**** BEGIN LOGGING AT Wed May 04 23:59:56 2005 **** BEGIN LOGGING AT Thu May 05 05:08:31 2005 May 05 08:08:49 crap! The latest openslug appears to kernel panic. May 05 08:08:51 What changed? May 05 08:09:24 I get: RAMDISK: couldn't find valid RAM disk image starting at 0 May 05 08:09:40 00:59 < rwhitby-web> NOTE NOTE NOTE: The current OpenSlug image will probably not boot, due to /mnt/tmpmnt being missing from the rootfs, and the new linuxrc May 05 08:09:43 depending on that. May 05 08:09:48 Might be that, dunno ;) May 05 08:09:54 Kernel panic - noy syncing: VFS: Unable to Mount root fs on unknown-block(1,0) May 05 08:10:03 I see. Thanks mate. May 05 08:10:34 on second thought, if it panics it shouldn't be that now, should it? May 05 08:10:40 that means I can't work on my slug until I get home though as I can't restart from work. damn May 05 08:10:48 no idea. May 05 08:11:06 Of course, rwhitby just went asleep :P May 05 08:13:44 hehehe May 05 08:14:17 I was expecting some problems though as I noticed linuxrc and other openslug changes when I pulled this morning. May 05 08:33:23 <[g2]> anybody use the ntp sw from oe ? May 05 08:43:28 Hi May 05 08:44:12 [g2]: What are the switchbox and linuxrc changes rwhitby is making to openslug? May 05 08:45:05 My NSLU2 is "off" :-( May 05 08:45:06 <[g2]> VoodooZ_Work, I think to fix it for OpenSlug was something relating to /mnt/tmpmnt May 05 08:45:47 yeah, Nail mentionned that but I was curious if he was adding any new feature. May 05 08:51:15 I noticed that the last few changesets said switchbox was being removed??? May 05 08:51:28 yeah May 05 08:52:37 removed it totally? replacing it with what? May 05 08:54:29 lunch time! May 05 09:01:58 From what I gather, it's basically gone. Something to do with booting directly to jffs2 May 05 09:02:14 Also saves one erase block in the process May 05 09:02:17 (not both) May 05 09:04:21 * NAiL hasn't really been listening. May 05 09:05:06 <[g2]> NAiL, are you familiar with NTP ? May 05 09:05:40 [g2]: Not much. I've set it up a few times May 05 09:05:54 <[g2]> do you know what the delay is measured in ? May 05 09:06:14 No idea May 05 09:06:51 hmm.. There's heaps of free space inside the slug case :) May 05 09:07:29 <[g2]> you should see inside the RV case :) May 05 09:08:19 I'm in the process of adding a LCD display to the slug, together with an internal usb-hub, usb->lpt converter ;) May 05 09:08:36 (and of course my trusty 512mb flash-stick) May 05 09:08:59 <[g2]> pls do tell me about the LCD May 05 09:09:18 <[g2]> how are you going to drive it ? May 05 09:09:21 heh... standard hd44780 display, paralell LCD May 05 09:09:39 No backlight though May 05 09:09:43 <[g2]> I didn't know there was such a thing May 05 09:09:45 <[g2]> url ? May 05 09:09:52 urm... hang on May 05 09:10:17 http://www.doc.ic.ac.uk/~ih/doc/lcd/ <-- LCD spec May 05 09:10:25 <[g2]> thx May 05 09:10:31 http://slug.repvik.info/lcd.png <-- halfway done cutting ;) May 05 09:10:48 <[g2]> what do they cost ? May 05 09:11:15 They aren't expensive, afaik. I got my 5 displays for free :) May 05 09:11:20 <[g2]> is that just driven by DIOs ? May 05 09:11:38 DIO? May 05 09:11:48 <[g2]> Digital Input / Output May 05 09:12:21 requires 5V as well, so I need to pull that from somewhere else May 05 09:15:21 http://lcdproc.sourceforge.net/docs/current-user-html/c424.html#hd44780-connections <-- how to wire from a lpt-port May 05 09:15:58 Hmm... wonder how much power the display draws May 05 09:21:03 [g2]: The slug is just wide enough for the display to fit May 05 09:21:21 <[g2]> cool May 05 09:21:48 And there's still plenty room for a usb-hub (Which I'll need) May 05 09:22:14 And if I'm lucky, there's even room for an additional usb-key in the bottom of the slug May 05 09:22:44 But by then I guess I'm pushing the power limit on the slug May 05 09:23:11 <[g2]> maybe not May 05 09:23:31 Haven't figured out how much the display draws yet, but I guess it's a little bit May 05 09:23:43 <[g2]> I'd guess the two usb