Multi-WAN gateway failover not switching back to tier 1 gw after back online

satadru

Just for clarification:

Default gateway switching is unchecked.
A single gateway group with Tier 1 gateway being highest priority, and Tier 2 being lower priority, and member down is the trigger.
Firewall rules use that gateway group.

And that works for failover? If you pull the cord for gateway 1 it switches to gateway 2? And if you reconnect gateway 1 it switches back?

matronyx

@satadru
Yes that is correct. The switch back and forth between tiers is fully automatic.

nleaudio

Although this is an older thread, I have the same issue happening with the very latest version, as of September, 2019. I have three WAN connections, and one of the gateways I have configured has two of the gateways on it. My PFSense will failover to my Tier 2 connection automatically; but when it comes back up, it will not go back to the Tier 1. I even tried clearing the states - no change. I tried changing the gateway set as Tier 2, and it just routed all the traffic thru that gateway, instead of the Tier 1. All gateways are up, and show as up.

What more can I do to debug this? I did not find the "Default Gateway Switching" option where indicated. Indeed, my "default" gateway is the Tier 1 gateway that seems not to be being used by the Gateway group.

My config is a bit complex, but I'm happy to try to debug this. Just need direction. Thanks.

Bob

gniting

I ended up writing a script and running it via cron to achieve the "switch." Yes, it is not elegant, but it gets the job done.

Here's what I have and I run this as a 5-minute cron job.

#!/bin/sh

# get active gateway and current time
CURRENT_TIME="$(date +"%c")"
CURRENT_GW="$(netstat -rn | grep default | awk '{print $4}')"

if [ $CURRENT_GW = "em2" ]; then
	#check if WAN1 is up or not
	WAN1_STATUS="$(pfSsh.php playback gatewaystatus brief | grep WANGW | awk '{print $2}')"
	if [ $WAN1_STATUS = "none" ]; then
		#WAN1 is back online, stop/start WAN2
		echo "$CURRENT_TIME: Bringing down WAN2"
		ifconfig em2 down
		echo "$CURRENT_TIME: Sleeping for 30s"
		sleep 30
		echo "$CURRENT_TIME: Bringing up WAN2"
		ifconfig em2 up
	else
		echo "$CURRENT_TIME: WAN1 is still down"
	fi
else
	echo "$CURRENT_TIME: Nothing to do!"
fi

C_C

Hey. Thanks @ibbetsion for the script.

Here is a slightly modified version that kills firewall states when there are connections remaining on WAN2 and WAN1 is back online.

Works great for my needs ( LTE failover ).

I set it as a cron, every minute:

*/1 * * * * /root/clear_state_back_from_failover_cron.sh >> /root/clear_state_back_from_failover_cron.log

I also checked "Flush all states when a gateway goes down" in System / Advanced / Miscellaneous.
The LTE gateway has monitoring disabled "Disable Gateway Monitoring" in System / Routing / Gateways. Otherwise states will be created on the interface and the script becomes wrong. Also, monitoring would consume data and I did not want that.

Code:

#!/bin/sh
# *** kills firewall states on failover WAN when WAN1 is up ***

WAN1_NAME="WAN_DHCP"
WAN2_IF=ue0
WAN2_GW_IP=192.168.3.1

CURRENT_TIME="$(date +"%c")"
WAN1_STATUS=`pfSsh.php playback gatewaystatus brief | grep "$WAN1_NAME" | awk '{print $2}'`

if [ "$WAN1_STATUS" = "none" ]; then
	# the following line may need to be tweaked depending on your needs
	WAN2_NSTATES=`pfctl -s state | grep "$WAN2_IF" | grep -v " -> $WAN2_GW_IP" | wc -l`
	if [ "$WAN2_NSTATES" -gt 0 ]; then
		echo "$CURRENT_TIME: WAN1 is online, but connections remain on $WAN2_IF. Killing states."
		pfctl -F state
	fi
else
	echo "$CURRENT_TIME: WAN1 is down"
fi

cmb991

I'm really surprised pfSense has nothing built in to handle this yet. This has been ongoing since 2017. In my case, my LTE modem (unlimited data) is still in gateway monitoring mode, so I'll be using @ibbetsion script. Thanks @ibbetsion

mo10

EDIT2
Issues not fixed. If i pull cable and put it right back in it will mess up Multi-WAN. Will not switch back correctly.

EDIT:
~~Resetting to defaults and setting everything up again seems to have fixed my issues.~~

Old:
I think i found the cause:

Seems Multi WAN is not working properly (or maybe dpinger) if the Interface goes down and back up (unplugging and re plugging).
In my tests i was just unplugging the cable on the WAN-Port.

