**** BEGIN LOGGING AT Sun Jun 17 02:59:59 2012 Jun 17 03:45:14 GeneralAntilles: anyway http://mg.pov.lt/maemo-irclog/search?q=**+ChanServ+sets+mode%3A+%2Bo Jun 17 03:46:13 (re "how much chan moderation been needed in the past?") Jun 17 06:37:16 w Jun 17 06:37:46 damn, that's the second time IRC commands got chopped by mIRC focusing new tabs Jun 17 07:20:05 Hurrian: that's what you deserve for using mirc Jun 17 07:24:05 kerio, I prefer XChat, but the styling just seems ever so off on Windows Jun 17 08:55:38 merlin1991, thanks for you help yesterday, navit built against CSSU qt libraries went smooth. Jun 17 10:53:01 hi all , where to read the ssu changelog? Jun 17 10:53:05 please Jun 17 10:53:59 also, is anyone aware of any tizen news? like, when can I buy a device with a working tizen? Jun 17 10:54:40 * SpeedEvil sighs. Jun 17 10:54:51 I am getting confused at dead projects. Jun 17 10:56:57 divide & conquer Jun 17 10:58:16 So... Maemo, harmattan, meltimi, meego, moblin, Jun 17 10:58:22 Is Limo dead? Jun 17 10:58:22 http://i.japh.se/n900/rp.png damdidam Jun 17 10:58:48 :) Jun 17 10:59:39 Ah - limo got eaten by tizen Jun 17 11:00:12 meh is anything of all that producing any code? Jun 17 11:00:41 Screw code. Jun 17 11:00:47 Released hardware. Jun 17 11:01:30 as the situation is I'm confident I'll stay with n900/n950/n9 for quite some years into the future, because vendors simply doesn't seem to care about us Jun 17 11:01:49 :/ Jun 17 11:02:03 http://arstechnica.com/gadgets/2012/02/preorders-begin-for-spark-the-open-kde-tablet/ Jun 17 11:02:08 sort-of Jun 17 11:02:44 sure but that's just a toy. or what's the purpose? Jun 17 11:03:33 And yes, there is that. Jun 17 11:03:55 To do anything more than 'toy' - you need much more investment. Jun 17 11:04:22 at least in this area yeah Jun 17 11:04:25 sadly Jun 17 11:05:57 I wonder how the talk is going in nokia hc ... I haven't seen or even heard of anyone who bought/were interested in their windows phones. Jun 17 11:06:03 maybe the situation is different elsewhere Jun 17 11:07:06 Sacking almost all developers isn't exactly positive. Jun 17 11:07:55 and they had such a great thing going :( Jun 17 11:08:47 I guess breaking up with ms now would cost nokia more than what the company's worth Jun 17 12:40:37 ~seen Estel_ Jun 17 12:40:42 estel_ <~Estel@Maemo/Community/council/Estel-> was last seen on IRC in channel #maemo, 16h 12m 35s ago, saying: 'side effect - camera doesnt work too, despite fact, that - apparently - camera module is ok'. Jun 17 12:56:23 In what channels, if any do you ask your advanced Linux questions? I am finding that the signal to noise ration has become so low, it's much more efficient to figure out everything myself. Jun 17 12:56:34 ratio* Jun 17 12:56:50 :/ Jun 17 12:56:55 hahaha Jun 17 12:57:08 fasta: i'd have to agree with you there ;) Jun 17 12:57:31 most of it is "My printer works in windows but not in Linux can someone help me figure it out?????" Jun 17 12:57:52 Even some developer channels are now getting infected with such issues. Jun 17 12:57:52 "linux sucks because my $10 lexmark printer doesn't work! Jun 17 12:58:19 Macer: no, more like "I stated information elements X, Y and Z, and then people ask whether Z holds." Jun 17 12:58:38 well.. there is option b.. .start your own # and lay down iron fisted control of it Jun 17 12:58:48 Macer: or asking 'what's the problem' after 3 minutes, which means that they don't know how to setup IRC properly. Jun 17 12:59:34 There are IRC logs too, so if people really wanted to help, they could just have read that. Jun 17 12:59:38 most of the time people just don't know how to fix the issue you're having... i agree with you that it is usually easier to simply figure it out yourself Jun 17 12:59:54 fasta: yeah but who wants to go find logs or scroll up for someone else's problem? ;) Jun 17 12:59:55 Macer: right, but instead of those people just shutting up, they 'try to help'. Jun 17 13:00:36 heh.. what makes it worse is that most of the ones trying to help are 2 week windows converts who think because they run ubuntu they are linux experts Jun 17 13:00:48 Macer: or very recently, I had a Gentoo user. Jun 17 13:01:02 gentoo users are a special breed Jun 17 13:01:06 Macer: he thought me using Ubuntu was a very bad insult. Jun 17 13:01:15 yeah. that is typical. Jun 17 13:01:18 Macer: of course, everyone who works at Google is also retarded. Jun 17 13:01:24 (by the same logic) Jun 17 13:01:38 ubuntu is great for the end user.. personally i'd rather use kubuntu Jun 17 13:01:48 but as far as a linux distro that simply works after installed it is the closest thing Jun 17 13:01:55 Macer: well, I installed Ubuntu, but use Kubuntu packages. Jun 17 13:02:07 heh.. then why not just use kubuntu ;) Jun 17 13:02:11 personally i love kde Jun 17 13:02:18 Macer: I didn't know that when I installed, and I tried all of them. Jun 17 13:02:28 Macer: I think I still have all of them installed. Jun 17 13:02:33 Macer: I have the resources, so why not? Jun 17 13:02:38 they're all the same anyways just a different ui Jun 17 13:02:48 Macer: yeah, the terminal is the most important. Jun 17 13:02:56 heh... don't tell that to the converts ;) Jun 17 13:03:11 if you can't click to fix it just isn't worth it Jun 17 13:03:22 I am very much passed the point where I care whether other people use it. Jun 17 13:03:33 cityLights: /join #maemo-ssu see /topic Jun 17 13:03:36 well. that makes 1... about 2 billion more to go Jun 17 13:04:01 Anyway, this channel is still not infected, although people might regard this as noise, which I suppose it is in some way. Jun 17 13:04:24 not really Jun 17 13:04:49 speaking of which.. if anybody wants to buy my n900 with a jacked up sim reader let me know :-P Jun 17 