Joined: 30 Jan 2015 Posts: 676 Location: Texas, USA
Posted: Thu Feb 15, 2018 8:03 Post subject:
labo wrote:
@Kong,
BTW, your build has the correct lime in the log file. However, BS build (this thread) has incorrect time in log. ( Router time is correct but not logger time)
Test write command to log file:
Written about the same time, R7800 shows correct clock in log.
R7800: (v3.0-r34900M kongat (02/12/18 ) )
Feb 13 20:26:00 d-r7k8 user.notice root: print time stamp in syslog
R9000: (v3.0-r34929 std (02/12/18 ) )
Feb 14 02:26:07 d-x10 user.notice root: print time stamp in syslog
@Kong Thanks for fixing this in R9000.
Your latest build DD-WRT v3.0-r35000M std (02/14/18 )
Now has the correct sys logger time. _________________ ASUS GT-BE98 PRO Main: Fiber 5gbps up/down
ASUS AXE16000: AI Mesh node
2 X ASUS RT-AX89X: AI Mesh nodes
QNAP QSW-1208-8C 12-Port 10GbE Switch
XS712T ProSafe 12-Port 10GbE Switch
3 X R9000 DD-WRT Mesh
I have spent some time looking at the mini-ntpclient.c code
I think that where the ntp problem occurring is
using rev 34895 line numbers
What is happening is that on line 87 when getaddrinfo is called result is a linked list of IPV4 and/or IPV6 addresses since on line 78 the ai_family was set to AF_UNSPEC (so the code does not care which ip version).
When line 96 is reached it is a for loop (because I am assuming the intent was to read through the list). When the "if" statements are reached (depending on if the ai_family is AF_INET or AF_INET6) there is a "break" that exits the loop. This means that only the first entry in the list is evaluated and copied. If this happens to be an AF_INET6 (as you know is IPV6) then that is what is used even if the user has ipv6 disabled. The DNS lookup does not care if it is IPv4 or IPv6 (for example do a nslookup on say google it returns IPv4 and IPv6 addresses).
Those of us that happen to select a server that can be resolved by either IPv4 or IPv6 seem to be drawing the IPv6 address as the first in the list. Obviously, if one is on IPv4, the IPv6 address will not work and the system logs an error and keeps repeating this bug.
There are multiple ways to fix this and since you guys maintain the code, I would rather you choose the method. If you need ideas though ask, I will suggest.
I hope this helps.....
One other note on line 79, it should not matter but since it is known as UDP why not set this to SOCK_DGRAM?
well, i'm in bootloop trying to upgrade to this version
Help
i tried 30-30-30
tried tftp-ing the factory image of this version, of the next one and of an old one
The router goes back to boot-loop
netgear 3700-v2
Help!
Edit: i can tftp the stock firmware and the router starts, but as soon as i put a ddwrt firmware (34411 to 350xx) it starts loop-booting again.
i can't understand this _________________ --
Netgear WNDR3700 v.2 - 26081
Tp-Link TL-WR841N v.9.2- 25934
Posted: Sun Feb 18, 2018 21:46 Post subject: TL-WR1043ND upnp
Router Model
TP-Link TL-WR1043ND v1.8
Firmware Version
DD-WRT v3.0-r34929M std (02/12/18 )
Kernel Version
Linux 3.10.108-dd #48865 Mon Feb 12 04:47:34 CET 2018 mips
i had to flash with tftp since the version 34080 broke the web flash functionality. So far my observations is that is usable. Some minor problems is that ntp receives the time but complain about it and upnp seems not to function properly.
update: via tftp
status: operational
errors: reports errors with ntp and upnp
Jan 1 02:00:12 dd syslog.info syslogd started: BusyBox v1.28.0
Jan 1 00:00:12 dd user.info : httpd : http daemon successfully stopped
Jan 1 00:00:12 dd user.info : httpd server shutdown
Jan 1 00:00:12 dd daemon.debug ntpclient[919]: Connecting to 2.pool.ntp.org [2001:678:3c8:1::226] ...
Jan 1 00:00:12 dd daemon.err ntpclient[919]: Failed connecting to 2.pool.ntp.org [2001:678:3c8:1::226]: Address family not supported by protocol
Jan 1 00:00:12 dd daemon.debug ntpclient[919]: Connecting to 212.18.3.19 [212.18.3.19] ...
