[Fork] Asuswrt-Merlin 374.43 LTS releases (Archive)

[john9527]

There is no 192.168.20.31 28:ef:01:4b:e3:4a either in the log.

I'm at a loss.....28:ef:01 is listed as a 'Private' mac address, so no clue on what device it might be.

The only device I know of that exhibits that behavior is Apple TV, when connected via wireless DHCP. Connecting via ethernet or setting a static ip on the Apple TV fixes it.

 

[il2]

I'm at a loss.....28:ef:01 is listed as a 'Private' mac address, so no clue on what device it might be.

The only device I know of that exhibits that behavior is Apple TV, when connected via wireless DHCP. Connecting via ethernet or setting a static ip on the Apple TV fixes it.

I've never seen neither 28:ef:01:4b:e3:4a, nor 192.168.20.31 given at my LAN. Though it is possible to get it from DHCP range .31-.50.

 

[ColinTaylor]

There is no 192.168.20.31 28:ef:01:4b:e3:4a either in the log.

This post suggests it could be an Amazon Kindle. (I seems to remember my wife's old Kindle had a strange MAC address as well)

 

[john9527]

This post suggests it could be an Amazon Kindle. (I seems to remember my wife's old Kindle had a strange MAC address as well)

Nice google-fu
I was aware of the Fire TV 192.168.49.1 address (for casting), but hadn't heard of this one.

 

[ColinTaylor]

This post suggests it could be an Amazon Kindle. (I seems to remember my wife's old Kindle had a strange MAC address as well)

Confirmed: Dredging through my backups (from 2013) I see that our original Kindle Keyboard had 28:EF:01:29:4C:6C

 

[Gopnique]

pptp vpn server is not working on ac68u running 31E6, 30E3 and 30E2 firmware. The client connects successfully, but has access to LAN only, can not resolve any internet site, router log is full with "compressor dropped pkt" messages. Had to downgrade to 29e2, which works fine.

 

[il2]

Confirmed: Dredging through my backups (from 2013) I see that our original Kindle Keyboard had 28:EF:01:29:4C:6C

Indeed that was Kindle Keyboard 3G plugged in just recently, I didn't know its WiFi is ON. Is it possible that it caused DNS havoc?
Thanks ColinTaylor for clarification!

 

[john9527]

Indeed that was Kindle Keyboard 3G plugged in just recently, I didn't know its WiFi is ON. Is it possible that it caused DNS havoc?
Thanks ColinTaylor for clarification!

Well, it was causing havoc with the DHCP portion of dnsmasq which also provides DNS . The sequence is right, it just appears as if the Kindle never accepts the address provided by the router and just kept looping requesting an address.

 

[john9527]

pptp vpn server is not working on ac68u running 31E6, 30E3 and 30E2 firmware. The client connects successfully, but has access to LAN only, can not resolve any internet site, router log is full with "compressor dropped pkt" messages. Had to downgrade to 29e2, which works fine.

I had turned on CTF for PPTP connections starting with V30. Looks like some piece may be missing. I'll take another look and disable it for the next release if I can't track it down.

 

[oso2276]

Well, it was causing havoc with the DHCP portion of dnsmasq which also provides DNS . The sequence is right, it just appears as if the Kindle never accepts the address provided by the router and just kept looping requesting an address.

Try using an static dhcp address for the kindle


Enviado desde mi iPhone utilizando Tapatalk

 

[ml70]

Extremely high sirq use, AC66 fully utilized at 100 Mbps, previously load only 10-20%. Upgraded from 23E to 30E and then 31L, now router is fully used at only 100 Mbps line speed, 0-5% idle.
Did factory reset and input all settings again manually. Tried without customs scripts. No help. Hardware acceleration is on.

At net ingress 80 Mbps, egress 40 Mbps,

$ watch -n1 -d cat /proc/interrupts
4:   68721224            MIPS  eth0
# ^this grows by ~10,000 per second

Mem: 75912K used, 163652K free, 0K shrd, 5996K buff, 27972K cached
CPU:  0.6% usr  2.1% sys  0.0% nic  3.6% idle  0.0% io  1.3% irq 92.1% sirq
Load average: 1.38 1.46 1.21 5/63 2870
  PID  PPID USER     STAT   VSZ %VSZ CPU %CPU COMMAND
 1306     1 admin    R     4408  1.8   0 11.9 httpd
  521     1 admin    S     1376  0.5   0 11.3 protect_srv
 2870   606 admin    R     1468  0.6   0  9.3 top
 1245     1 admin    S      868  0.3   0  3.3 miniupnpd -f /etc/upnp/config
  552     1 admin    S     2796  1.1   0  2.6 watchdog02
  515     1 admin    S     2796  1.1   0  1.6 /sbin/wanduck
    3     2 admin    SWN      0  0.0   0  0.8 [ksoftirqd/0]

I don't know anything about soft irq's but googling shows there should be some kind of relationship between ksoftirqd and sirq%. Not much correlation here!

