Hi all,
I performed a test on anoter ptp link (1km) using two SXTsq 5 ac. Ther's no way to use NV2, the link drops every 5 minutes. Only 802.11 seems to work.

Hi all,


No Answer from MT in the last 2 weeks,
we have no hope for next RC´s, problem is known for 8 Month now, and only ARM based AC devices available.

we have to build some new Sectors, looks like we have no another solution as use UBNT or Mimosa.
We don´t won´t to stick at a/n.

mistry7

Any update?

I hope Mikrotik works to improve ...

Hi all,
I performed a test on anoter ptp link (1km) using two SXTsq 5 ac. Ther's no way to use NV2, the link drops every 5 minutes. Only 802.11 seems to work.

Hi wispman74,
How many Mbps could you get for PTP link (1km) with 802.11 protocol?

Ping latency between SXTsq 5 ac devices?

Thanks!

Hi all,
I performed a test on anoter ptp link (1km) using two SXTsq 5 ac. Ther's no way to use NV2, the link drops every 5 minutes. Only 802.11 seems to work.

Hi wispman74,
How many Mbps could you get for PTP link (1km) with 802.11 protocol?

Ping latency between SXTsq 5 ac devices?

Thanks!

depends on noise and EIRP, if you are in noise area, don't buy Mikrotik go with something shielded and with GPS sync...

For the other Questions.......
- wait for ARM fix
- wait for spectral scan
- wait for ROS 7

or change Brand......

Hi binh,
sorry, but unfortunately there is another issue on my bridge. For some reason I can't perform bandwidth test, ping or access winbox from the wifi side. I noticed that service ports, in Firewall menu are all red. I hope it is not a RouterOS bug, but a wrong configuration. At the moment I can't figure out what is wrong.
Any suggestion is appreciate.

@mistry7: thanks for your advisory and information.

@wispman74: I think you got the problem with configuration. Lockup MikroTik wiki for how to configure PTP with MikroTik devices.

For the other Questions.......
- wait for ARM fix
- wait for spectral scan
- wait for ROS 7

or change Brand......

We pationally wait! , but as with everything one can not wait forever, currently we have very few customers affected, maybe 2-3.

Any update?

@mistry7: thanks for your advisory and information.

@wispman74: I think you got the problem with configuration. Lockup MikroTik wiki for how to configure PTP with MikroTik devices.

Hi binh,
thankyou for your advice, but this is not the first ptp I configured (tens before now), but is the third using ARM devices and I foud a lot of problems not only regarding nv2 and wifi connection.
I had to replace my ptp with another using MIPSE devices and everything gone well as always before.
Now I tested again the ARM ptp to understand why with the same configuration I have bad behaviour.
What happen is that sometimes, after finished to configure the devices, it becomes invisible from winbox. The reason is that RouterOS change itself the default value of neighbor discovery-settings.
Also, I was not able to reach master from slave and viceversa and I couldn't start a bandwidth test without any apparent reason.
After reset configuration and did it again, everything start to work fine.
I like Mikrotik, but is really unbeliveable to see so much issues in new ARM devices. Is it possible Mikrotik didn't see them during their tests?

@mistry7: thanks for your advisory and information.

@wispman74: I think you got the problem with configuration. Lockup MikroTik wiki for how to configure PTP with MikroTik devices.

Hi binh,
thankyou for your advice, but this is not the first ptp I configured (tens before now), but is the third using ARM devices and I foud a lot of problems not only regarding nv2 and wifi connection.
I had to replace my ptp with another using MIPSE devices and everything gone well as always before.
Now I tested again the ARM ptp to understand why with the same configuration I have bad behaviour.
What happen is that sometimes, after finished to configure the devices, it becomes invisible from winbox. The reason is that RouterOS change itself the default value of neighbor discovery-settings.
Also, I was not able to reach master from slave and viceversa and I couldn't start a bandwidth test without any apparent reason.
After reset configuration and did it again, everything start to work fine.
I like Mikrotik, but is really unbeliveable to see so much issues in new ARM devices. Is it possible Mikrotik didn't see them during their tests?

