You are here

Pre-Nightly Build testing of upgrade to Openwrt 19.07

53 posts / 0 new
Last post
AE6XE
AE6XE's picture
Pre-Nightly Build testing of upgrade to Openwrt 19.07

Device Test status is here:  https://drive.google.com/file/d/1IyDD_gPC4LLtZGERDQhS8Ix7EDiNjK12/view?u...

Below is the link to the images and packages for pre-nightly build ready for early testing.   Please assist as able to help test.  We are looking for confidence to put into the nightly build.  Please only consider testing if you are confident with loading AREDN images and using basic linux shell command line to return diagnostic information.  

Warnings
* Do not put these images on devices at tower sites or other production sites
* If a device does not respond after trying an image, then follow the instructions to load the previous image via tftp or revert back to vendor's firmware and recovery procedures
* tunnel features are not available to install (necessary server side won't be setup until put into nightly build)
* packages can be installed manually, by finding all dependencies in the folders at the link below and manually uploading
* images are split between targets/ar71xx and targets/ath79 folders.  If you can't find the image, look in the other folder.   Some of the images have upgraded to ath79 target,  which incorporates the most current linux features.

Known issues:
* ath79 nanostation-m-xw  <-  does not recognize boardID, so doesn't know power settings, but otherwise is functional

What is not yet known, to find out:
* the ath79 bullet-m image works on both the Bullet M2 and the NanoBridge M5.   Any device that previously loaded the bullet-m image should also try this image. 
* the ath79 bullet-m-xw image may support devices that previously loaded the Rocket-m-xw image.  Give these devices a try, e.g. some PBE devices.
* please read the posts to see what has already succeeded.  No need to post another "it succeeded again", but do post if a problem is found.

Capture this information
* Before upgrading a node, reboot, let stand 5 minutes on the local mesh, and then capture the free memory on the status page
* After the upgrade, do the same and post the "before / after" results.   We are answering the question how improved linux has become in size and memory management.

https://drive.google.com/drive/folders/1Pm466OqLPTARUuh_ux3xSoD2123H5EOw?usp=sharing

Joe AE6XE

Ai6bx
Preparing to test
I will try these on a few of my easier to reach nodes and report back. It s=
ounds like these are definitely riskier builds than the standard nightly so =
will build a few test nodes first to run on with local links before placing =
anything on the production side of life.
K6CCC
K6CCC's picture
Are these the same

Are these the same firmware that you linked on the SoCal e-Mail list last Monday?  I note that the filenames are shorter than what I tested earlier this week (but the revision number appears to be the same).
I have not had time to see if I can bring back my MikroTik RouterBOARD LHG 2nD-XL that was bricked when I updated it.  Maybe later this evening.
My MikroTik RouterBOARD RB952Ui-5ac2nD (hAP AC lite) is very much not stable with the code from last Monday.  Won't stay connected the vast majority of the time.  Went back to Nightly 1181 a couple days ago and it was stable.  Upgraded to 19.07-1272c57 again today and it is again not staying connected.  Interesting part is that the hAP it should be connected to is showing it there on charts, but not connecting.
 

AE6XE
AE6XE's picture
The images under the ar71xx
The images under the ar71xx folder are essentially the same.  The images under the ath79 folder are new.    I'll take a look at the hAP AC lite to reproduce.

Joe AE6XE
AE6XE
AE6XE's picture
I think I have an LHG XL 2GHz
I think I have an LHG XL 2GHz in a box around here.  I'll see if I can reproduce.

Joe AE6XE
AE6XE
AE6XE's picture
I must have given back the
I must have given back the LHG 2nD XL, was on loan.  But I had a successful upgrade on a LHG 2nD (the smaller low power one).   Thus, whatever info can be found; support dump, capture of output from serial console, etc. would be helpful.

Joe AE6XE
AE6XE
AE6XE's picture
I've loaded the test image on
I've loaded the test image on the hAP ac lite and reproduced the problem others have found.   The problem is with the 5GHz radio.  After loading the .elf, there is short time, if you type on the command line, "wifi down", this stopped the node from becoming unresponsive and allows the process to continue.  Then after loading to flash and rebooting, in basic setup, do not enable the 5GHz for LAN or Wifi Client.   

