BK7231N constant crash after "Error:MAIN:Low heap warning!"

[OoSTARTERoO]

Describe the bug
A clear and concise description of what the bug is.

Firmware:

• Version 1.17.692
• Device LSPA9
• Chip/model: BK7231N
• Device config
  startup command backlog startDriver BL0937; PowerSave 1; loglevel 6; is set

Logs

Info:MAIN:Time 13, idle 73866/s, free 52048, MQTT 0(0), bWifi 1, secondsWithNoPing -1, socks 4/38 POWERSAVE
Info:MQTT:mqtt_host empty, not starting mqtt
Info:MAIN:Time 14, idle 75616/s, free 40352, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 5/38 POWERSAVE
Error:MAIN:Low heap warning!
Info:MAIN:Time 15, idle 67394/s, free 16960, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 7/38 POWERSAVE
Debug:HTTP:TCP Client 192.168.0.13:43305 thread creation failed! fd: 8
Error:MAIN:Low heap warning!
Info:MAIN:Time 16, idle 71537/s, free 5264, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
Debug:HTTP:TCP Client 192.168.0.13:34089 thread creation failed! fd: 8
Debug:HTTP:TCP Client 192.168.0.13:34601 thread creation failed! fd: 8
Error:MAIN:Low heap warning!
Info:MAIN:Time 17, idle 68771/s, free 5264, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
Debug:HTTP:TCP Client (PC's ip):35113 thread creation failed! fd: 8
Debug:HTTP:TCP Client (PC's ip):35625 thread creation failed! fd: 8
Debug:HTTP:TCP Client (PC's ip):49449 thread creation failed! fd: 8
Debug:HTTP:TCP Client (PC's ip):49705 thread creation failed! fd: 8
Debug:HTTP:TCP Client (PC's ip):49961 thread creation failed! fd: 8
Debug:HTTP:TCP Client (PC's ip):36137 thread creation failed! fd: 8
Error:MAIN:Low heap warning!
Info:MAIN:Time 18, idle 49883/s, free 5264, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
Error:MAIN:Low heap warning!
Info:MAIN:Time 19, idle 59020/s, free 5264, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
bk_reboot
wdt reboot

Error:HTTP:TCP Client is disconnected, fd: 2
Info:MAIN:Time 13, idle 67354/s, free 63712, MQTT 0(0), bWifi 1, secondsWithNoPing -1, socks 3/38 POWERSAVE
Info:MQTT:mqtt_host empty, not starting mqtt
Info:MAIN:Time 14, idle 68115/s, free 63728, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 3/38 POWERSAVE
Info:MAIN:Time 15, idle 73756/s, free 28624, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 6/38 POWERSAVE
Error:MAIN:Low heap warning!
Info:MAIN:Time 16, idle 75687/s, free 16928, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 7/38 POWERSAVE
Error:HTTP:TCP Client is disconnected, fd: 7
Error:MAIN:Low heap warning!
Info:MAIN:Time 17, idle 71973/s, free 5232, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
Debug:HTTP:TCP Client 192.168.0.13:28985 thread creation failed! fd: 8
Error:MAIN:Low heap warning!
Info:MAIN:Time 18, idle 63971/s, free 5232, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
Error:MAIN:Low heap warning!
Info:MAIN:Time 19, idle 66038/s, free 5232, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
Error:MAIN:Low heap warning!
Info:MAIN:Time 20, idle 69648/s, free 5024, MQTT 0(1), bWifi 1, secondsWithNoPing -1, socks 8/38 POWERSAVE
bk_reboot
wdt reboot

opening the web interface seems to make the problem even worse.
It happens when the web interface is open during startup.

[openshwprojects]

Does it happen without loglevel setting?

[OoSTARTERoO]

Yes, without any startup command

[MaxineMuster]

Are there any other open browser windows open?
Or is something else trying to open a network connection to this device?

Something is eating your memory very fast (see free memory decreaseing from second to second):
52k
40k
16k
5k

63k
28k
17k
5k

Loosing over 30k of free RAM in a second is much...

[openshwprojects]

http server again, per socket malloc....

[OoSTARTERoO]

Oh... Is there any way that I can help you to solve this issue?
I can dig up deeper If you can't replicate it.

Also, Is there any versions without this issue?

[openshwprojects]

Please make a 2MB flash backup (strongly adviced) and try doing "restore RF partition".

This issue seems to happen very rarely, I've seen it like 3 times in like, I don't know, it's 2 years of OBK dev? It seems like something in RF (which is not overwritten by OBK) causes TCP sockets to don't finish so each HTTP call keeps the buffer allocated without freeing it