2014-03-01 11:37:35 clandmeter they were here before, last one i posted is with newer version 2014-03-01 11:45:56 crow: i didnt have one yet afaik 2014-03-01 11:48:36 it would be to catch this with gdb but it seems grsec is there before gdb.. [425933.118058] grsec: From XXXXXX: Segmentation fault occurred at 000000d8 in /usr/bin/oscam[wr05-hdplus:6964] uid/euid:1000/1000 gid/egid:27/27, parent /usr/bin/oscam[oscam:6619] uid/euid:1000/1000 gid/egid:27/27 2014-03-01 11:52:53 crow: can you replicate it? 2014-03-01 11:55:57 not realy 2014-03-01 12:28:09 crow: did you try edge-musl? 2014-03-01 12:55:09 clandmeter no i did not, i am on 2.7.4 2014-03-01 15:55:45 clandmeter you are on edge-muslß 2014-03-02 05:16:10 hm 2014-03-02 05:16:14 does alpine's iptables not have --comment 2014-03-02 05:17:40 oh 2014-03-02 05:17:42 I was doing it wrong 2014-03-02 05:17:42 nvm 2014-03-02 11:18:32 clandmeter regarding oscam.conf nice option, it seems it cant be disabled from webif (as default value is 99) commenting it while oscam not running should disable it. 2014-03-02 11:18:50 http://www.streamboard.tv/wiki/OSCam/en/Config/oscam.conf#nice 2014-03-04 14:32:03 "apk fix linux-grsec" used to regenerate a new initramfs (with LUKS included) - now this gives "(1/1) [APK unavailable, skipped] Reinstalling linux-grsec (3.10.28-r0)" ?? 2014-03-04 15:11:02 apk update ? 2014-03-04 15:11:25 maybe apk version says that some package is not being upgraded? 2014-03-04 15:17:20 it was caused my manually installed openvswitch packages 2014-03-04 15:17:58 removing ovs & running apk fix linux-grsec worked 2014-03-04 15:19:44 I'm going to revise the openvswitch patch - i've done a bit more testing & amended the init scripts so it now starts before networking - LXC guests now autostart on boot 2014-03-04 15:21:04 my /etc/apk/world had these entries from manually installed apk's: 2014-03-04 15:21:11 openvswitch-grsec> openvswitch> openvswitch-doc> & was what stopped "apk fix" working 2014-03-04 15:43:00 ACTION curiously runs apk fix and stuff to see if it'll fix the xenpv boot 2014-03-04 15:44:48 no :) 2014-03-04 15:45:04 at least i have hvm with serial console now, that's pretty comfortable 2014-03-04 15:51:16 i think i found the issue 2014-03-04 15:51:17 LINUX vmlinuz-3.10.30-0-grsec 2014-03-04 15:51:33 (from update-extlinux 6.02-r0) 2014-03-04 15:51:45 vs. KERNEL vmlinuz-3.10.25-0-grsec (from update-extlinux 5.10-r0) 2014-03-04 15:58:59 python code rule #1: no error handling 2014-03-04 15:59:05 python code rule #2: no documentation 2014-03-04 15:59:16 smart people won't need it and others shouldn't touch it 2014-03-04 15:59:31 let's just assume it has no errors, just edge cases 2014-03-04 16:53:10 clandmeter today i got new segfault, and grsec message, this time RLIMIT_CORE insted RLIMIT_NICE (as I put nice to default value): '[495455.760354] grsec: From IP_ADDRESS: denied resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 for /usr/bin/oscam[wr05-orf:18568] uid/euid:1000/1000 gid/egid:27/27, parent /usr/bin/oscam[oscam:1106] uid/euid:1000/1000 gid/egid:27/27' 2014-03-04 21:37:49 crow: can youtry ulimit -c unlimited and generate a core dump? 2014-03-04 22:54:44 crow: i added debug support to oscam 2014-03-04 22:55:10 apk add oscam-dbg 2014-03-04 22:55:31 and you can read here about splitted debug symbols 2014-03-04 22:55:42 http://www.technovelty.org/code/separate-debug-info.html 2014-03-04 22:55:44 http://www.technovelty.org/code/split-debugging-info-symbols.html 2014-03-05 07:43:03 morning 2014-03-05 07:43:16 basically, apk add oscam-dbg 2014-03-05 07:43:29 gdb --core core /usr/bin/oscam 2014-03-05 07:43:32 then type 'bt' 2014-03-05 07:43:37 for a backtrace 2014-03-05 10:44:24 clandmeter ncopa thank you, will try these 2014-03-05 10:54:24 is there a way that apk shows me where it did put files from an pacakges? (or what files are with that packages)? 2014-03-05 10:54:44 apk info -L 2014-03-05 11:04:36 ncopa thank you that indeed work, now how did you came to -L ? i checked apk help and that switch is not there or i am blind. 2014-03-05 11:07:52 apk info --help 2014-03-05 11:08:08 clandmeter sees that bin is not ok: /usr/lib/debug/usr/bin/oscam.debug --help give me this error oscam.debug: line 1: syntax error: unexpected "(" , same switch --help works fine on oscam package 2014-03-05 11:08:29 :) 2014-03-05 11:08:33 its not a bin 2014-03-05 11:09:04 those are the debug symbols and stuff 2014-03-05 11:09:11 its seperated from the binary 2014-03-05 11:09:32 clandmeter so it do owerwrite bin from oscam packages. 2014-03-05 11:09:42 no 2014-03-05 11:09:55 i think gdb will find them automatically 2014-03-05 11:10:02 like ncopa explained 2014-03-05 11:10:16 well he didnt really explain it 2014-03-05 11:10:22 but said how y ou should go about 2014-03-05 11:10:22 aha 2014-03-05 11:10:42 i never used it like this before 2014-03-05 11:10:51 clandmeter should not oscam packages be on same rev as this debug symbols? 2014-03-05 11:11:06 we (ncopa) added this to alpine recently 2014-03-05 11:11:11 clandmeter and thank you for this hint about info --help (every day something new) 2014-03-05 11:11:39 yes, but i guess they should 2014-03-05 11:11:51 apk update && apk upgrade should make them the same version 2014-03-05 11:12:12 this means the oscam need to be start with option to generate core, as i did not see any generated core. 2014-03-05 11:12:22 they are always generated from the same apkbuild 2014-03-05 11:12:32 ulimit -c unlimited 2014-03-05 11:12:40 that allows the core to be created 2014-03-05 11:13:27 but then file need to be startet manualy or it work also over init.d ? 2014-03-05 11:13:44 it should always work 2014-03-05 11:13:52 im not sure where it puts it then 2014-03-05 11:14:03 but you can find / -name core 2014-03-05 11:14:30 and if really needed, just run it manually 2014-03-05 11:19:01 crow: this seems like some usefull core infomration: http://www.moythreads.com/wordpress/2011/10/06/linux-core-dumps/ 2014-03-05 11:19:14 [66324.924634] serve_process[18250]: segfault at 0 ip 1143ec12 sp 4bfe6960 error 4 in oscam (deleted)[113e6000+fd000] 2014-03-05 11:19:31 i am not sure when did that happend 2014-03-05 11:19:55 as this (time after boot is not something i can say when that segfault did happend) 2014-03-05 11:23:04 If i get it correctly from that page, i need to make change in init script, 'If you are starting the process from an init script or something like that, the init script has to do it.' 2014-03-05 11:37:25 or set the core location manuall 2014-03-05 12:25:04 clandmet1r i did set it manuall, as on that page, and did run oscam -b from this shell where i executed ulimit -c unlimited 2014-03-05 12:26:17 and you got a core? 2014-03-05 12:39:21 still no segfault so i need to wait. 2014-03-05 12:39:32 i cant reproduce it myself 2014-03-05 14:46:13 i'm getting this while compiling libboost: g++: internal compiler error: Killed (program cc1plus) 2014-03-05 14:46:19 anyone found that problem? 2014-03-05 14:46:47 mljsimone: anything in dmesg? 