RPi 3A+ Update Issue

[N0XRV]

Let me start by saying that Pi-Star ROCKS! Keep up the good work!

For the last two years or so I have had very few problems working through the various Pi-Star Updates and Upgrades. However, this past week I seem to have hit a snag. So, I’ll share some info.

Individually, I run 3 RPi 3A+’s and 5 RPi0-W’s in DMR-only or YSF-only modes. The RPi’s are paired primarily with ZUMspots, a couple of N5BOC’s and one GD-77 (OpenGD77). — Don’t ask -hah! I keep them fairly up to date within a day usually with respective Updates and Upgrades. All RPi’s are running 4.1.1 and were running flawlessly until the latest Update.

The three RPi 3A+’s, after initial boot up and successful Master login, would allow me to log into Admin or Configuration but would soon thereafter lock up and become unresponsive. Additionally, the units would not respond to transmissions or Keeper commands through the radio. I also noticed that if I just allowed the units to just boot to Master login and then delayed a few minutes or so, they would lockup as just described. Nothing worked except rebooting. The whole process is unfortunately repeatable.

Interestingly, none of the RPi0-W’s have exhibited any anomalies whatsoever.

Guess it’s time to re-flash some images for the RPi 3A+’s.

73,
Malcolm, N0XRV

[MW0MWZ]

Thank you for the feedback, the upgrade to 4.1.1 only adds some WiFi firmware, so I don't *think* thats an issue, but I will check it out...

[MW0MWZ]

OK first some good news, I can re-create the issue, managed to get a 3A+ to lock up here, it locks up before upgrading to 4.1.1 also, so I am still pretty sure its not related to the 4.1.1 upgrade.

The unit was working pretty good until it started installing updates, and then the database gets toasted, and now every time it runs the unattended-upgrade process, it locks the unit up, I can see how that would be annoying - and I don't think Pi-Star causes that...

Testing a fix for this at the moment, more news for you in a few mins...

[N0XRV]

I forgot to mention that I was also not successful in accessing the hotspots through the web page interface using either the host name or IP Address. When the issue occurs, their presence on my network also disappears.

Malcolm, N0XRV

[MW0MWZ]

Yes I see the same - it's odd, only happens when the Pi is running updates AND you hammer the CPU, outside of that it seems to be OK.
Only managed to get mine to fail and lock up once - still working on it.

[YT5HOK]

OMG! Finally! I thought I was going crazy...

Having the same problem since, I would say, final version 4.1.0 came out. Mentioned it here a month ago: viewtopic.php?f=3&t=2567#p12757

I have two dual mmdvm hats, three Pi's 3A+ and two Pi's 3B+, plus several quality micro SD cards (never mind simplex hats on Pi's Zero). Any combination with Pi 3A+ exhibits same behavior, but works fine on 3B+, with any micro SD card.

But, as I mentioned in that other thread, system doesn't seem fully frozen, just very very unresponsive. Sometimes you can issue commands, if you were already connected via SSH. Also, sometimes it returns back to normal after few hours. Yeah, don't ask me how I know that... I've already went nuts.

[MW0MWZ]

Yes, from what I see here, it looks like the CPU is very busy, and/or the filesystem locks up... - very odd

[YT5HOK]

In last couple of days I did "apt-get update && apt-get upgrade" few times on raspbian buster full image on 3A+ and it passed fine.

Just an update... If it means anything...

[MW0MWZ]

In last couple of days I did "apt-get update && apt-get upgrade" few times on raspbian buster full image on 3A+ and it passed fine.

Just an update... If it means anything...

Yeah I have seen the same here, its all OK again now, I couldn't find the exact cause, but it looks VERY like a memory leak issue that has shown up on the 3A+ units in the past too, I've been holding off the next build waiting for this to be all fixed up, so I'll re-test it all again and if we're past the issue now, I will roll the new build out..

[N0XRV]

Hammered my new 3A+’s images with Updates and Upgrades. All better now. Thanks, Andy!

Malcolm, N0XRV