Site A (Leased Line)<--- pptp --->(FTTx; PPPoE) Site B
RB 4011 is at site B. The Traffic usage is only about 200Mbps but it consumes 100% on cpu0 and slow down overall traffic.

Any suggestion?

yes, right click cpu0, profile, sort, screenshot, upload here.

yes, right click cpu0, profile, sort, screenshot, upload here.

Edit: sorry. profile wrong cpu. will post the right one later when it occurs again.

I may be wrong here but does pptp spread over cores or is it only one core for all traffic?

I doubt anybody will fix PPTP even if there is some bug.

yes, right click cpu0, profile, sort, screenshot, upload here.

any suggestion?

Are you natting at your RB4011?

I assume that this only happens during high traffic over the pptp connection?

I maintain the idea that this is they way pptp works. The interface will only run over one CPU core and that is the effect you see now. I can be wrong but think this is the way it work. So not a bug but by design.

Still think it is wrong then do a supout file during the issue and send to support with a explanation of the issue.

Are you natting at your RB4011?

Yes. RB4011 dial PPPoE (FTTx) and NAT all client to Internet IP for accessing Internet. But no natting for traffic in vpn tunnel.

I assume that this only happens during high traffic over the pptp connection?

I maintain the idea that this is they way pptp works. The interface will only run over one CPU core and that is the effect you see now. I can be wrong but think this is the way it work. So not a bug but by design.

Still think it is wrong then do a supout file during the issue and send to support with a explanation of the issue.

Thank you. I'm planning to change vpn connection to wireguard ... and will check the result on high traffic again.

I assume that this only happens during high traffic over the pptp connection?

I maintain the idea that this is they way pptp works. The interface will only run over one CPU core and that is the effect you see now. I can be wrong but think this is the way it work. So not a bug but by design.

Still think it is wrong then do a supout file during the issue and send to support with a explanation of the issue.

Thank you. I'm planning to change vpn connection to wireguard ... and will check the result on high traffic again.

Very interested to see the effect of this.

Thank you. I'm planning to change vpn connection to wireguard ... and will check the result on high traffic again.

Very interested to see the effect of this.

Wireguard Tunnel between MikroTik must be specific ip (endpoint) on both side?

My situation is
MikroTik at Site A: has fixed public ip address.
MikroTik at Site B: has dynamic public ip address.

Still finding the way to do it without specific Public IP at site B (Treat Site B like road warier).

No, it's not mandatory to specify an endpoint for both sides.

No, it's not mandatory to specify an endpoint for both sides.

Thank you. Able to do the tunnel now.

just specific endpoint at site B and leave endpoint blank at site A.
and use persistence handshake at site B.

Will wait to see the CPU usage when high traffic.

Will wait to see the CPU usage when high traffic.

From Max Bandwidth 190 Mbps with PPTP ... to 490 Mbps with Wireguard.

Hmm, I think RB4011 can do a little more than that, unless you have a funky config.
What MTU did you set for the Wireguard interfaces?
And what MTU do you have on your WAN at BOTH sides (tested)?

Hmm, I think RB4011 can do a little more than that, unless you have a funky config.
What MTU did you set for the Wireguard interfaces?
And what MTU do you have on your WAN at BOTH sides (tested)?

I thought so.. My Internet b/w is 1Gbps but can use about 500Mbps with RB4011

I did not set the MTU .. just leave it to default value (1420) for both side on wireguard.

Site A: MikroTik CHR is directed connect to ISP with static IP (Interface MTU is 1500) ... cpu only about 10% on b/w 500Mbps
Site B: RB4011 connect Internet with PPPoE and default MTU (1492) ... firewall was only 2 rows. and no L7 protocol.

Anyway. I'm satisfied with the current result. 500Mbps is better than 200Mbps as before.