AP One AC Mini and Balance 20 - WAN connection dropping

Lewis_Kong · March 12, 2018, 2:44am

Hi all,

Unfortunately, we have confirmed a firmware bug in RC4 and RC5 which causes AP running out of resource gradually. That bug could be the reason why Ryan reported his IoT devices suffered after running RC4 for several days.

RC6 and corresponding AP Controller firmware pack 1155 are ready for download. We wish this is the last round of 3.6.0 RC.

Thanks,
Lewis

Ryan_Finger · March 12, 2018, 5:34am

I will download 1155 and apply it today.

scuba_steve · March 12, 2018, 4:16pm

My B1 Core is claiming that firmware pack is invalid.

Lewis_Kong · March 12, 2018, 4:23pm

Hi Steve,

Please try to remove an unused firmware pack from AP Controller Toolbox.
That should be the AP Controller front end bug which displays an incorrect error message when firmware pack exceed limit.

Thanks,
Lewis

scuba_steve · March 12, 2018, 4:44pm

That did the trick. Thanks for the prompt response Lewis. 1155 is pushing out to my AP One Es now.

Steve

Lewis_Kong · March 15, 2018, 4:07am

Hi All,

Unfortunately again… we found RC6 may get incorrect client IP addresses in some occasions. As incorrect client IP can affect captive portal functionality, we decide to release another RC7 before G.A.

RC7 is here.
AP Controller Firmware Pack is 1156

Thanks,
Lewis

Ryan_Finger · March 15, 2018, 6:30am

Just applied RC7.

Travis · March 20, 2018, 9:40am

Running the latest RC, I see the 5GHz clients show up as all using 6Mb/sec Tx rate (from the client list). I am guessing this is a UI bug, not an accurate reading of the PHY rate?

djk · March 24, 2018, 11:32pm

Unfortunately again…

3.6.0 RC7 1793 does not fix the problem. After 5 days of up time the APO-RUG units stop routing packets completely, but the performance is lost much sooner. A reboot fixes the problem temporarily. You have a memory leak. Seems to be a running theme in many of the bug reports. Suggest you find and fix it.

nslookup · March 30, 2018, 1:26pm

Installed AP Controller Firmware Pack is 1156 and now the VLAN’s are forced to 0: Untagged and getting pushed through the .0 DHCP which is for management stuff only.
Balance 305 HW1(single) and HW2(HA)
11 AC One Mini’s

adeebag · April 2, 2018, 9:02pm

@Lewis_Kong any updates? when do you expect the GA to be released?

Lewis_Kong · April 2, 2018, 9:52pm

Hi DJK,

Does the issue still happen? Could you please open a support ticket?
We cannot produce your issue in our lab which have been running for more than 2 weeks with 15 active Wi-Fi clients.

Hi Stephen,

I am not sure about what problems you are facing.
Could you please open a support ticket or PM me more information?

Hi Adeebag,

We are prepare related document for 3.6.0GA. Please stay tuned.

Thanks,
Lewis

nslookup · April 3, 2018, 1:10pm

Lewis,

No ticket is needed as we have reverted the firmware pack on the Routers for the time being.
We are having issues with the AP’s dropping users on the current release firmware.
However, this is what was happening for other users who might use the beta firmware.

When the beta firmware 3.6.0 RC7 (1156 Firmware Pack) was installed, all users that were handed DHCP under ANY SSID were handed out un-tagged IP addresses. e.g. 192.168.0.200 instead of 192.168.40.200
When we reverted to Firmware Pack 1145 and the DHCP started working correctly.

I will test on the one I have at the house later and if I can re-create the issue I will let you know.

Ryan_Finger · April 3, 2018, 2:13pm

I’m running RC7 managed by a Balance One running 7.1 and I’m not experiencing this. I have 4 VLANs, 3 of which are trunked to my two AP mini’s and two AP Enterprise’s. All devices are getting the proper DHCP addresses and functioning correctly.

djk · April 3, 2018, 2:26pm

Yes, issue is still happening. We have set our routers to reboot every day and that seems to help. I think the routers need to rebooted every 12 hours or so. Could you add that to the reboot menu until you find the memory leak? So far, I have not seen any problems within 12 hours of the router being powered on.

Ryan_Finger · April 5, 2018, 4:40am

@djk

I can report I’m experiencing something very similar. I am running RC7 and I have had to reboot two of my access points that have just stopped accepting connections. One of the APs has a dedicated SSID that is for two IoT devices. I discovered that those devices went offline. Resetting the devices made no difference. Trying to connect my phone to the dedicated SSID failed. I rebooted the AP (AP One Mini) and it fixed the issue and I can now connect. Earlier in the week I had something similar happen with an AP One Enterprise. In that case I rebooted the AP and again was able to connect.

@Lewis_Kong

For the last reboot, I grabbed a diag report for the AP and the Balance One. I am PMing the report to you.

For now, I have my APs rebooting weekly.

Regards,
Ryan

Travis · April 21, 2018, 11:18am

Is there also an RC build for the 11n products (AP One In Wall)?

Charles_Steel · April 25, 2018, 2:32am

This thread seems to have gone quiet. Is there any update?

djk · April 27, 2018, 5:58am

Problems continue. Haven’t heard anything back. It helps if you reboot the unit every few hours but the firmware only supports reboots once a day. ]

TK_Liew · April 29, 2018, 5:16pm

@djk, please follow up with us in the ticket if you still face the issue. We have no chance to access the AP to take a closer look.

For those still face issue on v3.6.0 RC7, please open ticket and do us favor below.

Provide device Serial number
Enable Remote Assistance
Download Diagnostic Report from the affected AP when the issue happens.