**** BEGIN LOGGING AT Wed Apr 13 02:59:57 2011
Apr 13 04:58:36 <tinker-f595> a beowolf cluster of sheevaplugs ?
Apr 13 06:07:38 <elFidel> hi. Do we have some dreamplug users here?
Apr 13 06:07:54 <elFidel> if so - how's the experience so faR?
Apr 13 06:07:56 <kerio> not yet, i think
Apr 13 06:08:06 <elFidel> hi ;)
Apr 13 06:16:33 <Tootoot222> elFidel: dmb has one
Apr 13 06:16:35 <Tootoot222> uhm
Apr 13 06:16:40 <Tootoot222> there are others too, i think
Apr 13 06:18:17 <elFidel> do you remember any dreamplug related experience/problem report so far?
Apr 13 06:19:57 <Tootoot222> only thing i know of atm is that there isn't any "stock" firmware downloads, some of the dreamplugs accidentally got shipped with 400mhz ram, 800mhz processors and ubuntu 9.04 on them (but they've fixed that now, and give a full replacement for any that were)
Apr 13 06:20:07 <Tootoot222> the wifi chip is different than in the guruplug
Apr 13 06:20:10 <Tootoot222> which could be good or bad
Apr 13 06:20:19 <Tootoot222> since the guruplug has problems with being a stable wifi client
Apr 13 06:20:33 <elFidel> ok
Apr 13 06:20:47 <elFidel> any heat issues known?
Apr 13 06:21:54 <Tootoot222> nope, heat issues are totally solved afaik; they put the PSU inline with the board, which freed up the wall
Apr 13 06:22:03 <Tootoot222> then they attached the hot components directly onto the wall
Apr 13 06:22:10 <Tootoot222> and there are large ventillation ports
Apr 13 06:22:13 <elFidel> good move
Apr 13 06:22:21 <elFidel> at least it sounds like a good move ;)
Apr 13 06:22:24 <Tootoot222> lol
Apr 13 06:22:35 <elFidel> remembering the guru-issues ;)
Apr 13 06:23:01 <Tootoot222> i think they learned their lesson from the guruplug and actually did some /quality/ testing this time around
Apr 13 06:23:04 <Tootoot222> =p
Apr 13 06:26:38 <elFidel> Tootoot222: thanks for the input.
Apr 13 06:28:44 <Tootoot222> np
Apr 13 06:45:05 <elFidel> where would you guys buy a dreamplug? newIT, globalscale or x? (coming from .de)
Apr 13 11:46:02 <dmb> Tootoot222: i sent my in for now as an RMA
Apr 13 11:46:08 <dmb> because of mine being clocked at 400mhz
Apr 13 11:56:59 <elFidel> hehe
Apr 13 11:57:35 <elFidel> dmb: i'm interessted in your experiences - do you mind being contancted inhere again after time x with some questions regarding the dreamplug?
Apr 13 12:10:31 <dmb> sure
Apr 13 12:10:47 <dmb> if i don't respond though, i'm probably at work
Apr 13 12:15:03 <elFidel> k - thanks
Apr 13 15:23:13 <tinker-f595> elFidel: Here is globalsclae's  dreamplug download page http://code.google.com/p/dreamplug/downloads/list
Apr 13 15:27:08 <elFidel> tinker-f595: thanks
Apr 13 16:15:01 <kusznir> Hi all: we got our dreamplug yesterday, but am having lots of issues indicating they suffer from the same (or worse, actually) heat problems as the guru.  For example, we can't create a tarball of the factory install without it segfaulting, unless we put it in the freezer.
Apr 13 16:15:42 <elFidel> hi
Apr 13 16:15:45 <kusznir> We recently let it boot all the way up (one gig-e interface connected, no other accessories), and about 10 minutes later tried to ssh in, but found it hard locked.  We then tried to reboot it, but it failed to boot successfully, always yielding a different error (most recent being a kernel Oops)
Apr 13 16:16:22 <elFidel> what os is installed? ubuntu/debian?