WISP hardware is not longer Focus for Mikrotik, look at Switches / Routers they release in the past 6 Month....... Hap Ac2 without POE out, SXTsq, LHG and all new
AC devices are build on Low Cost Platform, all with the same chipset, not features are the skill, only price.....
look at "KidControl" etc, these are thinks important for MT but not working P2P in noisy places and P2MP with much Clients....
And now really new, CWDM Hardware

@Normis

I need an answer: NV2 arm devices 'll work with NV2? If yes when?

If no we 'll have new AC hardware full compatible with NV2? If yes when?

We have 6000 CPE with 200 AP all mikrotik based switch to a new vendor is a big problem.

I hope to have soon an answer I have to decide how to build new towers.

Thank You in advance Giuseppe

this discussion will be the same as the one about Spectral Scan

A never ending story


we do not like ubiquiti but we change, because Mikrotik does not give any information about the further development. the next standard is coming up and the previous one is supported only rudimentarily. All my competitors have Spectral Scan, GPS Sync, working TDMA, to stay competitive we need that too, MT has none of that!

a Mimosa C5C in PTP mode, 40 MHz channel, modulation @ 400MBit over 5.5km with 27dBi Dishes from RF elements with max EIRP of 36 dBm, means about 280-290 MBit real TCP (iperf and 2 Laptops), for a short test we swapped the C5Cs against MT (RBM11 G + R11e-5HacD + shielded twist port adaptor) same channel same 40 MHz, same 36dBm, modulation varies between 324 and 400, with Iperf was only 170-180 MBit possible, with Nv2 even less.

Mikrotik has lost there competitors and is standing alone in the dark, and I believe more and more that it is there new Business case, go away from wisp wireless, but they don’t say it, they just go with low cost devices for costumers who don’t need speeds above 100MBIT, for selling 5 or 10 MBit plans, you don’t need any innovation......

Wisp with around 1000 clients here.

We are planning to installa new tower with 6 Netmetal on 2 triple pannels to migrate around 100 CPEs, using new Disc Lite AC (drm), and after reading all the problems posted here, we are really thinking it over.
It's a very noisy scenario and we really need NV2 to get it working.
Last routeros version 6.43rc66 (2018-Aug-28 13:36) says this:
*) wireless - improved Nv2 reliability on ARM devices
Is this statement real? Can we trust that Mikrotik will fix this really BIG problem? Or should we return the allready aquired Mikrotik equipment and switch to another vendor?

Thanks in advance and excuse my poor english.

can't really believe NV2 is the best choice in noisy area...
Every time i have interference issues and degraded thruput - its much better after switching to nstreme...

Wisp with around 1000 clients here.

We are planning to installa new tower with 6 Netmetal on 2 triple pannels to migrate around 100 CPEs, using new Disc Lite AC (drm), and after reading all the problems posted here, we are really thinking it over.
It's a very noisy scenario and we really need NV2 to get it working.
Last routeros version 6.43rc66 (2018-Aug-28 13:36) says this:
*) wireless - improved Nv2 reliability on ARM devices
Is this statement real? Can we trust that Mikrotik will fix this really BIG problem? Or should we return the allready aquired Mikrotik equipment and switch to another vendor?

Thanks in advance and excuse my poor english.

nv2 and nstreme are not working with ARM based Devices

@Normis

hundreds of WISP are in trouble cause of Mikrotik new devices and no answer comes from you. We would all appreciate a few words of explanation, because to be silent means to admit the existence of big problems that can not be solved, it means to say to everyone "Mikrotik is not the right product for you, change brand!". This is really frustrating for those who have invested and believed in Mikrotik for years and it is unthinkable that you want to give up the WISP market. Please let us know what's going on and what should we expect. Thank you

@wispman74

Don´t expect any answer, with the last "tries" to fix the issue we see that it is not an easy one.
That happened if you go with Low end chipsets.