2014-03-05 14:46:56 mljsimone, sounds like out-of-memory 2014-03-05 14:47:06 boost requires insane amount of memory to build 2014-03-05 14:47:10 damn 2014-03-05 14:47:26 memory problem :P 2014-03-05 14:47:36 Out of memory: Kill process 5612 (cc1plus) score 837 or sacrifice child 2014-03-05 14:48:00 yeah. been there. 2014-03-05 14:48:09 what to do now ummm 2014-03-05 14:48:10 my arm builder with only 2GB of physical ram did not make it 2014-03-05 14:48:13 needed to enable swap 2014-03-05 14:50:32 did you, by any change, set -jX? 2014-03-05 14:52:44 yes, 2014-03-05 14:52:46 to four 2014-03-05 14:56:13 mmm I don't think that with ~450mb of memory i'm going to build it. 2014-03-05 16:19:49 ncopa: are you around? 2014-03-05 16:22:57 most probably he already left office. 2014-03-05 16:30:28 oh 2014-03-05 16:31:19 tomorrow morning you'll be the first one 2014-03-05 16:31:23 9CET 2014-03-05 16:33:50 clandmet1r: do you know if there is someone else I can talk to related to db/db-c++ packages? 2014-03-05 16:34:42 you can send an email to our ML or use forums. most of them are gone. 2014-03-05 16:34:57 ML? 2014-03-05 18:37:21 mljsimone mailing list http://lists.alpinelinux.org/ 2014-03-05 18:37:49 oh yeah! I forgot about that 2014-03-05 18:38:02 crow: beer++ 2014-03-05 18:40:45 np 2014-03-06 07:06:24 mljsimone: hi 2014-03-06 07:26:24 clandmet1r this mornin arround 1 AM, was an segfault 2014-03-06 07:26:35 let me prepare what i find out. 2014-03-06 07:59:06 ncopa clandmet1r here i put all together, 2014-03-06 07:59:07 http://gw.gd/u4W0 2014-03-06 08:00:18 what weird is, is this this 'Reading symbols from /usr/bin/oscam...Reading symbols from /usr/lib/debug/usr/bin/oscam.debug...(no debugging symbols found)...done.' even file is there (see bottom of page) 2014-03-06 08:00:27 yeah 2014-03-06 08:00:31 that looks strange 2014-03-06 08:01:05 what is oscam? 2014-03-06 08:01:31 card reader software? 2014-03-06 08:01:35 what is it used for? 2014-03-06 08:01:48 yea card reader software http://www.streamboard.tv/wiki/OSCam/en/What_is_OSCam 2014-03-06 08:02:00 The most important: OSCam is Open Source 2014-03-06 08:02:01 heh 2014-03-06 08:02:07 that does not tell me much :) 2014-03-06 08:02:44 it some kind of card server and client? 2014-03-06 08:02:49 creditcards? 2014-03-06 08:02:50 client and server 2014-03-06 08:02:52 memory cards? 2014-03-06 08:02:54 no dvb 2014-03-06 08:03:02 well you could share your subscribe card in your home, let say one resiver with card could share it card to other resiver in your sleepingroom without taking the card and putting there 2014-03-06 08:03:18 tv decoder cards? 2014-03-06 08:03:25 yes 2014-03-06 08:03:46 Comman acces module 2014-03-06 08:03:49 that it is about tv decoding is apparently not so important (but that it is opensource is...) 2014-03-06 08:03:50 :) 2014-03-06 08:04:02 ncopa ofcourse that is most important :) 2014-03-06 08:04:07 i like the idee 2014-03-06 08:04:20 well there have been many closed source solutions 2014-03-06 08:04:34 thats why the emphesise opensource now i guess 2014-03-06 08:04:37 they should put on the "wht it is" that it is about tv decoding 2014-03-06 08:04:57 and not about memory cards, credt cards or ID cards 2014-03-06 08:05:01 or poker cards... 2014-03-06 08:06:22 seems like gdb does not find the symbols 2014-03-06 08:06:33 and it seems like it is related threading 2014-03-06 08:06:37 it might be a bug in uclibc 2014-03-06 08:06:44 or it might be a bug in oscam 2014-03-06 08:06:47 ncopa i am not sure what should be there, but if you have some stuff to add to that wiki, maybe you could :) 2014-03-06 08:07:16 yeah its a wiki... (and suddenly i regret i said anything... :) 2014-03-06 08:07:16 from new corner perspective 2014-03-06 08:08:05 Diese Seite enthält momentan noch keinen Text, du bist auch nicht berechtigt diese Seite zu erstellen. Du kannst ihren Titel auf anderen Seiten suchen oder die zugehörigen Logbücher betrachten. 2014-03-06 08:08:14 i suppose i could google translate it 2014-03-06 08:08:31 but at this point i kinda lost the interest of contribute... :) 2014-03-06 08:08:40 page not found 2014-03-06 08:09:22 ah dDiskussion wiki page 2014-03-06 08:09:24 so, 2 things. the -dbg package appears to be broken 2014-03-06 08:09:42 and either oscam or libpthread in uclibc is broke 2014-03-06 08:16:45 crow: are you sure those gdb is the same version as regular pkg? 2014-03-06 08:16:48 oscam packages is compiled with debug option: 'Debug mode: yes' 2014-03-06 08:17:25 clandmet1r well there is no new version avaiable in alpinelinux edge 2014-03-06 08:17:51 i mean, they are the same version? oscam and oscam-dbg? 2014-03-06 08:18:00 i remember you said something about it 2014-03-06 08:18:06 and you are mixing repo's afaik 2014-03-06 08:20:13 i have 2 versions of oscam runing and never got them to crash on me. 2014-03-06 08:20:21 are you doing anything special with it? 2014-03-06 08:21:01 clandmeter nothing special, start over init.d, running and serving like it should 2014-03-06 08:21:22 x86? 2014-03-06 08:21:50 clandmeter yes, ALIX.3D3 hw (not much resources but it is working) 2014-03-06 08:22:17 ok, thats probably the difference from me. 2014-03-06 08:22:41 although i have it had running on alix for a long time before 2014-03-06 08:23:39 ncopa: how can i check if symbols are actually present? 2014-03-06 08:24:10 maybe oscam strips it themsels 2014-03-06 08:24:47 clandmeter regarding missing repositorys, yes i use your oscam from edge, but rest ist from stable /etc/apk/repository: http://mirror1.hs-esslingen.de/pub/Mirrors/alpine/v2.7/main and @edge http://nl.alpinelinux.org/alpine/edge/main 2014-03-06 08:25:23 missing== mixing 2014-03-06 08:25:58 apk info oscam shows me both version avaiable (stalbe and edge) how to know which one is installed 2014-03-06 08:26:14 apk info only shows installed pacakges 2014-03-06 08:26:30 apk info |grep oscam 2014-03-06 08:26:36 and check if both have same version 2014-03-06 08:27:07 apk info -v |grep oscam 2014-03-06 08:28:52 clandmeter http://gw.gd/TbDi 2014-03-06 08:29:04 apk info oscam shows me both version 2014-03-06 08:29:22 its ok 2014-03-06 08:31:47 so the version are ok i guess 2014-03-06 08:32:03 yes 2014-03-06 08:32:08 well version nr are same. 2014-03-06 08:32:23 you mention it before so i wanted to be sure 2014-03-06 08:35:09 clandmeter: readelf i suppose 2014-03-06 08:35:21 there should be a reference to the symbols in the elf header 2014-03-06 08:35:38 yes, could be that ocaml strips things 2014-03-06 08:36:13 http://sprunge.us/eLEh 2014-03-06 08:36:18 does that say anything? 2014-03-06 08:36:35 i guess i need to run readelf on the debug file 2014-03-06 08:47:04 i think oscam builds a seperate debug version 2014-03-06 08:47:12 oscam.debug 2014-03-06 08:49:47 ncopa: should i overide dbg and create a package with oscam.debug? or just create a debug subpkg and not strip it? 2014-03-06 08:51:30 overide i mean put oscam.debug in dbg pkg 2014-03-06 08:54:31 .gnu_debuglink 2014-03-06 08:54:46 i think a .gnu_debuglink section should be visible 2014-03-06 09:05:43 clandmeter readelf is part of elfutils? 