w60g link stability and erratic rate selections since ROS 6.47

I have been having stability issues with all my w60g devices since 6.47. The rate selection is erratic and have frequent disconnects. Going back 6.46.5 improves a lot the stability. I had similar issues even on 50m links

I think this completely cocked up the stability.
*) w60g - improved rate selection in low traffic conditions;

Has anyone noticed strange rate selections on links with 90-100 signal quality? Any disconnects?

No? Nothing? None is using them? No strange issues for you?

I using on two links it is seems to me less stable links, before I don’t have disconnection for weeks after update to 6.47.1 have every 24-48h disconnections for one seconds it is not critical but seems to me less stable that has been before.

we see a lot of disconnects since moving to 6.47 a well. did you try the latest 6.46.x version and its not stable and 6.46.5 is most stable?

Hi,

We’ve also noticed some erratic stability issues with the wAP 60G in PTMP setup. The weird thing is that, there’s a powerbox behind the wAP 60G station with a tunnel terminating to it. The tunnel never goes down nor the powerbox behind it. We can ping the powerbox even with the station supposedly down.

The radios are currently in the 6.47.2 version and we’ve turned on the “mdmg-fix” feature which helps in the PTMP negotiation. Does lowering the version down to the stable 6.46.x version helped you? And also, because it’s bridged network, did you include the w60G interface to the bridge where the w60G-station interface is?

Thank you. Any insight would help.

I installed a new relative short PtP link (about 600m) with two LHG 60G. With the latest stable version (ROS v6.47.4) the Signal, RSSI, and TX Sector are permanently changing. Also the MSC and PHY Rate are changing quite often and the link was dropping multiple times a day. With the highest frequency (64,8GHz) not as often as with the lower frequencies.
I did not have any of this problems with an older comparable link in the same town. On the LHG 60G of this older link the ROS version v6.46.6 was installed. Therefore, I decide to install the latest version from the stable channel (v 6.46.7) on the LHG 60Gs of the new link. However, this did not much improve the situation with the permanently changing w60g state parameters. During searching for an explanation for this unstable behavior, I found this thread. After reading this I downgrade the ROS on both LHG 60G of the new link to v6.46.6. Since the downgrade, about 12 hours ago, the link parameters are perfectly stable.

I come to a similar conclusion as NetTraptor. One of the two w60g related changes from the changelog of 6.46.7 did lead to regression bug.
w60g related changelog entries:

  • w60g - added “mdmg-fix” parameter for RBwAP60Gx3 (CLI only);
  • w60g - improved rate selection in low traffic conditions;

Hi,

we did run into the same issues here. Some weeks ago (2020-09-21) i updated a wAP60G Link to current firmware. Afterwards i observed a lot of beamforming events (tx-sector jumping all over the place), PHY-Rate/MCS drops and serval short links loss events. Beamforming seems to be more aggressive in the newer releases.
Going back to 6.45.7 solved the issues.

Link Frequency: 60480
Link Length: 140m
Hardware: wAP60G (both sides)

I attached some plots (tx-sector, signal quality and rssi) - might be helpful to hunt down the firmware issue. Firmware downgrade to 6.45.7 was done Wednesday evening (2020-09-23). The first half of the graps shows the problem.
rssi.png
signal_quality.png
tx-sector.png

Hi, I have the same issue on my P2P wAP 60G link with 30m distance. I noticed it yesterday when I upgraded to 6.47.4. The link was unstable and the connection dropped about 100 times during 24 hours. Today I downgraded back to 6.45.7 (as suggested by illuminator, thank you very much) and it’s working fine for a few hours. But downgrading shouldn’t be the solution. Does anyone have more information what is causing this? I’ve seen quite a few people complaining about this, but not a single statement from MikroTik that they’ve acknowledged this issue and are working on a fix.

this is fixed in 6.48beta48, fix will also be merged in next stable release

Hello. I think what is need to do, if you can on 60GHz setting of power and parameters for the radio because seems that here is on short distance (around 100 meters lot of noise and ranging of the power adjustment to Transmit power to the other side. And majority of issue is when you have 60GHz WAP to W60G . On all tested ROUTEROS version from 6.46 Long term 6.47 stable(what is not stable at all) and Test 6.48 what is lost stable but all have Link dow or link up issues on any distance and more than 100 times per day.

I’m seeing the same erratic behaviours on my most recent deployments of LHG60G PTP links running 6.47 and newer. My older ones remain very stable. Rolling back the newer systems to 6.46.7 tonight…

Found no improvement on 6.46.7. Trying the new 6.47.6 tonight, and moving one way back to 6.43 (I have some old radios that have been outstandingly reliable on that version).

I’ve had luck with 6.47.6 on both ptp and ptmp setups. Curious if others see the same?

Still problem:

routerboard: yes
board-name: wAP 60Gx3 AP
model: RBwAPG-60ad-SA
serial-number: AC8F0AE230F7
firmware-type: ipq4000L
factory-firmware: 6.43.6
current-firmware: 6.48beta48
upgrade-firmware: 6.48beta48
Screenshot (168).png
Screenshot (219).png
Screenshot (220).png

Updated one wap60g link to 6.47.8 today (point to point).
It took a few minutes to get a stable link at MCS 8 (this is okay, as they are poorly aligned). Now everything is running smoothly - no jumpy beamforming.

Note: this is a backup link with close to zero traffic. Did run a btest with 900MBit/s (each direction) → looked good.

If it runs stable for the next days tests will continue on our LHG60 links.

Wap 60Gx3 routerboard is pinging without any problems. Downgrade did not help. I come to the conclusion of a problem with electronics / wifi. The warranty is overdue for a month :(. When I replace the device, I will try Netinstall, but I don’t think it will solve the problem.

The problem cannot be repeated in a day. Works great.
I just don’t know what to do …
Has anyone had a similar problem?

Just had the opposite experience: Upgraded a LHG 60G PTP link (1.1 km) that was quite stable with ROS 6.47.1 (MCS 7-8 and “Signal” 65-80) to ROS 6.47.8 and got MCS 2 with “Signal” ~ 20 on both ends although the TX sector remained near center. Downgrading to 6.47.1 fixed it.

Hi all,

upgraded a LHG60 link (140m) to 6.47.8 yesterday. Looking good so far.

@TMA: If the last updated link runs smooth for a couple days, i will upgrade a link with similar length as yours (>1km) next. If it shows similar problems as yours, i will report back.

Hi team,

A quick update, we’ve updated all the radios with a firmware version that’s not lower than 6.47.6 and it has been stable ever since.

More than a month in the firmware version, we hadn’t had any downs on the radio. I hope everyone’s problem has been sorted. Thank you.