I think the same happens with PPPOE or anytime the link is down and up again (physically).

This should not happen In my opinion. If modems reboot and so on: MultiWan would stop working.

I use "Paket Loss" as trigger level on Gateway-Group.

Would love to hear from you, thanks.

serbus

Hello!

I have several sites using multi wan and gateway groups with a mixture of static, dhcp, and pppoe. They all behave as expected.

Are you policy routing all of your WAN bound traffic?

"Defining gateway groups is only part of the story. Traffic must be assigned to these gateways using the Gateway setting on firewall rules."

https://docs.netgate.com/pfsense/en/latest/routing/multi-wan.html#firewall-rules

My experience is that you cant depend on the system routing table having your "preferred" (tier1) default route.

John

mo10

@serbus said in Multi-WAN gateway failover not switching back to tier 1 gw after back online:

Hello!

I have several sites using multi wan and gateway groups with a mixture of static, dhcp, and pppoe. They all behave as expected.

Are you policy routing all of your WAN bound traffic?

"Defining gateway groups is only part of the story. Traffic must be assigned to these gateways using the Gateway setting on firewall rules."

https://docs.netgate.com/pfsense/en/latest/routing/multi-wan.html#firewall-rules

My experience is that you cant depend on the system routing table having your "preferred" (tier1) default route.

John

Thank you,

strange is: if i don't unplug a cable on testing and switch off internet without pulling the cable, everything works just as expected. Every time. Soon as i unplug and replug i have to save interface settings for example to get switiching back to default Tier back working.

I checked almost every configuration before and nothing really helped.

serbus

Hello!

Ahhhh, gotcha.

I am having a problems following the thread. It is long and old, and seems to cover different (resolved?) problems. Yours could be yet another issue. Maybe a new thread?

John

nleaudio

As far as I know, this is still problematic! Some PFSense boxes I have on dual-wan setups will switch from tier 1 to tier 2 connections without issue, but going back when the tier 1 is restored does not always work... At least not in the timeframe I would consider usable.

Bob

mo10

@nleaudio said in Multi-WAN gateway failover not switching back to tier 1 gw after back online:

As far as I know, this is still problematic! Some PFSense boxes I have on dual-wan setups will switch from tier 1 to tier 2 connections without issue, but going back when the tier 1 is restored does not always work... At least not in the timeframe I would consider usable.

Bob

i think i had those issues because i imported a configuration to different hardware. Did you do the same?
After i did a reset to defaults and set up everything again it is now switching back fine.

gniting

@mo10 to be clear... in your dual-WAN setup, if WAN1 (default gateway) goes down and pfsense ends up making WAN2 the default, then upon recovery of WAN1, pfsense automatically marks WAN1 as default?

mo10

@ibbetsion

This was never a problem for me. It maked the gateway as default fine but still was sending traffic the wrong way. Saving an interface fixed it until i unplugged (physically) a cable again.
Now after resetting everything everything runs as expected.

Do you have problems with Multi-Wan? What exactly?

gniting

@mo10 my problem is that post recovery, WAN1 never goes back to being default. I have to use a script to bring down WAN2 so that WAN1 becomes default again. Not an ideal solution but it works.

mo10

@ibbetsion

Would you be able to make a test an reset you pfsense (save configuration first) and just setup the multi-wan an try again?

serbus

Hello!

Assuming a pretty standard multiwan: WAN1 -> tier1, WAN2 -> tier2, PREFWAN1/PREFWAN2/BALANCE gwgroups.

Whether you have states left open on WAN2 after WAN1 comes back up (sticky connections?) , or the default in the system routing table doesnt switch back to WAN1 after it recovers (make sure you dont have BALANCE as the default gateway), I believe the best approach is to policy route everything.

After WAN1 comes back, does traffic routed to a PREFWAN1 gwgroup still go out WAN2?

John

nleaudio

@mo10

Yes, it's quite possible that I moved the config to a new box.

And yes, when WAN1 comes back, new connections still go out WAN2.

It does appear to recover some time later though - maybe by the following day? I've not looked into it carefully.

Bob

mo10

@nleaudio

Are you using DHCP on the WANs or what are you using?

idiotzoo

I have what looks like the same problem. Gateway group with three gateways. When the tier1 goes down (packet loss) tier2 is used. When tier1 comes back, it does not get used and requires manual reconfigure or reboot. No changes I'm aware of to trigger this behaviour. No hints in the logs.