2014-03-06 09:51:17 clandmeter if i compile oscam manuell on some other host i see this 2014-03-06 09:51:18 LINK Distribution/oscam-1.20-unstable_svn9564-x86_64-unknown-linux-gnu.debug 2014-03-06 09:51:18 STRIP Distribution/oscam-1.20-unstable_svn9564-x86_64-unknown-linux-gnu 2014-03-06 10:45:37 clandmeter readelf on both binarys is identical, but .debug is 4MB vs 1MB 2014-03-06 12:44:19 new alix board avaiable http://www.pcengines.ch/apu1c4.htm 2014-03-06 13:02:46 crow, interesting 2014-03-06 13:13:14 with this PSU http://www.pcengines.ch/ac12veur2.htm and Enclosure http://www.pcengines.ch/case1d2redu.htm and mSATA http://www.pcengines.ch/msata16a.htm optional http://www.pcengines.ch/wle200nx.htm 2014-03-06 13:13:19 fabian_a indeed 2014-03-06 23:29:50 hello.. can anyone help me running Apache as worker mpm? I uncomened HTTPD=/usr/sbin/httpd.worker in conf.d and comented out HTTPD=/usr/sbin/httpd but when I restart apache it says: /lib/rc/sh/runscript.sh: line 313: /usr/sbin/httpd.worker: not found 2014-03-06 23:30:15 and there is no httpd.worker in /usr/sbin only httpd 2014-03-07 09:23:24 amarc_:thats a bug 2014-03-07 22:24:09 is there any reason Apparmor is not in the repos ? (thinking of building it to work with LXC) 2014-03-08 08:44:52 clandmeter yesterday i did have again an segfault, this this there was no core,because i did reboot this machine an ulimit -a was not set up. Still this core which is generate i cant use as debug symbols are not there. do you have maybe any more idee what to do do get debug symbols work 2014-03-08 09:13:30 crow: can you build it local? 2014-03-08 09:13:36 then try to use oscam.debug 2014-03-08 10:13:26 i could try, yes 2014-03-08 11:20:14 I could add that to the default build, but that doesnt seem right. 2014-03-08 15:44:15 clandmeter well i understand what you mean with that, but then -dbg should somehow work with this one? 2014-03-08 15:46:23 and as on this alix i dont have much spece left, i will need to setup a vm with alpine to be able to build own packages 2014-03-09 23:19:43 Hi, anyone can help me on zol (zfs on linux) using edge ? I've been able to install and load all modules but can not find the zfs and zpool commands ! 2014-03-09 23:29:50 hi 2014-03-09 23:29:56 isnt there a zfs-utils pkg? 2014-03-09 23:30:18 nope 2014-03-09 23:30:49 hmm 2014-03-09 23:30:54 i make one before 2014-03-09 23:30:57 api search zfs -> report zfs-grsec and zfs-grsec-dev only 2014-03-09 23:31:24 maybe an issue on the repo I'm using ! let me check 2014-03-09 23:32:52 found it 2014-03-09 23:33:01 it has been disabled in zfs-grsec 2014-03-09 23:34:26 im trying to build it 2014-03-09 23:34:32 do you know why ? cause it is mandatory to be able to use the modules available ;-) 2014-03-09 23:35:25 probably because they didnt build 2014-03-09 23:35:51 make sense :) 2014-03-09 23:36:02 http://git.alpinelinux.org/cgit/aports/commit/testing/zfs-grsec/APKBUILD?id=c5cfd319b05aeed1e4a299cc134ec5a3628ebafa 2014-03-09 23:38:15 damn, that configure script is awful slow 2014-03-09 23:38:27 ok will track this one then 2014-03-09 23:38:53 yes it is : I tried ti build from source : no success 2014-03-09 23:39:38 looks like its fine 2014-03-09 23:41:42 configure or make ? 2014-03-09 23:42:14 package 2014-03-09 23:42:16 :) 2014-03-09 23:42:29 ouch ok :-) 2014-03-09 23:42:56 http://git.alpinelinux.org/cgit/aports/commit/?id=ec3d4bff3a8a9f0729ef619e081e37f873630fb5 2014-03-09 23:46:46 Yada: apk update && apk upgrade && apk add zfs-grsec-utils 2014-03-09 23:47:03 checking 2014-03-09 23:48:10 if you dont use the nl mirror, its probably not yet available 2014-03-09 23:52:08 yep : added it and all good, thank you very much : gonna keep going on ZoL and this great Alpine distro 2014-03-09 23:52:48 enjoy 2014-03-09 23:52:54 and lets us know if it works 2014-03-09 23:53:15 we like to know the status of packages in testing repo, so we can move them to main. 2014-03-09 23:54:51 sure I'll keep you guys posted and happy to help 2014-03-09 23:56:54 same if you guys plan to update the OVS pkg ;-) the one available (v1.11) is from Aug 2013 2014-03-09 23:57:54 ovs? 2014-03-09 23:58:28 openvswitch 2014-03-09 23:59:01 based on git : look like ncopa is on it (v1.7.3) 2014-03-10 00:00:04 but api search report 1.11 only and current version is 1.9.3 (LTS) or 2.0.1 for the most recent 2014-03-10 00:00:29 i remember there was talk about it 2014-03-10 00:00:35 not sure what was up with it 2014-03-10 00:00:50 its not normal for pkgs holding back that long 2014-03-10 00:01:05 +1 2014-03-10 00:03:51 it builds just fine 2014-03-10 00:04:21 yes should be 2014-03-10 00:12:30 Yada: there is some talk about it on our ML 2014-03-10 00:13:09 ok will check it out 2014-03-10 00:13:10 http://lists.alpinelinux.org/alpine-devel/3569.html 2014-03-10 17:04:55 ACTION is backing up some ubuntu vm (old management vm) so it can be reborn with alpine 2014-03-10 17:05:15 fabian_a: just found your blog, saw you (used to) use cobbler too :) 2014-03-11 00:04:46 i'm trying to configure the serial console for the kernel but also for the bootmenu 2014-03-11 00:04:55 http://wiki.alpinelinux.org/wiki/Enable_Serial_Console_on_Boot 2014-03-11 00:05:30 the update-extlinux.conf seems to have nothing that lets you set the SERIAL 0 115200 settings in the extlinux.conf 2014-03-11 02:20:15 ansible can't really add a user by itself, it needs useradd. is there some useradd clone or package for alpine? 2014-03-11 03:09:07 darkfader: try createuser or adduser? 2014-03-11 07:42:58 kaniini: do you know if its possible to set serial console with update-extlinux ^^^ 2014-03-11 07:43:29 darkfader: coudl you please create a ticket for it (feature request) otherwise bugs.alpinelinux.org 2014-03-11 11:45:22 ncopa: ok 2014-03-11 12:09:09 ncopa: i'm not sure why it wouldn't be 2014-03-11 12:09:26 ncopa: there's an area for extra commandline options 2014-03-11 12:10:14 kaniini: the missing bit is the config for extlinux itself 2014-03-11 12:10:20 the "SERIAL 0 9600" bit 2014-03-11 12:10:28 its not a kernel command line 2014-03-11 12:11:43 hmm it's a shell script 2014-03-11 12:11:45 *peeks* 2014-03-11 12:15:07 seems it doesn't know about serial config, is it a part of syslinux or a part of alpine? 2014-03-11 12:15:19 oh, it says Alpine in there for sure 2014-03-11 12:15:21 ;) 2014-03-11 12:15:39 ncopa: i'll open the bug and describe it 2014-03-11 12:18:25 good, thanks! 2014-03-11 12:19:12 oh, okay 2014-03-11 12:19:22 it should be possible to add 2014-03-11 13:24:45 Is it possible to run alpine-xen on macbook air? and then do vms 2014-03-11 14:48:16 how can i reconfigure the timezone? 2014-03-11 15:26:02 kocka: http://wiki.alpinelinux.org/wiki/FAQ#How_do_I_set_the_local_timezone.3F 2014-03-11 15:27:22 thanks 2014-03-11 19:08:19 kocka: setup-timezone [-z UTC | -z America/New_York | -p EST+5] 2014-03-11 19:08:31 kocka: http://wiki.alpinelinux.org/wiki/Alpine_setup_scripts#setup-timezone 2014-03-12 11:57:06 unbelievable, three weeks holiday passed... 2014-03-12 11:57:26 back to work... 2014-03-12 13:55:37 StarWarsFan: wb! 2014-03-12 13:59:06 ACTION spent a nice honeymoon... 2014-03-12 13:59:09 ncopa: thx 2014-03-12 13:59:23 honeymoon! congrats! 