I´m wondering the RB4011 hast QCA-9984 as Chipset, and not another IPQ Variant.
Hopefully we see support for QCA 9888 too, so we are able to use M11G with Wave2
Modules and if we are lucky with working nv2 Support.


Ruckus hast first AP with IPQ8074 (Enterprise 802.11ax)
When will Mikrotik start with 802.11ax?
In 2 Years, with old Kernel and bad driver support?
Or was 802.11ac Mikrotiks Windows Vista, and there is some moving forward with

- Newest Linux kernel
- Driver near ATh10k (speed, compatibility, chipset support)
- Channel width 5 / 10 / 30/ 60Mhz
- TDMA Protocol (maybe NV3)
- GPS Sync

There is a very welcome answer from Mikrotik. Can you comment the "ARM and NV2" situation?
@Normis - Or do you have a ban on answering this topic?

Hi, we are hope for this solution... tested in a LHG 5ac with version 6.43rc66 is better with nstreme than NV2...

Come on MT guys, you are the best!

Regards,

The 5 Ghz outdoor wireless market is too big to lose it.. There are thousand of WISP using MT worldwide we have 200 MT routers and 6000 MT CPE the difference beetween the 2 markets is huge.

@Normis please an answer give us info to continue our business also if the news are bad and we have to move to other vendors for wireless.

Thank You

I think we are waiting for a solution for a really long time now. We started to change PtP link for ubnt as soon as 400iso came out. It works so much better than anything from mikrotik. Last week we changed some 911+jrc24dBi for iso400 and few lhg5 for litebeams and in the same channel throughput went up 4 times. Only think that keeps us with mikrotik are clients. It is hard to migrate everything. But i think thats only a matter of time when we do that...

The 5 Ghz outdoor wireless market is too big to lose it.. There are thousand of WISP using MT worldwide we have 200 MT routers and 6000 MT CPE the difference beetween the 2 markets is huge.

@Normis please an answer give us info to continue our business also if the news are bad and we have to move to other vendors for wireless.

Thank You

If you sell 5/10/15MBit plans, you should stay with MT we sell 50/100/200MBit plans, and this is impossible with MT
We are Surprised, 802.11ad clients with fast Ethernet, there are some engineers sleeping at work.....
They better thinking about combined 5Ghz/60 GHz Devices with GE and SFP, but this they want to sell at 2020 as big new solution ....

Nowdays a truck-roll, or 1-2hour of manwork more expensive than those device.
So i think most of us happily pay a few percent more, if the capacitor's lifetime much longer, and the chip is working well with software.

Please don't save few bucks in these!

6.43 Works only for TDMA period size 1ms. Try 3-5ms = disaster. Hardware driver/chipset/clock problem.

We have a LHG 5ac(Level4) like backbone concentrator of 4 others towers with LHG 5XL, Sextant-G, SXTsq, SXTsq... all are in a sector of 10° and the most longest link is at only 7km... all in band 5Ghz-only-N

We tested NV2 in TDMA 1ms, 2ms, 3ms, 4ms, auto, dynamic-downlink, fixed-downlink, etc etc and nothing was better then Nstreme...

Now the max troughput total is 80Mbps of uplink to theses others towers, before with a Sextant-G(Level4) in NV2 in the place of the LHG 5ac the traffic was 120-150Mbps...

Tomorrow i go up to the top of the tower to change back to the old Sextant-G... hoping NV2 and 802.11ac work fine in a near future

Now the max troughput total is 80Mbps of uplink to theses others towers, before with a Sextant-G(Level4) in NV2 in the place of the LHG 5ac the traffic was 120-150Mbps...