ports are rated at .5A@5V as required by USB May 05 09:23:45 together with the usb->lpt converter, and the internal hub May 05 09:24:45 IIRC, the slug without anything connected pulls ~300mah? The power brick delivers ~2A May 05 09:25:57 Oh well.. I'll find out later whether or not it goes flaky on me ;) May 05 10:18:30 Hmm.. who was building openslug with -Os and uClibc again? May 05 12:12:00 <[cc]smart> [g2]: NTP delay as in ntpq -p ? May 05 12:13:29 <[cc]smart> ms that would be May 05 12:37:35 <[g2]> [cc]smart, thx May 05 12:38:13 <[g2]> [cc]smart, do you use ntp much ? May 05 12:38:28 <[cc]smart> usually yes, but on the slug it causes me pain May 05 12:38:43 <[cc]smart> for running a real server, you cannot go without ntp or similar May 05 12:38:58 <[g2]> I'd like to get ntp working properly on OpenSlug May 05 12:39:02 <[cc]smart> but i have not seen the slug survive for long with ntp running May 05 12:39:04 <[g2]> what are we missing ? May 05 12:39:17 <[cc]smart> functionally it seems to work perfect May 05 12:39:24 <[cc]smart> once tickadj has been set May 05 12:39:42 <[cc]smart> but somewhere after 12-24 hours, the slug dies May 05 12:40:11 <[cc]smart> btw, the nslu HP is wrong on its tickadj 10000 May 05 12:40:13 <[g2]> what do I have to do to setup tickadj properly ? May 05 12:40:21 <[cc]smart> needs to be 10100 in my case at least May 05 12:40:32 <[cc]smart> otherwise NTP doesnt get stable May 05 12:40:51 <[cc]smart> just call "tickadj 10100" before starting ntpd May 05 12:41:01 <[cc]smart> actually, best take three steps one after the other May 05 12:41:04 <[cc]smart> tickadj 10100 May 05 12:41:11 <[cc]smart> ntpdate May 05 12:41:14 <[cc]smart> then ntpd May 05 12:41:48 <[cc]smart> ntpd should the clock in May 05 12:41:55 <[cc]smart> you can verify with ntpq -p May 05 12:42:07 <[cc]smart> that should shown an asterisk in front of a good server May 05 12:42:15 * [g2] hugs [cc]smart May 05 12:42:15 <[cc]smart> and the other one ntptime May 05 12:42:36 <[g2]> ntpq -p May 05 12:42:36 <[g2]> remote refid st t when poll reach delay offset jitter May 05 12:42:36 <[g2]> ============================================================================== May 05 12:42:36 <[g2]> ns3.oit.unc.edu 0.0.0.0 16 u - 64 0 0.000 0.000 4000.00 May 05 12:42:36 <[g2]> PARSLEY.SRV.CS. GPS0.net.cmu.ed 2 u 3 64 1 67.397 -0.026 0.004 May 05 12:42:36 <[cc]smart> ntptime line 5 May 05 12:42:37 <[g2]> brian.cede.psu. 0.0.0.0 16 u - 64 0 0.000 0.000 4000.00 May 05 12:42:44 <[cc]smart> you just started May 05 12:42:56 <[cc]smart> see the poll column May 05 12:42:57 <[g2]> yeah, but it was way off before May 05 12:43:00 <[cc]smart> its 64 secs May 05 12:43:19 <[cc]smart> when you repeat ntpq -p you'll see rising values in when column May 05 12:43:26 <[cc]smart> thats seconds May 05 12:43:26 <[g2]> should it be different from 64 ? May 05 12:43:37 <[cc]smart> once when > poll the server is verified May 05 12:43:45 <[cc]smart> poo will raise over time May 05 12:44:05 <[cc]smart> once ntp trusts the servers and its own understanding of time it will reduce polling interval May 05 12:44:13 <[cc]smart> poo :) May 05 12:44:18 <[cc]smart> aha, poll that is May 05 12:44:43 <[g2]> Ok, so it's checking ever 64 seconds now until it has some confidence May 05 12:44:49 <[cc]smart> yes May 05 12:45:10 <[cc]smart> once it has accepted a server and identified the most trustworthy you'll se an asterisk in front od it May 05 12:45:23 <[cc]smart> remember using ntptime, too May 05 12:45:50 <[cc]smart> once ntp trusts its time and has a feeling for your local clock drift two things will happen May 05 12:46:05 <[cc]smart> ntptime now shows error for you May 05 12:46:14 <[cc]smart> in line 1 May 05 12:46:18 <[cc]smart> paste that here May 05 12:47:21 <[cc]smart> have it ? May 05 12:47:42 <[g2]> ntptime May 05 12:47:42 <[g2]> ntp_gettime() returns code 5 (ERROR) May 05 12:47:47 <[cc]smart> exactly May 05 12:47:54 <[cc]smart> it doesn't understand drift yet May 05 12:48:09 <[cc]smart> once