JAX1337
Regular Contributor
tried everything from RTRmon to formatting usb to reinstalling WF script , still seeing the constant reapplying of WAN in debug logs . All that is left is complete factory reset . anything else that I'm not aware off ?
WANFailover Dual WAN Failover ***v2 Release***
- Thread starter Ranger802004
- Start date
Ranger802004
Very Senior Member
I hate to tell you to do that but that may be the only option, I have the same exact set up as yours with the AX-11000 and develop this script on it so something else is going on here. What other scripts are you running?
BikeHelmet
Regular Contributor
Shaw (Canadian Cable ISP) has done it 3 times in a decade. (Lost the route to the entire Google IP space, breaking most of the internet for the majority of a day each time.) And now that Rogers owns them (one of the worst ISPs in Canada), their network reliability is likely to decline.
As a Canadian, I'd guess that the odds of this feature being useful are roughly 100% within 5 years of today. Typically when Google is blocked, other providers like CloudFlare are still accessible. Having both 8.8.8.8 and 1.0.0.1 would be pretty ironclad. Just food for thought.
One of my ISPs sporadically doesn't initialize from Cold Standby without a modem reboot. If it's in hot standby, then months later one line can go down, and it fails over properly. I too don't see the reason... I am glad that the script initializes WAN1 before WAN0, to get it into Hot Standby mode.
I was encountering this on the latest version - v2.0.6 - on my RT-AX56U router running the latest Merlin 388.2_2 firmware. But it ended up being bad options on the Dual WAN page - Ping and DNS Query were enabled. That's because it didn't warn me adequately during install. See below.
May I ask how I change the assigned DNS for the fallback WAN connection? Mine seems stuck on automatic ISP DNS? I'd prefer to specifically assign it.
I'm also wondering about QOS - despite it being enabled and having aggressive low limits (4mbit, since my fallback line is 5mbit), speedtests come in at 6mbit, the sync rate without QOS enabled. That makes me think that QOS isn't enabling properly, despite the values being changed in the proper fields. Maybe FlexQOS needs to be restarted or something? I did wait 5 minutes before testing.
The install flashed 3 or 4 red lines past for about 200ms, then skipped back to the menu. I saw errors, but no idea what they were. Reading this thread helped me to figure out that it was Ping and DNS Query being ticked. If there's a timeout/pause there, maybe add 10-30 seconds on so that there's time to read. My router thought it wanted to proceed RIGHT NOW, lol. Whatever the timeout is, my RT-AX56U did not respect it. I read quickly and played through Super Meat Boy, but that red text was gone in an instant.
I just tested the failover a few times by unplugging one modem, and I must concur - the failover happens extremely quickly compared to ASUS's Dual WAN. Very nicely done.
Ranger802004
Very Senior Member
I understand it's within the realm of possibility and does happen occasionally and it is an extremely rare occurence, back to my point though the issue you stated is caused by your ISP and not a Google outage so that would technically would be a WAN issue and not a Google issue. I think having 2 monitor targets per WAN connection is a little excessive for these routers. I'm not saying never but it's not a priority to try and make work as it would require some serious overhaul. The router will always boot with WAN1 as the primary ISP, it uses the factory failback to switch it to WAN0 which it's only validation is a physically plugged in connection vs the intricacy of WAN Failover script actually monitoring both interfaces. When it does that it puts the passive WAN interface in cold standby (AKA stopped via service). The DNS Query option being enabled was recently added under the latest beta and the as far as warning messages it logs them in your System Log for later review as well as every time WAN failover starts up it will check these things and log them for you so there is a way to see what is going on. You would change it in the Router GUI under WAN > Internet Connection > (Select WAN1 Physical Interface in drop down menu) > WAN DNS Settings. It is possible for your throughput to exceed the max setting set by QoS, my experience with FlexQoS in the past has been mixed results and unreliable so I had to forfeit use of it all together.
