Hello everyone.

We met with very strange situation. We use Netgear 10G Base-T switches for NAS network (Storage > Switches > Servers > VMs). Now we start slow migration to Mikrotik SFP+ switches.
Physical:
- 2 Netgears connected one to each other and all servers\storages connected to each Netgear.
- we place new Mikrotik CRS-317 (latest firmware, SwOS mode, RSTP enabled) connect few servers and storage to it and connect to SFP+ port of one of the Netgear switch (not to both, just one, there is no physical loopback).
Logical:
- There is no VLAN, L3, ACL etc.extra settings configured on Netgear or Mikrotik. Just stupid L2 switch.

After Mikrotik was connected all ports start flapping and NFS datastores goes to Down\Innactive mode in vSphere. When Mikrotik was disconnected in few minutes almost everything goes to normal state.
Also we try to disable Flow control on Mikrotik and reconnect it - same situation.

Whats wrong we did or what we miss?
Could it be RSTP and Sw-Costs problem?
We have no idea for now.

Thanks in advance for answers.

Oh, 2 days and we concede.
We rebooted to RouterOS and now everything is ok.
As we understood SwOS is for very small and easy infrastructure. Just use normal RouterOS

Отправлено с моего SM-G930F через Tapatalk

Hello,
Did you notice actual interface disconnections on devices connected to CRS317 or were they just blocked connections?
Blocked ports with running link would be more common to RSTP, but actual disconnections may be related to some SFP compatibility issue.
Could you send screenshots from SwOS SFP menu or just supout.rif file from RouterOS to support@mikrotik.com for inspection?

Hello,

As it`s (Mikrotik) part of production system already we can`t reboot it to SwOS. In RouterOS it work`s perfect just like must work
Probably in couple of weeks we buy another CRS and could try to make some tests.
support file I`ll try to send later.
Thanks