13:04:55 checks not accepted Jun 17 13:05:07 i'll give it up for $100 Jun 17 13:05:15 (plus shipping) Jun 17 13:05:29 other than the flakey sim reader it works great Jun 17 13:05:55 one of you electronic nerds might be able to fix it Jun 17 13:05:59 hmm Jun 17 13:09:36 fasta: WUT? this channel has almost worst S/N I ever seen Jun 17 13:09:50 * ShadowJK doubts it's actually sim reader Jun 17 13:10:12 DocScrutinizer05: which channels do you think have high S/N then? Jun 17 13:10:18 because on my devices also emergency calls, which work without sim, got disabled :P Jun 17 13:10:48 fasta: would you be surprised when I got reasons not to mention their names? Jun 17 13:10:50 ShadowJK: weird Jun 17 13:11:00 did you actually try calling 911 with it? lol Jun 17 13:11:27 crossed out sim icon is just generic "modem failure" icon it seems Jun 17 13:11:42 DocScrutinizer05: well, you can send me in private message, unless you think I am also one of those morons. Jun 17 13:11:46 yeah but my logs seem to show the sim reader failing Jun 17 13:12:05 as though it keeps losing it over and over... maybe it was some software issue but i haven't tried to flash it yet to stock Jun 17 13:12:08 oh? Paste? Jun 17 13:12:14 it's CSSU's fault! Jun 17 13:12:16 * Macer hides Jun 17 13:12:30 ShadowJK: i really can't now... i honestly don't even know where my n900 is.. it's in my house somewhere collecting dust Jun 17 13:12:51 but i noticed it resetting over and over in the logs Jun 17 13:12:53 i've never seen logs showing sim reader failing Jun 17 13:13:04 don't fall for macer, he's assuming it#s SIM holder though odds are it's usual modem chip defect Jun 17 13:13:13 Yeah I've only seen modem crashing. Jun 17 13:13:27 DocScrutinizer05: oh :-/ my bad.. figured it was a reader error Jun 17 13:14:04 well.. i think it's still under warranty if someone wants to buy it for $100.. it's a UK based phone tho so nokia will only accept it from the UK :-/ Jun 17 13:14:17 which totally sucks Jun 17 13:14:32 last time I checked, device bought in eu has eu wide warranty Jun 17 13:14:32 there is no way for me to ship it to the UK through their methods in order to honor the warranty Jun 17 13:14:46 ShadowJK: when i contacted nokia they said it was UK not EU Jun 17 13:14:50 (plus norway and switzerland, iirc) Jun 17 13:14:51 you should swap it for a XY-7 and sell that other shite, or keep it Jun 17 13:15:12 don't dis my lumia 900 :) Jun 17 13:15:19 it's actually a lot better than most people think Jun 17 13:15:25 you're just spoiled by the n900 heh Jun 17 13:15:40 Making distinction between "UK" and "EU" is weird anyhow, there's no EU version of the phone Jun 17 13:16:00 ShadowJK: i guess... either way... going from US to UK to send in is impossible Jun 17 13:16:15 it has to originate from the UK Post Jun 17 13:16:47 Well for sure both "UK" and "EU" aren't accepted by us nokia care Jun 17 13:16:51 i forgot the exact method but it has to be brought to the UK Post with a printed label.. it sucks. Jun 17 13:17:00 ShadowJK: nope :( Jun 17 13:17:11 free trade in action Jun 17 13:17:35 i would guess that nokia is nokia but i guess geological positions matter to them Jun 17 13:18:13 i mean it really doesn't matter to me.. i left tmobile anyways so i most likely wouldn't be using the n900 even if it did work. it was the only thing keeping me with tmobile Jun 17 13:18:19 So in other words, you never tried to excercise the eu wide warranty, and this entire screenful of debate was entirely pointless? Jun 17 13:18:20 so maybe its for the better ;) Jun 17 13:18:54 ShadowJK: i did.. i asked nokia if i could ship it overseas but they said that it would be impossible since you had to send it out using some UK post type of thing Jun 17 13:19:11 their answer was the phone originated from the UK and it would have to be shipped from the UK to nokia uk Jun 17 13:19:18 Ah, so you didn't Jun 17 13:19:26 that clears things up Jun 17 13:19:29 i tried.. but there isn't a way to go USPS to UK PO Jun 17 13:19:57 i can probably dig up the email with the rma instructions.. but it really is pointless.. i just wrote the n900 off Jun 17 13:20:03 as a defective device (2 of 2) Jun 17 13:20:16 Trying to send eu device from outside of eu to eu for service is same thing as trying to service eu device outside of eu Jun 17 13:20:16 my first was the usb port then this one the sim reader/modem Jun 17 13:20:42 ShadowJK: you can take a mercedez in to be repaired at a US based mercedez shop :-P Jun 17 13:20:54 That's their policy Jun 17 13:21:13 i guess.. so nokia's is regions are different Jun 17 13:21:21 But as for the original claim that a UK device has no warranty in the EU outside of UK, that's false Jun 17 13:21:31 it's not the EU tho Jun 17 13:21:40 nor did the EU ever come up in my conversation with customer support Jun 17 13:21:49 it's US for a UK originated phone Jun 17 13:22:06 and i was told it had to be sent via UK based PO Jun 17 13:22:24 there was no way to send it to them via USPS to their UK warranty center Jun 17 13:23:13 the EU/UK part is irrelevant Jun 17 13:23:31 given my circumstance Jun 17 13:23:54 so i guess you were right earlier.. it was pointless :) Jun 17 13:24:30 anyways... if anybody wants it let me know.. i'll take $100 for it.. maybe you can send it in or something and get somthing else Jun 17 13:24:55 otherwise i'll put it up on ebay and see if ther are any takers for my 2nd broken n900... NOKIA FTW Jun 17 13:25:12 lets see if the lumia lasts longer than 1 year :) Jun 17 16:30:18 <_PanzerSajt> ping freemangordon Jun 17 16:30:30 _PanzerSajt: pong Jun 17 16:30:33 <_PanzerSajt> hy Jun 17 16:30:36 hi Jun 17 16:30:41 <_PanzerSajt> So Jun 17 16:30:49 <_PanzerSajt> yes the kernel is woking fine Jun 17 16:30:54 but? Jun 17 16:31:05 <_PanzerSajt> but some modules are still missing Jun 17 16:31:12 for example? Jun 17 16:31:19 <_PanzerSajt> and some are present but nor working in the way we need Jun 17 16:31:43 do you have some list? Jun 17 16:31:45 <_PanzerSajt> here is a documentation how the stock kernel works Jun 17 16:31:46 <_PanzerSajt> http://natisbad.org/N900/n900-commented-hardware-specs.html Jun 17 16:32:04 <_PanzerSajt> and most of these modules are present but not working correctly Jun 17 16:32:13 <_PanzerSajt> for example the accelerometer Jun 17 16:33:13 <_PanzerSajt> ./sys/class/i2c-adapter/i2c-3/3-001d/coord file doesn't exist Jun 17 16:33:33 <_PanzerSajt> but the folders exist Jun 17 16:33:37 hmm, I see Jun 17 16:33:51 <_PanzerSajt> and the vibra also Jun 17 16:33:59 did anyone tried to build newer kernel? Jun 17 16:34:04 <_PanzerSajt> the led module is woking but not driving the vibra Jun 17 16:34:24 <_PanzerSajt> I think yes Jun 17 16:34:42 <_PanzerSajt> but I think there were problems with the size Jun 17 16:35:04 <_PanzerSajt> it was too big for the U-boot bootloader I think Jun 17 16:35:16 <_PanzerSajt> but if you could compile a newer kernel Jun 17 16:35:28 <_PanzerSajt> or at least provide me soma useful information Jun 17 16:35:52 _PanzerSajt: well, I think I could help, but cannot be the driving force behind that. Jun 17 16:36:09 if you have the questions, I could try to answer Jun 17 16:36:28 but as I see it, it is a matter of forward-poering the drivers Jun 17 16:36:33 *porting Jun 17 16:36:44 <_PanzerSajt> ok first how good are you in licensing? Jun 17 16:37:00 you mean are kernel drivers FOSS? Jun 17 16:37:13 <_PanzerSajt> yes Jun 17 16:37:14 yep, the whole kernel is GPLEed Jun 17 16:37:29 no problems with licensing for the kernel Jun 17 16:37:36 AFAIK Jun 17 16:39:34 <_PanzerSajt> and the kernel used in maemo contains only free drivers? Jun 17 16:39:46 <_PanzerSajt> for example these: http://mxr.maemo.org/fremantle/source/kernel/drivers/ Jun 17 16:39:49 yes, there is no other way Jun 17 16:40:48 <_PanzerSajt> and how can we compile these modules and use them later? Jun 17 16:41:06 <_PanzerSajt> just to point out: I have never compiled any kernel Jun 17 16:42:24 _PanzerSajt: you need to forward-port them. I.e. make them compilable and compatible with whatever changes there are between 2.6.28.10 ( maemo5 kernel) and the kernel version you are forward-porting to. Jun 17 16:44:15 <_PanzerSajt> currently we are using 2.6.37 Jun 17 16:44:59 <_PanzerSajt> so the reason why some drivers are not working correctly is that the newer kernel works in a different way? Jun 17 16:45:11 yes. and it is Nemo one actually. Jun 17 16:45:26 <_PanzerSajt> oh thanks Jun 17 16:45:54 <_PanzerSajt> then I think I should learn how linux kernel works Jun 17 16:46:02 _PanzerSajt: :D Jun 17 16:46:17 <_PanzerSajt> It's too complicated for me right now Jun 17 16:47:15 <_PanzerSajt> and what about the omap-kernel Jun 17 16:47:19 _PanzerSajt: you may start with setting up a build environment (SB2 should do the job) and cloning the 2.6.37 kernel tree (Nemo one) from here Jun 17 16:49:51 hmm, cannot find it right now, let me check in Ubuntu thread on TMO Jun 17 16:50:34 <_PanzerSajt> ok Jun 17 16:50:55 <_PanzerSajt> and what is the main difference between scratchbox and sb2? Jun 17 16:51:22 the way it works Jun 17 16:51:58 <_PanzerSajt> and I assume the sb2 is bettet, right? Jun 17 16:52:05 <_PanzerSajt> *better Jun 17 16:52:36 _PanzerSajt: here https://meego.gitorious.org/meego-device-adaptation/n900_kernel Jun 17 16:53:01 the difference is you wont be able to compile 2.6.37 with gcc 4.2.1 AFAIK Jun 17 16:57:00 <_PanzerSajt> and sb2 exists only in source code on git? Jun 17 17:00:10 _PanzerSajt: NFC. I used gcc 4.6.2 (the one I am using to build -thumb stuff) to build nemo kernel in SB1. But you should setup a new environmemnt in SB, read #maemo logs from yesterday for howto do it Jun 17 17:01:28 <_PanzerSajt> one question remains: how I can read the log from yesterday? Jun 17 17:03:23 <_PanzerSajt> sorry for these noob questions Jun 17 17:03:52 _PanzerSajt: http://mg.pov.lt/maemo-irclog/ Jun 17 17:06:34 _PanzerSajt: http://mg.pov.lt/maemo-irclog/%23maemo.2012-06-16.log.html#t2012-06-16T18:28:44 Jun 17 17:07:39 <_PanzerSajt> freemangordon, yes I have found thanks Jun 17 17:21:21 <_PanzerSajt> freemangordon, different topic, but can you help me how can I remap the keyboard layout on N900 with ubuntu so that pressing two specified key would emulate a TAB being pressed Jun 17 17:22:25 _PanzerSajt: sorry, cannot help. Jun 17 17:22:51 <_PanzerSajt> no problem Jun 17 17:23:21 <_PanzerSajt> so now I have installed the -core and -libs package Jun 17 17:25:27 <_PanzerSajt> freemangordon, nevermind I will try it later Jun 17 17:25:33 <_PanzerSajt> and thanks for your help Jun 17 17:26:57 np Jun 17 17:48:18 _PanzerSajt: ctrl-i traditionally is meant to mean Jun 17 17:49:12 <_PanzerSajt> DocScrutinizer, oh really thanks Jun 17 17:50:39 of course this will only work on char level, not on keycode level Jun 17 17:50:56 i.e usually on any line discipline Jun 17 17:51:26 <_PanzerSajt> my initial purpose was to use it in terminal Jun 17 17:51:32 <_PanzerSajt> and it works great Jun 17 17:51:48 apps that scan the keycodes will not recognize ctrl-i as Jun 17 17:51:59 <_PanzerSajt> ok that's fine Jun 17 17:52:12 yes, shell is the prototypical case where it should work Jun 17 17:53:07 except for fullscreen apps, like e.g. htop, which may or may not work with ctr-i as tab-substituate Jun 17 17:53:16 <_PanzerSajt> since we don't have any tab key on N900 and I can't live without it I was always ssh-ing just to have TAB Jun 17 17:54:06 well, you got a true TAB key on the xterm softkeys Jun 17 17:54:41 (at least I thought