Aimesh goes down daily or every few hours

[Morris]

Can I 'reopen' this discussion to ask if a particular port in a node can hinder speed or otherwise go south? One of the nodes has been connected to a local router for wired devices in one room. When I plug the router into port 1 of the node, I get speeds <100 Mbps. When I plug it into port 4 in the node, I get >300 Mbps. Is this a known issue, has it happened before, and any way to fix? My Aimesh router has 386.7_2 and all 4 nodes have latest stock fw at this point.

This sounds like a hardware issue. What happens if you plug into port 2 and 3? Port one may be bad or you may have a cable issue.

 

[ldeveraux]

This sounds like a hardware issue. What happens if you plug into port 2 and 3? Port one may be bad or you may have a cable issue.

It seems to get progressively worse as I move from 4 to 1. I'm getting:
4 - >300
3 - 250
2 - 230
1 - 220
So I am getting 220 on port 1 at this point, a far cry from port 4. Weird!

 

[Morris]

It seems to get progressively worse as I move from 4 to 1. I'm getting:
4 - >300
3 - 250
2 - 230
1 - 220
So I am getting 220 on port 1 at this point, a far cry from port 4. Weird!

Try replacing jumper cables. If a run in the wall, move the nodes together and try with known good jumper cable. If jumper proves all ports good, re-terminate the cable in the wall one end at a time. Still bad, replace outlet one at a time. There are cable testing tools that could make this easier yet there expensive.

 

[ldeveraux]

Try replacing jumper cables. If a run in the wall, move the nodes together and try with known good jumper cable. If jumper proves all ports good, re-terminate the cable in the wall one end at a time. Still bad, replace outlet one at a time. There are cable testing tools that could make this easier yet there expensive.

But this is the same cable. I literally moved it from port to port on the router, then ran a test. If the cable was bad, it would be consistently bad across all port, no?

 

[Morris]

But this is the same cable. I literally moved it from port to port on the router, then ran a test. If the cable was bad, it would be consistently bad across all port, no?

There are many reasons all related to tolerances. Best to test and identify the source.

 

[ldeveraux]

There are many reasons all related to tolerances. Best to test and identify the source.

I'm not even sure what you're suggesting to do, considering certain ports on a node seem to be wonky. I tested the one cable I'm using with a cable tester, which passes. Again, all my nodes are wirelessly connected to the router; it's not the placement of the node. I can move the node around the room and still get the same speeds I've shown above.

 

[Morris]

I'm not even sure what you're suggesting to do, considering certain ports on a node seem to be wonky. I tested the one cable I'm using with a cable tester, which passes. Again, all my nodes are wirelessly connected to the router; it's not the placement of the node. I can move the node around the room and still get the same speeds I've shown above.

if your cable tester only tests continuity it is important yet there are many other factors including and not limited to cross talk, capacitance, resistance, echo etc. The minimum cost is about $600 for such a tester by a reputable brand such as the Fluke Networks MS2-100.

Move your node to the router. Use a known good jumper cable and test again on each port. What are the speeds you get? If you don't have a known good jumper, then try two and see if the results are the same.

 

[ldeveraux]

If a run in the wall, move the nodes together

Move your node to the router.

What do these even mean?

 

[Morris]

What do these even mean?

The router is the device that runs your AI Mesh. The nodes are the devices that attach to it

 

[ldeveraux]

The router is the device that runs your AI Mesh. The nodes are the devices that attach to it

I know that. What do your suggestions mean? What do you mean 'move the node to the router'? You want me to move the node physically closer to the router? What does that have to do with port performance on the same node?

What does 'If a run in the wall, move the nodes together' mean? Same as above, move the nodes closer together for some reason?

My question is about the ports on the back of the nodes and why they might have different speeds. It has nothing to do with the communication between nodes and router.

 

[Morris]

I know that. What do your suggestions mean? What do you mean 'move the node to the router'?

Yes

You want me to move the node physically closer to the router?

Yes

What does that have to do with port performance on the same node?

It lets you test your uplink

What does 'If a run in the wall, move the nodes together' mean?

Type of uplink cable

Same as above, move the nodes closer together for some reason?

My question is about the ports on the back of the nodes and why they might have different speeds. It has nothing to do with the communication between nodes and router.

Possibly it has noting to do with the uplink, possibly it dose. Do these tests and you find out. You need to use a process of elimination. If you don't want to do these tests, and feel that what I'm suggesting is a waste of time, then replace the node as you seem to be declaring it the issue. Have you done a full reset of the node?

 

[ldeveraux]

OK I give up. I'll just assume that some ports perform worse than others and call it a hardware issue.

 

[OzarkEdge]

OK I give up. I'll just assume that some ports perform worse than others and call it a hardware issue.

Are you speed testing with a fast PC wired directly to the node LAN, not through some other router (disconnect that other router)?

OE

 

[ldeveraux]

Are you speed testing with a fast PC wired directly to the node LAN, not through some other router (disconnect that other router)?

OE

I've tested with a laptop hardwired directly to the node, as well as a desktop connected through a router, then the node. Both are relatively current builds, same results each way.

 

[OzarkEdge]

I've tested with a laptop hardwired directly to the node, as well as a desktop connected through a router, then the node. Both are relatively current builds, same results each way.

Wire the suspect node to the AiMesh router (move it close enough), enable Ethernet Backhaul Mode to disable all wireless backhauls, turn OFF the other nodes, and then speed test that node's ports again with a fast wired PC.

OE