you see the asterisk that should change to May 05 12:48:10 <[cc]smart> ntp_gettime() returns code 0 (OK) May 05 12:48:16 <[g2]> ok May 05 12:48:23 <[cc]smart> then drift is decided May 05 12:48:27 <[g2]> do you have a driftfile configured May 05 12:48:33 <[cc]smart> certainly May 05 12:48:47 <[cc]smart> it's a long process for NTPd to understand its own drift May 05 12:48:51 * [g2] needs to add one May 05 12:48:55 <[cc]smart> line 5 of "ntptime" May 05 12:49:06 <[cc]smart> frequency 114.997 ppm May 05 12:49:16 <[cc]smart> that's what my ntpd knows about its drift May 05 12:49:32 <[cc]smart> this should also appear once a server was elected May 05 12:49:47 <[cc]smart> if this would become greater than 512 ntpd gives up May 05 12:50:15 <[cc]smart> this value, once calculated is saved in the driftfile and reused when you restart ntpd May 05 12:50:33 <[g2]> how often does the driftfile get written ? May 05 12:50:43 <[g2]> should it be in ramdisk area ? May 05 12:50:49 <[cc]smart> once you have this value in ntpdate line 5 and the driftfile has then not be created and contains this number, you know it has a problem, not beig able to write the file May 05 12:51:11 <[cc]smart> drift file needs to be stored so that it can be reread when ntpd is restarted May 05 12:51:29 <[cc]smart> it gets more precise the longer ntpd is/has been running May 05 12:51:44 <[g2]> I'm wondering howw often it gets written so as not to burn out the flash May 05 12:52:04 <[cc]smart> i'm not 100% sure May 05 12:52:11 <[cc]smart> but i think it was in order of 30 inutes May 05 12:52:42 <[cc]smart> certainly not faster than polling interval, which goes to 1024 (seconds) May 05 12:52:59 <[cc]smart> but the actual value should be verified May 05 12:55:18 <[cc]smart> [g2]: actually you're missing local clock setting May 05 12:55:22 <[cc]smart> in you ntpd.conf May 05 12:55:27 <[cc]smart> just npticed May 05 12:55:45 <[cc]smart> add these: May 05 12:55:47 <[cc]smart> server 127.127.1.0 May 05 12:55:47 <[cc]smart> fudge 127.127.1.0 stratum 14 May 05 12:56:17 <[cc]smart> 127.127.1.0 is no real ip. this is a "logical" ip. this specific one references bultin HW clock May 05 12:56:46 <[cc]smart> fudge stratum 14 sets it to low trust May 05 12:59:38 <[cc]smart> just found, the update rate is once per hout May 05 12:59:56 <[cc]smart> probably depending on the implementation, but ntp.org on is once per hour May 05 13:00:07 <[cc]smart> http://info.ccone.at/INFO/NTP/miscopt.htm May 05 13:01:52 <[g2]> thx for the link ! May 05 13:52:38 <[g2]> dyoung, hey! May 05 13:57:44 hello. May 05 17:27:41 <[g2]> VoodooZ_Work, ping May 05 17:28:30 <[g2]> hey gerdi_ May 05 17:28:47 <[g2]> ~seen gerdi_ May 05 17:33:20 [g2]: :) May 05 17:33:59 <[g2]> :0 May 05 17:34:00 <[g2]> :) May 05 17:35:17 * NAiL pings VoodooZ_work as well May 05 17:35:49 rwhitby-away: ping? May 05 17:35:59 <[g2]> he probably forgot to change _work and is away May 05 17:36:06 <[g2]> rwhitby-away, is away May 05 17:37:02 hmm May 05 17:38:25 VoodooZ's latest build of openslug failed to boot (kernel panic). I'm guessing it has something to do with the linuxrc-changing-thingy. May 05 17:38:42 Just trying to figure out whether or not it's worth building ;) May 05 17:38:48 <[g2]> actually I was going to bug him about the RTC May 05 17:39:58 for ntp? May 05 17:40:14 <[g2]> NTP is the 3rd issue May 05 17:40:30 oh May 05 17:40:32 <[g2]> The first is there an error where the clock starts off a month off May 05 17:40:44 <[g2]> the 2nd is the resolution of the time May 05 17:40:51 <[g2]> and finally there's fixing NTP May 05 17:40:58 ah... yes. May 05 17:41:26 <[g2]> NTP is running OK, butas ccsmart says it locks up between 12-24 hours May 05 17:41:47 Yeah, I seem to remember some reports about that from a long time ago May 05 17:42:16 glc: ping? May 05 17:42:45 Btw, do you know where to set CFLAGS? May 05 17:46:07 Going to build with uClibc and -Os, hopefully... May 05 21:07:57 NAiL **** ENDING LOGGING AT Thu May 05 23:59:56 2005