it's there by default) Jun 17 17:55:32 <_PanzerSajt> yeah but I'm using ubuntu on N900 right now Jun 17 17:55:43 aah, yes. Forgot Jun 17 17:56:00 <_PanzerSajt> finally we managed to get dsp-working Jun 17 17:56:31 <_PanzerSajt> now I'm searching for a graphical player that can use it Jun 17 17:59:12 well, I'm no friend of the approach to come from a bare bones standard system like ubuntu or open-embedded, and then try to add in all the stuff needed to support the platform. I prefer to come from an already working "board support package" like maemo fremantle and add-in/replace the stuff I'm missing and can cherrypick on e.g ubuntu Jun 17 18:03:56 <_PanzerSajt> yes but the problem for me that in mamemo some packages are outdated and the whole system is customized so it is not so simple to run your desired app Jun 17 18:08:06 well, it seems way simpler than making your desired distro replace maemo Jun 17 18:08:32 outdated packages can get updated (something we do constantly over at CSSU) Jun 17 18:11:03 while even for open components in maemo there are often inferior upstream implementations, esp when it comes to kernel driver modules etc Jun 17 18:13:07 <_PanzerSajt> yes that's right Jun 17 18:13:33 <_PanzerSajt> currently the thumb-2 cssu looks very interesting Jun 17 18:14:38 nevertheless it breaks ABI compatibility so you can't expect cssu-thumb to play nice with standard apps Jun 17 18:15:00 DocScrutinizer05: tend to disagree Jun 17 18:15:09 no ABI is broken Jun 17 18:15:37 you can execute your ARM stuff in thumb2 environment Jun 17 18:15:47 if you manage to compile a kernel that's loading your random custom closed source module binary just fine, I'd agree Jun 17 18:16:25 if you can't, it's just as incompatible as powerkernel is Jun 17 18:16:31 <_PanzerSajt> with ubuntu we have armhf kernel with thumb2 enabled Jun 17 18:18:05 DocScrutinizer05: aah, since when Nokia customised omap1 kernel is the standard for ABI? Jun 17 18:18:39 well, ABI maybe not the exact term. OK Jun 17 18:19:07 anyway, I am in gdb again, I'd better focus on it :) Jun 17 18:20:51 anyway as long as apps like BlessN900 camera app come with drivers that don't work under that patched kernel, I think that's a compatibility issue Jun 17 18:21:09 and the issue is with kernel, not with the app Jun 17 18:21:30 and nobody knows how many apps are out there that do stuff like that Jun 17 18:22:10 well, for somebody running ubuntu rather than maemo on N900 it's probably massively irrelevant Jun 17 18:22:45 <_PanzerSajt> yes but I'm still interested in everything in connection with n900 Jun 17 18:28:45 <_PanzerSajt> do you know any other smartphone or tablet that runs linux and it can be considered as modern (max 2 years old) Jun 17 18:29:25 some think android==linux Jun 17 18:30:09 <_PanzerSajt> no android is just based on linux Jun 17 18:30:27 <_PanzerSajt> it has some differences in kernel also Jun 17 18:30:31 tell that my coleagues at work ;-D Jun 17 18:30:44 <_PanzerSajt> but basically it is just a virtual machine Jun 17 18:31:58 at my work actually linux means android Jun 17 18:32:15 nobody ever says android there Jun 17 18:32:33 <_PanzerSajt> you work on a wrong place :) Jun 17 18:32:46 sometimes I think you're right ;-) Jun 17 18:36:10 well, usually I don't touch android either at work (no linux either), but recently that changed, since they "heard I know a bit about linux" (SIC!) Jun 17 18:36:55 DocScrutinizer05: you shouldn't let slip that you know a bit about linux Jun 17 18:37:00 you know a bit about *proper* linux ;) Jun 17 18:37:06 for all that matters in that case, it's actually a correct proposition Jun 17 18:38:11 but now I'm feeling constant nausea while makeing friend (or trying to do so) with android build environment at large Jun 17 18:51:35 DocScrutinizer05: Euro cup's been going well for you guys :P Jun 17 18:52:22 merlin1991: soccer? Jun 17 18:52:29 err MohammadAG^^^ Jun 17 18:52:53 It's called football Jun 17 18:53:00 see? Jun 17 18:53:03 Americans can call theirs handball Jun 17 18:53:59 I'm honestly not very much excited. What happened to make to think it's going well? Jun 17 18:54:26 s/to think/you think/ Jun 17 18:54:27 DocScrutinizer05 meant: I'm honestly not very much excited. What happened to make you think it's going well? Jun 17 18:56:30 You won/tied all games Jun 17 18:56:40 mhm Jun 17 18:56:41 There's a game right now, 0-0 Jun 17 18:56:50 ooh there is? Jun 17 18:56:55 * DocScrutinizer05 zaps a bit Jun 17 18:57:04 You and Denmark Jun 17 18:57:36 indeed Jun 17 18:57:50 * DocScrutinizer05 zaps back to NCIS Jun 17 18:58:27 * MohammadAG wants ncis Jun 17 19:04:03 ...Wait, what? Is Estel's email seriously "twilight312@gmail.com" or am I misreading something? Jun 17 19:08:42 DocScrutinizer05: You scored :p Jun 17 19:11:18 And Denmark scored back Jun 17 19:20:51 robbiethe1st: it is :/ Jun 17 19:20:58 XD Jun 17 19:21:20 I find that incredibly funny, for some reason Jun 17 19:47:36 Hello. I have an N900. I try to connect it to a bluetooth GPS (TomTom, SirfstarIII). It never locks, and eventually gives a connection failed message. Jun 17 19:47:49 Google makes me think I'm the only one with this problem. Jun 17 19:48:00 Does anyone else use an external GPS with their N900? Jun 17 19:48:50 I haven't, sorry Jun 17 20:44:42 ~seen pali Jun 17 20:44:48 pali <~pali@unaffiliated/pali> was last seen on IRC in channel #maemo-ssu, 9d 20m 2s ago, saying: 'DocScrutinizer, what do you think about it?'