OpenThread
-
|
We have a network with three BRs which all register the NAT64 route. The NAT64 route of one of the BRs now disappears within a few hours of restarting it (its still listed in ot-br-posix commit: d77ab33 BR 6060BR 6092BR 6061 (this ones netdata is lost)Network setup script#!/usr/bin/env bash
# Thread network setup extracted from HandleFormNetworkRequest():
# https://github.com/openthread/ot-br-posix/blob/main/src/web/web-service/wpan_service.cpp#L184
set -eEuo pipefail
# When ot-ctl returns anything other than the string "Done" it failed
run_check_ot-ctl() {
echo $1
ot-ctl $1 | grep -z "^Done[[:cntrl:]]*$"
}
# see FactoryReset() at
# https://github.com/openthread/ot-br-posix/blob/main/src/web/web-service/ot_client.cpp#L300
echo "performing factory reset"
ot-ctl factoryreset 2>/dev/null || true
sleep 4
ot-ctl version | grep -q "OPENTHREAD" || (echo "factory reset failed"; exit -1)
eui64=$(cat /sys/class/net/eth0/address | sed -E -e 's/://g' -e 's/(.{6})/\1ffff/')
run_check_ot-ctl "extaddr $eui64"
run_check_ot-ctl "txpower 8"
run_check_ot-ctl "networkname igi-ot-net"
run_check_ot-ctl "extpanid dada1234affe5678"
run_check_ot-ctl "panid 0xdada"
run_check_ot-ctl "networkkey fc1cbb6e94fb7590615f0cd66ee8ff9f"
run_check_ot-ctl "channel 23"
run_check_ot-ctl "pskc -p 123456"
run_check_ot-ctl "ifconfig up"
run_check_ot-ctl "thread start"
run_check_ot-ctl "prefix add fd01::/64 pasor"
run_check_ot-ctl "route add 64:ff9b::/96 s med"
run_check_ot-ctl "netdata register"
exit 0 |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 10 replies
-
|
@neye-zhaw , can you provide more information on your network topology? How many router-capable devices do you have? How many devices are active routers? One thing I noticed is that BR 6061 is a child (its RLOC16 is |
Beta Was this translation helpful? Give feedback.
-
|
@neye-zhaw , can you provide the output of:
Thanks. |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
|
Thanks @neye-zhaw. Can you please also paste One thing I noticed is that there seems to be multiple instances of same route in local network data. I do not expect this to be possible though. The question is how the device ended up in such a state (apparently having multiple of same route prefix I think this may be the root cause. If I recall correctly, the leader code integrating network data will reject it if the requested new network data from a node has duplicate entries. So that is why the registrations from this node are not accepted and not seen in partition (non-local) network data. |
Beta Was this translation helpful? Give feedback.
-
|
I agree that the duplicate prefixes are problematic. I actually submitting a fix for this issue several days ago in #7653. @neye-zhaw , can you see if #7653 helps address your issue? |
Beta Was this translation helpful? Give feedback.
-
netdata show localnetdata show local -xnetdata shownetdata show -xThanks for the answer. It will be some weeks/months before I can test that fix in that network. |
Beta Was this translation helpful? Give feedback.
-
|
I have been testing ot-br-posix commit e19e60b for over a month now. Seems to be munch better. |
Beta Was this translation helpful? Give feedback.
-
|
Great to hear! Thanks for the update. |
Beta Was this translation helpful? Give feedback.
I agree that the duplicate prefixes are problematic. I actually submitting a fix for this issue several days ago in #7653. @neye-zhaw , can you see if #7653 helps address your issue?