NWA90AX High CPU Usage
Hello,
There is a high CPU usage bug on the wifi access point NWA90AX. The wap is using the last stable firmware.
The CPU usage is stalled at about 30% without any activity.
I did not manage to found which process on the wap use CPU (with cli).
It is the second time i see this issue, the only way to return to a normal CPU usage is to reboot the device.
I tried to disable/enable SSID networks, but no way, the CPU usage persist.
You could found some informations in attachments.
Thanks.
Accepted Solution
-
Hi @ErgoSan
The reboot issue with the NWA90AX PRO was addressed in firmware version V6.55P1, released in April 2023.
Upon reviewing the Firmware V6.65 Patch 1 - the latest firmware's release notes for the NWA90AX PRO, there are no references to reboot problems, either for bug fixes or known issues. Additionally, there have been no reports from users of V6.65 Patch 1 regarding reboot issues.
Since the NWA90AX PRO is no longer in your possession, it is challenging for us to determine the cause of the reboots. If we exclude the AP itself as the cause, the issue might be related to the power system, such as an insufficient overall PoE budget. Note that the NWA90AX PRO requires PoE (802.3at) with a power draw of 20.5 W.
It is also possible that the PoE-supported device's reboot caused the NWA90AX PRO to reboot as well.
0
All Replies
-
Hi @cyayon ,
It's important to note that a CPU usage of around 30% is generally considered normal. This is because the CPU usage encompasses several factors, not just the wireless traffic.
I did not manage to found which process on the wap use CPU (with cli).
Execute the debug system show cpu status command to know that.
From the image you've shared, it appears that the CPU usage increases during business hours. This could be due to more clients connecting to the Access Point (AP) during these hours.
If you're not experiencing any issues with your wireless connection, there's no need to worry about the CPU level.
0 -
Hi,
I opened a support ticket (#378183). it seems that there is currently an firmware 11ax issue. There is currently a hotfix in preparation about this.
Support ask me to disable 11ax in 2GHz and 5GHz, but yesterday, i reboot the wap and CPU usage returned to normal.
Before rebooting, i launched a full diag, here is the result of show cpu status :
CPU utilization: 54 %
CPU utilization for 1 min: 68 %
CPU utilization for 5 min: 40 %I have to wait the issue back again and try to disable 11ax to see if CPU usage return to normal.
Disabling 11ax is a real problem because loosing WIFI6 is not a good deal… :-(
thanks.
0 -
@cyayon Disabling 11ax temporarily could help identify the cause of the high CPU usage, I understand losing WiFi6 Capabilities is not ideal.
0 -
Hi @cyayon ,
After investigating the issue with the hotfix mentioned in ticket #378183, I can confirm that it was implemented to address a connection issue, not CPU issue, with the NWA90AX. As of now, we haven't received any reports regarding high CPU usage with the NWA90AX. By the way, note that disabling 11ax likely won't have an effect on your CPU status.
Based on the information you've provided and from our previous communications, it appears that your CPU is operating normally. If you're interested in investigating further and want to see which processes are utilizing the CPU on your AP, you can use the following command: debug system show cpu status.
I hope this clarifies your concerns.
0 -
hi,
The CPU usage is definitely not normal.There is a loop process which consume 25% of cpu without doing anything.
In the diag file and debug command we can see the wan_module process which use 25%.
A reboot solve the issue. But issue come back again after a few days.Please check the ticket 378183, you can check the diag file attached.
thanks--------------------------------------------------------------------------------
[H[JMem: 145020K used, 100876K free, 4924K shrd, 9160K buff, 46468K cached
CPU: 7.6% usr 21.1% sys 0.0% nic 71.1% idle 0.0% io 0.0% irq 0.0% sirq
Load average: 3.46 2.69 2.43 2/178 18852
[7m PID PPID USER STAT VSZ %VSZ CPU %CPU COMMAND[0m
4763 4750 root R 2376 0.9 0 24.9 {wlan_module} wlan module: b8:ec:
18852 18851 root R 1388 0.5 1 3.8 /usr/bin/top -n1
26617 26616 admin S 13328 5.4 2 0.0 -zysh
18851 31468 root S 13328 5.4 0 0.0 zysh -p 100 -e debug system top i
2075 2071 root S 13040 5.2 3 0.0 /bin/zyshd
2175 1 root S 12740 5.1 3 0.0 /usr/local/bin/redis-server /var/
6679 1 root S 10144 4.1 1 0.0 /usr/sbin/radiusd -t -d /var/zyxe
26616 26328 admin S 9128 3.7 2 0.0 sshd: admin@pts/0
26328 13753 root S 9128 3.7 3 0.0 sshd: admin [priv]
13137 1 root S 6524 2.6 3 0.0 /usr/local/lighttpd/sbin/lighttpd
1814 1 root S N 5692 2.3 3 0.0 /usr/sbin/uamd
4424 1 root S 5192 2.1 2 0.0 hostapd -B -P /var/run/wlan-1-3.p
13753 1 root S 5188 2.1 0 0.0 sshd: /usr/sbin/sshd -f /etc/ssh/
4275 1 root S 5172 2.1 3 0.0 hostapd -B -P /var/run/wlan-1-1.p
7462 1 root S 5168 2.1 1 0.0 hostapd -B -P /var/run/wlan-2-5.p
4340 1 root S 5076 2.0 3 0.0 hostapd -B -P /var/run/wlan-1-2.p
7398 1 root S 5072 2.0 1 0.0 hostapd -B -P /var/run/wlan-2-2.p
3850 1 root S 4988 2.0 0 0.0 /usr/sbin/fauthd
2322 1 root S 3916 1.5 0 0.0 /usr/local/bin/wireless_hal
1739 1 root S 3876 1.5 3 0.0 /usr/sbin/generic_timer -F0 -
Hi @cyayon ,
After analyzing the diagnostic file, here are the detailed findings:
There were clients connected to both the 2.4GHz and 5GHz bands, the wlan_module is shown to be approximately 25%. Note that when there are clients connect / disconnect the network or there is communication between clients and AP, the wlan_module will be counted.