Tomorrow i go up to the top of the tower to change back to the old Sextant-G... hoping NV2 and 802.11ac work fine in a near future

NV2 Never Performs good in ac, and there are no Informations why, all our old 802.11n
P2P links are performing better with old Hardware, until we began to use another Vendor too.

With Mimosa and UBnT we are seeing speed increases about 3-4 times
with same Channel / Width / Interferrence.

Now the max troughput total is 80Mbps of uplink to theses others towers, before with a Sextant-G(Level4) in NV2 in the place of the LHG 5ac the traffic was 120-150Mbps...

Tomorrow i go up to the top of the tower to change back to the old Sextant-G... hoping NV2 and 802.11ac work fine in a near future

NV2 Never Performs good in ac, and there are no Informations why, all our old 802.11n
P2P links are performing better with old Hardware, until we began to use another Vendor too.

With Mimosa and UBnT we are seeing speed increases about 3-4 times
with same Channel / Width / Interferrence.

Hi mistry7,
what does you mean exactly when you say 3-4 times? I understand you compared mikrotik with ubiquiti/mimosa with same frequency/width/interference. What about TX power, antenna gain, MIMO type, test in udp or tcp?
If everything is the same for two vendors and you are working using 802.11n, to reach 4 times mikrotik speed, means you have 75Mbps with Mikrotik and 300Mbps with other vendors.
Is that correct?

My experience says that with Ubiquiti AF in PtP we have from 2 to 4 times speed, in PtMP with the recent improvments mikrotik in NV2@20Mhz is from 10% to 20% slower than Ubiquiti but still acceptable. @40Mhz Miktotik ac hardware don't work well.

But the main problem is that new arm devices don't work with NV2 and we come back to N mipsbe

Now the max troughput total is 80Mbps of uplink to theses others towers, before with a Sextant-G(Level4) in NV2 in the place of the LHG 5ac the traffic was 120-150Mbps...

Tomorrow i go up to the top of the tower to change back to the old Sextant-G... hoping NV2 and 802.11ac work fine in a near future

NV2 Never Performs good in ac, and there are no Informations why, all our old 802.11n
P2P links are performing better with old Hardware, until we began to use another Vendor too.

With Mimosa and UBnT we are seeing speed increases about 3-4 times
with same Channel / Width / Interferrence.

Hi mistry7,
what does you mean exactly when you say 3-4 times? I understand you compared mikrotik with ubiquiti/mimosa with same frequency/width/interference. What about TX power, antenna gain, MIMO type, test in udp or tcp?
If everything is the same for two vendors and you are working using 802.11n, to reach 4 times mikrotik speed, means you have 75Mbps with Mikrotik and 300Mbps with other vendors.
Is that correct?

We are working in very noisy inmoriment, and with https://mimosa.co/products/a5c and GPS sync we see 3-4times Bandwidth.
With MT often 50Mbit are impossible.

Here:
viewtopic.php?f=7&t=136002&start=100#p684307

Mimosa C5C in PTP mode, 40 MHz channel, modulation @ 400MBit over 5.5km with 27dBi Dishes from RF elements with max EIRP of 36 dBm, means about 280-290 MBit real TCP (iperf and 2 Laptops), for a short test we swapped the C5Cs against MT (RBM11 G + R11e-5HacD + shielded twist port adaptor) same channel same 40 MHz, same 36dBm, modulation varies between 324 and 400, with Iperf was only 170-180 MBit possible, with Nv2 even less.

170-180 Mbit wit 802.11 does not help really, because of Ping jitter, so it is impossible to deliver good voip services.
We need a working TDMA solution, and with nv2 in such test you reach 100-120Mbit, mimosa ist 3 times higher, with a 95€ Radio....

