Edit: it turns out I the two test PC’s were firewalled, everything appears to be working thus far. I’ll leave this post in case someone else finds it useful.
Hi, this is definitely a “read the manual” issue but I’m seeking help getting a compass where to start. Having a problem and having no idea what to throw at it, and working with a new platform is proving to be a time sponge. The requirements follow, the concept is simple, connect three CRS switches in a ring configuration with all ports able to talk to any other port (any port as good as any other, just three separate places) but I’m experiencing trouble making it work, it’s unclear if my assumptions how it should work, my configuration or my test method is flawed.
What I’m doing is:
-3 CRS226 switches with dual SFP cages.
-Each switch has a fiber link to each other switch (this creates a redundant ring)
-Make each switch behave as a satellite of the main switch which is essentially configured to be a SOHO router with a plethora of LAN ports. (basic routing functions on RouterOS I’m familiar with and are working)
-whatever configuration is used, a failure of one of the fiber links will result in automatic failover to the other path. This could be as simplistic as a script that detects link failure and reconfigures the active fiber interface port to the backup.
I’m aware now (not so much back when the units were purchased) of the issue that to configure the units to use any “brains” will result in a drop of throughput because it will force the data through the CPU. This isn’t at this point going to be a deal breaker as throughput won’t be that high. I’m aware that a suitable failover script could avoid this issue however I’m thinking functional first, fancy future proof later!
Initially my question is:
-
STP or OSPF? (I do not want to deal with the intricacies of a script capable of detecting a flakey link at this time)
-
I’ve been playing with putting the two physical SFP cage interfaces into a bridge interface and turning on STP on the bridge interface, however I was having trouble getting connectivity as expected (but I was seeing the toggle between blocking mode as expected).
What are the steps to make a bridge with STP interface work, from the beginning what I did was:
-create a bridge interface
-remove the physical SFP ports from the switch group (forcing data processing via CPU??? For testing I’m just using a copper port instead)
-configure the ports to be within the bridge
-assign an IP address to the bridge interface
I was having trouble getting a computer plugged into one of the bridge ports to get to the other ports on the LAN. I figured it would go PC1->bridge interface port->Switch’s CPU-> Switch’s LAN HW switch->PC2. I figured plugging the PC1 into the other bridge port would result in failover to that alternate port and still it would function.
If someone can confirm the bridge setup to work then I can focus my efforts reading the manual to get it all tweaked as I’d like.
I’ve liked Mikrotik so much I bought an RB2011 for my SOHO router, The same OS and features on all products is a wonderful selling point! Thank you for helping getting past these initial familiarization lumps.