Help please: large project - several problems

]Hey there,

As I’m completly new to MikroTik, you can imagine that I’m overwhelmed with questions and troublehsooting.
I’m an Sysadmin doing my job since 2012 - but never had to go “deep” like this.

I’d love to take your proffesional advice, since I’m lost atm.

I’m stumbling since days through the RouterOS and trying to get even the basics to work.
My apologies if it looks like I’m dumb, and it’s embarrassing for me aswell, to not be able get things to work.

Here is the current setup, which will be extended by anouther draytek router for high availability.

TLDR:

  1. Draytek Router
  2. 2x CRS520-4XS-16XQ hooked up to same draytek (another will be implemented later)
  3. MLAG between all connections
  4. All servers have Broadcom P225p, 2x25Gbit/s, hooked up to MT_01 & MT_02
  5. Management Network untagged: 192.168.101.0/24 and VLAN 99 tagged
  6. Using the MikroTik XS+DA0003 and MikroTik XQ+BC0003-XS+ - QSFP28

What are the problems:

  1. Some Interfaces going up/down all 10 seconds
  2. Some Interfaces (QSFP) are not even going up, even if “module present”
  3. Only got Bonding_main running (2x 100gbit/s between CRS520), Rest of MLAG not working (802.3ad, layer3&4)
  4. DHCP on 192.168.101.0/24 not available behind MT_02, even if this not hooked up by mlag yet, only on ether1
  5. 4x LAG SPF to Zyxel 1920-48hpV2 not even connecting (auto negotiation turned off, since it can only handle 1gbit/s)
  6. Same for Zyxel GS1930-10, no Link between Mikrotik SFP and Zyxel SFP, going up/down even with auto-negotiation off

You would make me really happy, if you can assist me - I’m missing out on the basics here, I’m learning a lot at the moment - but I need your help, please :slight_smile:

[attachment=0]Screenshot 2024-09-05 112550.png[/attachment
Screenshot 2024-09-05 112550.png

Honestly at this level asking for help on a forum is not going to help you - you need a consultant with experience to look over the design, configuration and assist in troubleshooting issues.

My only suggestion would be to roll it back to the most simple parts and add the complexities 1-by-1 until it stops working, i.e. get rid of MLAG, get rid of bonding, try do single links between a device and a switch to ensure there’s not a compatibility issue etc, then try single switch bonding to make sure everyone happy with the LACP config, then add MLAG in etc.

Hey JoeGoldMan,

Thank you, for your reply!

Yes, you’re right. I’m going to fast into a system I dont know yet.

Gotta roll it back and then try to get the links up first, witohout any unneccesary complexity.