Mistry7, I have no experience on ubiquiti or mimosa, but it seems you are comparing not similar architecture. Even if you have all the same parameters in mimosa and mikrotik, C5C is a 4x4 MIMO, R11e is 2x2. It is normal you have a double bandwidth as in post
viewtopic.php?f=7&t=136002&start=100#p684307
Also if you hace GPS sync it is normal you get better results in noisy environment. You should compare a mimosa product with all the same features.
So, we can discuss Mikrotik lacks some features (4x4 MIMO or GPS Sync), NV2 is not really a stable protocol, or Mikrotik works well in clean environment, but brobably with the same architecture, feature and protocol, maybe the results are not so far each other.
I think is important to be clear talking about different vendors because some of us make their decition based on other opinion.

Someone tested in NV2 how is working the RBDynaDishG-5HacDr3 which is 802.11ac but has MIPSBE?

Mistry7, I have no experience on ubiquiti or mimosa, but it seems you are comparing not similar architecture. Even if you have all the same parameters in mimosa and mikrotik, C5C is a 4x4 MIMO, R11e is 2x2. It is normal you have a double bandwidth as in post
viewtopic.php?f=7&t=136002&start=100#p684307
Also if you hace GPS sync it is normal you get better results in noisy environment. You should compare a mimosa product with all the same features.
So, we can discuss Mikrotik lacks some features (4x4 MIMO or GPS Sync), NV2 is not really a stable protocol, or Mikrotik works well in clean environment, but brobably with the same architecture, feature and protocol, maybe the results are not so far each other.
I think is important to be clear talking about different vendors because some of us make their decition based on other opinion.

C5C is normal 2x2 SMA Radio, not more then NetMetal etc....
C5C has not gps

Mistry7, I have no experience on ubiquiti or mimosa, but it seems you are comparing not similar architecture. Even if you have all the same parameters in mimosa and mikrotik, C5C is a 4x4 MIMO, R11e is 2x2. It is normal you have a double bandwidth as in post
viewtopic.php?f=7&t=136002&start=100#p684307
Also if you hace GPS sync it is normal you get better results in noisy environment. You should compare a mimosa product with all the same features.
So, we can discuss Mikrotik lacks some features (4x4 MIMO or GPS Sync), NV2 is not really a stable protocol, or Mikrotik works well in clean environment, but brobably with the same architecture, feature and protocol, maybe the results are not so far each other.
I think is important to be clear talking about different vendors because some of us make their decition based on other opinion.

C5C is normal 2x2 SMA Radio, not more then NetMetal etc....
C5C has not gps

In noise environment Mimosa is a shit! only work fantastic in the best conditions...

Mistry7, I have no experience on ubiquiti or mimosa, but it seems you are comparing not similar architecture. Even if you have all the same parameters in mimosa and mikrotik, C5C is a 4x4 MIMO, R11e is 2x2. It is normal you have a double bandwidth as in post
viewtopic.php?f=7&t=136002&start=100#p684307
Also if you hace GPS sync it is normal you get better results in noisy environment. You should compare a mimosa product with all the same features.
So, we can discuss Mikrotik lacks some features (4x4 MIMO or GPS Sync), NV2 is not really a stable protocol, or Mikrotik works well in clean environment, but brobably with the same architecture, feature and protocol, maybe the results are not so far each other.
I think is important to be clear talking about different vendors because some of us make their decition based on other opinion.

C5C is normal 2x2 SMA Radio, not more then NetMetal etc....
C5C has not gps

OK, sorry but I follow your link of A5C

Mistry7, I have no experience on ubiquiti or mimosa, but it seems you are comparing not similar architecture. Even if you have all the same parameters in mimosa and mikrotik, C5C is a 4x4 MIMO, R11e is 2x2. It is normal you have a double bandwidth as in post
viewtopic.php?f=7&t=136002&start=100#p684307
Also if you hace GPS sync it is normal you get better results in noisy environment. You should compare a mimosa product with all the same features.
So, we can discuss Mikrotik lacks some features (4x4 MIMO or GPS Sync), NV2 is not really a stable protocol, or Mikrotik works well in clean environment, but brobably with the same architecture, feature and protocol, maybe the results are not so far each other.
I think is important to be clear talking about different vendors because some of us make their decition based on other opinion.