Apr 13 16:16:53 <kusznir> Gentoo.
Apr 13 16:16:57 <kusznir> With stock kernel
Apr 13 16:16:59 <kusznir> (presently)
Apr 13 16:17:47 <kusznir> Same build we run on the guru's and sheeva's.  (We have 5 sheevas in production for a few years now, and have been working with a pre-fan guru with this build.
Apr 13 16:18:00 <elFidel> k
Apr 13 16:18:26 <elFidel> and right now you have just 1 test-dreamplug right?
Apr 13 16:18:48 <elFidel> just wondering if you can reproduce that problem with > 1 device
Apr 13 16:19:00 <Tootoot222> yeah that's what i was thinking
Apr 13 16:19:05 <Tootoot222> could be a bad one
Apr 13 16:20:08 <kusznir> Yea, right now we have one.  We just opened it up and checked some stuff; if this doesn't fix it, we'll probably go grab my personal one (which just arrived y esterday and hasn't been powered up yet) and give it a try.
Apr 13 16:20:29 <kusznir> We can say that in order to successfully tarball, placing it in the freezer (while powered on and connected) did stabilize it.
Apr 13 16:20:41 <elFidel> hehe
Apr 13 16:21:00 <elFidel> any chance you got a temperature cam available?
Apr 13 16:21:14 <kusznir> Actually, yea, I can go get a fluke FlIR.
Apr 13 16:21:30 <elFidel> hehe - we got a flir too
Apr 13 16:21:32 <kusznir> Got them downstairs in the lab.  But I don't have a baseline to comapair to.
Apr 13 16:21:41 <kusznir> I'll go get it, though.
Apr 13 16:21:44 <kusznir> bbiab
Apr 13 16:21:48 <elFidel> well - it was a stupid idea
Apr 13 16:22:06 <elFidel> out of curiosity. but you are for sure right
Apr 13 16:31:13 <kusznir> Ok, we put it back together and got it booted up, but when I tried to run "make menuconfig" on my kernel tree, on the second line, I got:scripts/basic/fixdep.c:398: internal compiler error: in set_lattice_value, at tree-ssa-ccp.c:471
Apr 13 16:31:34 <kusznir> second run gave me: scripts/basic/fixdep.c:221: internal compiler error: Segmentation fault
Apr 13 16:31:50 <kusznir> How do I know if it has the 400Mhz issue?
Apr 13 16:31:53 <Tootoot222> kusznir: is the unit getting hot to the touch?
Apr 13 16:31:59 <Tootoot222> kusznir: cat /proc/cpuinfo
Apr 13 16:32:24 <kusznir> I don't see 400 or 800Mhz listed there.
Apr 13 16:32:30 <kusznir> Its reading 99*F
Apr 13 16:32:50 <Tootoot222> kusznir: pastebin the whole output, i don't remember the name of the value that identified it
Apr 13 16:33:24 <kusznir> http://pastebin.com/1ZKMM651
Apr 13 16:36:12 <Tootoot222> well that looks normal
Apr 13 16:37:23 <kusznir> ok.
Apr 13 16:37:47 <kusznir> Its even segfaulting on boot with the origional install.
Apr 13 16:45:18 <kusznir> Status Update: so far, it seems to be running OK with the network unplugged.
Apr 13 16:47:36 <Tootoot222> kusznir: is your network switch/router gigabit?
Apr 13 16:47:41 <kusznir> Yep.
Apr 13 16:48:10 <Tootoot222> does it start segfaulting again if you plug the gigabit back in while its running? and does it do the same on a 10/100?
Apr 13 17:31:33 <kusznir> It actually crashed before we got there...but it ran longer.
Apr 13 17:34:07 <kerio> an even better question: can you check the temp with gigabit and with 10mbit?
