XT8 Packet Loss / Ping spikes

[030basti]

Hey all,

I am using the Asus XT8's in a Smart Connect Mesh on Firmware 42095 for quite some time.
So far I have been very happy.

Recently however I am having a couple of issues with Packet Loss/Ping spikes.
I noticed it while playing Rocket League. Once in a while I get a ping spike which is obviously an issue in online gaming.

From a setup perspective I have my ISP Modem connected via Ethernet Cable to my Asus XT8 set up as Router (running double NAT).
Then I use the dedicated wireless backhaul to connect to my second node which stands pretty close by and is simply separated by a wall.

Based on Pingplotter it looks like my Ping spikes/Packet losses happen due to may backhaul connection however I can not figure out why.
Is there any way I can test my backhaul even further and are there any specific suggested settings for it?

Thanks in advance

 

[bubbagump]

I am having the same issue, and have tried all available firmwares. When using only one XT8, no mesh, there are no issues. Once I create the mesh, I have hourly slowdowns and disconnects. Until recently, there was nothing in the logs. The latest beta firmware and newest 45879 firmware seem to have more info in the logs. I am running a syslog server to collect everything.

Now, exactly every hour, pingplotter shows latency spikes as high as 3500ms for 10-15 seconds. In the logs at the exact same time I find:

2021-09-30 09:33:36 User.Warning 192.xxxxxxxx Sep 30 09:33:37 ZenWiFi_XT8xxxxxxxxxx kernel: CSIMON: CSIMON[1.1.0] Initialization
2021-09-30 09:33:36 User.Warning 192.xxxxxxxx Sep 30 09:33:37 ZenWiFi_XT8xxxxxxxxxx kernel: CSIMON: M2M usr already registered ...

Does anyone know what this means? It clearly has something to do with AIMesh. This is the 2nd batch of XT8's I have used that exhibit the same problem.
Currently running 3 XT8's with wireless backhaul to the 2 remote nodes.

 

[pantsontoast]

I have been experiencing similar problems, packet loss (20-70%) when simply pinging the main node or mesh node directly from a client (not going to the internet at all). I have ASUS XT8 with one mesh node and wireless backhaul. I opened up a case with ASUS and provided them detailed logs and configuration file, but they have not been helpful so far.

Here is what I have learned by testing various configurations, packet loss occurs...

• using firmware 45898 and 43181 but not 42095
• on two mac book pros but not on an android phone and several ubuntu machines
• when connected to either the mesh node or main router

For now I'm sticking with firmware 42095, I have had no problems with it and very consistent and high speeds.

 

[tgl]

Happened across this thread just now ... I see some of these "CSIMON: M2M usr already registered" messages in my syslog, but not for the last couple of weeks. It looks like they stopped when I began to use a different SSID name for the hidden 5GHZ-2 backhaul channel than what I use for the non-hidden channels. If you have them set the same, might be worth changing.

 

[keith_ca]

My remote Xt8 has started to exhibit this behaviour in the past week, on f/w 46061. (I have two XT8, connected via MoCA, and with an AX92U also in AI Mesh2.0). Up until this point, it was quite stable, but now I am experiencing packet loss throughout the work day, including full outages of up to 10s. Nothing has changed in my network. All devices connected to the remote/satellite XT8 go offline, whether wired or WiFi.

I tried changing ethernet cables for my wired connections, with no luck. Since all the clients fall off, it's not wiring.

I see the same CSIMON errors during the outage:


Apr 6 08:18:29 kernel: CSIMON: CSIMON[1.1.0] Initialization
Apr 6 08:18:29 kernel: CSIMON: M2M usr already registered ...
Apr 6 08:18:30 kernel: CSIMON: CSIMON[1.1.0] Initialization
Apr 6 08:18:30 kernel: CSIMON: M2M usr already registered ...

I was seeing link up/down messages before on the wired interface, but they have gone away.


Apr 5 08:40:38 kernel: eth3 (Ext switch port: 2) (Logical Port: 10) (phyId: 2) Link DOWN.
Apr 5 08:40:42 kernel: eth3 (Ext switch port: 2) (Logical Port: 10) (phyId: 2) Link Up at 1000 mbps full duplex
Apr 5 08:40:49 kernel: eth3 (Ext switch port: 2) (Logical Port: 10) (phyId: 2) Link DOWN.
Apr 5 08:40:51 kernel: eth3 (Ext switch port: 2) (Logical Port: 10) (phyId: 2) Link Up at 1000 mbps full duplex

I tried reflashing the firmware on my XT8 -- no change.

Next step will be a ticket or RMA with Asus I guess...

 

[Mr. J]

I have started to experience the same thing. No changes to the router, settings or firmware for quite a while. However i recently started to get packet loss occasionally while gaming. Initially up to 50% on CSGO, which goes to zero after 20-30 seconds. The packet loss occurs again 1-2 times an hour. Firmware 46061.

 

[JasonMAP]

My 3 XT8 in mesh (HW ver 1) just started exhibiting this behavior the last week despite no changes to configs or firmware for at least a year, and thus I was drawn to this forum.

tl;dr: HW ver 1 has a new FW v3.0.0.4.388.21617 that might fix sporadic latency issue.

Pings to the gateway would go through the roof for wifi devices and they would lag out

I tried switching from wireless backhaul to wired and it didn't make a difference.
I upgraded the firmware to the latest 3.0.0.4.388.21099 and beta 9.0.0.4.388.21101 and it still was happening.

I tried changing a variety of settings on the wireless professional tab to see if there was a settings clash but to no avail. This included disabling Wifi 6 mode, which had always worked just fine previously.

ICMP response times over the last month and week:

Finally, yesterday 3.0.0.4.388.21617 was dropped (post: https://www.snbforums.com/threads/asus-zenwifi-ax-xt8-firmware-version-3-0-0-4-388-21617-2022-11-21.81888/)

The only things I can guess might be relevant for me are the 2.5G port usage and the starlink router (but that's behind a firewall and has been installed for several months now, so I just don't know). The changelog explanations are a little slim for zeroing in on the exact problem.

Nevertheless, this new firmware seems to have resolved the problem. Over the last 12 hours there was only a very short latency spike to around 13ms and the rest has been 2ms or less and smooth sailing.

While trying to troubleshoot all this, I did notice that when I moved my Android phone (Pixel 6 Pro) to different locations with me - notably, when placing it on the wireless charger on my desk in the kitchen (which is approximely equal distance between two of the nodes), the issue seemed to happen almost immediately after the screen shut off. Once I noticed that, I turned off the phone's wifi and the spike would drop away almost every time. I thought perhaps this was due to EITHER the WMM APSD feature having problems OR the RSSI (<-65dBm) I have set but changing these didn't seem to make a difference and I see no note of that in the changelog.

Ah well. At least it works again.

-Jason

 

[030basti]

Thanks Jason. Somehow I had fixed it previously via an older firmware. All of a sudden however the issue came back with Pingplotter giving me pretty much excatly 50% packet loss between node and router. I have then updated to the newest firmware like you suggested but that didnt help. Any other smart idea of what could have fixed it for you?

 

[JasonMAP]

Nah, since posting that on the latest firmware I've definitely had stability problems. The main mesh node has just...died a couple of times for seemingly no reason at all, whereas on older firmware it had been rock solid.

Using all ethernet backhaul did improve performance during normal functionality but I can't say that my problems are fully resolved either.

I hope asus gets their game together.