**** BEGIN LOGGING AT Mon Jan 11 02:59:57 2010 Jan 11 15:40:46 after reading buffers of this channel, I would ask what kind of port mapper brandano needs in fact? the best kind of it is nmap, if I understand the task properly. Jan 11 17:18:07 so angry Jan 11 17:18:14 samba still crashes my openrd Jan 11 17:18:30 writing files via samba Jan 11 17:18:47 havent yet gotten it to generate any kind of errors, it just locks up Jan 11 17:19:10 i dont know the first thing about debugging the kernel, but that appears two be in my not distant future Jan 11 18:14:11 debian, on 2.6.32. i also tried compiling the deb-src on the openrd.... two hours of compiling and still no diff Jan 11 18:15:00 Fabian6: ping Jan 11 18:15:41 Fabian6: saw your msg on the openrd list. Jan 11 18:15:53 It could be that your x config is different Jan 11 18:37:33 Hi, got a weird Linux prob, hoping a guru might be able to help. I moved my /var of my nand filesystem to sdcard (to keep a nice buffer of free space). Some applications occasionally whinge for example when I run screen after it reboot, it says that /var/run/screen needs to be chmod to 777 (which I do, and it works, after a reboot it needs doing again). Weird? Jan 11 18:48:10 eFfeM-home: pong Jan 11 18:48:23 bootlogo is blurry too Jan 11 18:49:49 hm, that is odd, for me the bootlogo was ok as far as I recall, the openembedded image does not have a boot logo Jan 11 18:51:15 my bad Jan 11 18:51:18 Fabian6: are you on auto res Jan 11 18:51:18 there is nog bootlogo Jan 11 18:51:40 but the black screen is blurry too :) Jan 11 18:51:53 Jan 11 18:52:46 one sec Jan 11 18:53:25 black screen has white 'jitters' too without X running. Jan 11 18:53:41 i think it's on 1280x1024@60Hz by default Jan 11 18:54:21 was trying to find out mine, x is running but i get no visual any more, will reboot, takes a minute Jan 11 18:57:02 my monitor when starting up says 1024x768@60hz not sure if that is the default, as soon as I get picture it is 1280x1024@75.2 Jan 11 18:57:08 that is what ti tells me Jan 11 18:57:22 you could try setting the moni to 75 hz Jan 11 18:57:45 hm Jan 11 18:57:49 X won't let me Jan 11 18:58:06 it's on 800x600@60Hz right now Jan 11 18:58:47 are there any major differences compared to the stock kernel ? Jan 11 18:59:08 not really, I am reusing the x settings from the .22 kernel (did you try x with that one?) Jan 11 18:59:18 jup Jan 11 18:59:22 stock kernel works fine Jan 11 18:59:27 both shell and X Jan 11 18:59:47 jitters begin when the kernel is loaded Jan 11 19:00:32 here is my xorg.conf Jan 11 19:00:46 as far as i understood from Tanmay this is the very same driver Jan 11 19:00:59 screen receives no signal until after Ucompressing linux. the de screen comes up (stays black now without the bootlogo) and starts jittering Jan 11 19:01:33 i'm having the same issues with the uImage.openrd from the developers CD Jan 11 19:01:43 only the stock kernel works Jan 11 19:01:58 it's not an X issue Jan 11 19:02:06 yeah for some reason in OE someone decided that the bootlogo is always off, it overrides the logo even if you specify it in the x config Jan 11 19:02:29 the only thing that could be different is u-boot Jan 11 19:02:30 bootlogo in the x config ? :) that isn't even loaded until after init Jan 11 19:02:42 i mean the kernel bootlogo Jan 11 19:02:47 small penguin Jan 11 19:02:56 bootlogo in kernel, that is off in openembedded kernels Jan 11 19:03:13 http://groups.google.com/group/openrd/browse_thread/thread/9432b69171a15629 Jan 11 19:03:18 kernel in last post works for me Jan 11 19:03:55 my u-boot is U-Boot 1.1.4 (Apr 22 2009 - 20:18:28) Marvell version: 3.4.16 Jan 11 19:04:26 U-Boot 1.1.4 (Apr 22 2009 - 20:18:28) Marvell version: 3.4.16 Jan 11 19:04:42 same ... Jan 11 19:05:04 i was trying a new kernel becouse of disappointing results in X Jan 11 19:05:08 100% cpu and such Jan 11 19:05:40 i needed a newer kernel, just applied the patch Jan 11 19:06:10 you might want to compare the sources from the url above and the one from Tanmay Jan 11 19:06:17 and/or maybe he can help Jan 11 19:06:25 actually out of ideas right now Jan 11 19:06:25 if X-speed doesn't improve with a newer kernel Jan 11 19:06:30 i'll just go back to the stock kernel Jan 