. Jun 17 20:44:59 dang Jun 17 20:45:53 anybody else got a script to do proper backup of $maemo to a local PC? Jun 17 20:47:16 where $maemo=the repos and other relevant stuff Jun 17 20:47:36 maybe even plus wiki Jun 17 20:47:53 DocScrutinizer05, yet another game won by you :P Jun 17 20:48:04 not by me :-) Jun 17 20:48:53 honestly I'd prefer to do some serious stuff, rather than watch football Jun 17 20:50:06 doesn't hurt to have it in the background :P Jun 17 20:50:20 NCIS been the better background Jun 17 20:51:21 but I just zapped and switchet to videotext, when some dudes outside shouted "...SCHLAAAAND" Jun 17 20:58:19 my favorite part of fight club is when he is in his bi-monthly sickle cell circle Jun 17 20:58:22 and he's the only white guy in the whole room Jun 17 21:31:54 MohammadAG: it does hurt! everytime someone hits a football god kills a kitten! Jun 17 21:34:04 who cares about football anyway? Jun 17 21:35:43 +1 Jun 17 21:36:52 chem|st: you happen to have a scriptie to mirror the repos? Jun 17 21:38:15 DocScrutinizer05: nope as they never setup rsync as requested 2.6 years ago Jun 17 21:38:26 indeed Jun 17 21:38:28 DocScrutinizer05: but well could write one.. Jun 17 21:38:45 want a mirror? Jun 17 21:38:54 DocScrutinizer05: I think I still have the script itself on my server Jun 17 21:38:54 there's that stuff from pali, but... Jun 17 21:39:06 so they're shutting down maemo.org, as I hear Jun 17 21:39:31 DocScrutinizer05: found it in my ff history Jun 17 21:39:39 http://pastebin.com/4GVQVZxL Jun 17 21:39:42 http://pastebin.com/SWeDqgZJ Jun 17 21:39:58 apply the 2nd to the ubuntu tarball of apt-mirror Jun 17 21:40:20 thanks Jun 17 21:40:29 but sorry, what ubuntu tarball? Jun 17 21:41:40 Venemo_N9: McDonalds and Calsberg's beer apparently Jun 17 21:42:52 DocScrutinizer05: I think this should do: http://archive.ubuntu.com/ubuntu/pool/universe/a/apt-mirror/apt-mirror_0.4.8.orig.tar.gz Jun 17 21:45:11 hmm, changelog and TODO, no manpage, no readme, no friggin nuttin Jun 17 21:45:36 merlin1991: why not just provide a patched .deb? Jun 17 21:45:43 * DocScrutinizer05 is temped to type apt-mirror_0.4.8.orig.tar.gz --help Jun 17 21:45:44 chem|st: because I have none? Jun 17 21:46:11 all I have is the diff and list in my ff history and a patched script somewhere on my server :D Jun 17 21:46:40 and some background what that stuff is meant to do, while I don't Jun 17 21:46:58 apt-mirror is a perl script that mirrors an apt repo Jun 17 21:47:08 DocScrutinizer05: for apt-mirror cfg you have http://wiki.maemo.org/User:Jebba/Mirror Jun 17 21:47:14 pali patched it to send the maemo specific headers Jun 17 21:47:23 sudo mirror-a-repo-for-me Jun 17 21:47:50 no, he didn't Jun 17 21:48:02 merlin1991: erm maemo.nokia.com isn't maemo.org Jun 17 21:48:17 [jr@lagrange scripts]$ ./download.sh Jun 17 21:48:19 chem|st: it mirros sdk tool nokia binaries and nokia ssu repos Jun 17 21:48:19 --2012-06-17 23:15:31-- https://downloads.maemo.nokia.com/fremantle/ssu/apps///Release Jun 17 21:48:28 2012-06-17 23:15:31 FEHLER 403: Forbidden. Jun 17 21:48:32 ... :-/ Jun 17 21:49:05 merlin1991: will they be gone too? Jun 17 21:49:24 possibly I can't know Jun 17 21:49:42 DocScrutinizer05: untar that tar, apply the patch and read the first 50 lines of it Jun 17 21:49:43 ffs Jun 17 21:50:19 I did that on 3 or 4 of pali's tarball Jun 17 21:50:28 preparations for doomsday after latest Nokia's exploits, eh? Jun 17 21:50:30 veeery enlightening Jun 17 21:50:45 DocScrutinizer05: do that for the apt-mirror from debian + patch from pastebin i posted Jun 17 21:50:51 because that stuff actually works Jun 17 21:51:00 see above, yeah Jun 17 21:55:06 merlin1991: patch this patch apt-mirror? Jun 17 21:55:11 yes Jun 17 21:55:47 so I just install apt-mirror from repo and patch is as it is perl?! Jun 17 21:56:02 that could work too Jun 17 21:56:20 I only got the tarball because it does not apply cleanly to my distrib apt-mirror Jun 17 21:56:22 well I mirror repo.maemo.org an then I'll try Jun 17 21:56:40 Estel_: yes Jun 17 21:56:59 chem|st: for maemo.org you could also use http://maemo.merlin1991.at/apt-mirror/ Jun 17 21:57:15 * chem|st wonders were his /ignore is gone.... Jun 17 21:57:57 merlin1991:have you all or just fremantle? Jun 17 21:58:06 just fremantle Jun 17 21:58:24 but if you give me the proper deb and deb-src lines I can add the others aswell Jun 17 21:58:57 just replace fremantle with diablo and the others... Jun 17 22:00:17 for extras it is diablo gregale chinook bora scirocco & mistral Jun 17 22:00:30 merlin1991: ^^^ Jun 17 22:01:12 huh Jun 17 22:01:19 how old are those Jun 17 22:01:23 in all that patches and tarball I fail to spot a single "maemo" Jun 17 22:01:33 I'm wondering whether anyone's using them Jun 17 22:01:41 DocScrutinizer05: because it's a friggin generic mirror this apt repo tool Jun 17 22:01:48 have a look at http://repository.maemo.org/pool/ Jun 17 22:01:49 the patch adds nokia(tm) headers Jun 17 22:02:03 and you supply it with the list of repos from the first pastebin to get the stuff Jun 17 22:02:23 uhuh, suply like in? Jun 17 22:02:37 you know, there's no manpage Jun 17 22:02:47 drop them in a file and find the comman line parameter that is in the first few lines of the script Jun 17 22:02:55 there is a friggin embedded helptext there Jun 17 22:02:57 and I have no friggin clue how this stuff is meant to be used Jun 17 22:03:17 coooool Jun 17 22:03:20 where? Jun 17 22:04:01 everything from the beginning till =head1 CURRENT AUTHORS Jun 17 22:04:17 in which file? Jun 17 22:04:25 DocScrutinizer05: you patch the apt-mirror file and download the OTHER then you edit the path in OTHER and start 'apt-mirror $PATH_TO_OTHER' Jun 17 22:04:40 aaaah Jun 17 22:05:19 DocScrutinizer05: and make path somewhere +60GB Jun 17 22:06:03 oh it is with sources... make it 120GB Jun 17 22:06:25 and with all versions.. are you guys crazy? Jun 17 22:06:29 extras devel + extras + cssu + cssu testing (including sources) is 109 gb Jun 17 22:06:30 path like "set base_path /var/spool/apt-mirror" ? Jun 17 