Closed edrikk closed 2 years ago
In my network everything is running smooth. Did you add the correct timeout value in your config?
config times
option con_timeout '60'
Yes I had noted that from the commit message and added:
config times option con_timeout '60' option update_client '10'
In case it makes any difference, my setup:
All devices are exhibiting the same behavior still this morning.
Same issue.
Devices:
Redmi-AX6
Error logs:
Fri Jul 22 19:36:02 2022 daemon.err dawn: client_to_server_state()=tcpsocket.c@106 eof!, pending: 0, total: 33816578 Fri Jul 22 19:36:02 2022 daemon.warn dawn: Connection to server closed Fri Jul 22 19:36:03 2022 daemon.err dawn: client_to_server_state()=tcpsocket.c@106 eof!, pending: 0, total: 33816578 Fri Jul 22 19:36:03 2022 daemon.warn dawn: Connection to server closed Fri Jul 22 19:36:07 2022 daemon.err dawn: client_notify_state()=tcpsocket.c@76 eof!, pending: 0, total: 0 Fri Jul 22 19:36:07 2022 daemon.warn dawn: Connection closed Fri Jul 22 19:36:16 2022 daemon.err dawn: client_to_server_state()=tcpsocket.c@106 eof!, pending: 0, total: 33816578 Fri Jul 22 19:36:16 2022 daemon.warn dawn: Connection to server closed
Proof:
AP-1
AP-2
Can you maybe go through the latest 3 commits and check which commit it breaks?
I guess it is because of this: https://github.com/berlin-open-wireless-lab/DAWN/blob/bb362db2facd8ce7a39c430a353b6413ec24d70d/src/network/tcpsocket.c#L108-L109
Not sure anymore why, but we close the connection if no more data is pending. I would say we just remove it.
Can you check https://github.com/berlin-open-wireless-lab/DAWN/pull/191 ?
Please provide the output of ubus call umdns browse
on each AP
Please provide the output of
ubus call umdns browse
on each AP
Here are the results of each APs. AP1-ubus.txt AP2-ubus.txt AP3-ubus.txt
Please provide the output of
ubus call umdns browse
on each APHere are the results of each APs. AP1-ubus.txt AP2-ubus.txt AP3-ubus.txt
looks good in umds
and try run lsof -ni :1026
on each AP
if no lsof
, try netstat -tpn | grep :1026
looks good in
umds
and try runlsof -ni :1026
on each AP if nolsof
, trynetstat -tpn | grep :1026
Note:
AP1 is 192.168.1.101 AP2 is 192.168.1.102 AP3 is 192.168.1.103
AP1
root@AP-LivingRoom-AX:~# netstat -tpn | grep :1026
tcp 0 0 192.168.1.101:1026 192.168.1.103:44736 TIME_WAIT -
tcp 0 0 192.168.1.101:1026 192.168.1.103:43784 TIME_WAIT -
tcp 0 0 192.168.1.101:1026 192.168.1.103:47638 ESTABLISHED 3694/dawn
tcp 0 0 192.168.1.101:1026 192.168.1.103:44134 TIME_WAIT -
tcp 0 0 192.168.1.101:1026 192.168.1.103:55462 TIME_WAIT -
tcp 68825 0 192.168.1.101:1026 192.168.1.102:51622 ESTABLISHED 3694/dawn
tcp 0 0 192.168.1.101:1026 192.168.1.103:37430 TIME_WAIT -
tcp 0 69057 192.168.1.101:44408 192.168.1.102:1026 ESTABLISHED 3694/dawn
tcp 0 96459 192.168.1.101:43318 192.168.1.103:1026 ESTABLISHED 3694/dawn
tcp 0 0 192.168.1.101:1026 192.168.1.103:41062 TIME_WAIT -
AP2
root@AP-Room1-AX:~# netstat -tpn | grep :1026
tcp 0 104571 192.168.1.102:51622 192.168.1.101:1026 ESTABLISHED 4294/dawn
tcp 71694 0 192.168.1.102:1026 192.168.1.103:47700 ESTABLISHED 4294/dawn
tcp 0 74461 192.168.1.102:49324 192.168.1.103:1026 ESTABLISHED 4294/dawn
tcp 72353 0 192.168.1.102:1026 192.168.1.101:44408 ESTABLISHED 4294/dawn
AP3
root@AP-Room2-AX:~# netstat -tpn | grep :1026
tcp 0 79538 192.168.1.103:47700 192.168.1.102:1026 ESTABLISHED 3765/dawn
tcp 72182 0 192.168.1.103:1026 192.168.1.101:43318 ESTABLISHED 3765/dawn
tcp 68461 0 192.168.1.103:1026 192.168.1.102:49324 ESTABLISHED 3765/dawn
@XiaoliChan and what about logread -e dawn
looks like you use diff version of dawn?
@XiaoliChan and what about
logread -e dawn
Lots of errors & warning messages.
likely one of your AP using old version dawn.
looks like you use diff version of dawn?
Nope, I used the newest commit of the main upstream.
likely one of your AP using old version dawn.
Every dawn on each APs is up to date.
likely one of your AP using old version dawn.
Every dawn on each APs is up to date.
or something like one of your dawn keep restart?
watch the pid of your dawn on each AP
pidof dawn
or something like one of your dawn keep restart?
Not sure what happened.
One thing I'm sure that is version "e596ff131735821684f7ecea73d7634733319f94" works fine.
your device is mt7621?
watch the pid of your dawn on each AP
your device is mt7621?
No, they are all Redmi AX6 which is working with robimarko's firmware.
Link: robimarko 5.15
try capture pcap on tcp port 1026 like
tcpdump -i br-lan -nvv tcp port 1026
enable dawn log to see more info:
config local
option loglevel '1'
try capture pcap on tcp port 1026 like
enable dawn log to see more info:
Here are the logs of each APs in log level 1.
dawn-moreinfo-ap1.txt dawn-moreinfo-ap2.txt dawn-moreinfo-ap3.txt
try #192
Issue solved by this commit.
Thanks for testing.
Confirmed as well. Closing issue. Thank you all!
I just installed snapshot version 20220721 of DAWN on Master Head.
I noticed that each AP (4 in total) after this upgrade results in the Network Overview page showing only the clients that are connected to that specific AP only. Clients from other APs (on same SSID) are not longer shown.
I noticed this after the 1st AP was upgraded, and watched the others as I upgraded them, and after each upgrade the page which previously showed Clients connected to each of the other APs then showed only Clients connected to itself.
Usually when I see this a restart of umdns followed by a restart of DAWN 'resolves' this; However this time it appears to be due to the recent changes.
I do not believe the July 18th commits caused it as that was the version I upgraded from. It could be ap_array_unlink_entry always returns NULL ?