11 19:07:14 mihgt be a matter of initialisation Jan 11 19:07:34 btw do you get X if you cold boot the board without a monitor being connected Jan 11 19:07:47 i'll try Jan 11 19:07:59 so monitor disconnected Jan 11 19:08:03 hard power cycle Jan 11 19:08:07 off Jan 11 19:08:11 on Jan 11 19:08:27 how do you test x speed ? Jan 11 19:09:43 GNOME started, firefox is started Jan 11 19:09:45 reconnecting monitor Jan 11 19:09:54 jup, up Jan 11 19:09:57 (with jitters) Jan 11 19:10:17 i'm not realy testing. just clicking around Jan 11 19:10:30 it sucks gevoelsmatig ;) Jan 11 19:10:58 for me cold boot with no monitor = no X Jan 11 19:11:15 guess there is something fishy wrt initialisation Jan 11 19:11:28 different chipsets/releases ? Jan 11 19:13:21 it's not a buggy monitor ? Jan 11 19:13:42 could be or regs taht are not really properly intiialised and not always have the same defaults Jan 11 19:14:20 ubuut regs ? Jan 11 19:14:33 or device regs that are dangling Jan 11 19:14:41 just set all my env settings back to default Jan 11 19:15:02 if you have one you could try a differnt monitor (i' Jan 11 19:15:18 already tried to different tft's Jan 11 19:15:21 btw i am using an iiyama 24" Jan 11 19:15:22 differend models/makes Jan 11 19:15:43 dell 24", or benq 17-somthing 16:9 screen for testing Jan 11 19:16:11 what are you using your openrd for ? Jan 11 19:16:47 just playing, little bit nas related functionaltiy Jan 11 19:17:10 so you don't realy need x ;) Jan 11 19:17:38 no, not really, perhaps i want to have a browser at some point in time Jan 11 19:18:27 did a diff between the url above and the patch from Tanmay, there is an additional line in vb_setmode.c in the one from the post above and some different initialisation in vb_table.h Jan 11 19:18:33 afk for 10-15 mins Jan 11 19:27:57 3rd monitor connected Jan 11 19:27:58 same result Jan 11 19:38:22 back Jan 11 19:39:20 wb ;) Jan 11 19:40:30 here is the diff between the old version (from aug or so) and the one from Tanmay: http://www.pastebin.ca/1747287 Jan 11 19:42:08 hm Jan 11 19:42:16 i'm going to boot my compilebox again Jan 11 19:42:56 btw, should it be possible to output a non standard resolution like 640x360 @60 or 50Hz ? Jan 11 19:44:37 i've no idea Jan 11 19:45:15 you could try to use the diff and change the differing lines from Tanmay's patch back to Chong's version and see if that helps Jan 11 19:45:59 btw if i may ask: where do you live ? Jan 11 19:46:06 hilversum Jan 11 19:46:57 why ? :) Jan 11 19:47:15 too far :-) Jan 11 19:47:29 from? Jan 11 19:47:38 i noticed you were from NL ": it sucks gevoelsmatig ;)" and also from your pic Jan 11 19:47:43 ;) Jan 11 19:48:08 I also live in NL, if you were closer to me we could perhaps meet and compare hw or so Jan 11 19:48:25 but I live near the Belgian border, sw of Eindhoven Jan 11 19:48:54 too bad. i'm also in groningen very often Jan 11 19:48:59 but that doesn't help ;) Jan 11 19:49:06 not really Jan 11 19:49:32 i'm going to use this openrd as carpc Jan 11 19:49:33 i've also posted the diffs to the mailing list, let's see if Tanmay picks up things Jan 11 19:49:44 oh cool Jan 11 19:49:46 displaying buienradar and others to my indash screen Jan 11 19:50:17 for which a vga break-in box is availible, but it's "16:9 VGA" Jan 11 19:50:21 so not 640x480 Jan 11 19:50:54 ah ok, was about to suggest ldvs or dvi Jan 11 19:51:00 not with openrd though Jan 11 19:51:28 i have to work with the volvo indash screen, so i'm limitted to vga or composite video Jan 11 19:52:01 just read one of your post, also mentioning X speed and 100% CPU Jan 11 19:52:12 is there some lightweight X availible ? Jan 11 19:52:16 i only need to run firefox Jan 11 19:52:31 no idea, i'm not really into x Jan 11 19:52:44 btw if you want lightweight perhaps better look at midori browser Jan 11 19:53:09 it's X that is hogging up all the cpu cycles, not firefox :) Jan 11 19:54:14 is X at 100-% or so ? Jan 11 19:55:04 my system is now displayign the login screen, x takes < 1% cpu load Jan 11 19:55:18 PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND Jan 11 19:55:18 1637 root 20 0 23132 9760 3884 R 99.9 1.9 27:50.26 X Jan 11 19:55:18 1925 root 20 0 2524 1140 916 R 0.3 0.2 0:00.05 top Jan 11 19:55:45 