2014-03-12 13:59:52 thx a lot 2014-03-12 14:49:01 congrats! 2014-03-12 19:45:07 StarWarsFan: congrats 2014-03-12 19:45:28 thx@all :-) 2014-03-12 21:10:05 n8@all 2014-03-12 21:16:39 congrats StarWarsFan|afk 2014-03-13 07:31:27 Moinmoin 2014-03-13 09:56:22 Morning all 2014-03-13 11:12:08 good afternoon 2014-03-13 13:11:07 Does Alpine runs on the Raspberry pi? 2014-03-13 13:20:34 mljsimone: Linux rpi 3.11.4 #2 Thu Oct 10 13:06:33 GMT 2013 armv6l Linux 2014-03-13 13:20:59 cool :) 2014-03-13 13:21:39 but its not official 2014-03-13 13:22:33 mljsimone: http://dev.alpinelinux.org/~tteras/rpi-with-networking.tar 2014-03-13 13:23:14 what is that? the img? 2014-03-13 13:23:18 copy it to a primary fat partition on the sd 2014-03-13 13:23:35 ok 2014-03-13 13:23:57 repo: http://nl.alpinelinux.org/alpine/edge-musl/main/ 2014-03-14 07:55:45 hi! 2014-03-14 07:56:44 anyone ever tried to virtualise al x86_64 on a i686 win machine with virtualBox (or any other vm host sw)? 2014-03-14 08:09:51 cool, got it :P 2014-03-14 13:05:27 to build an package for edge, the guest system need to be also on edge? 2014-03-14 13:05:45 masked in: @edge 2014-03-14 13:07:05 crow: yes, that way it's linked against the right libs 2014-03-14 13:50:50 jbilyk_ ok, then i need to switch to it. 2014-03-14 17:29:27 where is package located after building with abuild -r ? 2014-03-14 17:31:34 current directory 2014-03-14 18:03:52 jbilyk actually not, it is in my home directory same level as aports , there is an directory packages 2014-03-14 18:15:12 crow: /me usually just runs abuild -r within aports tree, and package gets dumped into working directory 2014-03-14 18:15:20 so ymmv 2014-03-14 18:15:22 :) 2014-03-14 18:19:20 jbilyk can you give me one example from that ? 2014-03-14 18:34:16 crow: sure - I ssh jbilyk@buildbox && cd ~/aports/main/kamailio && abuild -r 2014-03-14 18:34:52 that gives me ~/aports/main/kamailio/kamailio-4.1.0-r0.apk 2014-03-14 18:34:58 and a bunch of subpackages 2014-03-14 18:35:32 only thing custom in /etc/abuild.conf is "PACKAGER="Jeff Bilyk "" 2014-03-14 18:43:42 does one of you run netconsole? 2014-03-14 18:43:52 i thought about adding it everywhere, for years 2014-03-14 18:50:19 ACTION is doing straight ssh -> BB shell 2014-03-17 08:56:06 crow: im not sure about pax protection, maybe we should disable some of its protection in oscam. 2014-03-17 08:56:50 ncopa: any advise about pax? 2014-03-17 09:21:09 hi 2014-03-17 09:21:31 thats just a warning 2014-03-17 09:21:34 for gdb 2014-03-17 09:21:46 i thnk you should be able to get a backtrace anyways 2014-03-17 09:22:09 otherwise: https://wiki.gentoo.org/wiki/Hardened/Debugging 2014-03-17 10:06:52 ncopa if it is jut a warning it is ok, but in dmesg, i see a lot of these now when runing oscam in gdb. 2014-03-17 10:07:34 maybe it works if you run it in soft mode? 2014-03-17 10:08:35 echo 1 > /proc/sys/kernel/pax/softmode 2014-03-17 10:49:03 i will try 2014-03-17 10:52:02 ncopa is reboot needed? 2014-03-17 11:32:37 no i dont think so 2014-03-17 11:32:42 that would even clear it again 2014-03-17 12:14:31 yes no reboot 2014-03-17 15:03:36 Hi All, is there any official way how to run a kernel with KLIPS support? (Because of openswan) Google was not helpful this time. Thanks. 2014-03-17 15:35:16 x00r: i think we have KLIPS in our default kernel? 2014-03-17 15:35:46 i expect it to just work? 2014-03-17 16:20:59 ncopa: OpenSWAN was complaining about not having KLIPS and desperatly fallbacking to Netkey... But i may have made a mistake and probably i was booting an xen dom0. I will boot a deafult alpine system and test it! Thanks for the info. 2014-03-17 17:12:38 i am on kde and i would like the power button to work regardless of kde. i disactivated kde power buttton handling and now pressing it does nothing. yet i would like systemd to catch the signal and poweroff. how to? 2014-03-17 21:29:33 anyone ever booted alpine via pxe? 2014-03-17 21:45:28 clandmeter befor 2 hourse there was an segfault, no core was created, even i run oscam inside gdb. 2014-03-17 21:49:01 clandmeter log http://gw.gd/MohS 2014-03-17 21:59:37 clandmeter ncopa segfault which could not be loged to gdb http://gw.gd/Gp77 2014-03-17 22:02:04 and gdb-start and run files: http://gw.gd/UJp6 2014-03-17 22:14:56 fbettag: maybe a good saturday project, but still busy on the coming weekends 2014-03-17 22:58:55 darkfaded: i think i got it 2014-03-17 22:59:09 i added mtdblock and memdiskfind to my ramdisk 2014-03-17 22:59:14 wioll see in a few 2014-03-18 07:30:11 crow: i think it forks 2014-03-18 07:30:19 and that the child segfaults 2014-03-18 07:30:29 but 2014-03-18 07:30:47 [ 114.522915] PAX: From YYYYYYY: execution attempt in: , 51dc2000-51dc5000 51dc2000 2014-03-18 07:31:18 that could indicate that it has a JIT 2014-03-18 07:32:58 crow: what happens if you paxctl -c -m /usr/bin/oscam ? 2014-03-18 07:33:06 will it still segfault? 2014-03-18 08:03:16 ncopa let me try it and start gdb with oscam again 2014-03-18 08:06:13 try without gdb first 2014-03-18 08:06:42 if it has a JIT and that is the problem, then that might be the problem 2014-03-18 08:07:06 ncopa file /root/bin/oscam.debug had a PT_GNU_STACK program header, converted 2014-03-18 08:08:28 ncopa i only saw that in dmesg when i use gdb 2014-03-18 08:08:43 ok 2014-03-18 08:09:01 but i start it now over init.d 2014-03-18 08:09:46 ncopa des this with paxctl is reboot safe? 2014-03-18 08:10:02 yes, as long as the binary is on disk 2014-03-18 08:10:08 it modifies the file itself 2014-03-18 08:10:37 ok thanks 2014-03-18 08:11:26 is this "denied resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 for /usr/bin/gdb[gdb:1272] " because of from bash gdb-start script 'ulimit -n 4096' 2014-03-18 08:30:12 crow: thats possible 2014-03-18 08:30:25 you can set it to unlimited of you start it afaik 2014-03-18 08:30:40 s/of/after 2014-03-18 08:32:27 crow: JIT *would* make sence. 2014-03-18 08:32:57 that would explain the differences between yours crashing and mine not (i dont use grsec) 2014-03-18 08:34:19 that reminds me i need to upgrade my kernel and switch back to grsec again. 2014-03-18 08:36:11 lol, that irc logger gets on google fast. searching for oscam and jit results in this converstation... 2014-03-18 11:29:10 clandmeter yes i saw that irc logger to. well do upgrade so we can test if you have such stuff also. 2014-03-18 11:29:25 i changed ulimit -c unlimited instead that ulimit -n 4096 2014-03-18 11:29:38 i dont know what JIT is, let me google for it 2014-03-18 11:48:42 I created group "Alpine Linux" on Linkedin.: http://lnkd.in/bFuMCtD 2014-03-18 12:05:37 i noticed :) 2014-03-18 12:06:00 :) 2014-03-18 14:32:09 clandmeter what for is JIT, 2014-03-18 16:43:27 crow: http://wiki.gentoo.org/wiki/Hardened/FAQ#Why_is_the_jit_and_orc_flag_disabled_in_the_hardened_profile.3F 2014-03-18 22:27:57 Hello , can anyone point me on how to create a alpine-linux bootable usb that runs custom pkgs? ( preconfigured ) 2014-03-19 06:09:23 hi, http://wiki.alpinelinux.org/wiki/Configure_Networking means stateful DHCPv6 not supported by alpine linux? :-( 2014-03-19 06:09:28 *out of the box 2014-03-19 06:12:43 iface eth0 inet6 dhcp -> ifup/ifdown: unknown method "dhcp" 2014-03-19 06:22:41 anyone? 