22:06:34 who needs that obsolete stuff? Jun 17 22:06:39 yes Jun 17 22:06:48 sdk + tools + all ssu repos is 19gb Jun 17 22:07:03 130GB even... Jun 17 22:07:27 gasp Jun 17 22:07:52 ok, RIP mirror-idea Jun 17 22:08:05 merlin1991: so you have everything but nokia? Jun 17 22:08:25 chem|st: I have everything but pre fremantle stuff Jun 17 22:08:36 and I have to save the sb rootstraps and packages Jun 17 22:08:56 merlin1991: so why do I need to mirror it again? Jun 17 22:09:06 I have nfc Jun 17 22:09:07 do you have an rsync-mirror setup? Jun 17 22:09:13 nope Jun 17 22:09:23 and the nokia stuff isn't public because (C) Jun 17 22:09:33 would you mind or is this killing your traffic? Jun 17 22:09:34 who cares? Jun 17 22:10:12 chem|st: my traffic would be fine, only $nokia lawyer scares me slightly Jun 17 22:10:32 my traffic doesn't matter, but my storage is waaay to low Jun 17 22:10:40 where there is no lawyer for maemo as there is no more staff in $maemo Jun 17 22:11:05 what did the lawyer say? Jun 17 22:11:06 DocScrutinizer05: got plenty so no worries Jun 17 22:11:31 Venemo_N9: he's drowned lately Jun 17 22:12:44 anybody without any prior or current Nokia device here? no? so what's the issue with all of us downloading Nokia(C) stuff? Jun 17 22:13:37 you guys have no problem Jun 17 22:13:41 I mean, if Nokia wouldn't want us to download it, they wouldn't offer it for download, no? Jun 17 22:13:44 I on the other hand am distributing it Jun 17 22:14:17 that's not what I asked for ;-) Jun 17 22:15:00 (though it seems there are some private copies of nok stuff on my server which I accidentally haven't set -r) Jun 17 22:15:24 of course I will fix thjat as soon as Nokia complains Jun 17 22:16:09 for now I guess it's only been a prank somebody played on me, as I can't find any such stuff Jun 17 22:16:24 nah, good night guys Jun 17 22:16:52 but as mentioned before, only ~30GB free storage on that box Jun 17 22:17:32 * DocScrutinizer05 should get a phatter server Jun 17 22:18:53 a goddamn 130GB? incredible Jun 17 22:20:25 DocScrutinizer05: running... Jun 17 22:20:40 merlin1991: my folder looks somehow weird Jun 17 22:21:02 ... Jun 17 22:21:42 chem|st: you gotta be a bit more specific ;) Jun 17 22:21:51 i know Jun 17 22:22:00 Proceed indexes: [SSSSSSSSSSSSPsh: 1: cannot open downloads.maemo.nokia.com/fremantle/apps///.//Packages.gz: No such file Jun 17 22:22:03 apt-mirror: can't open index in proceed_index_gz at /usr/bin/apt-mirror line 458. Jun 17 22:22:16 same I got Jun 17 22:22:24 see above Jun 17 22:22:31 I have no clue about perl... Jun 17 22:22:33 though with pali's scripts Jun 17 22:22:42 +1 Jun 17 22:22:51 chem|st: and you applied the patch? Jun 17 22:23:26 sure, and only two lines got rejected as they were the same Jun 17 22:24:44 hm Jun 17 22:27:58 http://paste.debian.net/175018/ Jun 17 22:29:46 I think I spotted some magic in one of the scripts to extract ssl keys from somewhere Jun 17 22:30:07 which would suggest this script menat to run *on n900* Jun 17 22:30:17 DocScrutinizer05: found it, missing @args in 216 Jun 17 22:30:21 or some Jun 17 22:30:50 otoh last line of same script been > find /somewhere -name foobar -exec rm {} ; Jun 17 22:30:54 merlin1991: 157.8GB Jun 17 22:31:04 and that doesn't work with messybox as we all know Jun 17 22:33:30 chem|st: that includes what? Jun 17 22:34:46 merlin1991: that includes the mirror.list you posted Jun 17 22:35:01 * merlin1991 wonders why he only has 19 gb Jun 17 22:35:34 cause you already got 130 from maemo... Jun 17 22:36:33 ah yeah Jun 17 22:36:43 because it has extras extras-testing and extras-devel Jun 17 22:36:51 I only copied the ssu sdk and tools repos Jun 17 22:36:51 EEEEK Jun 17 22:37:00 DocScrutinizer05: what? Jun 17 22:37:09 and to -devel + extras independent Jun 17 22:37:18 at least testing might die unnoticed behind a hill Jun 17 22:38:14 AIUI there's hardly any relevant version in tetsing, no? Jun 17 22:38:44 DocScrutinizer05: ehrm for about a year only testing holds relevant versions^^ Jun 17 22:38:55 umm Jun 17 22:39:11 as noone pushed anyhting Jun 17 22:39:21 till last week Jun 17 22:39:28 I somehow thought that's -devel Jun 17 22:39:55 devel is like killing my device if I upgrade from it Jun 17 22:40:15 apt-get upgrade is a bad idea anyway xD Jun 17 22:40:17 sure, but the versions in devel and testing usually should be same Jun 17 22:40:50 50% of my post-flashed-cssu_installed software is or was from testing Jun 17 22:41:03 if they're not, then I'd rather forget about testing than devel Jun 17 22:41:23 got like 30 upgrades if a set devel active in ham Jun 17 22:41:55 well, anyway seems there's plenty of room for symlinks, hardlinks, and bzip2 Jun 17 22:42:55 yes Jun 17 22:43:11 there's OTOH plenty of arguments why a decent server should have at least 1TB storage ;-D Jun 17 22:44:03 well, this silly thing I got been planned as a bouncer, nuttin else Jun 17 22:44:14 for that the 50GB are plenty Jun 17 22:45:15 I again start to consider getting a decent one for hosting a few maemo related things Jun 17 22:45:18 I download it and will shove it up to uni-ftp if I get permission, at least for extras* Jun 17 22:45:34 good plan Jun 17 22:46:05 the admin just didn't want to download 400GB if only 10GB upload will happen ever Jun 17 22:46:07 the proprietary crap might already fit on those 30GB free Jun 17 22:46:29 DocScrutinizer05: 19GB prop. Jun 17 22:46:30 DocScrutinizer05: the proprietary crap is 16 gb Jun 17 22:46:42 chem|st: I checked again it's actually only 16 Jun 17 22:47:00 k Jun 17 22:47:53 chem|st: hetzner? Jun 17 22:48:28 me Jun 17 22:48:32 ? Jun 17 22:48:37 your box Jun 17 22:48:41 no I got no hetzner server Jun 17 22:48:47 :nod: Jun 17 22:49:07 one is amazon the other two are selfhosted Jun 17 22:49:31 