Openwrt has switched to use Candela Tech (ct) 802.11ac driver instead of the upstream linux one (takes longer to get updates) and our build does not appear to be inheriting that change.   I'll investigate and should have new images out shortly.

Joe AE6XE
AE6XE
AE6XE's picture
The test matrix of devices is
The test matrix of devices is relatively large.  Your efforts are best spent on devices that have not yet been tested:

https://drive.google.com/file/d/1IyDD_gPC4LLtZGERDQhS8Ix7EDiNjK12/view?u...

Joe AE6XE
K6CCC
K6CCC's picture
Did not have time to do
Did not have time to do anything last night (I'm hot and heavy in Christmas light show setup this weekend), but will attempt to bring the MikroTik RouterBOARD LHG 2nD-XL back to life.  After the 19.07rc update last Monday, it is left with nothing but the power LED.  No communication at all.  The plan will be to TFTP flash it with a recent nightly (yea with the two step process as if it was a new AREDN install).  If that works, I will let it run a while to make sure it's stable and then try the update again.
 
KB9OIV
Joe:
Joe:

Rocket M5 XW:

I reboot node first, then wait 5 minutes, then applied 'aredn-ae6xe-19.07-1272c57-ubnt_bullet-m-xw-sysupgrade.bin' as suggested on the spreadsheet.  Web interface showed 100% upload, and the 'Please Wait' pages.

A few minutes later, the node was back.

However, the firmware version was NOT newer.  All indication is that the firmware above did not apply.

I started with '1181-c17f565' for 'rocket-m-xw'.

Should I use an alternate method for the ath79 firmware?

Or would you like me to try to upgrade with '​aredn-ae6xe-19.07-1272c57-ubnt-rocket-m-xw-sysupgrade.bin'?

Matt
AE6XE
AE6XE's picture
Matt,  it was worth a try. 
Matt,  it was worth a try.  Yes, revert to try the ar71xx rocket-m-xw image.   I'll update the spreadsheet accordingly.

Joe AE6XE
KB9OIV
'​aredn-ae6xe-19.07-1272c57

'​aredn-ae6xe-19.07-1272c57-ubnt-rocket-m-xw-sysupgrade.bin' upgraded fine on my rocket M5 xw.

Some differences to note.  The 19.07 firmware has these additional packages stock that were in the firmware that the older version did not have built in:

cgi-io - 14
getrandom - 2019-06-16-4df34a4d-3
kmod-hwmon-core - 4.14.152-1
kmod-ipt-offload - 4.14.152-1
kmod-nf-flow - 4.14.152-1
rpcd-mod-file - 2019-11-10-77ad0de0-1
rpcd-mod-luci - 20191114
urandom-seed - 1.0-1
urngd - 2019-06-17-c057e177-1


Also, the older firmware had iperf 3 installed with these packages:
iperf3 - 3.5-1AREDN
iperfspeed - 0.5

So, my attached screenshots have those installed, too, to be fair:
iperf3 - 3.7-1AREDN
iperfspeed - 0.5


The attached screenshots are my rocket 5 minutes after reboot before the firmware upgrade, and 21 hours after installing the 19.07 firmware.



 

Image Attachments: 
AE6XE
AE6XE's picture
KB9OIV,   perfect.   Can you
KB9OIV,   perfect.   Can you capture the support data download and upload?   Something needs fixed since it shows untested.   Also, can you telnet or ssh into the device and see if this file exists?   "cat ​/sys/devices/pci0000:00/0000:00:00.0/subsystem_device".

Joe AE6XE
KB9OIV
The path:
The path:

/sys/devices/pci0000:00/0000:00:00.0/subsystem_device

does not exist.

I have attached a list of the files/paths that do exist under /sys/devices/

I have attached the support file as well.

Matt
File Attachment: 
kj6dzb
kj6dzb's picture
support data from  a
support data from  a NanoStation M2 1190-1272c57. 

