That's safe mode afaik. That doesn't mount /overlay so you are back to default packages & settings. From there, you could mount /overlay, remove the default packages, install the newer mt76 ones and modprobe them, and see if that breaks your build.
If I am not wrong then that's U-boot and kernel doesn't have ssh so I am not sure how can interact with it, is there anyway to connect to it so that I can mount and test?
On the stripped down version (Snapshot) when i tried installing kmod-mt76x2 packages from the link mentioned by you, I got several dependencies error. For few of the package (common, core) I tried it installing dependent pack but still having many errors therefore I dropped it. I think i need to compile it which will require lot if time and I don't have much knowledge
Actually, I have tried all the possible builds - Snapshot, LEDE, full firmware with LUCI, few of them compiled by @julianocs but unfortunately none of them worked for me
I'm unable to reproduce your problem. I loaded custom-built snapshot r9071-27be78ef46 some weeks ago from master. 18.06.2 was released after this snapshot.
My WR841 v13 (192.168.2.1) hangs from a second one, so by pinging 192.168.1.1 it has to go trough WAN interface. After more than 1000 pings got 0 timeouts. My laptop connects to WR841 v13 via wifi, with good signal strength
Sorry for late reply, I got busy in office work and was doing some more testing.
First off all I have installed the TP-Link original firmware to confirm if there is any hardware issue, but on 3 days usage I didn't observe any packet drop and worked as expected.
Then I have flashed OpenWrt 4.14.98 with SNAPSHOT, r9293-e3311cb on 18.06.2 branch and as you might have guessed it came back to original packet drop issue.
One more I would like to mention is, earlier when I was using 18.06.1 then it was working but sometimes it will simply disconnect all WiFi clients so I wrote a script to restart every time client count was zero. But currently its not even working for me.
I can relate few issue with my router too but SSH connection never lost in my case, but if I ping google.com then in every 15 icmp request next 5 packets will be dropped and then again network will be restored.
Sorry, all I can say is -as before- that I'm using custom build of r9071-27be78ef46
Yesterday I noticed (as reported by twinkleLED) that open of web pages was very slow, DNS looked as stalled but then there was not ping at all. So I guess connections were lost. After restart, i got connection loss again in very short time. Before that, I have been working with that router for more than a week without noticing any problems.
I'm using snapshot r9258-420376a.
I think that the only issue that I can relate to is webpages not opening or taking too long in some cases (as if the DNS took more than the average time to respond). Other than that, it has worked reasonably well. No ping drops (in a 1000 ping sample), no ssh connection drop or problem establishing said connection.
To be fair, all the things that work well are using a wired connection, the webpage load lag or failure happens sometimes using wifi. Also, I'm not using the device as a router, rather as an AP.
Here's an example of a 48h ICMP monitoring session to a WiFi connected printer:
To test I am pump data to an upstream server over the wifi connection. It works fine for a while, and then there are periods when it stalls, as shown in the chart below.
Interestingly the tx retry rate is quite high. Below is from the router's prespective.
# iw dev wlan0 station dump
Station xx:yy:zz (on wlan0)
inactive time: 0 ms
rx bytes: 3387487657
rx packets: 2893513
tx bytes: 106736789
tx packets: 1433994
tx retries: 949590
tx failed: 3478
rx drop misc: 12
signal: -56 [-56, -58] dBm
signal avg: -55 [-55, -58] dBm
tx bitrate: 115.6 MBit/s MCS 13 short GI
rx bitrate: 115.6 MBit/s MCS 13 short GI
expected throughput: 51.360Mbps
authorized: yes
authenticated: yes
associated: yes
preamble: short
WMM/WME: yes
MFP: no
TDLS peer: no
DTIM period: 2
beacon interval:100
short preamble: yes
short slot time:yes
connected time: 762 seconds
What's the tx retry correlated with? tx packets? If it is, then it shows a 66% retry rate.
I am not carrying a test as intensive as this, but one example from normal usage is this:
Station xx:yy:zz (on wlan0)
inactive time: 3030 ms
rx bytes: 4182446
rx packets: 37507
tx bytes: 43979387
tx packets: 40300
tx retries: 21449
tx failed: 154
rx drop misc: 366
signal: -77 [-77, -78] dBm
signal avg: -77 [-77, -78] dBm
tx bitrate: 26.0 MBit/s MCS 3
rx bitrate: 6.0 MBit/s
expected throughput: 20.49Mbps
authorized: yes
authenticated: yes
associated: yes
preamble: short
WMM/WME: yes
MFP: no
TDLS peer: no
DTIM period: 2
beacon interval:100
short preamble: yes
short slot time:yes
connected time: 11259 seconds