Only thing i haven't tried is going back to 23E, there's been too much downtime already due to 2 fw upgrades and the manual settings input.
Edit: should have mentioned wan internet is with pppoe.

 

[il2]

dnscrypt running fine on 30L2!

That was too soon. dnsmasq/dnscrypt stopped resolving on 30L2. I forgot to mention that the only setting that was changed after upgrading from 29L2 to 31L6 was forcing master browser. I didn't change it when downgrading to 30L2 and the problem with dnsmasq/dnscrypt persisted. I've turned it off now and switched dnscrypt from cisco OpenDNS to cisco OpenDNS familyshield for a change.
I'll let you know if trouble happens again.

 

[john9527]

That was too soon. dnsmasq/dnscrypt stopped resolving on 30L2. I forgot to mention that the only setting that was changed after upgrading from 29L2 to 31L6 was forcing master browser. I didn't change it when downgrading to 30L2 and the problem with dnsmasq/dnscrypt persisted. I've turned it off now and switched dnscrypt from cisco OpenDNS to cisco OpenDNS familyshield for a change.
I'll let you know if trouble happens again.

Don't know about this one....I run with dnscrypt active all the time for both my VPN and non-VPN clients. I do have two servers defined, as the dnscrypt servers do sometimes go up and down (look for unable to retrieve certificate messages in the syslog).

 

[il2]

Don't know about this one....I run with dnscrypt active all the time for both my VPN and non-VPN clients.

Do you have force master browser set ON?
Right now I suspect this setting interfere with the DNS resolution on the router. It's very strange to me, that it may happen quick or after quite some time after reboot ~12 hours. And sometimes web pages are accessible after retries. John, may I send you router's log if it happens again?

I do have two servers defined, as the dnscrypt servers do sometimes go up and down (look for unable to retrieve certificate messages in the syslog).

I'll try with both dnscrypt servers enabled.

 

[il2]

dnscrypt servers do sometimes go up and down (look for unable to retrieve certificate messages in the syslog).

No such messages in the log. They were in the past when router didn't have time set correctly on dnscrypt service start (from entware), but hwclock service resolved this.

 

[Gravityz]

He John,

i turned log DHCP queries on some time ago.
no i noticed that turning it of does not stop the enties in the logfile.(is still logs the DHCP entries)

i selected NO and hit apply.

 

[john9527]

Do you have force master browser set ON?
Right now I suspect this setting interfere with the DNS resolution on the router. It's very strange to me, that it may happen quick or after quite some time after reboot ~12 hours. And sometimes web pages are accessible after retries. John, may I send you router's log if it happens again?

Yes, both Master Browser and WINS server set to on.

Sure, be happy to look at a log.

 

[john9527]

He John,

i turned log DHCP queries on some time ago.
no i noticed that turning it of does not stop the enties in the logfile.(is still logs the DHCP entries)

i selected NO and hit apply.

I can turn off the messages without any problem.
Have you modified any of the other syslog option log levels?

 

[john9527]

@il2

You can also run the following command logged in to the router
/usr/bin/killall -s USR1 dnsmasq

This will dump the dnsmasq stats into the syslog. A 1-2% failure rate on each server is normal (these are primarily tracking sites that set an invalid TTL). Anything much higher, and I'd look to select a different server.

 

[il2]

After /usr/bin/killall -s USR1 dnsmasq:

Mar 27 23:43:22 dnsmasq[552]: time 89079
Mar 27 23:43:22 dnsmasq[552]: cache size 1500, 0/19 cache insertions re-used unexpired cache entries.
Mar 27 23:43:22 dnsmasq[552]: queries forwarded 11209, queries answered locally 390
Mar 27 23:43:22 dnsmasq[552]: DNSSEC memory in use 88, max 88, allocated 149996
Mar 27 23:43:22 dnsmasq[552]: server 127.0.0.1#65053: queries sent 11209, retried or failed 3872

with

# cat /etc/dnscrypt-proxy1.conf
Daemonize yes
LocalAddress 127.0.0.1:65053
LogLevel 6
ResolversList /jffs/etc/dnscrypt-resolvers.csv
ResolverName cisco-familyshield
SyslogPrefix dnscrypt-proxy1
PidFile /var/run/dnscrypt-proxy1.pid
IgnoreTimestamps no
BlockIPv6 no

Btw, applying settings after enabling both dnscrypt servers doesn't add new server option to /etc/dnsmasq.conf
Reboot is still needed to rebuild dnsmasq.conf