Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

2017.50 2275226

This site may earn commission on affiliate links.
Software version 2017.50.1 ceae3b8 was just detected on a Tesla Model S P100D with Autopilot 2 hardware. The highest previous version was 2017.50 2275226. You were alerted per your settings on TeslaFi.com.

I believe this is the fix for the fix
Yes, it is as I got my car (X) yesterday delivered with 2017.50_ and it was fixed last night with 2017.50.1 -- created this thread on 50.1 -- 2017.50.1 ceae3b8 is out per TeslaFI
 
After 400km drive today
I notice the autopark is greatly improved
It detects the parking spot much better than .48
However, when I tried the autopark today, the car scratch the curb, good thing is little damage on the bottom of the bumper.

I have AP2.5 and EAP didn't break for me, instead, the breaking is one bit smoother than .48

I don't think autopark detects parking spots. It just puts itself between existing vehicles. That's why you can't use autopark unless there are vehicles present.
 
I don't think autopark detects parking spots. It just puts itself between existing vehicles. That's why you can't use autopark unless there are vehicles present.

Yeah, that is how it works now, generally speaking. Though I’ve seen auto park useable in situations where there is a vehicle on one side and say a wall (like in a parking structure) on the other side. So it seems it will auto park your car between vehicles or other solid obstacles like walls. I’ve had it be hit and miss whether it will function with a column on one side as opposed to a solid wall. In most cases, it has not worked with a column...but in a few cases it has. I think it may depend on how deep (i.e. how far back into the parking space the column or half-wall goes) it is.
 
I actually spent some time in a service center and got some backend information about the updating process. Once the vehicle has been staged for an update, Tesla can see the download status near-real time as it is downloading to the car. He said that download size is larger depending on how many updates back you are. Also, the real reason wifi is preferred is because it takes SO darn long through LTE (especially since the LTE appears to be throttled). I had mine on the SC wifi, and it was taking quite a while, so I ended up leaving (the car had already downloaded 3/4 of the package, around 700 megs or so) and I was driving to Chicago (about a 5 hour drive). Both me and the tech assumed I'd have it fully downloaded within 2 hours, where I could install it while Supercharging. Well, I didn't get it by the Supercharger, so I called Tesla Support, and they were able to check the status of it. The saw that it was 87% downloaded, and still going, so I had to just be patient... It didn't fully download until I was pulling into my subdivision. 5 hours later. So, that's why wifi is that much better - it's just a lot quicker than the allegedly throttled LTE.
 
the car had already downloaded 3/4 of the package, around 700 megs or so
No way! the full package is 820 or so megs. They do a binary diff so typical download is in tens of megabytes at most.
if you have ap2, that adds about 310M for the ap firmware download.

The longest part of the firmware download process is to apply the binary diff and write out full 800+ megs onto the emmc and then replicate that same write onto IC computer and then read it back to calculate integrity of the whole thing - needless to say neither of those depend on your download speed.
 
No way! the full package is 820 or so megs. They do a binary diff so typical download is in tens of megabytes at most.
if you have ap2, that adds about 310M for the ap firmware download.

The longest part of the firmware download process is to apply the binary diff and write out full 800+ megs onto the emmc and then replicate that same write onto IC computer and then read it back to calculate integrity of the whole thing - needless to say neither of those depend on your download speed.

Then perhaps it was 70mb out of 90mb, rather than 700mb out of 900mb. I was looking over the tech's shoulder at his screen. I can tell you 100% though the download speed was crap, both on wifi (which I was on for over 45 minutes, watching it download via his laptop) and on LTE, which 2 hours after I left the SC had still not completed when I called Support while Supercharging.

I DO have AP2, btw.

EDIT: On second thought, I know it had to be more than that, because the tech commented how big the package was... and it took forever.
 
Then perhaps it was 70mb out of 90mb, rather than 700mb out of 900mb. I was looking over the tech's shoulder at his screen. I can tell you 100% though the download speed was crap, both on wifi (which I was on for over 45 minutes, watching it download via his laptop) and on LTE, which 2 hours after I left the SC had still not completed when I called Support while Supercharging.

I DO have AP2, btw.

EDIT: On second thought, I know it had to be more than that, because the tech commented how big the package was... and it took forever.
Well, without knowing the exact wording what is it you saw it's a bit hard to comment on it.
It is possible mothership would opt to do a full download instead of a diff, but it is usually only when what you are running is different from what you are getting which does not appear to be the case with you.

Same version to same version is only ~13k, I don't know how far back you were behind on the update of course, but unless something drastically changed in 17.50 it still should not be all that big of a diff.
 
I drove by my SC this evening on my way home when they were closed and sat outside the service bay enterance to see if I could connect to WiFi (since my car had been there twice and received updates each time). A separate article I read said that some believe simply hooking up to a SC WiFi will trigger an update. Well, there was no magic moment. There was no unsecured wifi to connect to.

My 50 reported failed but everything seems to work. Maybe they know that and I won't be pushed the .1 version.
 
Well, without knowing the exact wording what is it you saw it's a bit hard to comment on it.
It is possible mothership would opt to do a full download instead of a diff, but it is usually only when what you are running is different from what you are getting which does not appear to be the case with you.

Same version to same version is only ~13k, I don't know how far back you were behind on the update of course, but unless something drastically changed in 17.50 it still should not be all that big of a diff.

I went from .42 to .48, so I gained Easy Entry and Chill mode, plus whatever other behind the scenes updates were there.

I can tell you 100% I was at the SC connected to the wifi, and it made it to around 75% after about 45 minutes. We were both surprised at how long it was taking, but we stood there and BS'd for a while. In fact, I actually saw a Model 3 pull up - the second one they had come through - and got some pics of it while I was waiting. They wouldn't let me touch it, but I saw it. :)
 
Well, I had 17.48->17.48 "upgrade" - download size: 13K (no ape firmware download because I already had it).

From the moment it was pushed to "your firmware update is ready to install" it took about 40 minutes.

Code:
2017-12-15T10:43:06.257268-08:00 cid : cid-updater:23312: HTTP Client sid 95 using session->xfer_filename = /var/spool/cid-updater/handshake-response.part-89
...
2017-12-15T10:51:02.780041-08:00 cid : cid-updater: 9226: handshake command=patch http://...
...
2017-12-15T11:19:03.347204-08:00 cid QtCarServer[1671]: [SoftwareUpdateAvailableNotification] INFO Alert UPDATE_AVAILABLE triggered with value true - sending
 
Updated overnight from 2017.44 to 2017.50 2275226 without any issues. My MX 100D is only 1 month old, no clue whether that is of any influence to the upgrade process (other hardware?).
Perhaps your are on to something. I had problems with my new X but went from .44 (Thu test drive) to .50_ (Sat pickup) where I had no AP. Then to a successful AP usable .50.1 (Sat night).
 
Well, I had 17.48->17.48 "upgrade" - download size: 13K (no ape firmware download because I already had it).

From the moment it was pushed to "your firmware update is ready to install" it took about 40 minutes.

Code:
2017-12-15T10:43:06.257268-08:00 cid : cid-updater:23312: HTTP Client sid 95 using session->xfer_filename = /var/spool/cid-updater/handshake-response.part-89
...
2017-12-15T10:51:02.780041-08:00 cid : cid-updater: 9226: handshake command=patch http://...
...
2017-12-15T11:19:03.347204-08:00 cid QtCarServer[1671]: [SoftwareUpdateAvailableNotification] INFO Alert UPDATE_AVAILABLE triggered with value true - sending


I wonder if there is a way I could get a copy of my logs just to see what took it so long?!? You ever heard of that happening??