SoftEther VPN User Forum

Hi Therem,
Need help- while trying to deploy soft-ether VPN in cluster mode using openvpn client , seems like I am getting error notsupported. Not sure if softether vpn does support openvpn 1194(tcp) in cluster mode ?
While I am workingwith standalone mode it works. Could you please help ? I ran "OpenVpnMakeConfig" command and it fails saying Unsupported.

Thanks
Suddhasil

No

Hi There/team,
I would like to know if softether vpn client does support open-source recursive DNS resolver ? It works with openvpn integration but would like to know if it supports softether client as well ?
Please let me know if anyone resolve this requirement with softether.

Thanks

Hi There,
While deploying 500 vpn tunnels to a server running on aws EC2, its based of ARM and having 16vCPU and 32 GB RAM and 50Gbps constant throughput. I see around 490 connection establish but still 10 conenction failed getting DHCP ( I am using dnsmaq) . these connection stabled after multiple try. And I see some connections having very slow throughput like 1Mbps, 2Mbps.
Is there any document or if anyone can suggest me any good configuration and OS tuning if required to get best performance ? Also I am using openvpn client and thats a hard limit our side. So I can not use cluster , please suggest ?

Thanks
Suddhasil

SE in L2 mode supports everything.

Consider this research viewtopic.php?f=7&t=68100&p=98117#p98117

Hi There and team,
If any one can guide me here please. I am setting up 300 tunnels to softethervpn server running on AWS/EC2(16vCPU , 32 GB RAM and 50Gbps network throughput). I am using local bridge networking and DNSMASQ for the DHCP ip allocation to client. What I have have been observing is , for 300 connection I get ip allocated for 280-290, now if I try to connect 500 connection i see around 470-490 getting connection established. I see lots of retry and disconnect with "auth failed" message but actually in the log its showing unable to get ip from dhcp.
Now I am using openvpn client to conenct to softether over tcp 1194. I am using openvpn because certain limitation in application /client side.

Now with this I am wondering if there is any tuning I need to perform ? Any thought please ?

This has been fixed, hence closing this thread. Just had to adjust parameter in dnsmasq config :

dhcp-rapid-commit # NEW - Fast DHCP
no-ping

Hi There,
Need help- I am curious to know if SE client identify local network's default gateway? so my goal is once I connected to SE vpn tunnel from my client, the local network traffic should go through local gateway and it will be identifiable automatically by the SE client, is that possible ? If i use openvpn client it works i have tested that , but I have large case requirement and may be need to use SE client instead oipenvpn since it does not support cluster.
Any help around this would be great and please let me know if any quesitons

Hi There,

I am setting up SE cluster for VPN architecture this will be on AWS and i will be having NLB in front and all tunnels will go to VPN /SE server via the NLB. My question is in cluster mode , do i need to send all traffic to SE vpn controller ? or just to the workers node and controller only will manage load distribution and sessions ? that means traffic flow would be " SE client --> AWS NLB--> SE CLuster controller --> SE vpn Server/s" ?
OR
"SE clients -->AWS NLB--> SE cluster worker nodes "?

Some document mentioned it would be the controller behind the loadbalancer, curious to know how the load will distribute then ? I am working on a very large scale implementation.
Please let me know if any questions

Thanks

suddhasilsarkar wrote: ↑

Mon Feb 23, 2026 9:05 pm

I am working on a very large scale implementation.

For business support contact SoftEther Corporation.

Thanks Solo ! will do. Do you have any recommendation on the actual ask? or any documentation I can follow ?

Hi ALl,
need help- I have implemented SE cluster connection distributed across multiple nodes . Now I emulate a failover scenario and terminated one of the node, what I found is session got rebalanced across multiple running nodes but they are not getting DHCP ip and seems like they are pointing to old dhcp ip ( which was pertaining to that terminated node).
I am using dnsmasq in local bridge setup . Trying to understand what I am missing here , could you please help?

Thanks

Setup your DHCP server on the LAN's PC/VM, not the terminated node nor any other node, of course.

Hi Team,
QQ: did anyone implemented HA of the controller node in SE cluster configuration ? If SE controller goes down the whole solution will be down, I see some blogs referring to build own solution for recovery/HA of the controller node, any suggestion on this would be great help