Also with Huawei CloudEngine 6860-48S8CQ-EI and Huawei WTN 910D-A showing same result like you guys.

Is anyone can tell slow and unstable autoneg with XL710 already fixed?

IS-IS towards Cisco no longer works on ptp , only broadcast

Also with Huawei not working on p2p Level 2.

Hi, Per-VLAN L3HW counting is neither a bug nor an intended behavior. It is an unimplemented feature. Since traffic is routed via hardware, VLAN software counters are not updated (because the software does not see the traffic). We need to find a way to gather per-VLAN stats from the hardware to upd...

You missed this?
viewtopic.php?t=204932

Wow thanks!
Too busy checking RoS, thought Swos will never got update for hundred years.

Swos is very obsolete in my opinion, never got any updates. Try ROS latest and try disable negotiation. Im using molex, aoc, sr and lr without problem to XL710.

Dont forget for latest ROS, enable all 4 ports (its needed for auto negotiation).

Try to set nexthop force-self. I think its because multihop and the nexthop / gateway that your peer send to you not reachable in your routing table. In Huawei it can be set connect-interface or nexthop peer address. Or you can also set in your FILTER-IN additional set gw <some-reachable-gateway-ip>...

Thanks for your reply. As i understand this ways i can achieve ecmp. But actually what i need is to always route 10.0.0.0/24 traffic (both inbound and outbound) through ISP1 and 10.0.1.0/24 through ISP2 unless one of the peers fails. i assuming 10.0.0.0/24 is your downlink network owned by you and ...

Theres some ways to achieve that. 1. You can peer both isp without Route Rule and VRF, then edit your bgp peer connection with different router-id (will go with different instance) and you will see + sign on routing table showing multipath. 2. Also without vrf and Route Rule. You can add single dumm...

Couldn’t agree more

To MikroTik team :

Please allow redistribution of IS-IS routes from other routing protocol

And after IS-IS is complete, hopefully there will be EVPN and SR-MPLS with TI-LFA

And hopefully will be SRv6 also..

Please bring IS-IS ipv6 :(

Have you even tried?

Ofcourse not.
There is no changelog for IS-IS.
And still flaged as red here,
https://help.mikrotik.com/docs/display/ ... l+Overview

If its ready, at least tell us in changelog.

Please bring IS-IS ipv6 :(

Is it for ARM Ampere cpu? If yes, is it coming soon.

Edit cost on 5Ghz to more than default. 50, 100, 150, anything. Then it will flow to lower cost (60Ghz).

And some feature maybe, add bfd to interface 60Ghz and 5Ghz, then check ospf "use bfd" for faster failover than default dead timer ospf.

Try to connect switch first, then connect to 2 different port on your Router (different mac). But, because Mikrotik doesnt support BGP ECMP, it will flow upload only 1 port and download only 1 port too. Maybe same port, maybe not. Its just my opinion, maybe that ixp only accept 1 mac for 1 ip. Just ...

Hope IS-IS ipv6 come with 7.14 stable.

Is anyone trying IS-IS cli? is it good?

Try to using cpu affinity for dedicated CHR cpu and SR-IOV to bypass esxi kernel for using dedicated CHR NIC.

Virtualization always be a nightmare for networking if not using ideal configuration.

Use combine with filter in tag community and local preference will works. We also using it and transiting our cdn, ggc, ixp to our downstreams. Will happy to help you. Reach me out

Networking with 2 CPU(s) is bad. Because the pcb lane on pcie are directly to 1 CPU socket. If all cpu plugged but only 1 pcie, the traffic will flow to another cpu and go back again through a lane called QPI. QPI is natural enemy for networking, latency and jitter. Please read about Networking with...