better get my image :-) Jan 11 19:55:56 it is based upon openembedded Jan 11 19:56:33 your rootfs image ? Jan 11 20:00:28 well it is not really a straightforward image, but there is an angstrom builder that can give you a better image Jan 11 20:00:37 not sure if it is fully working for client, was about to test taht Jan 11 20:01:02 i'll try to get a working screen again first ;) Jan 11 20:03:00 good plan Jan 11 20:03:14 feel free to ping or mail me if you further want to discuss things Jan 11 20:03:30 tnx!, would be nice if both issues are fixed at the same time :) Jan 11 20:03:35 hmm, wonder when those new plugs are available Jan 11 20:03:54 no info yet afaik guess it can take several months Jan 11 20:04:05 your no image on disconnected monitor thingy Jan 11 20:04:13 rabeeh - any idea there? Jan 11 21:12:56 is there a place to download a u-boot binary for openrd client with sd support, plugcomputer only seems to have src Jan 11 21:27:51 if anyone is interested, I've just got fw_setenv running (cfg = dockstar ~ pogoplug): http://fpaste.org/IS0t/ Jan 11 21:31:58 uups, that cfg was the testing one, that is the correct one: http://fpaste.org/rIKA/ Jan 11 21:38:36 calling it a day, cya Jan 11 21:38:45 the diff is against fw_env found in pogoplug-u-boot-1.1.4.tgz Jan 11 22:58:57 is wordpress too much for a sheevaplug to run? I know computingplugs.org runs on a sheevaplug and seems to run well, so it handles mediawiki well. Jan 11 22:59:46 * christoph_debian would bet it can run wordpress Jan 11 22:59:51 Reason I ask... I have apache running with wordpress on my tonidoplug (technically a sheevaplug) and it runs horribly slow. I don't think it's just from the tonido software running, as it seems my CPU is idle w/o accessing it, however apache eats the CPU and goes to crap when I start accessing wordpress Jan 11 23:01:08 I thought that's wordpress' MO Jan 11 23:01:27 but sarcasm aside, I suspect there's something bad going on, because 1.3GHz is plenty to run wordpress Jan 11 23:01:56 page loads take around 2 seconds, and my load average goes from .01 to ??? Jan 11 23:02:10 i'm up to 3.4 load average now (few tabs loading WP pages) Jan 11 23:02:25 maybe more like 4 seconds per page load Jan 11 23:02:25 oh Jan 11 23:02:30 that might be disk throughput Jan 11 23:02:35 are you storing the database on the NAND? Jan 11 23:02:40 [if you are, stop it. that's bad] Jan 11 23:02:47 no, booting and running via USB hard drive Jan 11 23:03:02 ahkay Jan 11 23:03:10 maybe still suffering throughput problems? Jan 11 23:03:21 ok, well, step back Jan 11 23:03:25 a 500Gb seagate... Jan 11 23:03:25 when you have a high load average Jan 11 23:03:36 do you have processes burning a lot of CPU ? Jan 11 23:03:48 or do you have very little CPU usage but a high load average? Jan 11 23:03:50 well, high load average means it's not waiting on IO though right? that means using the CPU Jan 11 23:03:57 No necessarily Jan 11 23:04:09 high load average can often be caused by bad IO throughput Jan 11 23:04:14 when the load average goes up, it's from apache burning CPU... normally (when load average is low) the cpu is listed as about 95% idle in top Jan 11 23:04:34 ok. You probably need to figure out what's burning up the CPU and make it stop, because that's not normal Jan 11 23:05:04 when i start hitting apache, apache jumps up to 50% (per thread) when 2 threads are up there (i.e. 2 connections), basically however many apache threads end up being used, eat the CPU up to 100% Jan 11 23:05:10 but even when it's only 1, it's slow Jan 11 23:06:49 let me re-phrase that last sentence for clarity. when it's only one apache thread running, the page-load is slow, and the CPU usage is ~100% Jan 11 23:07:21 but now, sitting idle, CPU=98% idle.... how can i test disk throughput? hdparm? Jan 11 23:07:58 it's ok, if you're literally seeing apache using 100% cpu usage, then it's unlikely to be the io problem Jan 11 23:08:08 I would suggest googling for wordpress high cpu usage problems Jan 11 23:08:22 ok, thanks! Jan 11 23:08:41 also, isn't bogomips supposed to be twice the CPU? Jan 11 23:09:00 cat /proc/cpuinfo Jan 11 23:09:00 Processor : Feroceon 88FR131 rev 1 (v5l) Jan 11 23:09:00 BogoMIPS : 1192.75 Jan 11 