Hi all, I picked up the research project using the Stretch last Spring. The students before me encountered a massive error that corrupted the SSD. I believe it was traced back to powering off the stretch incorrectly, not shutting down fully before flipping the switch.
Thankfully the bad sectors were still readable and they got their progress back.
Well, they talked to support and eventually replaced the SSD. This is where I came on the project, and upgrading the OS to get ROS2.
At some point over the summer, the robot got powered on, or maybe unplugged and just depleted the battery over a few days/weeks, and after coming back this fall, the same thing happened - the SSD bricked itself.
To be specific, this is the error we recieved:
Medium critical error, dev nvme0n1, sector ##### op 0x0: (READ) flags 0x80700 phys_id #
Then the NUC just refused to power on after that.
Thankfully, after powering the NUC on manually, I loaded a boot drive and was able to recover my research.
But I have no clue what to do from here. This seems like a major issue, that can happen by accident.
Is it better to just use an external drive? Or even just a USB stick with 128Gb storage and load the OS into it? Would I need to worry about I/O read speeds if I go this route? I’m really not too familiar with hardware issues like this, or the best practices for it.
I really need to continue, but it’s just not feasible to continue replacing the SSD whenever a minor accident occurs or the bit is powered off a second too soon.
Thanks all for any help or advice you have!