Hey @David-TP,

 

Can you help check with the engineers on this matter I raised? I feel they have gone astray with the priority of the issue.

 

My issue where temperature is very high and deco kept going down have been plauging me since the day I bought the X50-5G a few months ago.

The Tapo issues are not the ones that are causing my CPU usage and memory usage to go max.

 

 

gnz wrote

  @David-TP Hey I actually requested clarification from the engineer about my issues that I raised. Somehow the two issues I raised got mixed up now.

 

There are two distinct issues now I am facing:

 

1. Deco X50-5G repeatedly goes down throughout the day (with error messages in log saying temperature hitting high levels) which I raised some time back. Now the version 1.20 update brought the CPU and memory usage percentage, which now confirmed my suspicious that CPU usage kept going to 100% most of the time and memory usage over 80% most of the time too.

 

2. Second issue is about Tapo matter smart devices (T110M and T304M) which cannot stay connected if they are added to Apple Home via mattter.

 

These are two separate issues but please refer to the mail reply here which the engineer has mixed up two different issues into one. The first issue happened months before I bought the Tapo smart devices, so they are unrelated. And I don't seem to be the only one experiencing very high CPU and memory usage. There are other people here who are seeing it too.

 

Please help pass on the message.

 

 

  @ShadowLabs 

 

 yes, that did the trick. 24 hours without internet is a struggle. Should have checked the community first... :D

 

ShadowLabs wrote

  @David-TP This firmware update just prompted me to install it via the Deco app, and after the update completely crippled my network. The CPU was stuck at 100% and reboots did not fix the issue. I found that the QoS setting had become enabled, though no devices or scenarios were selected. Before the update, QoS was not active on this network. As soon as I disabled the QoS feature the network became usable again and my devices were able to connect to the internet. It seems the update changed this setting. Not a good experience.

 

  @Grimschold glad my posting did the trick for you!

 

I gotta say this update completely ruined the stability of my X50-5G. It seems like every 24 hours the DHCP server dies and all my devices stop getting IP addresses until I reboot the damn thing. I really hope TP link comes out with a hotfix quickly because this is damn unusable.

  @David-TP Thanks for this update, it's been long awaited and I installed it straight when it was available. I'm going to go technical here, since I'm a software engineer myself with over 20 years of experience in the field in various positions.

 

First, I want to start with the good news, for me the update has been more stable than the 1.1.7 version that I was on before, especially when it comes to the 5G-NSA network stability. I don't use QoS and I don't use mesh networking, but I do need VPN to the network, so the introduction of both WireGuard and other VPN support is more than welcome.

 

However, this release has some very obvious bugs that I don't understand how they weren't caught in the QA process. The system boots with the build date of the firmware, which is 11 October 2024. There's a race condition which causes sometimes the system to record its time to get online before the first NTP requests are finished, meaning it thinks it got online on 11 October 2024 after the boot. This doesn't always happen and sometimes it records the time correctly, but honestly it doesn't take many boots for this problem to come up, and it seems very consistent in my case. Should have been caught by the QA before release.

 

Related to the above, the whole boot process of the device seems extremely chaotic, with cross-dependencies between the different services clearly not correctly handled and everything starting at the same time. There's lots of error messages showing up during boot that seem to be related to services being started in the wrong order, here's a couple most obvious ones:
cat: can't open '/tmp/is_binded': No such file or directory
/etc/rc.common: line 1: can't create /proc/sys/net/ipv6/conf//accept_ra: nonexistent directory
/etc/rc.common: line 1: can't create /proc/sys/net/ipv6/conf//autoconf: nonexistent directory
/etc/rc.common: line 1: can't create /proc/sys/net/ipv6/conf//sendrs: nonexistent directory
/etc/rc.common: line 1: can't create /proc/sys/net/ipv6/conf//autoconf: nonexistent directory
/etc/rc.common: line 1: can't create /proc/sys/net/ipv6/conf//sendrs: nonexistent directory
/usr/bin/apsd: apsd_check_eth_has_neigh:1553: Error: can't find ifname[eth2.1] in eth_ifname
ntp check error!!!; errno(2): No such file or directory