SELFhosted? dang, with FDDI backbone? Jun 17 22:50:14 misleading sry Jun 17 22:50:22 hosted but my own hardware Jun 17 22:50:27 aaah Jun 17 22:50:31 colo Jun 17 22:50:33 one is down atm Jun 17 22:50:51 one is mining bitcoins 24/7 Jun 17 22:50:57 hehe Jun 17 22:51:02 I love powerflat! Jun 17 22:55:34 chem|st: so could you share the fixed script, so eventually I could run it when I got a server with more sotrage? Jun 17 22:55:54 or strip down the list to exclude the fat stuff Jun 17 22:56:30 i have a mirror Jun 17 22:56:34 for maemo repo Jun 17 22:56:45 oooh, and anyody got an idea how to mirror wiki? Jun 17 22:56:49 http://maemo.jacekowski.org/mirror/repository.maemo.org/ Jun 17 22:56:58 DocScrutinizer05: you can download whole DB Jun 17 22:57:01 ahhh Jun 17 22:57:03 that wiki Jun 17 22:57:03 nfc how to mirror wiki properly Jun 17 22:57:10 jacekowski: how? Jun 17 22:57:12 wget -r? Jun 17 22:57:19 jacekowski: it's getting a popular game now to get more mirrors Jun 17 22:57:33 merlin1991: normal wikipedia publishes their db dumps daily Jun 17 22:57:42 :D Jun 17 22:57:56 gotta contact nemain then :/ Jun 17 22:58:08 indeed Jun 17 22:58:17 wget is no brilliant idea Jun 17 22:58:31 http://dumps.wikimedia.org/backup-index.html Jun 17 22:58:37 you get all special pages, which adds up to a *100 explosion Jun 17 22:59:35 iirc it's around 1TB in total Jun 17 22:59:42 compressed Jun 17 23:00:34 but plain text articles are under 10G compressed Jun 17 23:09:57 jacekowski: the wiki.maemo.org? Jun 17 23:10:27 damn that shouldn't be more than 10GB, maybe 20 Jun 17 23:11:25 hmm, with all edit history it might be some GB actually Jun 17 23:14:58 http://wiki.maemo.org/index.php?title=Special:Allpages&from=100Days plus http://wiki.maemo.org/index.php?title=Special:Allpages&from=Maemo_Summit_2009/Schedule/Day_3 should convert to a wget script easily Jun 17 23:15:50 and shouldn't be that much actually Jun 17 23:16:03 DocScrutinizer05: i was talking about wikipedia.org Jun 17 23:16:30 can we ask X-Fade for db dumps? Jun 17 23:16:39 ~x-fade Jun 17 23:16:48 sure Jun 17 23:17:10 asking is for free ;-D Jun 17 23:18:14 btw the above linked two pages don't contain the /user:* pages it seems Jun 17 23:18:15 DocScrutinizer05: patch brakes with adding @args in line 54 Jun 17 23:19:04 my system is now load 31 31 28... ram 16GB used swap 8GB used... is it need be wget? Jun 17 23:20:13 wget shouldn't be using that much Jun 17 23:20:27 what are you running exactly Jun 17 23:20:40 90% wget in ram Jun 17 23:20:49 apt-mirror Jun 17 23:20:58 chem|st: does this script start a new background shell for each wget? ;-) Jun 17 23:21:56 (would probably run out of TCP ports eventually) Jun 17 23:22:45 no but it is 3 wgets with --no-cache... Jun 17 23:23:00 what's --no-cache? Jun 17 23:23:59 umm, that's server side Jun 17 23:24:11 while your server is the client on that wget Jun 17 23:24:54 a nd I wonder if that's actually needed (no-cache) Jun 17 23:25:12 cannot write the line as I have no more response Jun 17 23:25:40 ?? o.O Jun 17 23:25:54 is rendered useless for 22h as it writes with 2MB/s to disk Jun 17 23:26:10 why? I have no clue Jun 17 23:26:54 vserver? Jun 17 23:28:17 no proper 4core 16GB ram 4TB-hdd Jun 17 23:31:38 ok retry.... cannot be it can it! Jun 17 23:32:06 just exited anything writing or reading from that disk *fingers-crossed* Jun 17 23:32:28 check smart Jun 17 23:32:39 and dmesg Jun 17 23:33:31 tweak vfs Jun 17 23:34:02 jacekowski: how exactly? Jun 17 23:34:57 http://pastebin.com/1tT3E79y Jun 17 23:35:00 that's what i've got Jun 17 23:35:35 total used free shared buffers cached Jun 17 23:35:35 Mem: 24164 23936 227 0 0 18897 Jun 17 23:35:35 -/+ buffers/cache: 5039 19125 Jun 17 23:35:57 this laptop/distri killing me with those random 10..100s lags on starting any app, particularly browser Jun 17 23:36:17 get ssd Jun 17 23:36:20 and windows Jun 17 23:36:36 i never thought my laptop can start so fast untill i bought an ssd Jun 17 23:36:39 a ssd* Jun 17 23:37:05 jacekowski: increase or decrease? Jun 17 23:37:19 well, I think there's not even any read from HDD involved when those lags occur Jun 17 23:37:32 chem|st: some are increased some are decreased Jun 17 23:37:40 jacekowski: I mean vfs Jun 17 23:37:59 i niced it for now... 40% ram usage Jun 17 23:38:06 look at my paste Jun 17 23:38:14 http://pastebin.com/1tT3E79y Jun 17 23:38:23 oooh you run that as root maybe? Jun 17 23:38:24 first 4 are the important ones Jun 17 23:38:37 DocScrutinizer05: what? Jun 17 23:38:38 chem|st: ^^^ Jun 17 23:38:54 jacekowski: apt-mirror Jun 17 23:39:29 DocScrutinizer05: did not want to run as user^^ Jun 17 23:39:35 lmc Jun 17 23:39:49 there's no excuse for a proper 4core box writing to a 4TB drive @ 2MB/s Jun 17 23:40:21 i that's pretty fast Jun 17 23:40:53 hm? Jun 17 23:40:56 15krpm drive can deliver 1M/s for fully random writes Jun 17 23:41:35 anyways, i've got my own apt-mirror setup Jun 17 23:41:40 if anybody is interested Jun 17 23:41:54 definitely yes Jun 17 23:42:09 http://maemo.jacekowski.org/mirror/ Jun 17 23:43:04 err, I thought about the setup, not the mirror :-D Jun 17 23:43:21 ahm Jun 17 23:43:29 well, Jun 17 23:43:56 http://paste.ubuntu.com/1046446/ Jun 17 23:44:01 that's my setup file Jun 17 23:44:11 apt-mirror /home/maemo/mirror.list Jun 17 23:44:15 that's all it takes to run it Jun 17 23:44:23 jacekowski: you're assuming 1 seek per rev Jun 17 23:44:38 jacekowski: As I understad it, ost drives can't quite do that. Jun 17 23:45:02 Actually - no drives can do that - it's physically impossible. Jun 17 23:45:13 well, drives can seek pretty fast Jun 17 23:45:18 Sure. Jun 17 23:45:27 and home drives can deliver around 80 iops Jun 17 23:45:32 wait what? you're talking about writing one random block ber