JohnSmith
Senior Member
On my AX88U, you go under WAN | Internet Connection | WAN Type (My choices are WAN(WAN0) or Ethernet LAN(WAN1), then under "WAN DNS Setting", click the "Assign" button, which gives you options of a "DNS List", which at the bottom is "Manual Setting", which you can assign IP addresses for DNS Server1 and DNS Server2
BikeHelmet
Regular Contributor
Thank you for the detailed responses! I understand. 
Makes sense, not a priority - and it probably is better to fallback to the other connection when Shaw does stupid stuff.
Ahh, I had missed that assign button - although oddly enough, even if I assign it, some of the time it doesn't "stick" and I'll see it on ISP DNS servers later... weird. Would using identical DNS servers for both WAN connections screw things up by having conflicting routes during failover, or something like that? Or is that all handled gracefully? I didn't notice any weirdness loading websites - as mentioned, it fails over very quickly. Yank a cable, run back to computer, everything is already loading fine.
I very much appreciate all the work that you've done to enhance the capabilities of Dual-WAN Failover on these routers. You've done a fabulous job!
I just found out that my RT-AX56U on 388.2_2 is EOL as of last month. But I'm happy to use your script as long as I can on this version. Let me know if any testing needs to happen on future betas, to ensure compatibility.
Makes sense, not a priority - and it probably is better to fallback to the other connection when Shaw does stupid stuff.Ahh, I had missed that assign button - although oddly enough, even if I assign it, some of the time it doesn't "stick" and I'll see it on ISP DNS servers later... weird. Would using identical DNS servers for both WAN connections screw things up by having conflicting routes during failover, or something like that? Or is that all handled gracefully? I didn't notice any weirdness loading websites - as mentioned, it fails over very quickly. Yank a cable, run back to computer, everything is already loading fine.
I very much appreciate all the work that you've done to enhance the capabilities of Dual-WAN Failover on these routers. You've done a fabulous job!
I just found out that my RT-AX56U on 388.2_2 is EOL as of last month. But I'm happy to use your script as long as I can on this version. Let me know if any testing needs to happen on future betas, to ensure compatibility.
Ranger802004
Very Senior Member
It can cause some conflicts during failover when reassigning DNS servers but probably would be fine, it definitely will cause route conflicts for the routes created by firmware to each DNS Server and same if you use one of your assigned DNS Servers as a Target IP for one of the WANs.Thank you for the detailed responses! I understand.
Ahh, I had missed that assign button - although oddly enough, even if I assign it, some of the time it doesn't "stick" and I'll see it on ISP DNS servers later... weird. Would using identical DNS servers for both WAN connections screw things up by having conflicting routes during failover, or something like that? Or is that all handled gracefully? I didn't notice any weirdness loading websites - as mentioned, it fails over very quickly. Yank a cable, run back to computer, everything is already loading fine.
I very much appreciate all the work that you've done to enhance the capabilities of Dual-WAN Failover on these routers. You've done a fabulous job!
I just found out that my RT-AX56U on 388.2_2 is EOL as of last month. But I'm happy to use your script as long as I can on this version. Let me know if any testing needs to happen on future betas, to ensure compatibility.
BikeHelmet
Regular Contributor
Okay. Well Google and CloudFlare only have so many IPv4 and IPv6 DNS servers. Maybe I should cut it down to one each, and avoid overlap with the target IPs? In your opinion, should the target IPs be different for different WANs? So that each one has a route out a different WAN connection? That's how I currently have it set up.
Oh - I have the DNS over TLS override set to strict - do you know if that takes precedent, or are the Assigned DNS servers somehow senior to that, and get used by the router itself? Maybe assigning them doesn't matter with DNS over TLS enabled and in that mode?
I appreciate all the extra insights into how this all functions!
Ranger802004
Very Senior Member
***v2.0.7-beta4 Release***
Enhancements:
- Added metric values to IP Routes created for target IPs.
- Added additional debug logging to WAN Switch function.
- Added 386.12 to supported firmware list.
- Minor optimizations to increase performance.
- Added CRLF argument to email.
- Added restart option to Status Console.
- Major performance optimization for NVRAM Check function.
- Parent PID is now displayed on Status Console with Dev Mode enabled.
