You are here

3.19.3.0 / Powerbeam M5 going haywire

9 posts / 0 new
Last post
VA7NIC
VA7NIC's picture
3.19.3.0 / Powerbeam M5 going haywire

Has anyone around experienced any issues with PowerBeam M5's (in our case the 400's) and 3.19.3.0 where after a period of time, they go offline for no apparently reason? 

The one time we were able to get onto a unit after it went down via a dtdlink and using an IP address that was a bridge from the dtd device, any command ran via ssh resulted in an out of memory error.

If you don't catch it right away, you have to power cycle it to come back online.

We have had 4 of these units now go sideways since 3.19.3.0 came out.  We've backed down to 3.18.9.0 to see if that cures the issue.

I hadn't seen any traffic regarding this but I figure we can't be the only ones with this issue...

Anyone?

Nick
VA7NIC

 

K6AH
K6AH's picture
Force Ethernet Speed to 100Mbps

Out of the box these devices support 300Mbps throughput, so their Ethernet port runs 1000 Mbps.  You need to manually configure your node to not negotiate and force a 100Mbps connection. 

Andre, K6AH
 

AJ6GZ
See
VA7NIC
VA7NIC's picture
I had forced the EdgeRouter-X

I had forced the EdgeRouter-X its plugged into to run at 100Mbit since 3.17.something for a similar issue.    I've had 4 nodes now locally "croak" on 3.19.3.0, and I think they ran for the longest time OK using 3.18.

 

AE6XE
AE6XE's picture
I'm trying to recall, but I

I'm trying to recall, but I think the PBE M5 still needed the workaround, even though the switch it is plugged into was locked at 100Mbps.  (or at least it's not confirmed that only the switch side link rate settings can resolve the symptoms.)   

There's some guessing currently, if/when the device has symptoms in future, be sure to try and access through all methods,  LAN device, DtDLink, RF mesh and see if a support download can be obtained.   We could be off on a tangent and not know it...

Joe AE6XE

AJ6GZ
My testing

In my testing both the switch and node need to be set to 100/Full with autonegotiate off. The problem exists in the node but the switch must match.

The node will need the workaround reapplied after each upgrade/downgrade since the setting gets overwritten.

I haven't seen any lockup issues with 3.19.3 since it came out on the 400's here. Do they recover after an hour or so or are they down for the count? Checked PoE voltage, power supply? Are they all at the same site or different?

Has 3.18 mitigated the problem?

Ian

KL5T
I have 2 M5-300 units here in

I have 2 M5-300 units here in Anchorage and over the course of the last week or 10 days, each has lost its mind completely and I've had to physically restart each node.

AE6XE
AE6XE's picture
Do you have the workaround

Do you have the workaround applied on issue 165? 

AJ6GZ
300

Note: Haven't seen the need for the 165 workaround on the M5-300.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer