starred/WiFiManager

Fork 0

mirror of https://github.com/tzapu/WiFiManager.git synced 2026-04-27 00:55:52 +03:00

[GH-ISSUE #1240] OTA fails with 1MB ESP8266 #1062

New issue

Open

opened 2026-02-28 01:28:21 +03:00 by kerem · 5 comments

kerem commented

2026-02-28 01:28:21 +03:00

Owner

Originally created by @BernhardSchlegel on GitHub (Apr 12, 2021).
Original GitHub issue: https://github.com/tzapu/WiFiManager/issues/1240

Basic Infos

OTA update fails (out of memory?) with WRITE ERROR 4.

Hardware

WiFimanager Branch/Release: Master (pulled as of today)

Esp8266:

Hardware: Sonoff TH 16 (which is why I cant go for more Flash)

Core Version: 2.4.0, staging

esptool.py --port /dev/ttyUSB0 flash_id yields (probably good practice for the issue template?)

esptool.py v3.0
Serial port /dev/ttyUSB0
Connecting....
Detecting chip type... ESP8266
Chip is ESP8266EX
Features: WiFi
Crystal is 26MHz
MAC: 60:01:94:07:01:2e
Uploading stub...
Running stub...
Stub running...
Manufacturer: e0
Device: 4014
Detected flash size: 1MB

Description

OTA update fails

sleep disable
...........................................................................................................................................
*wm:[0] [ERROR] OTA Update WRITE ERROR 4
.*wm:[0] [ERROR] OTA Update WRITE ERROR 4

Settings in IDE

platformio.ini

[env:sonoff_th]
platform = espressif8266
board = sonoff_th
framework = arduino
lib_deps = 
	paulstoffregen/OneWire@^2.3.5
	bblanchon/ArduinoJson@^6.17.2
	WiFiManager=https://github.com/tzapu/WiFiManager.git
monitor_speed = 115200

[common]
upload_port = /dev/ttyUSB0

Sketch

To big to print, but it breaks down to:

Program size output

   text    data     bss     dec     hex filename
 468804    5572   26256  500632   7a398 .pio/build/sonoff_th/firmware.elf

Build ouput

RAM:   [====      ]  38.8% (used 31824 bytes from 81920 bytes)
Flash: [======    ]  62.3% (used 474372 bytes from 761840 bytes)

Project inspect

RAM 39% (31.3kB)
Flash 48% (494.8kB)

Top 5 files

139.7 KB	unknown
30.5 KB	…WiFiManager/WiFiManager.cpp
8.0 KB	…ssl/bearssl/src/ec/ec_p256_m15.c
7.1 KB	…Json/JsonDeserializer.hpp
7.0 KB	…src/WiFiClientSecureBearSSL.cpp

Problem

I don't know why all programm size numbers are different (at least I dont get them aligned). What I do know: The 139.7kB seem to be inevitable, as well as the 139.7kB from the WiFiManager - thus reducing bin size is not really an option. Is there a way I can tweak the partition scheme using board_build.partitions? It would be absolutely awesome to get OTA to work!

Originally created by @BernhardSchlegel on GitHub (Apr 12, 2021). Original GitHub issue: https://github.com/tzapu/WiFiManager/issues/1240 ### Basic Infos OTA update fails (out of memory?) with `WRITE ERROR 4`. #### Hardware WiFimanager Branch/Release: Master (pulled as of today) Esp8266: Hardware: Sonoff TH 16 (which is why I cant go for more Flash) Core Version: 2.4.0, staging `esptool.py --port /dev/ttyUSB0 flash_id` yields (probably good practice for the issue template?) ``` esptool.py v3.0 Serial port /dev/ttyUSB0 Connecting.... Detecting chip type... ESP8266 Chip is ESP8266EX Features: WiFi Crystal is 26MHz MAC: 60:01:94:07:01:2e Uploading stub... Running stub... Stub running... Manufacturer: e0 Device: 4014 Detected flash size: 1MB ``` ### Description OTA update fails ``` sleep disable ........................................................................................................................................... *wm:[0] [ERROR] OTA Update WRITE ERROR 4 .*wm:[0] [ERROR] OTA Update WRITE ERROR 4 ``` ### Settings in IDE platformio.ini ``` [env:sonoff_th] platform = espressif8266 board = sonoff_th framework = arduino lib_deps = paulstoffregen/OneWire@^2.3.5 bblanchon/ArduinoJson@^6.17.2 WiFiManager=https://github.com/tzapu/WiFiManager.git monitor_speed = 115200 [common] upload_port = /dev/ttyUSB0 ``` ### Sketch To big to print, but it breaks down to: Program size output ``` text data bss dec hex filename 468804 5572 26256 500632 7a398 .pio/build/sonoff_th/firmware.elf ``` Build ouput ``` RAM: [==== ] 38.8% (used 31824 bytes from 81920 bytes) Flash: [====== ] 62.3% (used 474372 bytes from 761840 bytes) ``` Project inspect ``` RAM 39% (31.3kB) Flash 48% (494.8kB) ``` Top 5 files 139.7 KB | unknown -- | -- 30.5 KB | …WiFiManager/WiFiManager.cpp 8.0 KB | …ssl/bearssl/src/ec/ec_p256_m15.c 7.1 KB | …Json/JsonDeserializer.hpp 7.0 KB | …src/WiFiClientSecureBearSSL.cpp ### Problem I don't know why all programm size numbers are different (at least I dont get them aligned). What I do know: The 139.7kB seem to be inevitable, as well as the 139.7kB from the WiFiManager - thus reducing bin size is not really an option. Is there a way I can tweak the partition scheme using `board_build.partitions`? It would be absolutely awesome to get OTA to work!

