search

oarcher / piotech

my custom components for esphome
18 stars 10 forks source link

tested PR branch results in error, #19

Closed tomatensaus closed 8 months ago

tomatensaus commented 8 months ago

I recompiled known working yaml file on the latest esphome (Version: 2023.12.9) I recently upgraded esphome (was on 10.6 before)

Same yaml code that used to work. I tried looking at the branch and do not see many changes.

Relevant log below:

19:34:22    [I] [ota_http:111]  Progress: 0.1%
19:34:22    [E] [ota_http:103]  Error code (131) writing binary data to flash at offset 4000 and size 1493968
19:34:22    [E] [ota_http:150]  Abort OTA backend
19:34:22    [E] [ota_http:153]  Abort http con
19:34:22    [E] [ota_http:155]  previous safe mode didn't succed. ota_http skipped

Perhaps changes in OTA backend that are not compatible with ota_http?

oarcher commented 8 months ago

error 131 is 083x and is OTA_RESPONSE_ERROR_WRITING_FLASH from ota_component.h

I can see that some change has been made to OTA 3 weeks ago at https://github.com/esphome/esphome/pull/6066

Can you try with:

ota:
  version: 1

I will try to have a closer look in the next days.

oarcher commented 8 months ago

I was not able to reproduce using esphome 'release' (2023-12-9) or 'dev' branch. I've tested http on esp32dev board, with arduino framework. Can you give me more details about your config ?

tomatensaus commented 8 months ago

Yes you tested with the correct framework, these are the relevant pieces of config (all urls are https)

esp32:
  board: esp32dev

external_components:
  - source: github://pr#5586
    components: [ ota_http ]

ota_http:
  safe_mode: fallback
tomatensaus commented 8 months ago

version is not valid for this version of esphome

Failed config

ota: [source <unicode string>:96]
  password: !secret ota_password

  [version] is an invalid option for [ota]. Please check the indentation.
  version: 1
tomatensaus commented 8 months ago 
09:08:40    [E] [ota_http:096]  Incorrect file size (-1) reported by http server (http status: -7). Aborting

Also getting this error. 

[09:52:16][I][ota_http:111]: Progress: 0.1%
[09:52:16][E][ota_http:103]: Error code (131) writing binary data to flash at offset 4000 and size 1493968
[09:52:16][E][ota_http:150]: Abort OTA backend
[09:52:16][E][ota_http:153]: Abort http con
[09:52:16][E][ota_http:155]: previous safe mode didn't succed. ota_http skipped

Tried all three the options below

safe_mode: false
safe_mode: true
safe_mode: fallback

Using ESP32 devkitC that I use for all my testing with 4MB flash

INFO ESPHome 2023.12.9
INFO Reading configuration bootstrap.yaml...
INFO Detected timezone 'Africa/Johannesburg'
INFO Generating C++ source...
INFO Compiling app...
Processing inverterdemo (board: esp32dev; framework: arduino; platform: platformio/espressif32@5.4.0)
------------------------------------------------------------------------------------------------------------------------------------------------------
HARDWARE: ESP32 240MHz, 320KB RAM, 4MB Flash
 - toolchain-xtensa-esp32 @ 8.4.0+2021r2-patch5
tomatensaus commented 8 months ago

I added more logging. I also cleaned all downloaded source to force a refresh from github. Cleaned all compiled binaries. changed url to example.com

11:34:09    [D] [button:010]    'Firmware Upgrade' Pressed.
11:34:09    [V] [ota_http:048]  restored pref ota_http_state: 40
11:34:09    [V] [ota_http:052]  setting mode to progress
11:34:09    [D] [esp32.preferences:114] Saving 1 preferences to flash...
11:34:09    [V] [esp32.preferences:126] sync: key: 99380598, len: 261
11:34:09    [V] [esp32.preferences:126] sync: key: 99380598, len: 261
11:34:09    [D] [esp32.preferences:143] Saving 1 preferences to flash: 0 cached, 1 written, 0 failed
11:34:09    [D] [ota_http:054]  Trying to connect to https://example.com/firmware.bin
11:34:09    [D] [ota_http:054]  Trying to connect to https://example.com/firmware.bin
11:34:09    [V] [ota_http:072]  http begin successfull.
11:34:09    [V] [ota_http:080]  http headers collected.
11:34:11    [V] [ota_http:086]  http GET finished.
11:34:11    [D] [ota_http:101]  firmware is 1493968 bytes length.
11:34:11    [V] [ota_http:079]  md5sum from received data initialized.
11:34:11    [V] [ota_http:087]  OTA backend begin
11:34:11    [I] [ota_http:111]  Progress: 0.1%
11:34:11    [E] [ota_http:103]  Error code (131) writing binary data to flash at offset 4000 and size 1493968
11:34:12    [E] [ota_http:150]  Abort OTA backend
11:34:12    [E] [ota_http:153]  Abort http con
11:34:16    [E] [ota_http:155]  previous safe mode didn't succed. ota_http skipped
oarcher commented 8 months ago

firmware size looks good, but can you double check that it's correctly handled by the http server ?

If you fetch it by wget and make an hex dump (for ex with ghex), does it start with byte E9 ?

tomatensaus commented 8 months ago

Just downgraded to 10.6 and it failed with the same error.

If you fetch it by wget and make an hex dump (for ex with ghex), does it start with byte E9 ?

lol.. I don't know if I should laugh or cry... I copied from a shell script and was using firmware-factory.bin which is obviously the wrong file. Should have known when the previous version also gave the same error.

Is there any way you could give an error message that gives you a hint that the file flashed does not have the correct header.

Thank you so much for the help. I am sorry I wasted your time

oarcher commented 8 months ago

ah!

I was checking for magic byte 0xE9 in the past, but I've stopped checking it, because the firmware.bin can ben zipped (so it doesn't starts with 0xE9).

I think that in the future, we will also be able to flash the firmware-factory.bin file.

Giving a more understandable message would be better, but I think it should be handled from ota backend, not ota_http.