Updating GoPiGo O/S 3.0.0 results in an unusable system

Greetings!

While doing some testing today, I discovered something very interesting:

  • Upgrading the GoPiGo O/S 3.0.0 with whatever the current updates are as of today, the result is a GoPiGo O/S that is unusable because it believes that the robot is a Grove Pi.

It doesn’t matter if you do a “plain vanilla” upgrade or a “full-upgrade” the result is the same. A 'bot that doesn’t work.

I do not know why, nor how to correct this at the present time, so I will simply warn people to avoid upgrading the GoPiGo O/S until things get figured out.

Hi Jim,
Can you send that to customer support please?

generally speaking doing a full-upgrade on GoPiGo OS is playing Russian Roulette. It has the potential of destroying a lot of the customization we did. However a standard upgrade should be safe. I’ll take a look.

1 Like

I will do that, however messages to support have no visibility here so I wanted to make sure others knew of this problem.

If you want, I can send the output of apt-get upgrade to a file and send it to you.

I cannot prove it, but I strongly suspect that an update to Python might be the issue.

Update:
This is, apparently, not entirely true - but may also depend, (at least to some extent), on the boot method being used.

Theoretically, BOTH boot methods should produce:

  • Identical results with each other.
  • Results identical with booting directly from a SD card.
     

Apparently, this is not always true, nor can it be guaranteed.  I need to do additional research on the Raspberry Pi’s boot method.

Also note that this issue - boot method - appears to affect Raspbian for Robots, at lest on the Pi-4.

Update:
I have inquired at the usual Fountains Of Wisdom on the Raspberry Pi forum,
Viz.: Are "start[x].elf and fixup[x].dat unique to a particular distribution? - Raspberry Pi Forums

. . .and there does not seem to be anything obviously wrong with what I am doing so far.  I have extended the question with additional details and hopefully I will get some information that will help me narrow down these issues.

There are a number of things I want to do but cannot do them until I am absolutely sure that the individual OS’s are behaving in a uniform and expected manner.

1 Like

Update to add a necessary answer and complete this thread:

Despite the fact that the various startup files contained within the boot partition appear to be absolutely identical from version to version and update to update, even to the point of being the exact same size, they are NOT identical.

Being the same size, (depending on the file), may be a requirement of the GPU starting up the Raspberry Pi.

Kernel files are almost always different and are bound tightly to the version and update level of the system.

Therefore my original assumption that boot files were identical and interchangeable between versions is not true.  In fact, many of the issues I experienced while experimenting with boot configurations was caused by incorrectly mixing boot partition files.

2 Likes

Update:

This topic has been considerably extended and expanded here:

This issue appears to affect both versions of GoPiGo O/S - 3.0.0 and 3.0.1.

Likewise. . . .

Though important and valid, it is not the root cause of the problem with updating the GoPiGo.  Therefore the research into this problem continues in the thread noted above.

2 Likes