Jan 1 00:00:12 dd user.info : httpd server shutdown
Jan 1 00:00:12 dd user.info : httpd : https daemon successfully started
Feb 18 13:39:12 dd daemon.info ntpclient[919]: Time set from 212.18.3.19 [212.18.3.19].
Feb 18 13:39:12 dd daemon.err process_monitor[918]: cyclic NTP Update success (servers 2.pool.ntp.org 212.18.3.19 88.99.174.22)
Feb 18 13:39:12 dd daemon.debug process_monitor[918]: Restarting cron (time sync change)
Feb 18 13:39:12 dd user.info : cron : cron daemon successfully stopped
Feb 18 13:39:12 dd user.info : httpd server started at port 80
but im logged in
.
.
.
Feb 18 13:40:00 dd user.info : Maybe upnp had died, we need to re-exec it
Feb 18 13:42:00 dd user.info : Maybe upnp had died, we need to re-exec it
Feb 18 13:44:00 dd user.info : Maybe upnp had died, we need to re-exec it
Feb 18 13:46:01 dd user.info : Maybe upnp had died, we need to re-exec it
Feb 18 13:48:00 dd user.info : Maybe upnp had died, we need to re-exec it
Are you sure open vpn works ? its not working for me ! i flashed and did reset everything to default then tried to setup open VPN (PIA) but im still getting the client connected, disconnected error
[quote="kroolas"][quote="kroolas"]TPLINK TL-WDR3600 v1.5
DD-WRT v3.0-r33772 std (11/16/17)
Linux 3.10.108 #43307
Status: OK
Reset: NO
Errors: None so far (I'll test OpenVPN later)
I keep getting these errors in log and also the time in the log file is incorrect.
Feb 13 08:30:48 d-x10 daemon.err ntpclient[2705]: Failed connecting to 2.pool.ntp.org [2607:fcd0:107:3ab::1]: Address family not supported by protocol
Is that all you get in your log?
Works fine here, it tries ipv6, i do not have ipv6, then it falls back to ipv4:
Feb 13 16:06:37 kong daemon.err ntpclient[18777]: Failed connecting to 2.pool.ntp.org [2a02:c207:2010:6775::2]: Address family not supported by protocol
Feb 13 16:06:37 kong daemon.debug ntpclient[18777]: Connecting to 212.18.3.19 [212.18.3.19] ...
Feb 13 16:06:37 kong daemon.info ntpclient[18777]: Time set from 212.18.3.19 [212.18.3.19].
Feb 13 16:06:37 kong daemon.err process_monitor[1744]: cyclic NTP Update success (servers 2.pool.ntp.org 212.18.3.19 88.99.174.22)
It doesn't falls back to ipv4 to me?
Here is how I set it up:
<snip>
It doesn't falls to ipv4:
<snip>
<snip>
If I use Ip adress instead then it works...
<snip>
but it keeps updateing on every hour?!
DD-WRT should only query an NTP server once a day or per boot. Even then NTP will typically adjust ddwrt's systems clock by a very small amount (often less than a second) to compensate for drift.
Maybe fallback from IPv6 to IPv4 depends on if you have one DNS server for IPv4 and IPv6 or separate ones. IMHO one is better.
I think dd-wrt queries DNS on your named NTP server(s) once at boot and just keeps the results to use later. That's not ideal since *.pool.ntp.org change to point to different servers regularly, and old results are not guaranteed to stay working, though mostly they do.
Yes syncing every hour is probably overkill for time that's just for the router itself. Maybe there are plans to enable a NTP server soon? If so PLEASE can we have proper ntpd instead of whatever weird thing it is at the moment.
But at least syncing every hour is better than what I believe it was before, namely once at boot, then once more 2 minutes later if not yet synced, then never again!!
Posted: Tue Feb 27, 2018 0:28 Post subject: Netgear WNDR3700v4
Router Model: WNDR3700v4
Firmware: DD-WRT v3.0-r34929M std (02/12/18)
Kernel Version: Linux 3.18.94 #7509 Mon Feb 12 06:09:29 CET 2018 mips
Status: Excellent build
Reset: No
Errors: No errors seen except usual minor stuff that never works like UI color preview, IPv6 QoS etc.
This is my main router with PPPoE connection to modem, wired and WiFI clients, guest WiFi network on separate subnet on br1 and an OpenVPN server running in the router.
Later: I had httpd crash after downloading via bit torrent and viewing /Status_Conntrack.asp
Even later: The red 'Reboot Router' button on /Management.asp does not seem to reboot the router!
Much later: Occasionally when my internet connection goes down the router will crash and reboot!!