2014-03-19 09:36:21 crow: fix the crashing? 2014-03-19 09:49:48 clandmeter well one day was ok, still need to see if it will crash, (it was not able to reproduce it) 2014-03-19 09:50:12 nice 2014-03-19 09:50:23 i think you had crashed daily right? 2014-03-19 09:51:26 clandmeter yes it was daily, i dont see any grsec or PAX messages in dmesg. But like i wrote before, PAX was only when i use gdb. 2014-03-19 10:01:57 i still can't get stateful dhcpv6 working... this took me long enough, back to windows server... thanks for wasting my time alpine-linux 2014-03-19 10:08:04 haha 2014-03-19 10:08:20 ncopa: its all your fault! 2014-03-19 10:08:53 i suppose it is 2014-03-19 10:09:05 windows server sounds like a good alternative :) 2014-03-19 10:09:11 did he mention he wanted his money back? 2014-03-19 10:09:33 he will probably sue us soon 2014-03-19 10:09:48 we should start hire some lawers 2014-03-19 10:10:17 i'd rather spend time on fixing ipv6.... :) 2014-03-19 10:10:40 ok, maybe it was better for all that he went back to windows server 2014-03-19 10:12:37 http://www.rjsystems.nl/en/2100-dhcpv6-stateful-autocfg.php 2014-03-19 10:12:40 he can try to configure it with his metro interface ;-) 2014-03-19 10:13:02 to be honest 2014-03-19 10:13:13 i fell a bit bad that i simply dont have the capacity to help all 2014-03-19 10:13:17 i wish i could 2014-03-19 10:13:30 but thats reality... 2014-03-19 10:15:00 also i dont have ipv6 myself 2014-03-19 10:15:30 there are bigger issues in this world, so i wouldnt worry about these that much. 2014-03-19 10:15:37 true 2014-03-19 10:16:01 right now the bigger issue is musl and 3.0... 2014-03-19 10:17:59 what's not working with ipv6? 2014-03-19 10:18:09 i still can't get stateful dhcpv6 working... this took me long enough, back to windows server... thanks for wasting my time alpine-linux 2014-03-19 10:18:20 hi, http://wiki.alpinelinux.org/wiki/Configure_Networking means stateful DHCPv6 not supported by alpine linux? :-( 2014-03-19 10:18:20 *out of the box 2014-03-19 10:18:20 iface eth0 inet6 dhcp -> ifup/ifdown: unknown method "dhcp" 2014-03-19 10:18:20 anyone? 2014-03-19 10:18:41 Ah. I'm not using any dhcp for ipv6, though I do need to set it up at home. 2014-03-19 10:18:53 since I've got a /56 in colo I could tunnel a /64 to the house easy enough 2014-03-19 10:19:01 i found this: http://www.rjsystems.nl/en/2100-dhcpv6-stateful-autocfg.php 2014-03-19 10:19:07 we might want figure out how to do it 2014-03-19 10:19:19 i don't know if dhcpcd has support for it 2014-03-19 10:19:24 it might be it has 2014-03-19 10:20:58 I think it was my alpine boxes that'd stop responding over ipv6 after about a month unless I rebooted... might have been something else though. 2014-03-19 10:24:33 but it has been a few updates since then too. 2014-03-19 18:45:34 <_lb_> Hi, What is the alpine recommended way of deploying BGP? 2014-03-19 18:57:04 i dunno about the "alpine recommended", but I've used quagga's bgpd and it works quite well 2014-03-19 18:57:35 <_lb_> jbilyk_: it scares me a lot 2014-03-19 18:57:39 <_lb_> my scenario... 2014-03-19 18:57:40 ? 2014-03-19 18:57:53 <_lb_> We don't support BGP... But we are going to do this for a customer 2014-03-19 18:57:59 <_lb_> So... We have this 44 nic boxes... 2014-03-19 18:58:08 <_lb_> each nic with independent networking 2014-03-19 18:58:28 <_lb_> I am afraid quagga can affect the rest of the connections 2014-03-19 18:59:58 it can :) 2014-03-19 19:00:48 <_lb_> ACTION wonders if he could run quagga in a virtual server/lxc... 2014-03-19 19:01:56 <_lb_> just to keep the bgp isolated 2014-03-19 19:41:52 didn't know existed routing software like cisco ios and junos 2014-03-19 19:42:21 how do you fit 44 nics? 2014-03-19 19:44:09 <_lb_> Frosh: we built the boxes 2014-03-19 19:44:14 <_lb_> we also have a 120 nics one 2014-03-19 19:44:54 anywhere I can look at the price? 2014-03-19 19:45:37 <_lb_> If you are interested... You can contact our manufacturer... 2014-03-19 19:46:06 How well does it run? and what are the specs on the system? 2014-03-19 19:47:03 <_lb_> Frosh: Perfect uptime running alpine 2014-03-19 19:49:20 any kind of vlan joke is not in place now, right? 2014-03-19 19:57:44 darkfaded: that thought occurred to me too 2014-03-19 19:59:11 well, isolation, i guess :) 2014-03-19 19:59:26 no competing for throughput? _lb_? 2014-03-19 20:08:05 <_lb_> darkfaded: competing for througput? 2014-03-19 20:08:56 _lb_: what's the actual reason for using 120 nics instead of vlans 2014-03-19 20:09:06 i was just trying to figure it out 2014-03-19 20:09:13 <_lb_> darkfaded: 120 different topological endpoints 2014-03-19 20:09:22 <_lb_> 120 different networks 2014-03-19 20:09:38 <_lb_> darkfaded: some of those connections also have vlans ;) 2014-03-20 16:20:55 Hi Guys, I'm thinking of installing alpine-server on a Enterprise environment and I'm reluctant on the point that (with what I'v searched) there is no LTS version 2014-03-20 16:21:15 what's do you think? 2014-03-20 16:23:03 there is 2014-03-20 16:23:45 all major version is supported for a long term 2014-03-20 16:28:19 Arditi: iirc, each stable release is supported for 18months 2014-03-20 16:28:31 with a stable release every 6 months 2014-03-20 16:31:33 thanks jbilyk 2014-03-20 16:37:44 hmm 2014-03-20 19:22:18 Hi 2014-03-20 19:25:55 we are developping with python3 on Alpine-2.8 and we have small but annoying problems with time method in python, something does not woriking well in Alpine (maybe the uclibc), here is a simple python test to illustrate the issue : 2014-03-20 19:26:05 import time 2014-03-20 19:26:16 t = (2009, 2, 17, 17, 3, 38, 1, 48, 0) 2014-03-20 19:26:33 t = time.mktime(t) 2014-03-20 19:26:46 print("time : ", time.strftime("%b %d %Y %H:%M:%S", time.gmtime(t))) 2014-03-20 19:27:24 which work well on any distro/python version and give nothing on Alpine 2014-03-20 19:29:34 any help ? 2014-03-20 20:12:28 clandmeter segfault :( 2014-03-20 20:17:49 <_lb_> Francois__: Does it crash or is it simply not returning anything? 2014-03-20 20:18:23 clandmeter running oscam normaly not in gdb. http://gw.gd/dfbF 2014-03-20 20:19:26 now i started again the oscam.debug (self compiled) inside gdb and this is start output. http://paste.ie/view/6358999c 2014-03-20 20:50:45 clandmeter ncopa forgot to mention that there are still PAX messages in dmesg when i run gdb. http://gw.gd/SCsD 2014-03-20 22:06:14 <_lb_> Mmmm 2014-03-20 22:06:39 <_lb_> How can I get which package contains a particular file? 2014-03-20 22:45:33 crow: did you use paxctl -c -m on the debug version too? 2014-03-21 07:07:56 Francois__: could you please file abug on bugs.a.o for that? 2014-03-21 07:08:09 very good that you have a how to reproduce it 2014-03-21 07:12:53 seems like it is enough to: from time import gmtime 2014-03-21 07:14:00 ha 2014-03-21 07:14:08 i think its an uclibc issue only 2014-03-21 07:16:52 hm 2014-03-21 07:17:03 my testcase is wrong 2014-03-21 07:21:07 but yes 2014-03-21 07:21:13 its uclibc only issue 2014-03-21 07:29:30 clandmeter yes i did, in this link you see that i am doing paxctl http://paste.ie/view/6358999c and then in dmesg http://gw.gd/SCsD 2014-03-21 08:13:01 crow: thats probably just a detection of grsec kernel 2014-03-21 08:14:39 crow: im no gdb export, maybe ncopa can assist you in it, if he has time. 2014-03-21 08:18:48 its gdb thats get killed by kernel 2014-03-21 08:19:01 try paxctl -c -m /usr/bin/gdb 2014-03-21 08:19:53 or even better, consult https://wiki.gentoo.org/wiki/Hardened/Debugging on how to do debugging 2014-03-21 08:20:12 crow: is there some way I can reproduce it here, locally? 