I don't have a node to test this on, this week.
Support File Attachments: 
K6CCC
K6CCC's picture
Finally worked on the hAP
I finally had a little time to poke at K6CCC-hAP-Portable which had been updated to this 1907 version several times.  I plugged my laptop into port 2 of the hAP, and was a bit surprised to NOT get a DHCP address from the hAP.  Gave it time and retried several times without success.  Finally power cycled the hAP, and then I got an address.  Even directly connected, it would respond for a couple minutes and then fail for a while.  Sometime it would recover on it's own and sometimes I had to power cycle it.  One of the more interesting notes, was that the status page was showing LQ=100% , NLQ=0% on both of the other 2 GHz nodes in this room.  The other two nodes would sometimes show a connection, and sometimes not.  Pings between the nodes would be the same way - sometimes working and sometimes not.
Finally power cycled for about the 5th time and reloaded Nightly 1181.  Almost immediately connected solidly.  I have been pinging both directions for a while.  This computer (connected to the hAP-at-home) reported 2 dropped packets out of 1179 sent.
Off to bed...
 
AE6XE
AE6XE's picture
K6CCC, thanks.  I'll drill
K6CCC, thanks.  I'll drill into. I had ordered another hap ac lite to do tests on, and Amazon sent me a wrong model in a box with the right external label.   Another on a truck on the way.

Joe AE6XE
K6CCC
K6CCC's picture
I realized after I did it,

I realized after I did it, that I should have grabbed some more info before I reflashed it for 1181, but I was really tired...  If you need me to get some specific information, feel free to ask.  Should be able to get to it late tonight.
BTW, I reloaded 1181 because that is what I happened to have on the laptop :)
 

K6CCC
K6CCC's picture
Left this computer pinging
Left this computer pinging the hAP-Portable running 1181 all afternoon.  The path for this is:  PC > switch 1 > switch 2 > router 1 > switch 2 > switch 1 > hAP-at-home > hAP portable.  The same test to K6CCC-hAP-at-home (eliminating the RF path) is consistently <1ms.

Ping statistics for 10.32.72.9:
    Packets: Sent = 26365, Received = 26364, Lost = 1 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 1ms, Maximum = 3233ms, Average = 104ms
 
KE6WOB
CPE210 v2 test results

Hi -- I tested the TP-LINK CPE210 v2.8 with the latest image.  If I'm doing everything correctly, the image is not working properly.
I've loaded via the ftp method and then the webpage in the default application.  Both behave identically.  The image appears to load, but then upon reboot, the power light, and the LED furthest from power led come on.  No network activity.

It seems to me the next step would be to find and latch on to an ICSP header -- but not if someone knows something that I don't.  This is my first time tryying to load an AREDN image onto a device of any kind.

AE6XE
AE6XE's picture
KE7WOB,   To confirm, the
KE7WOB,   To confirm, the "latest image" is from these early build images in this forum post, link above.   If so, it appears the new flash chip is not defined for the ath79 target images.   I've added the cpe210v2 image to the other ar71xx folder in the link above.   Please give this ar71xx cpe10-v2 image a go.  This should be identically to the earlier and successful test.

Joe AE6XE

 
KE6WOB
Update

So -- I attempted to load "aredn-ae6xe-19.07test-2e35fec-cpe210-v2-sysupgrade.bin" from the UI -- and got the GUI error "Failure: The uploaded file is incorrect".
But -- I'm a troublemaker -- so I also tapped the UART interface.  and the device spits out "[utilities_error: swCheckFirm:303] image size is out of range"  I'm going to upload some files in a moment (either by editing this post or responding to it)

Edit:  Okay -- just uploaded the readout from the UART.
Default Firmware At Bootup uses default firmware.

Test Results Loading Image shows the readout while loading the ...test....factory.bin image
Test Results Bootup shows the readout while booting up after loading the factory.bin image.

The device shows those same LEDs -- no network LED and no activity on the RJ45 port.

Hope this is useful to you in some way.
Thanks!
Mark
 

AE6XE
AE6XE's picture
File:   DefaultFirmware
File:   DefaultFirmware_bootup_trial1.txt:  can't read this, looks like a binary file?

File:   DefaultFirmware_bootup_trial2.txt:  This is trying to boot some other firmware, I don't recognize it:​ Linux version 2.6.31 (jenkins@sohoiapbuild) (gcc version 4.3.3 (GCC) ) #1 PREEM

File:   TestFirmware_AtBootup.txt:   booting AREDN, don't see any errors, would need a data download to see more details (look for wifi SSID meshnode in first boot state if ethernet port is not responding?).  Booting:  ​Linux version 4.14.155 (joe@AE6XE-VM)

File:  TestFirmware_LoadingImage.txt:  looks like same as TestFirmware_AtBootup.txt plus it has the TFTP upload of recovery.bin.  I don't see any errors in this data ether, need support data download for more details.  Booting:  ​Linux version 4.14.155 (joe@AE6XE-VM)
 
KE6WOB
Testing Info
I can see it from the Wifi interface, and after a bit of tinkering, the wifi interface is up.  Thanks Joe!
KE6WOB
New CPE210s
AE6XE
I just got a couple refurbished nodes in -- They are listed as version 2.0 (whereas my original pair are 2.8).  Would it be of any benefit to try to load the images onto those?  Or should I wait until I hear from you?