C5C is normal 2x2 SMA Radio, not more then NetMetal etc....
C5C has not gps

In noise environment Mimosa is a shit! only work fantastic in the best conditions...

We don’t have fantastic conditions, on no place, 5 GHz is disturbt all
over the hole band, no place where we don’t see 15-20 ssids

We use maximum on antenna technology, we have small EIRP here in Germany 30/36 dBm,
And for us mimosa works great, but killer is AF-5xHD, but it is an another class of Radio.

We don’t have fantastic conditions, on no place, 5 GHz is disturbt all
over the hole band, no place where we don’t see 15-20 ssids

Off topic: this is not disturb - We have in Czech Republic more than 60 ssid on P2P link.
More then 100 SSID on 30degree horn antenna. In small city.

We have small EIRP here in Germany 30/36 dBm,

We have only 27-30dBm

We don’t have fantastic conditions, on no place, 5 GHz is disturbt all
over the hole band, no place where we don’t see 15-20 ssids

Off topic: this is not disturb - We have in Czech Republic more than 60 ssid on P2P link.

More then 100 SSID on 30degree horn antenna. In small city.

Here A5C really works great...

We have small EIRP here in Germany 30/36 dBm,

We have only 27-30dBm

But when you see all SSId with -60? Ubnt and Co are often used here form cameras etc, every IT company use them, and WISPs searching the Gaps...
My City Airfiber Show hole Band as red

For profesional cameras we use other band than 5Ghz. In CZ is problem with a lot of WISPs in cities... A lot of SSID, a lot of interference....

Hello Folks!

I see a lot of discussions in this thread now.

Has microtic corrected their routers now so older devices work with wireless again, so called arm devices, e.g. mipsbe devices like rb411, sxt, sextant etc. ?

We have very many such still in production ptmp 5GHz band using no higher than 6.41.3, if we upgrade beyond that version the wifi fully fails and the device need a full netinstall recover to 6.41.3 again to become usable.

I am having this same issue with a PTP link with two of the new SXT ac square units. We have simple bridge to station bridge setup with wlan and eth bridged together.

No traffic would pass through link on NV2. When I upgraded to 6.43 suddenly the link began to pass traffic through and work, but now my Cisco 3750g switch keeps receiving it's keepalive packet on the same port it sends it out on thus putting the port into err-disabled mode. This only happens when the SXT set to "station-bridge" connects to the SXT set to "bridge".

So it seems for whatever reason the bridge unit is looping traffic back to the CPE device (or should I say the other end of the PTP bridge.

Setup as follows:

Cisco 3750g 1 <-> RBSXTsq5HPnD-US ("station-bridge") <-> RBSXTsq5HPnD-US ("bridge") <-> Cisco 3750g 2

Cisco 3750g 1 shuts down the port that the RBSXTsq5HPnD-US is connected to because it receives the same keepalive packet (Sent from Cisco 3750g 1) that it sent out on the same port connected to the same RBSXTsq5HPnD-US. Clear as mud?

Someone tested in NV2 how is working the RBDynaDishG-5HacDr3 which is 802.11ac but has MIPSBE?

Has anyone tested the newer RouterOS with these higher versions? Does NV2 link work normal?

Also, I find it odd that nstream would work and NV2 would not given that NV2 is short for "nstream version 2"

Someone tested in NV2 how is working the RBDynaDishG-5HacDr3 which is 802.11ac but has MIPSBE?

Has anyone tested the newer RouterOS with these higher versions? Does NV2 link work normal?

Also, I find it odd that nstream would work and NV2 would not given that NV2 is short for "nstream version 2"

This is not strange. Nstreme (v1) do the things via board CPU, while NV2 do that on the wireless card.
Compare the CPU usage when set to nstreme and nv2! You will find, that using nstreme will consume 20-30% cpu even with low network load.