2014-03-21 08:20:25 do i need some special hardware to reproduce it? 2014-03-21 09:27:04 fabled: yesterday we have 840 visitors 2014-03-21 09:27:24 while day before it was 271 2014-03-21 09:27:39 ncopa: ^ 2014-03-21 09:28:30 and windows is still 50% of the visitors :) 2014-03-21 09:30:24 cool! 2014-03-21 09:39:02 nice 2014-03-21 09:39:09 ncopa, we really need to get the alpine-3.0 / musl email out 2014-03-21 09:40:04 i need clean up rboisver's mess 2014-03-21 09:40:33 i think he logged in to git.core.wtbts.net, and manually edited a file in the checked out repo for my scripts 2014-03-21 09:40:38 next git pull choke 2014-03-21 10:45:43 wrong win.. sorry 2014-03-21 10:51:00 ncopa well you have to have oscam running and user on it. but to reproduce this with gdb i think you can just start gdb with oscam 2014-03-21 10:52:50 what is the issue you try to debug? 2014-03-21 10:54:33 ok i have oscam running an webbrowser pointing to localhost:8080 2014-03-21 10:55:32 i think we should really post a message about musl 2014-03-21 10:55:43 on mailing list 2014-03-21 12:43:41 ncopa well after some time oscam will crash, but i have fev users and lokal cards, and then i would like to run oscam in gdb.. but ok i will do paxctl from gdb. 2014-03-21 13:45:23 ncopa clandmeter now i did this http://gw.gd/OP11 and only what i now see in gdb is warning: Unable to find libthread_db matching inferior's thread library, thread debugging will not be available., nothing in dmesg from PAX after doing paxctl -c -m /usr/bin/gdb 2014-03-21 13:47:03 i hope next segfault will generate nice dump for developers to fix it. 2014-03-21 13:59:35 crow: i think you have good chance of that 2014-03-21 13:59:37 good work! 2014-03-21 14:36:53 ncopa thank you . 2014-03-21 15:36:15 anyone know if iommu / multi function nics should work with a default alpine setup? 2014-03-21 15:36:37 i inserted one but so far only see the two phys. ports 2014-03-21 15:36:54 it's not tested working on another distro, i'm just starting to look into it 2014-03-21 16:06:54 wow 2014-03-21 16:06:57 2.8 released? 2014-03-21 16:07:08 should upgrade my 2.5 box :/ 2014-03-22 04:05:07 <_lb_> Dear all... Don't you think that enabling the framebuffer by default on boot could be a bad idea in production? 2014-03-22 04:05:25 <_lb_> IE, Serial Over Lan won't work properly 2014-03-23 15:20:59 has anyone used acf-provisioning ? 2014-03-23 22:40:35 hi 2014-03-23 22:42:34 could anyone help me, get squid running in transparent mode? when i add "http_port 8181 intercept" and the needed iptables rule ( redirect http to 8181 ) transparent is working, but wenn I add the proxy to the browser ssl does not work anymore untile I remove "intercept" agein :( 2014-03-23 22:45:41 in the acces log I found the following: "1395613339.555 1 192.168.1.1 NONE/409 3830 CONNECT www.google.de:443 - HIER_NONE/- text/html" 2014-03-23 22:54:25 nobody ? 2014-03-23 23:22:47 :( 2014-03-23 23:38:58 ! 2014-03-24 02:09:25 John_P: SSL in transparent mode doesn't work on purpose since that'd be a man in the middle attack... 2014-03-24 11:45:10 hi 2014-03-24 11:45:43 i've updated an edge-box and now wlan is not coming up 2014-03-24 11:45:55 output looks like this: 2014-03-24 11:47:00 http://pastebin.com/vBYQW1Xh 2014-03-24 11:49:17 any ideas from where the "invalid option" output comes? 2014-03-24 11:49:47 /etc/init.d/dhcpcd does not contain something like this... 2014-03-24 11:51:21 wlan worked before the update without problems 2014-03-24 11:51:55 the update updated 113 apk's, so i didn't this for some days... ;-) 2014-03-24 12:09:21 oh 2014-03-24 12:09:30 this looks like busybox ifup bug 2014-03-24 12:11:03 or maybe dhcpcd 2014-03-24 12:11:07 yeah probably dhcpcd 2014-03-24 12:11:19 StarWarsFan: do you think you could please file a bug on bugs.a.o? 2014-03-24 12:11:31 http://linux.die.net/man/1/logger 2014-03-24 12:11:44 my guess is, dhcpcd exec's logger -i 2014-03-24 12:11:52 and bb logger does not support it 2014-03-24 12:12:09 we can fix it by either implement -i on bb logger or edit dhcpcd to not use -i 2014-03-24 12:53:59 ct2001 is gone, but the answer is yes, I have used acf-provisioning 2014-03-24 12:54:03 that's why I wrote it :) 2014-03-24 13:01:44 StarWarsFan: I just pushed a fix for dhcpcd and busybox logger 2014-03-24 13:02:10 http://git.alpinelinux.org/cgit/aports/tree/main/dhcpcd/busybox-logger.patch 2014-03-24 13:02:43 you can manually edit dhcpcd-run-hooks to kickstart it again if you need 2014-03-24 13:11:34 re 2014-03-24 13:11:41 ncopa: cool, thx 2014-03-24 13:11:46 will have a look now... 2014-03-24 13:11:58 should i still file a bug on b.a.o? 2014-03-24 13:14:10 nah 2014-03-24 13:26:09 nice, the error went away 2014-03-24 13:26:19 now i have to figure out, why wlan is not working 2014-03-24 13:28:40 StarWarsFan: is it musl or uclibc? 2014-03-24 13:28:46 does it hang? 2014-03-24 13:29:05 * Starting networking ... 2014-03-24 13:29:05 * lo ... [ ok ] 2014-03-24 13:29:05 * wlan0 ... 2014-03-24 13:29:05 dhcpcd[3405]: version 6.2.1 starting 2014-03-24 13:29:05 dhcpcd: wlan0: starting wpa_supplicant 2014-03-24 13:29:06 dhcpcd[3405]: wlan0: waiting for carrier 2014-03-24 13:29:09 dhcpcd[3405]: timed out 2014-03-24 13:29:11 dhcpcd[3405]: allowing 8 seconds for IPv4LL timeout 2014-03-24 13:29:13 dhcpcd[3405]: timed out 2014-03-24 13:29:15 dhcpcd[3405]: exited [ !! ] 2014-03-24 13:29:23 it's uclibc 2014-03-24 13:29:27 ok 2014-03-24 13:29:29 interesting 2014-03-24 13:29:39 as long as the edge repo still contains uclibc content!? 2014-03-24 13:29:48 i didnt get dhcpcd work on my musl laptop either 2014-03-24 13:29:49 i just used 2014-03-24 13:29:54 i thought it was due to musl 2014-03-24 13:29:54 apk update && apk upgrade 2014-03-24 13:29:58 which updated 113 apk 2014-03-24 13:30:07 but then its probably dhcpcd thats broke somehow 2014-03-24 13:30:18 wired connection works 2014-03-24 13:30:28 with dhcpcd? 2014-03-24 13:31:58 oh, have to check. it was some month's ago as i setup this box... 2014-03-24 13:37:53 yes, should be hdcpcd: 2014-03-24 13:38:00 auto eth0 2014-03-24 13:38:00 iface eth0 inet dhcp 2014-03-24 13:38:00 hostname smallone 2014-03-24 13:38:12 right? 2014-03-24 13:49:03 looks right 2014-03-24 13:57:08 hm, strange... 2014-03-24 21:25:47 n8@all 2014-03-25 13:09:41 anyone have an idea when xen4.4 is coming to alpine? 2014-03-25 13:10:01 i have a new test box and am still trying to make up my mind what to put on it 2014-03-25 13:10:14 royger, kaniini ^^^ 2014-03-25 13:10:32 :) 2014-03-25 13:10:37 i must say, i am *very* happy with musl libc 2014-03-25 13:11:02 all kinds of old, unfixable bugs are suddenly fixable... 