rotation? Jun 17 23:45:35 15k drives can deliver over 2 of that Jun 17 23:45:35 per* Jun 17 23:45:43 jacekowski: Sure. Jun 17 23:45:56 NCQ helps a bit as well Jun 17 23:46:07 jacekowski: but that's not 250iops/s Jun 17 23:46:30 And reordering helps a little, yes. Jun 17 23:47:05 dang, on writing new data out of cache, reordering should help quite a lot Jun 17 23:47:17 jacekowski: something else was hogging it so it was 'random writes' Jun 17 23:47:44 30MB/s Jun 17 23:47:46 chem|st: tweaking vfs will help a lot Jun 17 23:47:53 chem|st: so it waits bit more to write it Jun 17 23:48:04 so it can order it a lot better Jun 17 23:48:22 with 16GB used up for buffers I'd think it already waits long enough Jun 17 23:48:24 so 100 isnt a good idea? Jun 17 23:49:19 i'm talking about all those numbers Jun 17 23:49:24 not only cache pressure Jun 17 23:50:00 dirty_background_ratio dirty_expire_centisecs dirty_ratio dirty_writeback_centisecs Jun 17 23:50:03 [18.06.2012 01:19:03] my system is now load 31 31 28... ram 16GB used swap 8GB used... is it need be wget? Jun 17 23:50:23 those 4 increased will help a lot Jun 17 23:50:38 help a lot on what exactly? Jun 17 23:50:46 with write performance on random writes Jun 17 23:50:54 however it will use more ram at the same time Jun 17 23:51:05 AHA Jun 17 23:51:13 see last quote Jun 17 23:51:22 ram 16GB used Jun 17 23:51:26 jacekowski: ram is gone by then... Jun 17 23:51:53 [18.06.2012 01:48:20] with 16GB used up for buffers I'd think it already waits long enough Jun 17 23:51:56 didn't notice the swap part Jun 17 23:52:21 atm it is 8GB ram and 200MB swap... without nice Jun 17 23:52:26 au contraire I think it swaps out too aggressive Jun 17 23:52:40 but I think it will take only a few more minutes^^ Jun 17 23:52:57 just a bit longer than before... Jun 17 23:53:13 9GB ram Jun 17 23:53:29 used by? Jun 17 23:53:46 buffers, cache...? Jun 17 23:53:55 50% by wget so7.x GB Jun 17 23:54:10 53% Jun 17 23:54:18 dafaq, wget can't use 50% of 9GB Jun 17 23:54:23 54 Jun 17 23:54:32 50%of 16GB Jun 17 23:54:49 55% of 16GB Jun 17 23:55:02 unless you invoke one wget instance to download one or several files summing up to >>9GB Jun 17 23:55:29 seems like that is what it is doing Jun 17 23:55:36 OMG Jun 17 23:56:11 wget per letter but max 3 at a time^^ Jun 17 23:56:32 ? Jun 17 23:56:33 61% by 2 of those 3 wgets Jun 17 23:56:43 well, 3 wgets then Jun 17 23:56:45 per letter in /pool Jun 17 23:56:51 does PID change? Jun 17 23:57:11 no Jun 17 23:57:26 this job is a brainfsck Jun 17 23:57:30 and the lowset PID is the smallest in RAM Jun 17 23:58:25 you shouldn't download 1M files with 150GB total size via one wget process Jun 17 23:59:09 now I seem to see what was the *intention* of --no-cache Jun 17 23:59:25 just --no-cache doesn't do what the intention been Jun 18 00:02:08 my 2 cent: wget mallocs new RAM for each chunk of data downloaded, and tries to write in background. It however seems to never reuse the allocated RAM Jun 18 00:02:22 and never flushes Jun 18 00:02:40 so if your download is faster than your disk write, you run into trouble Jun 18 00:04:02 unless you force flushing every now and then, by invoking a new instance of wget for the next file Jun 18 00:06:41 chem|st: does wget diagnostic output on terminal? Jun 18 00:06:53 if yes, a ^S might help Jun 18 00:08:06 or you throttle your NIC somehow Jun 18 00:09:13 or you refactor the job, so each file download has to finish writing to disk before next one gets started - I.E. invoke a new instance of wget for each file Jun 18 00:10:07 nthread to 2 and it writes with 8MB/s Jun 18 00:10:37 sorry? Jun 18 00:10:39 ok 5-8MB/s Jun 18 00:11:04 nthread? Jun 18 00:11:06 the apt-mirror list has a nthreads of 20 Jun 18 00:11:47 68MB ram Jun 18 00:12:13 wit 2 instead of 20 Jun 18 00:12:27 you lost me Jun 18 00:12:57 mirror.list has a value of 20 threads I reduced it to 2 Jun 18 00:13:27 it will take ages... but it wont kill the machine Jun 18 00:13:44 aaah there's nthreads Jun 18 00:13:50 whatever that does Jun 18 00:14:55 it distributes the indexing into threads, seems like stacking Jun 18 00:15:08 hmm Jun 18 00:15:51 nfc what's indexing here Jun 18 00:16:03 fileindex Jun 18 00:16:18 I thought the problem is with wget Jun 18 00:16:23 eating all the RAM Jun 18 00:16:39 t is what apt-mirror is doing with it Jun 18 00:18:16 didn't you say wget using 50%...100% of RAM, and wget PID never changes? Jun 18 00:18:24 yes Jun 18 00:18:33 then that's your problem Jun 18 00:18:46 well that too Jun 18 00:19:32 I need to get some rest Jun 18 00:19:37 there's one (or 3) wget instances running from a pipe with list of files to download, and it stuffs all the download to malloc'd RAM Jun 18 00:20:13 and hands that RAM to write() Jun 18 00:21:02 now your aggressive swappiness or whatever kicks in and maybe swaps out wget to make space for disk buffers/cache Jun 18 00:21:45 which isn't exactly helping for thruput to HDD of the downloaded stuff in wget's memory Jun 18 00:22:27 the job shouldn't run one wget instance form a command pipe but rather invoke a new wget instance for each file Jun 18 00:23:48 since then wget would quit only after flushing all buffers to write to disk, and thus the download speed gets throttled naturally Jun 18 00:24:04 to match the max HDD write speed Jun 18 00:24:57 this job is optimized for NIC thruput, assuming there can't ever be a bottleneck on the write-to-storage side Jun 18 00:25:31 when your download speed is higher than your write speed on HDD, it will fail Jun 18 00:25:54 which isn't exactly helping for thruput to HDD of the downloaded stuff in wget's memoryasically a flaw in wget itself Jun 18 00:26:06 oops Jun 18 00:26:36 basically a flaw in wegt itself **** ENDING LOGGING AT Mon Jun 18 02:59:58 2012