- Added error message if an invalid run argument is specified.
Fixes:
- Minor visual bug when WAN Failover kill command is being executed.
- WAN Failover will go to disabled state now if DNS Query or Failback are checked under Dual WAN Settings.
- Fixed issue causing PID File not to be deleted under /var/run/wan-failover.pid
- Failover will now properly timeout when the 30 second timeout timer has been reached.
Installation:
- Warnings for DNS Query or Failback being enabled will now alert and log during installation
Deprecated:
- WAN0 Route Table and WAN1 Route Table configuration options have been deprecated and are now pulled directly from the Route Table file.
Enhancements:
- Added metric values to IP Routes created for target IPs.
- Added additional debug logging to WAN Switch function.
- Added 386.12 to supported firmware list.
- Minor optimizations to increase performance.
- Added CRLF argument to email.
- Added restart option to Status Console.
- Major performance optimization for NVRAM Check function.
- Parent PID is now displayed on Status Console with Dev Mode enabled.
- Added error message if an invalid run argument is specified.
Fixes:
- Minor visual bug when WAN Failover kill command is being executed.
- WAN Failover will go to disabled state now if DNS Query or Failback are checked under Dual WAN Settings.
- Fixed issue causing PID File not to be deleted under /var/run/wan-failover.pid
- Failover will now properly timeout when the 30 second timeout timer has been reached.
Installation:
- Warnings for DNS Query or Failback being enabled will now alert and log during installation
Deprecated:
- WAN0 Route Table and WAN1 Route Table configuration options have been deprecated and are now pulled directly from the Route Table file.
Ranger802004
Very Senior Member
Yes, target IPs should be different, it is possible for them to be the same but some routers have issues with the IP rules configured this way and WAN Failover will compensate for that during status checks and if it has to fall back to creating a route for the target IP instead of an IP Rule specific to the WAN interface and WAN Route Table then a conflict can occur if the IP is used for another service. As far as your DNS questions I am not the most knowledgeable on the topic and would recommend to ask in the open forum about that question. WAN Failover only updates the resolv.conf file and nvram variables for DNS during failover events.Okay. Well Google and CloudFlare only have so many IPv4 and IPv6 DNS servers. Maybe I should cut it down to one each, and avoid overlap with the target IPs? In your opinion, should the target IPs be different for different WANs? So that each one has a route out a different WAN connection? That's how I currently have it set up.
Oh - I have the DNS over TLS override set to strict - do you know if that takes precedent, or are the Assigned DNS servers somehow senior to that, and get used by the router itself? Maybe assigning them doesn't matter with DNS over TLS enabled and in that mode?
I appreciate all the extra insights into how this all functions!
BikeHelmet
Regular Contributor
Okay, will do - much appreciated.
One final question - now that I've got your improved Dual-WAN going, I see that I'm getting lots of disconnect email alerts. I hadn't noticed anything particularly bad - maybe the occasional lag spurt in a game - but I'm getting quite a few of these every day since installing... like 4+
One of these followed by the second one 2 minutes later.
Code:
***WAN Failover Notification***
----------------------------------------------------------------------------------------
Hostname: xxxxxx.xxxxx.org
Event Time: Sep 20 20:24:58
WAN0 Status: DISCONNECTED
WAN1 Status: CONNECTED
Active ISP: TekSavvy Solutions, Inc.
Primary WAN: wan1
WAN IPv4 Address: xxx.xxx.xxx.xx
WAN Gateway IP Address: xxx.xxx.xxx.x
WAN Interface: eth4
DNS Server 1: 1.0.0.1
DNS Server 2: 8.8.4.4
QoS Status: Enabled
QoS Mode: Manual Settings
QoS Download Bandwidth: 4Mbps
QoS Upload Bandwidth: 448Kbps
QoS WAN Packet Overhead: 0
----------------------------------------------------------------------------------------
Code:
***WAN Failover Notification***
----------------------------------------------------------------------------------------
Hostname: xxxxxx.xxxxx.org
Event Time: Sep 20 20:26:06
WAN0 Status: CONNECTED
WAN1 Status: CONNECTED
Active ISP: TELUS Communications Inc.
Primary WAN: wan0
WAN IPv4 Address: xx.xxx.xxx.xxx
WAN Gateway IP Address: xx.xxx.xxx.x
WAN Interface: eth0
DNS Server 1: 1.0.0.1
DNS Server 2: 8.8.4.4
QoS Status: Enabled
QoS Mode: Manual Settings
QoS Download Bandwidth: 44Mbps
QoS Upload Bandwidth: 11Mbps
QoS WAN Packet Overhead: 0
----------------------------------------------------------------------------------------The line did have some issues a few months back, and got swapped by Telus to a slower alternate one. I am just wondering if my line is really so bad that 4 times per day it fails over, or if perhaps I have something tuned improperly, and need to adjust settings?
Currently both lines read 12ms ping 0% packet loss, targets 8.8.8.8 and 1.1.1.1...
Ranger802004
Very Senior Member
That means it had 100% packet loss trying to ping the total number of ping count you specified in the config with a timeout of whatever you specified in seconds in the config, so something prevented the traffic somewhere between your router and your ISP.
BikeHelmet
Regular Contributor
I'll bet on line noise. I can't test line noise, but given what the Telus tech said...
I have updated to your beta and set it to require 6 pings. I'll report back in a few days on whether it has quieted down, or is still doing it. Thanks.
Hi ...
Once the Asus Dual-WAN failover doesn't work properly, I'm planning to try your (@Ranger802004) V2 Dual WAN Failover Script, but first I'd like to confirm a couple of things is possibile.
Is this Dual Wan configuration OK in order to use your script ?
Is the AC88Urunning the 386.12 firmware version supported ?
Edit: Meanwhile found that latest beta version is required to support 386.12.
Does the script write the operating/debug messages into the router syslog or does it use some private log file ?
Edit: Looks like the messages goes into the router syslog and there's some special debug mode where a temporary file may be used as well.
I do have a couple of router equipment's (one fiber and the other coax) from my ISP's, both working on bridged mode.
Any configuration recommendation for such a scenario, or the default one should handle it fine ?
Thanks in advance.
Once the Asus Dual-WAN failover doesn't work properly, I'm planning to try your (@Ranger802004) V2 Dual WAN Failover Script, but first I'd like to confirm a couple of things is possibile.
Is this Dual Wan configuration OK in order to use your script ?
Is the AC88U
Edit: Meanwhile found that latest beta version is required to support 386.12.
Does the script write the operating/debug messages into the router syslog or does it use some private log file ?
Edit: Looks like the messages goes into the router syslog and there's some special debug mode where a temporary file may be used as well.
I do have a couple of router equipment's (one fiber and the other coax) from my ISP's, both working on bridged mode.
Any configuration recommendation for such a scenario, or the default one should handle it fine ?
Thanks in advance.
Last edited:
Meanwhile I gave it a try.
A few moments after being started, the script decided to fail over to wan1, but as far as I could see wan0 was working OK all the time.
May you please tell me what should I lookup for on the syslog to find out the reason why the fail over to wan1 occurred ?
I do have a 1Gbit symmetric fiber connection (wan0) while on coaxial connection (wan1) I have 200 Mbit download/ 20 Mbps upload.
Obviously I'd like that the fiber connection has an higher priority. I saw that configuration uses the default value of 100 for both.
To prioritize the fiber connection, should I lower/increase the priority value ?
Thanks.
Ranger802004
Very Senior Member
1 yes that is the correct options to turn off factory failover to use my script.
2. 386.12 should be fine, the support is just officially listed but probably can work with older versions of the script too.
3. WAN Failover writes to the system log with debug logging only showing up if you enable debug logging in the GUI.
Ranger802004
Very Senior Member
Capture the debug logs and let me take a look. If you run the capture command it will write just wan failover logs to a temporary file in /tmp/
Hi @Ranger802004 ...
Thanks for answering.
My main log setup is as follows:
Once the wan-failover scripts is enabled via cron, I see a lot of messages in there prefixed as follows:
"Sep 24 09:40:04 wan-failover: Debug".
May you please confirm that this is the expected behavior ?
As for the test, I just re-enabled the wan-failover cron job, restarted the router and captured the log during some time.
From what I've seen, it looks like an endless loop where the wan-failover script keeps failing over bettween wan0 and wan1.
Probably because the wan0 interface state is not being detected correctly I guess.
Thanks for your support.
Thanks for answering.
My main log setup is as follows:
Once the wan-failover scripts is enabled via cron, I see a lot of messages in there prefixed as follows:
"Sep 24 09:40:04 wan-failover: Debug".
May you please confirm that this is the expected behavior ?
As for the test, I just re-enabled the wan-failover cron job, restarted the router and captured the log during some time.
From what I've seen, it looks like an endless loop where the wan-failover script keeps failing over bettween wan0 and wan1.
Probably because the wan0 interface state is not being detected correctly I guess.
Thanks for your support.
Last edited:
Ranger802004
Very Senior Member
1. Based on the logs it appears you have some kind of custom version of iproute2 installed, perhaps from Entware? I would uninstall this package as I'm seeing a lot of errors with it trying to delete, recreate routes, and etc., the version installed with 386.12 should be 5.11.0
Code:
Sep 24 09:30:48 wan-failover: Debug - IPRoute Version: ss1502102. I would use a different set of Target IPs that are different than your DNS Servers shared by both of your WAN interfaces.
EDIT:
Run the output of this command for me:
Code:
/usr/sbin/ip -V
Last edited:
1. I do not have Entware nor any other addons installed; just your script and a couple of scripts self made for DDNS and local backup.
I've a couple of Asus routers running the 386.12 firmware version and both report the same output for the command /usr/sbin/ip -V :
ip utility, iproute2-ss150210
2. Considering that both of my wan connections are setup with Cloudfare DNS, would this values be ok as Target IP's ?
3. # /usr/sbin/ip -V
ip utility, iproute2-ss150210
# ls -g /usr/sbin/ip*
Code:
-rwxr-xr-x 1 root 111392 Sep 4 17:10 /usr/sbin/ipset
-rwxr-xr-x 1 root 7534 Sep 4 17:10 /usr/sbin/ipsec
-rwxr-xr-x 1 root 273716 Sep 4 17:09 /usr/sbin/ip
lrwxrwxrwx 1 root 13 Sep 4 17:09 /usr/sbin/ip6tables -> xtables-multi
lrwxrwxrwx 1 root 13 Sep 4 17:09 /usr/sbin/ip6tables-restore -> xtables-multi
lrwxrwxrwx 1 root 13 Sep 4 17:09 /usr/sbin/ip6tables-save -> xtables-multi
lrwxrwxrwx 1 root 13 Sep 4 17:09 /usr/sbin/iptables -> xtables-multi
lrwxrwxrwx 1 root 13 Sep 4 17:09 /usr/sbin/iptables-restore -> xtables-multi
lrwxrwxrwx 1 root 13 Sep 4 17:09 /usr/sbin/iptables-save -> xtables-multi
Last edited:
Similar threads
DomainVPNRouting
Domain VPN Routing
Similar threads
Similar threads
-
-
-
Unbound DHCP DNS same as WAN DNS based on merlin-dns-monitor.sh
- Started by salvo
- Replies: 0
-
-
DDNS does not update dyndns.org when WAN IP changes
- Started by jbodnar
- Replies: 11
-
-
spdMerlin WAN speed is correct...VPN is very low!!- Started by rwheaton1
- Replies: 20
-
wan-failover (2.0.5) showing 100% packet loss, but not failing over
- Started by monorailmedic
- Replies: 0
-
Latest threads
-
-
-
So much for VPNs - Except when used for Android devices
- Started by L&LD
- Replies: 5
-
Cant host Friday the 13th game lobby
- Started by Johnnyproduce
- Replies: 5
-
Sign Up For SNBForums Daily Digest
Get an update of what's new every day delivered to your mailbox. Sign up here!