I have a RB5009UG+S+ (RouterOS 7.15.1) connected to a standard OpenReach ONT and establishing a PPPoE session for my 550/75 FTTP package. This has all worked perfectly for 3+ years and I have seen solid 500Mbit+ download speeds. However, sometime last week (I think) max download speeds dropped to roughly 98-99 Mbits throughput on the WAN ethernet interface when running a test on speedtest.net from various wired ethernet clients on the LAN.

Given that suspicious Fast Ethernet-like speed, my first thought was cabling. However, Mikrotik reports a 1Gbps negotiated link on the WAN port. If I remove the WAN cable from the router and plug it into my Windows laptop directly, I can bring up a PPPoE connection and download at 500+MBits. Also, If I connect my laptop to the ONT end of the cable I can link at 1Gbps to the router.

Any suggestions welcome, as I'm not sure where to take this one next. The PPPoE debug logs (without packet-level info) are below, unfortunately I don't know enough about PPPoE to know if the PADO drops and LCP ConfReq state errors are significant. My sanitised config is attached.

---

16:45:35 pppoe,ppp,info pppoe-out1: initializing...
16:45:35 pppoe,ppp,info pppoe-out1: connecting...
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,ppp,debug pppoe-out1: LCP lowerup
16:45:35 pppoe,ppp,debug pppoe-out1: LCP open
16:45:35 pppoe,ppp,debug pppoe-out1: LCP opened
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:35 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:36 pppoe,ppp,debug pppoe-out1: LCP received ConfReq in opened state: restarting
16:45:36 pppoe,ppp,debug pppoe-out1: LCP closed
16:45:36 pppoe,ppp,debug pppoe-out1: LCP opened
16:45:36 pppoe,ppp,debug pppoe-out1: LCP received ConfReq in opened state: restarting
16:45:36 pppoe,ppp,debug pppoe-out1: LCP closed
16:45:36 pppoe,ppp,debug pppoe-out1: LCP opened
16:45:36 pppoe,ppp,info pppoe-out1: authenticated
16:45:36 pppoe,ppp,debug pppoe-out1: IPCP lowerup
16:45:36 pppoe,ppp,debug pppoe-out1: IPCP open
16:45:36 pppoe,ppp,debug pppoe-out1: IPV6CP lowerup
16:45:36 pppoe,ppp,debug pppoe-out1: IPV6CP open
16:45:36 pppoe,ppp,debug pppoe-out1: MPLSCP lowerup
16:45:36 pppoe,ppp,debug pppoe-out1: MPLSCP open
16:45:36 pppoe,ppp,debug pppoe-out1: BCP open
16:45:36 pppoe,ppp,debug pppoe-out1: CCP lowerup
16:45:36 pppoe,ppp,debug pppoe-out1: CCP open
16:45:36 pppoe,ppp,debug pppoe-out1: IPCP opened
16:45:36 pppoe,ppp,info pppoe-out1: connected
16:45:36 interface,info pppoe-out1 detect UNKNOWN
16:45:37 pppoe,ppp,debug pppoe-out1: IPV6CP timer
16:45:37 pppoe,ppp,debug pppoe-out1: IPV6CP opened
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:38 pppoe,debug ether1 (WAN): received PADO with unknown host-uniq, dropping
16:45:42 interface,info pppoe-out1 detect INTERNET
16:45:47 pppoe,ppp,debug pppoe-out1: MPLSCP timer

I have the same issue on a Hex S. Suddenly speeds have slowed to a crawl and bypassing the mikrotik yields results in line with the ISP line speed. Up dated firmware with marginal improvement but still under 50% of the 100Mbps

There is something very wrong with Your PPPoE connection. I'm running 7.14.3, and this is what I see from the router, on my remote log server:
It just goes like this, with nothing else. Sometimes (once every 20 - 40 days?) the connection drops and is reestablished. The it shows on the logs - but nothing like what you are getting.

I've (mostly) fixed it. MTU issue.

My ISP, IDNet, was claiming that nothing had changed at their end and that it must be failing hardware, so I remembered and dug out my old Hex S from the attic. This was my main router before the RB5009 and it turned out contained a 3 year old config on RouterOS 6.9 that still connected - a good opportunity to probe for hardware problems, OS regressions and configuration regressions.

It showed the same issues, with speed tests coming in just under 100Mbits, ruling out a hardware issue. However, when I reset the configuration and set up the PPoE WAN interface using WebFig's quick config, I got the expected 500MBps+.

The RB5009 ppoe-client configuration contained additional properties not in the fresh Hex S config: keepalive-timeout=60, max-mru=1480, max-mtu=1480, mrru=1600, service-name=internet. On removing these, actual MTU in the PPPoE status went down from 1520 to 1492, and download speeds were restored.

TBH I can't remember when/why I introduced these manually set properties - but I've been with IDNet for about 8 years and it was likely fairly early on in that period. Clearly some recent change in the PPPoE negotiation and/or routing path suddenly resulted in packet fragmentation and an >5x decrease in speed. I hadn't realised MTU problems could be that dramatic.

I am now seeing a log message about every 3 mins saying "invalid mtu 1500 on pppoe-out1 from fe80::262:ecff:fedb:c00". Advice welcome on whether I should tune the MTU up to 1500 or if it's safe to ignore it as per viewtopic.php?t=197367.

These messages are safe to ignore (it seems that IPv6 RA daemon on ISP side is not aware of PPPoE and it's lower MTU ... it's quite common or so it seems). You can simply stop logging them by setting
Make sure you run print command beforehand and verify that setting with index 2 is indeed the one with topics set to warning ...

You most probably should not increase MTU to 1500, not many ISPs support that over PPPoE connection.

Thanks. It looked similar to the forum answer I linked, but the MTU was order of magnitude bigger there so wanted to check it was still applicable.

IDNet have indeed separately recommended 1492. They're still claiming no change from their end, which is a bit hard to square with the results of the experiments above, but hey ho, it’s fixed now - even if I had to pull out most of my hair over the last few days to get there. I’ll pass on the possible RA misconfiguration to them.