 
AE6XE
AE6XE's picture
KE6WOB,   see overall status:
KE6WOB,   see overall status:  ​https://docs.google.com/spreadsheets/d/1IyDD_gPC4LLtZGERDQhS8Ix7EDiNjK12/edit#gid=1212766508

KK4ZUZ (Andrew is the contributor to get cpe devices working in AREDN) has tested cpe210 v2  using the "ar71xx" folder images.  We have this marked as working (the v2.8 is a different row, still needs confirmation or fixes).    You might repeat KK4ZUZ's test to confirm.

Joe AE6XE
AE6XE
AE6XE's picture
New hAP ac lite images
There is a fix to the hAP ac lite image.   Please proceed and give the device another try.   I am currently running with no issues so far. 

Joe AE6XE
K6CCC
K6CCC's picture
Initial test is good

Updated my hAP-Portable to:  AE6XE-19.07test-2e35fec
Connected right up to both 2GHz devices that it should have along with a DtD link.  We'll see what it looks like in the morning.
Before:


After:

 

K6CCC
K6CCC's picture
It's working
After about 12 hours, my hAP portable is working fine with the new release.
 
AE6XE
AE6XE's picture
release status update

The upcoming release of many new devices, bug fixes, and capabilities still has a few issues being worked out:

1) The openwrt release AREDN is based on, recently published another release candidate rc2, and there is a rc3 in the pipeline (this means we are still days/weeks away to see the openwrt official release).
2) There are 2 identified issues so far on Rocket M5 XW and Nanostation M5 XW devices still being worked.
3) Many models do not yet have any know testing to confirm all is working.

How can you help?   Check the status here:

https://drive.google.com/file/d/1IyDD_gPC4LLtZGERDQhS8Ix7EDiNjK12/view?usp=sharing

If you have a device that has not yet been tested, please down load the image and give a try here:

https://drive.google.com/drive/folders/1Pm466OqLPTARUuh_ux3xSoD2123H5EOw?usp=sharing

Joe AE6XE

nc8q
nc8q's picture
Nano loco M2 XW tested

Hi, Joe

I have 3 of the untested devices.

I tried uploading aredn-ae6xe-19.07-1272c57-ubnt_bullet-m-xw-sysupgrade.bin
with keep settings, 2x, device reboots to 1010-8177bab;
without keep settings, 2x, device reboots to 1010-8177bab.

Since this is a new kernel, should I be trying the 'factory.bin' via tftp ?

I have a NB-M9 XM to test next.

My PBE-M5-400 is 'in service'.
I'll test it last.

Chuck

AE6XE
AE6XE's picture
Loco M2 XW - the bullet-m-xw
Loco M2 XW - the bullet-m-xw image was from the "targets/ath79" directory?   If so, then revert to the targets/ar71xx directory and use the loco-m-xw image instead.     As long as AREDN is running, then always use the sysupgrade version to upload from the AREDN ui.  The factory flavor is only used with tftp  methods (and for UBNT from AirOS). 

Joe AE6XE
nc8q
nc8q's picture
Loco M2 XW - ar71xx loco-m-xw

"Loco M2 XW - the bullet-m-xw image was from the "targets/ath79" directory?"
Yes, I see no *bullet* in the /ar71xx/.
Trying this again with known working POE and cables. :-|

Tried 2x with  ath79/generic/aredn-ae6xe-19.07test-2e35fec-ubnt_bullet-m-xw-sysupgrade.bin.
Rebooted to 3.19.3.0.

Tried 2x with /ar71xx/generic/aredn-ae6xe-19.07-1272c57-ubnt-loco-m-xw-sysupgrade.bin
Worked the 2nd upload.
"!!!! UNTESTED HARDWARE !!!!".
Can only set part-15 frequencies. 2412 ... 2484.
Links with my Bullet-M2, hAP, and GL-ar300M.