Apr 13 17:37:51 <kusznir> A couple dreamplug questions: 1) how does one dump the boot loader (backup copy); and 2) is there any versions of uboot available to reflash with?
Apr 13 17:38:14 <Tootoot222> i think you can use nanddump/nandwrite
Apr 13 17:38:25 <Tootoot222> and there is a git repo if you want to compile the latest from source
Apr 13 17:38:38 <Tootoot222> not sure if they have a dreamplug config in there or not though
Apr 13 17:39:12 <Tootoot222> git clone git://git.denx.de/u-boot-marvell.git u-boot-marvell.git
Apr 13 17:39:14 <kusznir> (we're looking to replace the bootloader to support things like loading from ext2 fs, loading from second (external) sd cartridge.
Apr 13 17:39:25 <kusznir> We checked it out, but there doesn't appear to be a config for the dreamplug specifically.
Apr 13 17:40:28 <kusznir> btw: its our understanding that the dreamplug uses NOR instead of NAND for bootloader now.
Apr 13 17:40:32 <kusznir> Does that change anything?
Apr 13 17:42:09 <Tootoot222> hmm, idk
Apr 13 17:43:31 <kusznir> There doesn't appear to be any /dev/mt* devices listed in the stock firmware.
Apr 13 17:49:06 <kusznir> FYI: we had it booted without networking (which makes it more stable, but not stable).  We plugged in networking, and nothing happened (including not dhcp'ing an IP), then ran /etc/init.d/networking restart, and it started printing a stack trace, and hard-locked in the process.
Apr 13 17:49:15 <kusznir> We're now retrieving the second (spare) one.
Apr 13 17:49:36 <kusznir> well, not spare...my personal one (as opposed to the lab one we've been working with)
Apr 13 18:05:11 <kerio> you could have a botched cpu
Apr 13 18:38:34 <kusznir> So, I have my other dreamplug out, how can I stress test this one with the default OS?  There's no compilers, so my normal trick of compiling the kernel doesn't work :(
Apr 13 18:41:30 <kerio> install a compiler
Apr 13 18:41:41 <armin76> kusznir: there's the nice 'stress' tool :)
Apr 13 19:12:47 <Tootoot222> kusznir: it's debian lol, apt-get install gcc
Apr 13 19:16:09 <elFidel> "yes" might produce cpu stress as well
Apr 13 19:19:47 <kusznir> Well, its been creating a bzipped-tarball of the root filesystem and writing it to a USB hard drive for 30 min now without any issues...I think we're safe.
Apr 13 19:21:11 <kerio> so the other one is just botched?
Apr 13 19:22:30 <Tootoot222> ``dd if=/dev/zero of=/dev/null'' will max your processor too
Apr 13 19:23:28 <elFidel> regarding yes: ''yes > /dev/null''  iirc
Apr 13 22:26:45 <kusznir> Anything special required for a dream kernel over a guru one?  (Our guru kernel is hanging on initializing sata)
Apr 13 22:28:40 <kusznir> And anyone know anything about upgrading dream uboot to support ext3 fs and booting from second (external) sd?  There's no notes on how to build uboot for dream on the dream source tree.
Apr 13 22:39:37 <kusznir> It freezes at ata1:SATA....
Apr 13 22:52:46 <Tootoot222> kusznir: where'd you get your kernel from?
Apr 13 22:54:00 <kusznir> Myself.
Apr 13 22:54:07 <kusznir> I need a custom kernel for our application.
Apr 13 22:54:14 <kusznir> It works fine on a guru.
Apr 13 22:54:44 <kusznir> It starts to boot on the dream, and I thought the dream "looks" like a guru for the most part.
Apr 13 22:55:10 <kusznir> If you have a repo of known-good dream kernels, we can test with that.
Apr 13 22:57:52 <Tootoot222> have you tried the ones from sheeva.with-linux.com
Apr 13 22:57:52 <Tootoot222> ?
**** ENDING LOGGING AT Thu Apr 14 02:59:58 2011