AC86U x2 and AC66UB1 AI MESH unstable - 3.0.0.4.384_20648

[jyth_82]

Have found the solution to my problem, amid , that it only occurs in the AC86U and not like the previous AC68U and AC66U_B1.

My initial setup is AC86U (as main node) and then to 2 switches which I have a NAS@Asustor running Automatic Load Balance which I've configured Asustor's Port 1 to 1st switch and Port 2 to the 2nd switch. ALB means that I have dual throughput load balance and failover capabilities without needing special switches. It configured with a single virtual IP but using 2 ports.

After a very long while pondering and replacing the AC86U to the AC66U as the main node, I re-setup the AC86U now as the main node and was thinking whether the ALB on Asutor is causing the random crash. I took out the Port 2 out from the 2nd switch and.... VIOLA!!! My AC86U is stable without any random crashes!

JUST FYI, I previously run the same ALB setup in AC68U and AC66U_B1 and NEVER had a crash! Only the AC86U has the random crash so often.

So far, I've added the AC66U_B1 as the mesh node and it is still running fine without any crash..... will conclude next with the ALB setup test again.

 

[arthurlien]

Have found the solution to my problem, amid , that it only occurs in the AC86U and not like the previous AC68U and AC66U_B1.

My initial setup is AC86U (as main node) and then to 2 switches which I have a NAS@Asustor running Automatic Load Balance which I've configured Asustor's Port 1 to 1st switch and Port 2 to the 2nd switch. ALB means that I have dual throughput load balance and failover capabilities without needing special switches. It configured with a single virtual IP but using 2 ports.

After a very long while pondering and replacing the AC86U to the AC66U as the main node, I re-setup the AC86U now as the main node and was thinking whether the ALB on Asutor is causing the random crash. I took out the Port 2 out from the 2nd switch and.... VIOLA!!! My AC86U is stable without any random crashes!

JUST FYI, I previously run the same ALB setup in AC68U and AC66U_B1 and NEVER had a crash! Only the AC86U has the random crash so often.

So far, I've added the AC66U_B1 as the mesh node and it is still running fine without any crash..... will conclude next with the ALB setup test again.

Could you duplicate the random crash issue and send feedback with System log, Wi-Fi log by Administration -> Feedback function?

 

[jyth_82]

Could you duplicate the random crash issue and send feedback with System log, Wi-Fi log by Administration -> Feedback function?

Sure.... send the feedback back to Asus it it?

BTW, I can kinda of confirm that the ALB is causing the issue. I've just re-setup using port 3 and port 4 of the Asustor NAS to configure as ALB. Within a 24 hour period, the AC86U crashed. Am capturing the full logs now.

 

[jyth_82]

So, I've tested many modes from the NAS and seems like the Automatic Load Balancing features are causing plenty of errors in the log file :
Sep 8 18:25:55 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when adding flow net_p=ffffffc010c9f1c8 dir=1 old_key=0x200010ba new_key=0x200010d5
Sep 8 18:25:55 kernel: ^[[0m
Sep 8 18:26:13 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc01234fc18
Sep 8 18:26:13 kernel: ^[[0m
Sep 8 18:26:40 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc0121c68a8
Sep 8 18:26:40 kernel: ^[[0m
Sep 8 18:26:40 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc0137216f0
Sep 8 18:26:40 kernel: ^[[0m
Sep 8 18:26:40 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc010ca7538
Sep 8 18:26:40 kernel: ^[[0m
Sep 8 18:26:49 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc010c6c380
Sep 8 18:26:49 kernel: ^[[0m

When I disconnect them, the router is super stable - last tested was 5 days running straight till today.

Question - Anyone encountering the same problem or similar and have managed to fix it?

 

[Mitch]

So, I've tested many modes from the NAS and seems like the Automatic Load Balancing features are causing plenty of errors in the log file :
Sep 8 18:25:55 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when adding flow net_p=ffffffc010c9f1c8 dir=1 old_key=0x200010ba new_key=0x200010d5
Sep 8 18:25:55 kernel: ^[[0m
Sep 8 18:26:13 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc01234fc18
Sep 8 18:26:13 kernel: ^[[0m
Sep 8 18:26:40 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc0121c68a8
Sep 8 18:26:40 kernel: ^[[0m
Sep 8 18:26:40 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc0137216f0
Sep 8 18:26:40 kernel: ^[[0m
Sep 8 18:26:40 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc010ca7538
Sep 8 18:26:40 kernel: ^[[0m
Sep 8 18:26:49 kernel: ^[[0;33;41mBLOG ERROR blog_request :blog_key corruption when deleting flowfor net_p=ffffffc010c6c380
Sep 8 18:26:49 kernel: ^[[0m

When I disconnect them, the router is super stable - last tested was 5 days running straight till today.

Question - Anyone encountering the same problem or similar and have managed to fix it?

I don't have an Asusror NAS, I've got a QNAP TS-251+ with dual Nics and an ALB load balancing setup similar to what you described through a virtual switch setting on the NAS. I just got 2x RT-AC86U's installed in Aimesh last week and have had zero issues or errors in the logs relating to load balancing. However, my NAS is not wired directly to the router or the node - its running off of a cheapo 8 port gigabit dumb switch thats connected to the main router, and this same switch also provides the backhaul wire to the node (if thats relevant). If you have a spare switch try wiring a switch between the router and NAS just to see if it fixes the issue?

 

[jyth_82]

I don't have an Asusror NAS, I've got a QNAP TS-251+ with dual Nics and an ALB load balancing setup similar to what you described through a virtual switch setting on the NAS. I just got 2x RT-AC86U's installed in Aimesh last week and have had zero issues or errors in the logs relating to load balancing. However, my NAS is not wired directly to the router or the node - its running off of a cheapo 8 port gigabit dumb switch thats connected to the main router, and this same switch also provides the backhaul wire to the node (if thats relevant). If you have a spare switch try wiring a switch between the router and NAS just to see if it fixes the issue?

I have done the same - AC86U router -> 2 dumb 8 port switches -> ASUSTOR NAS 2 port to individual switches ....

The moment I connect the port either to different or same switches OR even directly to the router, the error will just keep popping up. Hence the Question on who has the same issue and how they fixed it.