Furthermore, the idle index indicates that 71.1% of the CPU is still idle, meaning that only 28.9% of the CPU's capacity is utilized for processing other applications.
So, we have confirmed that the CPU status is normal and there are no issues with your APs. You need not be concerned about the CPU index as it is sufficient for your service requirements.
0 -
hi,
Thanks for your answer.
But I do not understand why there is CPU usage when all clients are disconnected and there is no activity on wifi.
I made debug files just before disconnect all clients, this is why you see clients connected .
But this abnormal cpu usage continue even after disconnect everything.
I also sniffed traffic network on port switch and i confirm that there was no network activity.
Moreover, just after a simple reboot all connected clients reconnect as expected and the CPU usage stay at 1-4% as expected.
I have to wait some days/weeks to encounter again abnormal cpu usage stalled at 28%. Then reboot again, etc…
I am sure there is something wrong with wlan_module process. Is there something like strace or ptrace on wap debug shell to see what is hanging/locking on this process ?
thanks
0 -
Hi @cyayon ,
The CPU does more than just handle wireless connections and traffic. It also processes other tasks and broadcast and multicast traffic in the network as well. These factors can influence CPU usage. If there's no activity on the WiFi and you wish to understand what the CPU is processing, you can execute the above command to find out. You also can refer this CLI:
debug system top iterations 1
Please note that, a CPU usage of 28% is considered normal and it is sufficient for your service. Therefore, there's no need to reboot the Access Point (AP) for it to function effectively.
0 -
thanks but I already know which process (wlan_module) is concerned.
Note that the cpu is stalled at 29% indefinitely !
Until I reboot. I really doubt that it process multicast and broadcast or background tasks for hours / days / weeks … until I reboot !
I am looking to debug inside the process itself with a strace like command (just look at the internet to know what is strace).
I agree that 28% is not very high but I disagree to have this usage with no reason. Moreover the cpu usage do not vary between 1% - 29% it is exactly 29% indefinitely. If one of my Linux systems have this behavior I should be really concerned about this and do my best to find the issue (mutex, lock, loop, …)
I do not understand why you do not acknowledge that there is possibly an issue and refuse to find it…
I have another same AP (same firmware), in another room which don’t have the issue.the cpu usage is always between 1% to 10%. Never stalled at 29%.
Perhaps should I switch to a different vendor…0 -
Hi @cyayon ,
Thank for sharing with us your experience.
From the diagnostic file that you provided, we just can share with you that the CPU shows the process with wireless connection.
We're always ready to support you check the CPU in case there is no wireless connection with your APs. So please help to download the diagnostic file of 2 APs (APs with CPU stalls at 29% and CPU always between 1% to 10%) when there is no wireless connection. You can do it by connect your laptop to the Switch that APs connect to in the same subnet.
We’ll check and share with you which processes are different from them or any finds. You can send the diagnostic via the private message.0
Categories
- All Categories
- 415 Beta Program
- 2.4K Nebula
- 151 Nebula Ideas
- 98 Nebula Status and Incidents
- 5.7K Security
- 277 USG FLEX H Series
- 277 Security Ideas
- 1.4K Switch
- 74 Switch Ideas
- 1.1K Wireless
- 42 Wireless Ideas
- 6.4K Consumer Product
- 250 Service & License
- 395 News and Release
- 85 Security Advisories
- 29 Education Center
- 10 [Campaign] Zyxel Network Detective
- 3.6K FAQ
- 34 Documents
- 34 Nebula Monthly Express
- 85 About Community
- 75 Security Highlight