Is anyone else seeing serious performance issues on a Mikrotik hAP when running nightly builds 1300 or 1307? I am. First noticed performance issues on my hAP with nightly 1300, but I think it is worse with nightly 1307. Let me describe my hap installation.
This is K6CCC-hAP-at-Home.local.mesh or 10.32.147.197 (primary address) or 10.4.158.41 (LAN address). I live in an AREDN RF desert so my only connection to the rest of the mesh is via tunnel at this time. All five LAN ports on the hAP are in use as follows:
1) Internet connection via switches and a router to my spectrum cable internet.
2) Connects to one of my routers by way of switches that allows LAN connection from my "normal" home LAN to the AREDN network.
3) Connects to a Grandstream 1625 IP phone
4) Connects to a Raspberry Pi-4 that mostly does nothing, but occasionally used to access the AREDN network.
5) DtD link to a dumb switch that also has another hAP and a Rocket M3.
Mesh RF is currently turned off, and the hAP is operating as a WiFi access point on 5 GHz, but there is nothing connected to it.
There is a Raspberry Pi-3 running MeshChat is connected to a switch in my data cabinet that gets to the hAP by way of port 2.
At this time, there is only one tunnel expected to be connected - and right now that is down due to a networking issue at the far end. My live tunnel to the rest of the network is coming out of the other hAP because of the pre 1300 vs post 1300 tunnel issue.
I am observing that I frequently can not access the rest of the AREDN network from my desktop PC, and even accessing the hAP is at times VERY slow or non-responsive. Yesterday the hAP would report that OLSR is not running when selecting "Mesh Status". Manually restarting OLSR (from the advanced Configuration page) did not change that. I have observed CPU loading from the Node status page as high as 16.xx, although 3 to 5 is more common.
A power cycle or reboot restores service for a few hours and then it starts getting bad again.
This is K6CCC-hAP-at-Home.local.mesh or 10.32.147.197 (primary address) or 10.4.158.41 (LAN address). I live in an AREDN RF desert so my only connection to the rest of the mesh is via tunnel at this time. All five LAN ports on the hAP are in use as follows:
1) Internet connection via switches and a router to my spectrum cable internet.
2) Connects to one of my routers by way of switches that allows LAN connection from my "normal" home LAN to the AREDN network.
3) Connects to a Grandstream 1625 IP phone
4) Connects to a Raspberry Pi-4 that mostly does nothing, but occasionally used to access the AREDN network.
5) DtD link to a dumb switch that also has another hAP and a Rocket M3.
Mesh RF is currently turned off, and the hAP is operating as a WiFi access point on 5 GHz, but there is nothing connected to it.
There is a Raspberry Pi-3 running MeshChat is connected to a switch in my data cabinet that gets to the hAP by way of port 2.
At this time, there is only one tunnel expected to be connected - and right now that is down due to a networking issue at the far end. My live tunnel to the rest of the network is coming out of the other hAP because of the pre 1300 vs post 1300 tunnel issue.
I am observing that I frequently can not access the rest of the AREDN network from my desktop PC, and even accessing the hAP is at times VERY slow or non-responsive. Yesterday the hAP would report that OLSR is not running when selecting "Mesh Status". Manually restarting OLSR (from the advanced Configuration page) did not change that. I have observed CPU loading from the Node status page as high as 16.xx, although 3 to 5 is more common.
A power cycle or reboot restores service for a few hours and then it starts getting bad again.
Here is a simplified extract of my LAN drawing as it applies to AREDN at this time.
If/When it gets sluggish, grab a support download file from the LAN. This will tell if it is a RAM issue.
Joe AE6XE
Joe AE6XE
With 5Ghz on with no wifi clients, 2 Ghz on the live mesh:
Hours Free
0:30 27584 KB
1:00 27708 KB
2:00 27904 KB
6:00 24372 KB
7:00 20980 KB
9:00 14664 KB
20:00 12728 KB
24:00 12508 KB
30:00 10816 KB
(unit crashed overnight)
Joe AE6XE
Joe AE6XE