Hope this helps,
Chuck

Image Attachments: 
AE6XE
AE6XE's picture
Chuck,  I've updated status
Chuck,  I've updated status on ar71xx local-m-xw .  This has the same issue as rocket xw and Nanostation xw devices.   Common issue to find resolution for all these devices.

Joe AE6XE
nc8q
nc8q's picture
Ubiquiti M9-NB

Loaded 19.07/ath79/generic/aredn-ae6xe-19.07-1272c57-ubnt_bullet-m-sysupgrade.bin
on the 1st attempt.

Image Attachments: 
Support File Attachments: 
AE6XE
AE6XE's picture
NC8Q,  Success with bullet M2
NC8Q,  Success with bullet M2 XM.  I'll mark it on the status .xls.  - Joe
 
w6bi
w6bi's picture
CPE210 v 1.0
Joe, my CPE210 v.10 is running with no issues on AE6XE-19.07-1272c57.  I think you can mark it as passed.

Orv W6BI
AE6XE
AE6XE's picture
Orv,  updated status .xls for
Orv,  updated status .xls for CPE210 v1.x   Got it.  

Joe AE6XE
kj6dzb
kj6dzb's picture
I noticed a bug in the
I noticed a bug in the current firmware (I was unable to get the br-wan interface to come up ) on the node. It want until I flashed the upgrade image, that the problem went away. 

besides the bug... the node is stable with the nighty build!!!

73 Mathison KJ6DZB
AE6XE
AE6XE's picture
The support data shows the
The support data shows the LAN interfaces are up and 100k+ packets of traffic:
 
br-lan    Link encap:Ethernet  HWaddr 80:2A:A8:A9:65:16  
          inet addr:10.134.81.97  Bcast:10.134.81.111  Mask:255.255.255.240
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:108697 errors:0 dropped:6 overruns:0 frame:0
          TX packets:165277 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:37374554 (35.6 MiB)  TX bytes:141572501 (135.0 MiB)

Maybe a vlan issue or need to renew IP on the computer?

Joe AE6XE
kj6dzb
kj6dzb's picture
The support data file is from
The support data file is from the post issue node. The br-wan interface failed to come up not br-lan. I would loke the support data from the buggy node, altho its a bug with the main firmware not the nighty build.
AE6XE
AE6XE's picture
Mathison,   To confirm, 
Mathison,   To confirm,  "main" or "current" firmware is the 19.07 test firmware images available at the link on this thread?   Can you confirm the device is a Nanostation Loco M5 XW?   if so, how new?  before ~Nov 2017 or after --  test/manufacture date on box?   It takes different images depending on when it was manufactured. 
kj6dzb
kj6dzb's picture
the node was originally

the node was originally loaded with ​3.19.3.0. It was a (loco-m-xw)​ LocoM5 US test date of 7/25/16. the br-wan interface failed to load. If I miss matched the firmware I was unaware of the mismatch. 

after I flashed the current nighty build 1234-dcdd217 (support data suppled) the bug went away. 


should I test another firmware? 

 

AE6XE
AE6XE's picture
Test only with the loco-m-xw
Test only with the loco-m-xw image under the ar71xx folder.   After loading, if something is not functioning, please upload a support data file.   This has been reported to show a yellow banner, a known issue consistent on ~3 XW devices.  The channel #s are not showing up in basic setup as the model # isn't determined correctly.  If you also have the br-wan failing, this would be a new symptom. 
 
Joe AE6XE
kj6dzb
kj6dzb's picture
 (RBLHG5HPnD) LHG HP5 
 (RBLHG5HPnD) LHG HP5 

I need to flash the node back to its original ​routeros-mipsbe-6.39.3.npk inorder to get ​aredn-1234-dcdd217-mikrotik-rb-nor-flash-16M-sysupgrade.bin to take. Other wise the node gets caught in a boot loop and has to be network loaded again. 

Ive attached the support file from the Pxe .elf load. ​​routeros-mipsbe-6.39.3.npk

If you interested I can push the newest ​ ​​routeros-mipsbe-XXXXX.npk  and supply support data in the next post....

 
Support File Attachments: 
kj6dzb
kj6dzb's picture
 (RBLHG5HPnD) LHG HP5 
 (RBLHG5HPnD) LHG HP5 

Loaded routeros-mipsbe-6.46.1.npk netinstall.