2014-03-25 13:11:06 ncopa: is it in 2.8? i have a 2.8 on the test server, but i'd probably not even notice 2014-03-25 13:11:11 oh, nice :)) 2014-03-25 13:11:22 there will be no 2.8 2014-03-25 13:11:27 there will be 3.0 2014-03-25 13:13:04 oh 2014-03-25 13:13:08 *blinks* 2014-03-25 13:13:10 fine too 2014-03-25 13:13:20 what about it makes bugs more fixable in musl libc? 2014-03-25 13:13:45 darkfaded, though, current edge propably identifies itself in motd as alpine 2.8 2014-03-25 13:13:52 jzono1: there are some bugs that are not possible to fix with uclibc 2014-03-25 13:14:05 well, possible but not worth fixing 2014-03-25 13:14:15 like time zones 2014-03-25 13:14:23 but we did send out http://lists.alpinelinux.org/alpine-devel/3596.html just yesterday or so 2014-03-25 13:14:25 uclibc has only support for posix timezones 2014-03-25 13:14:42 fabled: *reads now* 2014-03-25 13:14:44 which cannot present all timezones 2014-03-25 13:15:04 where should feedback regarding armhf go? 2014-03-25 13:15:13 i have 2 cubietrucks and one raspi 2014-03-25 13:15:33 please send to alpine-devel@lists.alpinelinux.org 2014-03-25 13:15:36 (plus a nexus and a few zyxel nas, but those are too oddball systems) 2014-03-25 13:15:36 ok 2014-03-25 13:15:39 hen we get it in the archives 2014-03-25 13:16:26 jzono1: this is an example of a bug that we will be able to fix after musl switch: http://bugs.alpinelinux.org/issues/927 2014-03-25 13:16:45 darkfaded: I wouldn't jump too quickly to 4.4.0, I think we should at least wait until 4.4.1 2014-03-25 13:17:00 oh 2014-03-25 13:17:27 ok 2014-03-25 13:19:47 darkfaded: anyway, that always depends on whether there's something in 4.4 that you desperately need (which I don't think it's the case), updating doesn't seem that complicated, and the patch to add support for musl is much more minimal that what I had expected 2014-03-25 13:20:36 well the most regressions normally come from .X version to next one 2014-03-25 13:21:09 so if building a new server i'd rather step to the next major and see whats broken before starting to use it, than inherit issues later when i update 2014-03-25 13:21:39 ncopa: btw, there's a XSA in the Linux kernel, see http://marc.info/?l=linux-netdev&m=139575006018251&w=2 2014-03-25 13:21:50 but 4.3 is fine for quite some more time i hope 2014-03-25 13:22:14 royger [PATCH net V2] ... 2014-03-25 13:22:22 did it get accepted upstream? 2014-03-25 13:22:35 i suppose it will come with a stable release soonish 2014-03-25 13:23:08 ncopa: not yet AFAIK, no need to pick it now, just FYI 2014-03-25 13:23:38 darkfaded: I guess kaniini will update it when he feels confident it's stavble enough :) 2014-03-25 13:23:54 hehe 2014-03-25 13:43:32 Alpine Linux LinkedIn Group: http://www.linkedin.com/groups/Alpine-Linux-6648617 2014-03-25 15:45:12 alacerda: nice! 2014-03-25 15:45:49 didnt know we has so many brazilian users 2014-03-25 15:46:18 i suppose they like that we like Lua :) 2014-03-25 16:00:39 it is one reason... but alpine has it own merits ;) 2014-03-25 16:00:50 ist* 2014-03-25 16:00:52 its** 2014-03-25 22:31:12 Is an optical drive required to install alpine? I don't have one. 2014-03-25 22:31:57 I tried to boot the iso from usb and got "mounting boot media failed" and it drops to a recovery shell. 2014-03-25 22:37:19 piousminion: Optical drive is definitly notrequired to run alpine... Is the alpine_dev boot parameter specified correctly? 2014-03-25 22:37:44 like alpine_dev=UUID=ffeaccf0-1eee-4860-98bf-9e86fc0eb5ab:ext3 2014-03-25 22:38:50 x00r: not required to run, but what about install? I can only find isos. Not sure about the boot param, but I'll check. 2014-03-25 23:12:35 someone can tell piosminion about unetbootin if he comes back... 2014-03-25 23:12:51 or the old way of doing manual syslinux + copying over apks + boot folder 2014-03-26 02:16:27 Looking for help with Xen HVM (alpine 2.7.5), anybody available? 2014-03-26 02:46:12 13:09:03 darkfaded | anyone have an idea when xen4.4 is coming to alpine? 2014-03-26 02:46:25 darkfaded: 3.0 will ship with Xen 4.3.2, for selfish reasons. 2014-03-26 02:46:37 darkfaded: Xen is really something that you're better off waiting for a point-release on 2014-03-26 09:10:54 kaniini: see my reasoning back up. i need some time to find what what kinds of regressions they added, i find that easier to do when it's an all new release. but i guess i could pin the version on the prod systems? 2014-03-26 09:15:23 darkfaded: well, you could build 4.4 locally 2014-03-26 09:15:57 darkfaded: i'm going to keep xen on 4.3.2 though for the moment because we're tracking edge right now 2014-03-26 09:23:06 ;p now it's out 2014-03-26 15:29:50 https://www.irccloud.com/pastebin/9d9q1sZt 2014-03-26 15:30:24 what is alpine equivalent of systemctl start xend.service ? 2014-03-26 15:31:17 http://lists.xen.org/archives/html/xen-devel/2013-09/msg03081.html 2014-03-26 15:58:44 Frosh: Dom0 domain name is set in xenstore init script, did you start it? 2014-03-26 15:59:34 Frosh: ahh, this is not related to the xen-devel post you are referencing 2014-03-26 15:59:55 Frosh: this is not Dom0 not having a name set, this is a zombie domain 2014-03-26 16:00:43 Frosh: does `xl destroy 8` solve your problem? 2014-03-26 16:16:35 it was what started the problem 2014-03-26 16:16:55 I went and xl destroy 8 2014-03-26 16:16:59 and that showed up 2014-03-26 16:17:29 if you refere to /etc/init.d/xend start, it has some import error 2014-03-26 18:34:44 hmm after last apkl upgrade i have a certain lack of init 2014-03-26 18:35:02 openrc didnt get upgraded? 2014-03-26 18:35:47 no idea, i didn't keep an eye on it 2014-03-26 18:35:56 just noticed later on it won't boot 2014-03-26 18:36:06 i think i just give up on today 2014-03-26 23:18:02 hi 2014-03-26 23:18:24 so uh, I'm having an issue, I've upgraded my stuff to alpine edge so I can transition to alpine-musl 2014-03-26 23:18:34 http://kaniini.dereferenced.org/2014/03/23/upgrading-to-alpine-musl.html following this 2014-03-26 23:18:37 but it now says 2014-03-26 23:18:44 "Failed to execute /init" 2014-03-26 23:18:45 when booting 2014-03-26 23:18:59 wtf? :S 2014-03-26 23:22:03 (note I am upgrading from 2.7.5) 2014-03-26 23:37:43 Elizafox: prolly should've gone to 2.8 first, dunno how much of a difference it makes 2014-03-26 23:51:26 I did 2014-03-26 23:51:28 -.- 2014-03-26 23:51:32 I had these problems at 2.8 2014-03-27 04:36:04 anyone having issues with alpine xen domU running with pygrub? 