23:09:30 or am i used to dual core machines? Jan 11 23:09:40 http://en.wikipedia.org/wiki/BogoMips Jan 11 23:09:47 (reading now..) Jan 11 23:09:59 "the number of million times per second a processor can do absolutely nothing" Jan 11 23:10:28 so, "yes". It's because you're used to dual core machines, which can do twice as much nothing in the same amount of time Jan 11 23:11:14 yeah, thanks for the link & education, and not just giving me the "RTFM" answer I actually expected Jan 11 23:11:23 haha Jan 11 23:11:27 well, you have a legitimate question Jan 11 23:11:54 and load average isn't necessarily just "my CPU is too crispy", it can be caused by other things Jan 11 23:12:13 in your case ,it just sounds like a common wordpress-eats-too-much-cpu problem Jan 11 23:12:15 oh wow, looks like i'm not alone: http://www.google.com/search?q=wordpress+high+cpu+usage Jan 11 23:12:23 yes it does seem like that's the case! Jan 11 23:12:54 i have always run it on machines with much more power and never had problems... now I know why there's so much work done for caching plugins and PHP caching in the wordpress realm Jan 11 23:16:41 well, looks like I have some homework on caching, not only the wordpress side of things, but even mysql! thanks for all the help Chunky_Ks Jan 11 23:17:14 welcome Jan 11 23:18:14 this is an even better google: http://www.google.com/search?q=wordpress+%27digg+effect%27 Jan 11 23:21:59 anybody have any more info about the new sheevaplugs i've seen some CES posts about? Jan 11 23:22:08 i want to know when they're expected to ship and how much!!! :D Jan 11 23:22:44 don't know and don't know Jan 11 23:22:47 as in Jan 11 23:22:51 the design isn't finalised Jan 11 23:23:06 only 150 have been produced [which in manufacturing is a close approximation of zero] Jan 11 23:23:28 the price isn't finalised because there's some bbeefy stuff not decided Jan 11 23:23:38 the prototype picture you've probably seen with a metal ring around the outside Jan 11 23:23:50 that ring is a heatsink because no-one knew how much heat the 2GHz chip would produce Jan 11 23:23:55 turns out to not produce that much Jan 11 23:24:07 but that much metal, to machine or cast, is expensive Jan 11 23:24:31 oh wow Jan 11 23:24:44 next insurmountable question? Jan 11 23:24:48 i've seen 2 pictures... 1 that looks like a coffin, and 1 that looks like a space ship Jan 11 23:24:58 lol, how do you know that much? work for marvell? Jan 11 23:25:00 the coffin one Jan 11 23:25:11 http://plugcomputer.org/plugforum/index.php?topic=1200.0 Jan 11 23:25:14 i like it :D Jan 11 23:25:32 I know a lot because I spent a *lot* of time talking with the ionics guys Jan 11 23:25:53 ah Jan 11 23:26:46 omg, i need to get that running! i'm an avid off-roader and have been trying to figure out how to get my chumby doing something cool like this... any thoughts of doing this real-time instead of plotting it on google earth afterwards? :D Jan 11 23:27:01 heh. Jan 11 23:27:07 I have a crappy gui [look at the front page] Jan 11 23:27:13 It still logs stuff Jan 11 23:27:17 but it also spams dbus Jan 11 23:27:30 and if you launch via popen or similar, you can read and parse stdout Jan 11 23:27:43 so if you want to do real-time display, the toosl are there Jan 11 23:27:46 nice Jan 11 23:27:53 [look at my webpage and you can see screenshots of the shoddy gui] Jan 11 23:28:01 yeah, just pulled it up Jan 11 23:31:13 if you have questions or suggestions or anything, don't hesitate to ask Jan 11 23:31:19 I'm happy to talk about stuff :-) Jan 11 23:38:25 thanks a ton! Jan 11 23:44:59 welcome :-) Jan 12 00:02:23 so what did you win? i guess I didn't really keep up on the CES/marvell/sheevaplug thing... i was sick all last week Jan 12 00:02:37 just the chance to hang out with some of the marvell developers? Jan 12 00:02:58 i work with a bunch of software developers (and am one myself)... doesn't sound like too much of a prize... lol Jan 12 00:03:43 oh, nm, i see you got a new ionics plug! very cool! Jan 12 02:00:14 fuck a new kernel crash. Jan 12 02:00:27 tried checking out samba via git Jan 12 02:03:04 are there any watchdog drivers i can enable? Jan 12 02:03:13 watchdog devices* **** ENDING LOGGING AT Tue Jan 12 02:59:56 2010