[Xastir] Raspberry Pi with Xastir

[Lee Bengston]

On Wed, Apr 9, 2014 at 7:53 PM, Lee Bengston <lee.bengston at gmail.com> wrote:

> On Sun, Feb 23, 2014 at 4:17 AM, Ray Wells <vk2tv at exemail.com.au> wrote:
>
>> Hi Lee,
>>
>> It didn't competely dawn on me during early use of of the Pi that the
>> supply that was most reliable was the 5.22v mobile phone charger. Back then
>> I couldn't do extended running on that ps because it belonged to the phone
>> of SWMBO and she needed it every second day. Also back then I didn't
>> measure the voltage of that supply, I took its 5v on face value. It was
>> only when the DTV dongle crashed so soon and repeatedly on other supplies,
>> including clean, linear supplies,  that I checked the phone charger more
>> closely.
>>
>> The adsb Pi is now running from a Murata smps that delivers 5.0v but,
>> that Pi has the polyfuses shorted to keep USB voltage as high as possible.
>> The new Pi is running from a linear supply set to 5.1v and that puts an
>> adequate voltage on the USB sockets - 4.85v from memory.
>>
>> You might have a point about cpu loading being in the equation. FWIW
>> dixprs uses around 33% of cpu. Load average on that Pi is constantly around
>> the 1.0 mark but fpac contributes a bit to that. The adsb Pi on the other
>> hand shows load average of around 0.4.
>>
>> I await the outcome of you running higher supply voltage.
>>
>> Ray vk2tv
>
>
> I took the lazy way out and ordered a 5.1v power supply that is supposedly
> designed for the "special needs" of the Raspberry Pi.  I believe it's rated
> at 5.1v at 1.5 amps.  It took a long time because it was shipped from the
> UK, but it seems to be paying off.  I'm at 17 and a half days now using
> Dixprs as an Igate.  It never lasted more than 14 in that mode before, and
> typically it was only 10-11 days.  I'll run it a bit longer to be sure -
> then I'll add ldsped into the equation.  With ldsped running, it usually
> took just a few days to crash.  If that's stable, I'll finally be able to
> fire up Xastir and APRSIS32 on other computers in the network and share the
> TNC with the Pi - knock on wood - with no more crashing.
>
> ​Following up on this - intended to a long time ago, but then the list was
down due to the server move.  Only a day or two after I wrote the above on
April 9th, the Pi froze yet again.​

​ That meant I had gone longer than ever before with the Dixprs IGate using
the new power supply, but I still wasn't satisfied.  ​In late May I started
down the path of trying out Arch Linux, and during the course of playing
around trying to share a boot partition between Raspian and Arch, I ended
up booting into Raspian while running the Arch kernel (long story - won't
go into detail, but my boot partition is on the SD card while everything
else is on a hard drive in a USB enclosure).  I used apt-get dist-upgrade
to get the Raspian kernel back, and I think I ended up with a newer kernel
than the one I had started with.

The upgrade made me realize that I was about 6 months behind on updates for
Raspian.  After the updates were finished I started Dixprs again, added
ldsped into the equation, and ran Xastir, APRSIS32, and YACC on a separate
computer each with AGW interfaces into ldsped.  I had all 3 running for at
least 10 days, and I have had at least one of the 3 running for virtually
all of the 25 days plus that the Pi has been up.  Prior to this trial, I
had never seen a week of uptime when both ldsped and Dixprs were running.
Also when I SSH into the Pi now, the response is normal - it has not slowed
down like it used to after several days of uptime.

SO, either something in the package updates fixed the stability issue or
it's related to the LXDE desktop.  Previous trials involved starting all
aprs software on the Pi from a terminal running on the LXDE desktop.  This
time I started Dixprs and ldsped remotely via SSH.  Overall it finally
looks like I have stability with the Rasberry Pi, and I didn't have to
bypass any fuses.

Lee - K5DAT