Also, mixing nv2 and nstreme in the same board (on different cards) is not a good way. They would interfere somehow even if the used freq very far from each other.

Someone tested in NV2 how is working the RBDynaDishG-5HacDr3 which is 802.11ac but has MIPSBE?

Has anyone tested the newer RouterOS with these higher versions? Does NV2 link work normal?

Also, I find it odd that nstream would work and NV2 would not given that NV2 is short for "nstream version 2"

This is not strange. Nstreme (v1) do the things via board CPU, while NV2 do that on the wireless card.
Compare the CPU usage when set to nstreme and nv2! You will find, that using nstreme will consume 20-30% cpu even with low network load.

Also, mixing nv2 and nstreme in the same board (on different cards) is not a good way. They would interfere somehow even if the used freq very far from each other.

So nstream is using arm archtecture while nv2 is using wireless chip directly, awesome insight! Going back to the original issue, though, that is the core issue of nv2 and arm not working correctly. There obviously seems to be a issue with bridging or routing going on as we keep receiving random sent packages from our AP that we sent to it.

It almost seems like:

Broadcast message is created -> "CPE PTP bridge radio" receives packet in bridge (via eth1), sends out message on all bridge ports except port where message originated, aka wlan1 -> The "bridge PTP radio" receives the packet on its bridge interface wlan1 and should then send out the packet on all interface except the port where the message originated, aka eth1*


* = This seems to be the point where the "bridge PTP radio" will send the keepalive packets generated by the cisco switch back to the radio that sent it which in turn broadcast's the message back ot the port putting it into err-disabled mode and losing all connectivity between sites.

This seems to be a major issue with ARM based boards only. WE ALL NEED TO KNOW WHAT IS GOING ON HERE? CAN WE GET AN ACKNOWLEDGEMENT OF THE ISSUE, @ULDIS, @EMILS, @NORMIS?

Given the latest upgrade fixed some of my issues without any config change it sounds like some major changes are being made to arm based boards with wireless cards in them. Our company wants to buy a TON of the next ARM-based SXT's but we can not move forward until we know if this is a known issue being worked out or a configuration issue.

I am trying to hook my cell up to my laptop so i can get GPS on Network Stumbler and dont know how to go about doing this?

And what is the relationship to the topic?

I am trying to hook my cell up to my laptop so i can get GPS on Network Stumbler and dont know how to go about doing this?

piss off

Can someone with some sort of authority acknowledge this? I may have missed it but this is a big deal. Is this a known issue or am I crazy???

Traffic is being looped back for some reason - this can cause MAJOR network issues and will create an incredibly unstable network.

What do you recommend your clients do with ARM based boards for stable PTP communication? Use 802.11 for now? Any suggestions are welcome and anything I can do to help I will gladly do.

I can confirm the issue. We have had major throughput issues, so we replicated it in lab environment.

LHG 5 ac XL (bridge) -- LHG 5 AC (station bridge)
only chain 0 (we don't have enough free space for polarization to form, so used single chain to reduce other effects)
Throughpus is dead-straight at 78Mbit up + 78Mbit down, CCQ at 100%. That is exactly half of what it should be taking into account protocol overhead and conforms with the forth-and-back routing issue.

We also tried to downgrade to 6.40.3, but now the XL is rebooted with kernel panic if any wireless interface setting is changed. So there are major problems with ARM port of the latest stable.

We can not switch back to MIPSBE LHG 5 utilizing N standard as we are facing competition throughputing 400+Mbit.

With test runs with RB922-RB922 before switching to ARM we achieved 400+Mbit/s in distance of 60km. Now we are sitting at 20-80Mbit with very high jitter.

So yes, ARM devices miss NV2 slots and yes, ARM devices route traffic back after receiving reducing available timeslots.