Loaded aredn-1234-dcdd217-mikrotik-vmlinux-initramfs.elf with PXE server

Loaded thru the web interface aredn-ae6xe-19.07test-2e35fec-mikrotik-rb-nor-flash-16M-sysupgrade.bin

Doesn't WORK I get group of 3 light flashing on and then off, then the usr led comes on for 30sec, then the sycle starts again. 



 
Support File Attachments: 
kj6dzb
kj6dzb's picture
 (RBLHG5HPnD) LHG HP5
 (RBLHG5HPnD) LHG HP5 mikrotik loaded with aredn-ae6xe-19.07test-2e35fec-mikrotik-rb-nor-flash-16M-sysupgrade.bin

Node doesn't connect over RF.... ?

 
Image Attachments: 
Support File Attachments: 
kj6dzb
kj6dzb's picture
Im reading thru the support
Im reading thru the support data to compair the 2 (RBLHG5HPnD) ... It looks like a mismatched of listed Frequencies on two identical devices on the 2 units I received! 

It looks like 5ghz CH 166->184 are all compatible across the devices. Im guessing that the AREDN firmware dosent check the frequencies avalible in the device? 
 
* 5830 MHz [166] (28.0 dBm)
* 5835 MHz [167] (28.0 dBm)
* 5840 MHz [168] (28.0 dBm)
* 5845 MHz [169] (28.0 dBm)
* 5850 MHz [170] (28.0 dBm)
* 5855 MHz [171] (28.0 dBm)
* 5860 MHz [172] (28.0 dBm)
* 5865 MHz [173] (28.0 dBm)
* 5870 MHz [174] (28.0 dBm)
* 5875 MHz [175] (28.0 dBm)
* 5880 MHz [176] (28.0 dBm)
* 5885 MHz [177] (28.0 dBm)
* 5890 MHz [178] (28.0 dBm)
* 5895 MHz [179] (28.0 dBm)
* 5900 MHz [180] (28.0 dBm)
* 5905 MHz [181] (28.0 dBm)
* 5910 MHz [182] (28.0 dBm)
* 5915 MHz [183] (28.0 dBm)
* 5920 MHz [184] (28.0 dBm)


Im supplying the support data for the node that supports ch 149.
AE6XE
AE6XE's picture
KJ6DZB,  
KJ6DZB,  

The command "iw phy phy0 info" does not show all the channels the AREDN firmware can be set to, e.g. 133.  We can see the channel the node is set to with the "iw dev wlan0 info" command.  Note, it always shows 20Mhz regardless and is blind to channel width settings:
 
========== iw dev wlan0 info ==========
Interface wlan0
ifindex 8
wdev 0x2
addr 64:d1:54:ad:e8:fb
ssid AREDN-10-v3
type IBSS
wiphy 0
channel 133 (5665 MHz), width: 20 MHz, center1: 5665 MHz

The nodes would need to be on the same channel to link:

KJ6DZB-LHGHP5:
config wifi-device 'radio0'
    option type 'mac80211'
    option disabled '0'
    option channel '133'

KJ6DZB-LHGHP5-test:
config wifi-device 'radio0'
    option type 'mac80211'
    option disabled '0'
    option channel '149'

Joe AE6XE
kj6dzb
kj6dzb's picture
AE6XE I had a bad pice of
AE6XE I had a bad pice of hardware... It's been returned!



 
KB9OIV
From this info on the Google
From this info on the Google Docs sheet:
ubnt: nanostation M2/M5 XW
ath79: Y: partial
Notes: no channels showing, only freqencies.  unable to retrieve boardid and shows model is being tested with yellow banner.  But everything is working otherwise. 


I have a Nanostation M5 XW.  What firmware was used to confirm ath79 works for M5 XW?  I haven't found one that works for my M5 XW for the pre-nightly.  Specifically, I don't see one for nanostation-m-xw.  Nothing in the ar71xx folder either.  I have tried:
 
aredn-ae6xe-19.07test-2e35fec-ubnt_nanostation-m-sysupgrade.bin​
aredn-ae6xe-19.07test-2e35fec-ubnt_bullet-m-xw-sysupgrade.bin​

out of the ath79 folder, but every time my device ends up back to the original firmware when I wait for the update to complete.
 

Pages

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer