Solid magenta after a few minutes of flashing magenta?

Hmm I hadn’t realised that I’d have this same issue with every bit of code that I wanted to download :frowning: As it took me about 30 attempts the first time it means it just not viable currently for me.

My issue is not the power availability - it’s the robustness of the download protocol in busy networks. I understand there’s a fix coming for that too. My core will be receiving a lot of interspersed UDP broadcast packets too due to some devices on the network. When this fix is incorporated in the new firmware can you post and let me know and I’ll dig my core back out of the box.

K

Hi @xAPPO,

@satishgn committed a fix for the “stuck during flash” issue to the communications library here: https://github.com/spark/core-communication-lib/commit/a72692553ebeb66cca1623e1a67a20f260bf5983 , and that code should now be available when flashing from the web IDE. We’re also working on a fix for a more aggressive retry / recovery when faced with lots of lost packets. :smile:

Thanks,
David

Doesn’t that commit just cause a reset after 20 seconds if the update locks/fails ? My issue is that invariably I only get somewhere between 10 and 30 magenta flashes before it stalls. I dont know how much data each flash represents but I’m guessing I need at least 64 flashes (blocks) so I’m a away off being ‘almost’ there. I must have tried over 50 times in the last hour or so.

Something is broken in the TCP update and I think I need to wait for your other fix - although quite how I’ll be able to apply the update I’m not sure. Are others not seeing this ?

Hi xAPPO,

Yes, you’re totally right, that commit prevents the core from locking up forever after a dropped packet.

Networks with a lot of dropped packets are the most likely to have trouble flashing the core right now, but I’m also seeing a number of flashes fail here on my relatively stable network, so I’m looking forward to that fix I mentioned earlier. However that pausing and restarting is protecting you from corrupted firmware. :smile:

That over the air flash fix should hopefully be ready sometime this month, but if you’re seeing a lot of failed updates, you could always flash code over usb in the meantime, or wait for the command line tool, or local cloud which would also probably help.

Thanks,
David

For USB use can I download a successfully compiled bin file from the cloud compiler to my PC or must I install GCC ?

There is quite a bit of adjacent channel RF noise here with Sonos aside from the standard WiFI.

K

There should be a way to get the binary back from the API, but I think that endpoint is missing at the moment. The create binary endpoint used to serve this function, but we will definitely expose this again in the near future. I’m adding this to the top of the API backlog. :slight_smile:

1 Like

Hey @xAPPO , Just wanted to let you know we haven’t forgotten about this, I was hoping to get this “compiled binary” API endpoint implemented today, but the change is much more involved than suspected. Will let you know when we add this feature. Additionally, when we add this, we’ll mention it in our dev blog here:

blog.spark.io

The fix to the downloader firmware to better handle downloads on higher traffic networks is the thing I’m really waiting on as I can’t use my unit currently at all. I realise the cloud compiler API change provides another route but it’s not as click and go as the WiFi update.

Also just to mention again… I am surprised that my problem is so acute and yet others obviously are not having this problem or there would be more posts here. Differences here are that there is quite a bit of network traffic, several wireless AP’s and also that there is a lot of UDP broadcast traffic. I keep worrying about the latter inparticular. I have tried several AP’s with same results but hope to try a friends soon. It is nigh on impossible to get a complete download even with 100 attempts.

We have another improvement in the works which should make the firmware upload more robust and also faster. In the meantime, do keep us posted on what happens when you try a friend’s AP.

1 Like

I have had this problem ever since i received it. 9 out of 10 times i flash it, it gets stuck and i have to reboot.
It is getting very frustrating.

@seulater certain AP routers seem to make the OTA more unreliable. Did you get an opportunity to try it on a differrent router? We have been working hard to get the OTA more stable across all platforms.

1 Like

Good to know, i will try another router tomorrow and respond back on the results.

Just an FYI. I tried it now on a total of 3 different routers. Same results. Hangs 9 out of 10 times i try.
Also, tried it in 2 different browsers, and 2 computers for what its worth.

Just got my Spark and I’m getting the same thing…flashing green to flashing magenta, it repeats that a couple of times until it gets a solid magenta. Then flashing green back to solid magenta until I restart.

I’m plugged into a 2A wall-wart and it’s not helping. I’m only about 10 feet from my router.

Can the instructions here: http://docs.spark.io/#/connect/troubleshooting-by-color-flashing-yellow be used to manually flash the newest firmware instead of trying to download it? It sounds like you guys have a lot of fixes for solid magenta issues, but only in the newest versions of the firmware that I can’t get!

Must have been a firmware update today? it was happy till it tried to do the update, and it’s been barfing the reboot/connect/firmware push cycle. it’s on a battery, so should be no power issues.

Ooo, maybe it’s not my fault then. Bad firmware for everyone?

That’s what it looks like from here. maybe it gets fixed soon.

@mightymik @cannikin Sorry, we pushed out a bad firmware update - the issue is fixed now, if you do a factory reset it should be fine.

Hmm, still not working for me. I get flashing magenta and then solid. Then it cycles between blinking green, then solid magenta a few times, then will do a blinking magenta for a while, then repeat. Eventually it lands on a rapid flashing cyan. :frowning: I restart and it starts the whole process again…

I could not get my last remaining working a few hours ago Core to sync with the cloud ide and reflash… so I did a factory-reset at 12:39am 1/24 CST and it’s doing this cycling through various colors thing for me as well.