Re: Odroid heat issue and lock up

[Alan Corey <alan01346%gmail.com@localhost>]

You can probably check the SD with fsck -f <device> also badblocks.
Works best if it's not mounted so you'd need a reader.  I don't really
know NetBSD well, but those in are Open and Free, as well as Linux.
They're standard hard drive tools but they work on SDs too, I was
seeing wacky happenings for a month or more.  Badblocks told me 3% of
my SD was bad.  I've got a USB hard drive connected now.

Try the command "apropos sensors", on an i386 OpenBSD machine it gives me:
uguru(4) - ABIT temperature, voltage and fan sensors
sensorsd.conf(5) - configuration file for sensorsd
sensorsd(8) - hardware sensors monitor
sensor_attach(9) - sensors framework
sensor_detach(9) - sensors framework
sensor_find(9) - sensors framework
sensor_task_register(9) - sensors framework
sensor_task_unregister(9) - sensors framework
sensordev_deinstall(9) - sensors framework
sensordev_get(9) - sensors framework
sensordev_install(9) - sensors framework
Some of those could be installed packages, I'm not sure.

On a Pi under Raspbian there's
vcgencmd commands
commands="vcos, ap_output_control, ap_output_post_processing,
vchi_test_init, vchi_test_exit, pm_set_policy, pm_get_status,
pm_show_stats, pm_start_logging, pm_stop_logging, version, commands,
set_vll_dir, set_backlight, set_logging, get_lcd_info, arbiter,
cache_flush, otp_dump, test_result, codec_enabled, get_camera,
get_mem, measure_clock, measure_volts, scaling_kernel,
scaling_sharpness, get_hvs_asserts, get_throttled, measure_temp,
get_config, hdmi_ntsc_freqs, hdmi_adjust_clock, hdmi_status_show,
hvs_update_fields, pwm_speedup, force_audio, hdmi_stream_channels,
hdmi_channel_map, display_power, read_ring_osc, memtest,
dispmanx_list, get_rsts, schmoo, render_bar, disk_notify,
inuse_notify, sus_suspend, sus_status, sus_is_enabled,
sus_stop_test_thread, egl_platform_switch, mem_validate, mem_oom,
mem_reloc_stats, hdmi_cvt, hdmi_timings, file, vctest_memmap,
vctest_start, vctest_stop, vctest_set, vctest_get" but vcgencmd
probably came from Broadcom and is peculiar to Raspbian.  Seems like
there should be something though.  I can check my CPU and GPU
temperatures with

cpu=$(</sys/class/thermal/thermal_zone0/temp)
echo "GPU => $(/opt/vc/bin/vcgencmd measure_temp)"
echo "CPU => $((cpu/1000))'C"

I just got a DrOk on Amazon which shocked me by telling me that even
with all 4 cores cranking away mining Litecoin my Pi 3B only draws
0.33 amps, yet the "official" power supply is 2.5 amps.  I chased
problems for over a year with my first Pi, turned out to be a bad SD
card socket.

You need more instrumentation somehow.  A $6 Harbor Freight DMM might
help, at least for voltage and current.

On 6/25/17, manlyphall%manlymail.net@localhost <manlyphall%manlymail.net@localhost> wrote:
>> >So I'd like to verify that others are not experiencing this. This way I
>> >can hope that replacing the board will solve the problem. But since I
>> >haven't asked, I really don't know.
>>
>> I've got a bit of a different workload on mine (building from pkgsrc on
>> the
>> eMMC) but my ODROID-C1 (non plus), in the stock case, with all 4 cores
>> running at 1.5GHz hasn't seen any of these issues on 8.0_BETA.
>>
>> First thing's first -- have you tried a different SD / eMMC card, and/or
>> power supply?
>
> Using SDcards, and I have tested with a different one. I have also used
> more than one powersupply. The current one is 5.25V@2A. It was suggested
> by adafruit do to power loss throught cables. I have also used different
> cables.
>
> Webkit2.4-GTK built to completion without lock up using reduced
> settings (after disabling webkit-jit). I've done several tests on the
> same package, and without fail it locks up not long after enabling
> another core, and or increasing above 1200Mhz.
>
> I thought for certian this had to be a power issue. SD cards I have a
> few of and one is new. But power supplies I have many. All at
> least 5v@2A. But I really doubt I am reaching near the draw required to
> put 2Amps to the test.
>
> Webkit isn't the only thing that has locked up. But the temperature in
> the room seems to greatly affect it. During early spring and late
> winter, I never even thought to reduce the cpu settings. It only locked
> up a hand full of timed. And it could have been that usb2ethernet
> adapter had been bumped or something, so I never thought too much about
> it. Ever since it has warmed up it happens all the time. now nothing is
> connected to it other than what I mentioned before.
>
> Thanks for the reply
>


-- 
-------------
No, I won't  call it "climate change", do you have a "reality problem"? - AB1JX
Impeach  Impeach  Impeach  Impeach  Impeach  Impeach  Impeach  Impeach