#Thread / matter - OBTR MeshForwarder events

1 messages · Page 1 of 1 (latest)

cobalt anvil
#

Gents,
trying to understand if the thread transport layer is sufficient or needs more routing devices....

Currently running with only one border router (OBTR in HA) and ZBT-1 stick dedicated for Thread (HA in proxmox VM). 12 Thread devices where five of them are routing capable (Eve outlet + Aqara T2).
Devices are stable and online in HA.

Can see a few CASE errors in the matter server logs each day...

What bothers me is the frequent events in the OBTR log.

Its either "ReassemblyTimeout" from HA interface towards a thread device:
00:04:41.222 [N] MeshForwarder-: Dropping (reassembly queue) IPv6 UDP msg, len:1058, chksum:11a9, ecn:no, sec:yes, error:ReassemblyTimeout, prio:normal, rss:-81.0, radio:15.4
00:04:41.222 [N] MeshForwarder-: src:[fd34:cb0c:cbe1:1:4126:1745:5267:d4c3]:5540
00:04:41.222 [N] MeshForwarder-: dst:[fd34:cb0c:cbe1:1:8c67:60bd:44e0:b412]:60421

Or "failed" when src is a thread device towards the HA interface:
00:05:30.660 [N] MeshForwarder-: Failed to send IPv6 UDP msg, len:646, chksum:d4dc, ecn:no, to:0xdc00, sec:yes, error:NoAck, prio:low, radio:15.4
00:05:30.660 [N] MeshForwarder-: src:[fd34:cb0c:cbe1:1:8c67:60bd:44e0:b412]:60421
00:05:30.660 [N] MeshForwarder-: dst:[fd34:cb0c:cbe1:1:a194:4bee:827e:dcb8]:5540

I assume this would be transport layer issues? But is it considered normal to see 10-15 MeshForwarder events per minute... or would that indicate issues with the transport layer?

OBTR
BR ULA prefix: fd34:cb0c:cbe1::/48 (loaded)

ip a
19: wpan0: <POINTOPOINT,MULTICAST,NOARP,UP,LOWER_UP> mtu 1280 qdisc fq_codel state UNKNOWN qlen 500
link/[65534]
...
inet6 fd34:cb0c:cbe1:1:8c67:60bd:44e0:b412/64 scope global flags 02
valid_lft forever preferred_lft forever
...
inet6 fd93:6edb:e93:b19f:0:ff:fe00:fc38/64 scope global deprecated flags 02
valid_lft forever preferred_lft 0sec
inet6 fe80::203f:a741:110f:c741/64 scope link flags 02
valid_lft forever preferred_lft forever

wide dew
#

A couple of CASE errors a day is unavoidable, as long as your devices stay responsive and not go unavailable. If the errors however are more often and you see lot of timeouts, that may indicate either RF issues (so mesh issues) or issues with mdns.

#

The frequent events in OTBR log are also pretty much standard, depending on the log level. Its rather noisy by default

cobalt anvil
#

Thank you, will investigate further.

#

If adding another TBR (for an example HA green with another obtr instance ) should you avoid mixing sticks by using a SLZB-06M with the ZBT-1(usb connection)?

spice breach
#

at this point, i wouldn't recommend using a SLZB-06M for thread (unless you're using it via usb connection)

#

the actual radio hardware itself doesn't seem to make much difference.