anything breaks it resets [16:23] as it assumes somethings gone wrong| [16:23] ya [16:23] yep [16:23] if it was rebooting it would be fine..eveything comes back to lie..apache..vpn..iptables rules etc [16:24] but it just crashes/hangs from what i know..im not there..dad says its on..i cant reach it minute he unplugs and plugs back in works again;( [16:26] mapps: The other thing you can do is do things like write a program that just continously flashes an LED (I assume it has one you can do that to) - and when you invoke your dad to kick it, he can tell you if the LED is still flashing === Lcawte is now known as Lcawte|Away [17:50] there's a fun task for an arduino there .. a hardware watchdog. that if it hasn't had a 'blink' for long enough it powercycles [17:53] shauno: You don't need an arduino for that; I think you can do it with transistor+capacitor type of thing [17:53] shauno: Or relay+cap [17:55] well yeah. but arduinos are like Pis .. I have more than I'd like to admit :) [17:56] (and can evolve pretty easily .. eg to provide a soft-off) [17:59] I have a troubling amount of these knocking around .. http://oomlout.co.uk/collections/arduino/products/arduino-component-bundle-arcb [17:59] shauno: Oh that's quite a nice little set [18:00] a couple of quid more than it needs to be, but I'm lazy [18:01] yeh, I was thinking it's a bit more pricey than shrimpit === Lcawte|Away is now known as Lcawte [18:59] lol penguin42 re kicking it:D would be nice to know why it does it though [19:27] mapps: Right, but that's what your serial or network console type things help with if you can't get anything from the logs [19:34] not sure what you mean by serial or network console..and on a pi?? [19:38] Evening. I need to setup OpenVPN on a server in a remote office, but that's going to involve fiddling with network [19:38] what's the safest way to adjust settings? Any way I can do it and not lock myself out? [19:44] is there a way make it so that malloc will never reuse a virtual address? (other than never freeing memory at all) [19:45] with a 64 bit address space this should be at least feasible [19:45] camason: Give yourself another way in [19:46] camason: setup two vpn setups, only ever work on one at a time [19:46] penguin42 it's more I need to re-jig my eth settings, and I fear one mistake will take the network offline [19:47] ali1234: See man mallopt - but I don't think anything like that lets you do it [19:48] camason: Trust me, it will :-) [19:49] camason: Some type of watchdog where you arm the system to reboot itself/put it's self back to a good state is what you need then arm that before you make any change - if you really only have one way in [19:50] yes I was thinking perhaps a script that will execute after 5 mins, replacing /etc/network/interfaces with a known working config, then restarting it [19:51] camason: It is of course difficult to anticipate how you might screw it up [19:51] yup! [19:51] there are a myriad of ways when I'm involved :) [19:52] camason: Humans are creative [19:52] camason: Doing things like holding your config files is a good way to be able to roll back; but automating that I've not seen [19:52] ^your config files in git.... [19:52] reminds me of a line from The Thick Of It... that I could't replicate here for reasons of taste. [19:54] is it possible to tell valgrind i only care about memory referenced from a single source file? [19:55] ali1234: I know it has ways of marking areas of code, I'm not sure by source file [20:05] penguin42, what did yoiu mean by serial or network console [20:06] mapps: You can configure syslog to send all messages to another host, similarly you can configure the kernel to send console messages to another host; so if the SD card stops or a lot of other stuff fails you still see the messages [20:06] mapps: Similarly for a rs232/serial console [20:08] i dont get it..so how can i do this on the pi [20:09] 'network console'?:! [20:12] mapps: Is the pi the only machine there or is there another linux machine within spitting distance? [20:13] ther'es another machine a dell desktop running Linux that i could getr my dad to turn on [20:13] but can you explain what you mean when you say network console that i cant grasp..i know what the console is;p [20:14] mapps: Every piece of text that's output to the console gets sent over the network so that it can be logged [20:14] aha [20:15] surely errors and that wouldnt tho ..dont see them on console if they go to a log [20:16] mapps: Well it depends why it's dieing, if you're getting a kernel panic, some types of kernel panic will get to the console and over the network, some won't [20:17] so if i jhad a console up and was logged in locally with as GUI kernel panics would show in that terminal? [20:17] just be handy to find out so can fix it [20:17] i want it running 24/7 and noones gonna be there most the time [20:17] mapps: Best not to have the GUI running, and best to turn the screen blanking off with setterm, but yes [20:18] thanks [20:18] mapps: So you can ask someone just to turn a monitor on and take a picture of it before they powercycle it [20:18] yea [20:18] or network console could be a good idea? [20:19] but if i cant reach the pi ..surely its networks down so it wouldnt be able to reach the other machine? [20:19] well that depends how it died [20:19] mapps: Serial consoles are more primitive, they tend to be able to send things with almost everything else gone [20:19] so network could still unction but apache/vpn dead? [20:19] yea [20:20] cant access it locally tho;) [20:20] that's why you need another pi to monitor the first pi.... [20:20] for network console? ..i could do that with the dell linux machine [20:20] yes you could [20:21] mapps: See https://www.kernel.org/doc/Documentation/networking/netconsole.txt [20:21] thanks [20:26] could it be as simple as i have too much running on it? vpn/apache/ftpd /mrtg ? [20:27] if you're running out of memory, possibly, but it's unlucky to kill everything [21:40] £6.40 or a lite of stolichynia [21:40] :D [21:40] bargain [21:51] doesn't drink or smoke; what's the good chocolate like... [21:53] i smoke more now im here [21:53] as foolish as that is === Lcawte is now known as Lcawte|Away