Update to no power to car

[Percentage]

The car was in the Scottsdale SC since Saturday. I just received a call from SA. He said the latest OTA update before I picked up the car, did not download properly and one of the files were corrupted. This caused a problem with the charging from the HV to the 12V system. Engineering is sending the full file this evening and they feel once it is loaded correctly they can release the car tomorrow afternoon.

 

[Coho]

The car was in the Scottsdale SC since Saturday. I just received a call from SA. He said the latest OTA update before I picked up the car, did not download properly and one of the files were corrupted. This caused a problem with the charging from the HV to the 12V system. Engineering is sending the full file this evening and they feel once it is loaded correctly they can release the car tomorrow afternoon.

Did they offer why the download corrupted? Poor connection, interrupted connection, or stuff happens?

 

[HC_79]

May be time to activate those USB ports so you can download the update and install via USB in event of an update failure.

 

[hydbob]

May be time to activate those USB ports so you can download the update and install via USB in event of an update failure.

Lol this would probably cause way more issues...

 

[Percentage]

Did they offer why the download corrupted? Poor connection, interrupted connection, or stuff happens?

No, they just said the files did not load correctly.

 

[wolfepack]

The car was in the Scottsdale SC since Saturday. I just received a call from SA. He said the latest OTA update before I picked up the car, did not download properly and one of the files were corrupted. This caused a problem with the charging from the HV to the 12V system. Engineering is sending the full file this evening and they feel once it is loaded correctly they can release the car tomorrow afternoon.

Wow, I have not heard about software corruption yet in an update. Hope it’s a quick fix.

 

[stratus]

Curious, did you download it directly in the car or via the phone app? For tech experts, could that potentially make a difference?

 

[Rogue]

Curious, did you download it directly in the car or via the phone app? For tech experts, could that potentially make a difference?

There should not be a difference. Keyword being should. It is a little shocking though that a hash check is not run to ensure no corruption, tampering, or missing files. If the corruption occurred mid-way through an update... That is a little harder to resolve, but again, a hash should be present to confirm correct installation with a roll back procedure if not.

These edge cases are too often not accounted for in software, or it is a case of not knowing the edge case until it occurs. Google an article about a driver who registered their plate as null in an attempt to subvert the occasional ticket. The epoch fiasco with 32-bit systems is another.

 

[stratus]

There should not be a difference. Keyword being should. It is a little shocking though that a hash check is not run to ensure no corruption, tampering, or missing files. If the corruption occurred mid-way through an update... That is a little harder to resolve, but again, a hash should be present to confirm correct installation with a roll back procedure if not.

These edge cases are too often not accounted for in software, or it is a case of not knowing the edge case until it occurs. Google an article about a driver who registered their plate as null in an attempt to subvert the occasional ticket. The epoch fiasco with 32-bit systems is another.

I just read the article on wired.com; really interesting. Thanks for the explanation.
As somewhat of a tech neophyte, learning about these topics has become really enjoyable.

 

[Bishop]

These edge cases.....

There is nothing "edge case" about what happened here. These are the things we embedded engineers design and test. The hash of any upload is checked, there should be three versions in flash. The current version, the previous version and the golden image. These are really basic and you learn this in the first few years of your career (if not before) if any decent company hires you.

 

[Rogue]

There is nothing "edge case" about what happened here. These are the things we embedded engineers design and test. The hash of any upload is checked, there should be three versions in flash. The current version, the previous version and the golden image. These are really basic and you learn this in the first few years of your career (if not before) if any decent company hires you.

You have been among good embedded engineers and decent companies. I thought the days of no version control in projects had died, but I am at two consecutive companies where that has not been the case. Speaking of tests, all the companies I have worked at have either had little to no automated testing, regression testing, and system testing in place.

We are of the same mind. It should never be an edge case, but so should the presence of version control always be in place and testing of code.

Anyways, has there been any news on the cause or fix?

 

[Bishop]

You have been among good embedded engineers and decent companies. I thought the days of no version control in projects had died, but I am at two consecutive companies where that has not been the case. Speaking of tests, all the companies I have worked at have either had little to no automated testing, regression testing, and system testing in place.

We are of the same mind. It should never be an edge case, but so should the presence of version control always be in place and testing of code.

Anyways, has there been any news on the cause or fix?

lol, version control at the first company I ever worked for was "copy the project into a directory on the network with `_VXX` extension". Never again. And this was code for embedded computers in airplanes.