**** BEGIN LOGGING AT Sat Jan 10 02:59:57 2009 Jan 10 11:57:35 hi Jan 10 11:58:26 i cloned and resized my debian installation from hdd to usb stick Jan 10 11:59:15 then i changed the uuid of the usb stick partition and also in the fstab Jan 10 12:00:12 in the hope it would boot from the usb stick while the hdd from which i cloned is still available Jan 10 12:01:05 but it doesn't boot from the stick any more, no matter if the hdd is plugged in or not Jan 10 12:01:55 does someone know what else i need to change or do i have to clone / resize once more? Jan 10 12:03:22 maybe something in /boot ? Jan 10 12:05:03 i think i'll just clone it once more and change the uuid of the hdd instead xD Jan 10 12:23:47 Small question: does nslu have the 'daemon' user? and is it recomended to be used by third-party daemons? Jan 10 12:23:50 or is it prefered to create a dedicated user? Jan 10 16:10:46 does NSLU2 have the same init.d system as debian? ie. would this work: http://trac.transmissionbt.com/wiki/Scripts/initd ? Jan 10 16:25:45 What OS? Jan 10 16:27:30 ShadowJK: euhm, I'm not all that familiar with linux, so basicaly in general for the nslu project? Jan 10 16:27:49 Well you can run debian on it. Jan 10 16:27:55 Then you'd have the same init.d system as debian. Jan 10 16:28:19 unslung is a bit different Jan 10 16:28:42 SlugOS has a more normal init.d system Jan 10 16:28:50 dunno about openwrt and angstrom Jan 10 16:29:57 hmm, in your opinion, would it be ok to recomend NSLU2 users to use that init.d script? Jan 10 16:31:01 btw. "adduser", as in "adduser --disabled-password transmission" exists in all variants? Jan 10 16:32:05 the LSB stuff probably doesn't exist on anything for the nslu2 except debian Jan 10 16:33:10 lsb? Jan 10 16:34:55 in the script you pasted Jan 10 16:35:13 It looks a bit over-engineered Jan 10 16:38:00 oh... yea probably. I think the original author took it from a template or something... I simplified it as much I could (with my knowledge) Jan 10 16:40:29 So yeah, that thing would only work on debian Jan 10 16:41:02 If you remove the lsb line, and the lines that log, it might work on openslug Jan 10 16:41:19 though I dunno about busybox's start-stop-daemon :) Jan 10 16:42:33 lol, I'll just advice to try it, and wait for "doesn't work" feedback ;-) Jan 10 16:43:19 ShadowJK: oh, and "adduser"? Jan 10 16:43:41 * ShadowJK shrugs Jan 10 16:43:44 probably fine Jan 10 16:44:01 k thx a lot! Jan 10 18:47:26 after I extract an ipkg manually with 'ar -x' is there a way to manually install it? Jan 10 20:31:45 scant: not automagically, but you can just use "tar tvf data.tgz" (or data.tar.gz, whichever it is) to see what's in the packages, and extract to a tmp location and cp/mv the files in place. It's just that your opkg database will be out of date by then. Don't forget to run any postinstall script taht may be provided by the package. Jan 10 20:32:56 scant: I suggest, though, that you look for the root cause of your problem. If the issue arises with any ipk, then it would seem that you have a bad opkg utility, or perhaps you have corrupted/lost your opkg conf file. Jan 10 20:34:20 Build the latest stuff from OE, and compare what's on your dsmg600 rootfs with what's in the tmp/rootfs directory --- that's the easiest way I've found to see what's changed, or what I've messed up on my rootfs when things go wonky. Jan 10 20:49:40 mwester: can you install local ipkgs with 5.2-alpha with EABI? Jan 10 20:50:29 what exactly do you mean? My local autobuilder is in the feeds for my image; so it randomly downloads from either the official repos or my own. Jan 10 20:51:33 mwester: i mean, can you install an ipkg via 'opkg install my.ipk' Jan 10 20:53:26 Ok. I'll test that when I get a chance. Jan 10 20:55:24 mwester: i don't think my rootfs is messed up. i've come across this error before, it can mean opkg isn't passing parameters to opkg-cl correctly or tmp is full i think Jan 10 20:55:46 mwester: but i still get the error when using opkg-cl directly and my tmp is not full Jan 11 00:54:18 scant: corrected some error handling and sloppy file mode handling with tmp files that I introduced earlier; see if the fixes work for you. I think the newer libc just introduced is a bit more strict and won't permit the sloppiness that the original SlugOS versions (pre-EABI) would... Jan 11 00:55:44 mwester: i saw the commit... thanks, i'll try it out Jan 11 02:12:02 mwester: i think your fix solved my issue, thank you Jan 11 02:14:07 ah, good. I guess... I do wonder what other horrors lurk in opkg, though. Jan 11 02:14:51 rwhitby suggested that we take a look at the openwrt version of opkg; I think I'll do that. Jan 11 02:46:18 mwester-laptop: how'd you discover the issue was with that patch? Jan 11 02:48:09 I guessed. Problem appeared with the switch to EABI, but other distros are using opkg with EABI with no problems; there are 4 patches unique to SlugOS... Jan 11 02:48:57 Installing the full opkg confirmed that it worked, from there it was just a matter of backing out the patches one-by-one until the broken patch was found. Jan 11 02:49:56 Then adding some debug printfs showed up the point of the failure, and that's where nothing made any sense except that an error might be ignored, or something weird with the file open mode... Jan 11 02:50:32 turns out it's the file mode that made the difference. very unexpected, but technically it was wrong to begin with. Jan 11 02:51:59 cool Jan 11 02:53:15 my next issue is libusb-1.0-0 & 0.1-4 and saned Jan 11 02:54:03 should I install libusb_0.0.0? **** ENDING LOGGING AT Sun Jan 11 02:59:57 2009