There's also plenty of messages that seem to be clearly related to programming errors in your scripts, I'm sure there's more than these but just showing some examples:

Fri Oct 11 12:02:21 2024 user.emerg : logger: unrecognized option: >
Fri Oct 11 12:02:21 2024 user.emerg : BusyBox v1.22.1 (2024-10-11 08:36:16 UTC) multi-call binary.
Fri Oct 11 12:02:21 2024 user.emerg :
Fri Oct 11 12:02:21 2024 user.emerg : Usage: logger [OPTIONS] [MESSAGE]
Fri Oct 11 12:02:21 2024 user.emerg :
Fri Oct 11 12:02:21 2024 user.emerg : Write MESSAGE (or stdin) to syslog
Fri Oct 11 12:02:21 2024 user.emerg :
Fri Oct 11 12:02:21 2024 user.emerg :     -s    Log to stderr as well as the system log
Fri Oct 11 12:02:21 2024 user.emerg :     -t TAG    Log using the specified tag (defaults to user name)
Fri Oct 11 12:02:21 2024 user.emerg :     -p PRIO    Priority (numeric or facility.level pair)
Fri Oct 11 12:02:21 2024 user.emerg :

Fri Oct 11 12:02:38 2024 user.emerg : /etc/rc.common: /etc/rc.d/S48luarsa_keys_gen: line 10: syntax error: unexpected "}"

Fri Oct 11 12:02:39 2024 user.emerg : lua: (command line):1: bad argument #1 to 'execute' (string expected, got boolean)
Fri Oct 11 12:02:39 2024 user.emerg : stack traceback:
Fri Oct 11 12:02:39 2024 user.emerg :     [C]: in function 'execute'
Fri Oct 11 12:02:39 2024 user.emerg :     (command line):1: in main chunk
Fri Oct 11 12:02:39 2024 user.emerg :     [C]: ?

These things combined make me concerned about the overall attitude to quality at TP-Link, since you're clearly deploying syntax errors in the code and easy to find race conditions get through QA, both of which would be easy to catch with some basic quality processes. Yes, your app looks pretty and the Deco device looks pretty as well, but you could at least give us SSH access to the device to allow us to fix some of the most obvious bugs in your firmware.

 

Last, but not least, could we PLEASE fix the logging features of this device. When 5G is on, it's filling the whole log with heaps of GetEventType and PostEvent messages for unhandled events and temperature data (which is actually somewhat useful, but not with the frequency it's logging it), that fills the whole backlog and means the actually interesting and meaningful messages disappear in the matter of minutes. It might help if you didn't log everything with the warn level on a production system, this looks quite unprofessional. The logs this device gives are practically completely useless.

I thank you for this update, and I wish you keep the updates coming and fix some of the issues people are having with their QoS and mesh networks. But if you really consider yourself to be the "the world's #1 provider of consumer WiFi networking devices" as you advertise, I strongly suggest you invest just a little bit more into your firmware development division. Hope you pass this message on to your peers, thanks!

  @David-TP I'm now having a new nightmarish issue. So my reserved IP addresses for my smart home devices have been taken over by my iPhone randomly.

 

So now my iPhone is being shown as my reserved devices in the device list.

 

I don't know what to say now. It's so bad.

I tried the DOT and DOH features as well, but they don't seem to work; nothing happens, and my ISP still provides DNS. There may be bugs that need to be fixed. I attempt VPN functionality as well. In addition to being connected, my device also received a DNS from my ISP when I installed Warp using the Wireguard protocol. Other problem after update.. SSID is not being broadcast by the IoT network. Fixed after hard reset.

  @David-TP TP-Link support is non-existent here in my country.

Email them for exchange almost two weeks ago. No reply. Called up many times, no one picks up. Live chat? No agents available.