kerem added the

enhancement

Documentation

labels

2026-02-28 01:28:21 +03:00

kerem commented

2026-02-28 01:28:21 +03:00

Author

Owner

@tablatronix commented on GitHub (Apr 12, 2021):

I have been trying to reduce the size of wm back down , its just a little too big for ota on 1M ( you need double the size for ota )

62.3% is 12% too much
I will work on it some more, but you can remove the help strings and turn debug off that will help.

@tablatronix commented on GitHub (Apr 12, 2021): I have been trying to reduce the size of wm back down , its just a little too big for ota on 1M ( you need double the size for ota ) 62.3% is 12% too much I will work on it some more, but you can remove the help strings and turn debug off that will help.

kerem commented

2026-02-28 01:28:21 +03:00

Author

Owner

@BernhardSchlegel commented on GitHub (Apr 13, 2021):

Hi tablatronix, thanks so much for the reply ! Your efforts are really appreciated (I couldn't find a buy me a coffee link...). I was aware of the 50% rule, that's why I tried to generate the binary and library size metrics. I was just hoping, that the latter (48% (494.8kB)) was the correct one..

Adding #define WM_NODEBUG on top of my main.cpp resulted into the same size (only the build output, since that seems to be the relevant)

RAM:   [====      ]  38.8% (used 31824 bytes from 81920 bytes)
Flash: [======    ]  62.3% (used 474372 bytes from 761840 bytes)

Questions:

Did I do that right?
I'm not sure what you mean by "removing help strings" - all I found in strings_en.h seemed necessary.

@BernhardSchlegel commented on GitHub (Apr 13, 2021): Hi tablatronix, thanks so much for the reply ! Your efforts are really appreciated (I couldn't find a buy me a coffee link...). I was aware of the 50% rule, that's why I tried to generate the binary and library size metrics. I was just hoping, that the latter (48% (494.8kB)) was the correct one.. Adding `#define WM_NODEBUG` on top of my main.cpp resulted into the same size (only the build output, since that seems to be the relevant) ``` RAM: [==== ] 38.8% (used 31824 bytes from 81920 bytes) Flash: [====== ] 62.3% (used 474372 bytes from 761840 bytes) ``` Questions: 1. Did I do that right? 2. I'm not sure what you mean by "removing help strings" - all I found in `strings_en.h` seemed necessary.

kerem commented

2026-02-28 01:28:21 +03:00

Author

Owner

@tablatronix commented on GitHub (Apr 13, 2021):

hmm there is another issue discussing the same issue #1154 ?

@tablatronix commented on GitHub (Apr 13, 2021): hmm there is another issue discussing the same issue #1154 ?

kerem commented

2026-02-28 01:28:21 +03:00

Author

Owner

@BernhardSchlegel commented on GitHub (Apr 14, 2021):

I came across this issue before I opened this. I'm maybe lacking oversight, but I was not able to find to find concrete instructions that will get me near to saving 12% flash anywhere.

@BernhardSchlegel commented on GitHub (Apr 14, 2021): I came across this issue before I opened this. I'm maybe lacking oversight, but I was not able to find to find concrete instructions that will get me near to saving 12% flash anywhere.

kerem commented

2026-02-28 01:28:21 +03:00

Author

Owner

@tablatronix commented on GitHub (Nov 3, 2021):

Yeah not sure, I hope to revisit this at some point, but just have not had time to see how to slim it down

@tablatronix commented on GitHub (Nov 3, 2021): Yeah not sure, I hope to revisit this at some point, but just have not had time to see how to slim it down