2014-03-27 04:36:24 pygrub's extlinux parser does not seems to be able to parse the generated extlinux.conf file 2014-03-27 04:36:26 :( 2014-03-27 11:26:34 fbettag: i think you were right about openrc, i bet i only did a apk upgrade -a and not -U, too 2014-03-27 11:26:47 still a bit weird and it's nuked nonetheless 2014-03-27 11:39:19 =) 2014-03-27 11:39:30 i just had that while playing around with the xen hosts 2014-03-27 11:40:29 i suppose i could fix it somehow, but the ipmi will drive me crazy before i get there 2014-03-27 14:05:04 Elizafox: i think you got bitten by the bug that affects uclibc 2014-03-27 14:05:20 the initramfs image is broke 2014-03-27 14:05:34 we just found out how to fix it 2014-03-27 14:10:35 ah 2014-03-27 14:10:41 there we go :/ 2014-03-27 14:10:46 *grin grin* 2014-03-27 14:11:08 was my "improved" mkinitfs :-/ 2014-03-27 14:11:22 its faster though ... :( 2014-03-27 14:11:23 :) 2014-03-27 14:12:31 patching /usr/bin/lddtree with this: http://sprunge.us/QZKa 2014-03-27 14:12:33 well, living on the edge is dangerous at times ;) 2014-03-27 14:12:44 and regenerating initramfs image should fix it 2014-03-27 14:12:57 (I'm testing if it actually does) 2014-03-27 14:13:48 yup 2014-03-27 14:13:50 i think it does 2014-03-27 14:45:18 <_boot> Hi, getting "Booting kernel failed: Bad file number" from the 2.6.6 x86_64 disc under VirtualBox, has anyone else seen this? 2014-03-27 14:46:48 <_boot> oh, but after leaving it for long enough it's booted anyway :| nevermind! 2014-03-27 15:15:38 ncopa, oh phew, thanks 2014-03-27 15:19:07 Still working on it 2014-03-27 15:19:12 found other bug :-/ 2014-03-27 15:21:22 ok 2014-03-27 15:21:27 and i htink i have the fix 2014-03-27 15:21:36 this is what got my laptop unbootable i think 2014-03-27 15:21:51 kernel module dependency detection is broke 2014-03-27 15:35:31 ncopa: hehe ubuntu has that bug as a default feature 2014-03-27 15:36:32 ok i think it should be solved with mkinitfs-2.6.0-r3 2014-03-27 15:36:54 if you have your root on ext4 you can verify if kernel module deps are there with: 2014-03-27 15:37:04 gunzip -c /boot/initramfs-grsec | cpio -t | grep -E 'fs/(ext|mb)' 2014-03-27 15:37:16 if mbcache.ko is missing you'll have a problem 2014-03-27 15:37:27 if its there it should work 2014-03-27 19:47:16 ncopa: i suppose i can get out of the problems if i boot and chroot into the system and then apk update so i get your fix? :) 2014-03-27 19:47:24 i'll give it a shot 2014-03-28 06:46:29 ncopa, about yet 2014-03-28 06:46:35 er, are you about yet 2014-03-28 07:10:35 Elizafox: hi 2014-03-28 07:46:14 ncopa, sup 2014-03-28 07:46:30 ncopa, I've got a pretty big problem on my hands :/ 2014-03-28 07:46:57 localhost:~# mkinitfs 3.13.7-0-grsec 2014-03-28 07:46:58 returns 2014-03-28 07:47:21 cpio: --no-auto-root Do not automatically prefix input ELFs with ROOT: No such file or directory. 2014-03-28 07:47:24 and then CPIO says 2014-03-28 07:47:26 1 blocks. 2014-03-28 07:47:30 and immediately after 2014-03-28 07:47:44 depmod: can't change directory to lib/modules/3.13.7-0-grsec: no such file or directory 2014-03-28 07:51:17 cpio: --no-auto-root Do not automatically prefix input ELFs with ROOT: No such file or directory. 2014-03-28 07:51:20 thats a bad sign 2014-03-28 07:51:24 dont reboot 2014-03-28 07:51:28 is it in a chroot? 2014-03-28 07:51:55 is it uclibc or musl? 2014-03-28 07:52:09 i got that cpio error too while trying to rescue my laptop 2014-03-28 07:52:25 i figured out it comes from readlink -f being broken 2014-03-28 07:52:43 i thought it came from /proc not beeing mounted in my chroot 2014-03-28 07:53:11 cpio: --no-auto-root Do not automatically prefix input ELFs with ROOT: No such file or directory. 2014-03-28 07:53:28 means that lddtree failed to run, and spits out usage help instead 2014-03-28 07:58:57 ncopa, it was in the OS proper 2014-03-28 07:59:05 ncopa, but I sadly rebooted and alas, I am in a chroot 2014-03-28 08:00:01 it's for musl 2014-03-28 08:00:02 hm sec 2014-03-28 08:00:36 /lib/modules/3.13.7-0-grsec/kernel/drivers/block/cpqarray*: No such file or directory. Skipped 2014-03-28 08:00:44 okay, I think that fixed it 2014-03-28 08:01:21 oh god not this shit again 2014-03-28 08:01:27 Failed to execute /init (error -2) 2014-03-28 08:02:00 but /sbin/init doesn't work either 2014-03-28 08:02:01 crap 2014-03-28 08:02:48 (I tried init=/sbin/init) 2014-03-28 08:02:50 root 2014-03-28 08:02:51 oops 2014-03-28 08:02:54 wrong window 2014-03-28 08:03:03 trying to rescue from a chroot 2014-03-28 08:04:16 ok 2014-03-28 08:04:51 what worked for me was: http://lists.alpinelinux.org/alpine-devel/3607.html 2014-03-28 08:05:15 that mount of proc was wat fixed the readlink -f issue (causing cpio --no-auto-root) 2014-03-28 08:05:38 you dont really need to chroot in to your root 2014-03-28 08:05:46 just mount it on forexample /mnt 2014-03-28 08:06:04 and mount --bind /proc /mnt/proc 2014-03-28 08:06:11 and then: apk fix --root /mnt mkinitfs 2014-03-28 08:06:25 or: apk upgrade -U --root /mnt 2014-03-28 08:06:49 it works even if you boot from uclibc but has a musl system mouted on /mnt 2014-03-28 08:07:09 I still get that weird cpqarray issue 2014-03-28 08:07:43 that "Skipped." thing can be ignored 2014-03-28 08:08:14 unless you need the cpqarray drivers too boot. i think its HP server disk arrays 2014-03-28 08:08:23 I do not 2014-03-28 08:08:30 aha 2014-03-28 08:08:33 there we go! thanks ncopa 2014-03-28 08:08:56 : not such file or directory. Skipped. 2014-03-28 08:09:00 is not dangerous 2014-03-28 08:09:09 cpio: blahblah.. error. 2014-03-28 08:09:12 is dangerous 2014-03-28 08:09:23 i'm gonna make cpio error more verbose 2014-03-28 08:09:50 so its more clear when error will cause unbootable system 2014-03-28 08:11:44 i'm gonna rescue my edge vm now too 2014-03-28 12:02:19 ncopa: i restored this morning, went well 2014-03-28 12:02:59 i booted off an iso. needed to mount /proc and stuff, rsync'd /dev including devices and had to create a "lock" subdir in /run 2014-03-28 12:03:25 then apk update && apk upgrade -a -U ; apk fix also apk fix mkinitrd 2014-03-28 12:03:34 since i forgot to turn off grsec 2014-03-28 12:03:43 so, thats another step 2014-03-29 20:21:29 any better solutions than Tinc VPN to connect a proxy / web / mysql servers to each other ? (VPS provider does not have internal ip's) 2014-03-31 00:13:32 hm 2014-03-31 00:13:46 are you able to encrypt alpine's root partition during setup? 2014-03-31 06:25:34 BitL0G1c: ipsec in transport mode? 2014-03-31 06:25:45 maybe overkill w 2014-03-31 06:25:51 maybe overkill with opennhrp 2014-03-31 08:28:46 is there wol package in alpine linux (can not found it in repository) 2014-03-31 08:43:51 when i search for wol 2014-03-31 08:44:13 crow, is 'awake' good enough? 2014-03-31 08:44:34 i assume you mean wake-on-lan with wol 2014-03-31 08:49:03 ether-wake? 2014-03-31 08:49:22 ah, ether-wake is from busybox 2014-03-31 12:46:24 ncopa thank will try it 2014-03-31 12:46:39 fabian_a yes it was about wake on lan 2014-03-31 12:49:44 Mp5shooter: i dont think alpine has support for encrypted root from setup 2014-03-31 13:29:21 oh 2014-03-31 13:35:05 encrypted installs work 2014-03-31 13:36:43 just mount the partitions under /mnt & run an installation for LVM 2014-03-31 13:38:40 setup-disk -m sys /mnt 2014-03-31 13:42:23 notes for a LUKS encrypted install http://pastebin.com/nB4Hwb1a 2014-03-31 13:44:13 only boot needs to be unencrypted 2014-03-31 13:53:11 :o 2014-03-31 22:59:21 i hope i find a fix for the pv domU syslinux issue soon 2014-03-31 22:59:56 did some benchmarks, hvm eats up >60% of io throughput 2014-03-31 23:00:02 not ok that