I recently bought crs304-4xg-in as I upgraded my internet connection to 8Gbs and needed to replace my 1Gbe switch. I actually wanted unmanaged switch as I have no experience in networking but assumed that I can easily make it a ‘dumb switch’ (and got great price for the switch )
As I understand it should work as such out-of-the box but it doesn’t for me.
I also tried:
System > Reset Configuration
No Default Configuration
Do Not Backup
[admin@MikroTik] > /interface bridge add name=bridge1
[admin@MikroTik] > /interface bridge port add bridge=bridge1 interface=ether2
[admin@MikroTik] > /interface bridge port add bridge=bridge1 interface=ether1
[admin@MikroTik] > /interface bridge port add bridge=bridge1 interface=ether3
[admin@MikroTik] > /interface bridge port add bridge=bridge1 interface=ether4
[admin@MikroTik] > /interface bridge port add bridge=bridge1 interface=ether5
Your crs304-4xg-in is defective it has two ether4 ports
Now, seriously, how are you managing the Mikrotik? (I presume with Winbox via MAC).
A good idea is (unless you really-really need it) is to keep a port outside the bridge, for emergency access (it is not a problem in a “dumb” bridge but when/if you will want to make it more secure or play with some advanced feature it is easy to lock oneself out).
As a matter of fact if you look at the device layout: https://cdn.mikrotik.com/web-assets/product_files/CRS304-XG_240945.png
you will see that ether5 is designed to be not part of a bridge.
I would take port 5 out of the bridge, a 1 Gb port in a bridge with 4 10 Gb bridge may slow down everything (or worse), even if not used/not connected to anything.
Well, the “dumb” bridge should have an IP (L3) assigned to be able to ping a destination.
A “dumb” bridge is L2 and knows nothing of L3, while not becoming “smart”; it will become a “disadvantaged” switch if you give it a L3 address.
Once an address is assigned to the interface, a dynamic route for the whole subnet (192.168.1.0/24) will be automatically added to the “main” route table.
The status of “searching” means that the DHCP server is not found. Otherwise it would be “bound”, and you would see the assigned IP.
Better leave quick set alone.
You can add manually an IP address to the bridge to see if pinging to 192.168.1.1 works, let’s say 192.168.1.9/24.
But do check that the DHCP server works.
I tried with statically assigned IP and it didn’t work as well.
But I found the issue…
It was the cable. I did a sanity check and move with laptop and switch closer to router and connect new short cables and it worked like a charm
I still don’t know why my old TP-LINK had no issues with old cable, but Mikrotik did, hmm.
But I have one to replace it
Thanks for the support!
I did learn few things in the process!
Your previous switch was 1Gbps (if I understand your opening post right) and your CRS304 is 10Gbps. Which is a huge difference when it comes to UTP cable. Even though you’re using a cat7 cable (indicated on your chart), it might be of low quality, it might be improperly terminated, it might be (slightly) damaged (e.g. too tight curves). And if the cable is longer than around half of declared max length (which is 100m/328ft), anything not ideal can affect ability to operate connection at desired speed. Requirements regarding UTP cable for 10Gbps are much higher than for 1Gbps and minor problems might not break 1Gbps connection.
Still, if the ports are set to 10 Gb it should show “no link” (and not “running” status), this kind of “almost connected” (interface running, but not a real communication on it) seems like a tricky issue.
In the (good?) ol’ times there was a monthly column on Byte by Jerry Pournelle that had (at the time bordering science fiction) a lot of fast SCSI devices and the moral in every other article was “when something doesn’t work in SCSI it is often always the cable”.
Anyway, all is well that ends well .
Just to provide full picture.
I might have lied with my diagram, but now it is new and fresh cat 7 cable
The one that was installed was
cat. 6e 28AMG Network Flat Patch Cord CABLE ROHS MADE IN CHINA (30 meters) and it was indeed a bit damaged in one place.
Even if not damaged “flat”, “cat6a” and “30 meters” rarely can work at 10 Gb, personally I wouldn’t use the 28 AWG ones for more than 15-20 meters, maybe the 26 AWG for 30 m, but 23 or 24 are the “normal” ones, good for 100 m.
If I recall correctly good manufacturers give max 60 or 65 m for those and the rule of thumb to be on the safe side Is to not exceed half of the theoretical max.
Still in theory a damaged (but not interrupted) cable should have the devices negotiate a slower connection rate or have a high error rate, but preventing a DHCP lease it is AFAIK unusual.
Maybe the cable was actually damaged with one wire among 4,5,7,8 interrupted i.e. in a way that your previous 1 GB hardware down negotiated to 100 Mb whilst your current 10 Gb hardware cannot negotiate lower than 1 GB (i.e. all 8 wires are needed).
But then It should cause “no link”.
Yep, besides there are two kinds of “flat” cables, let’s call them “semi-flat” and “super-flat”.
The “semi-flat” have anyway twisted pairs, but the “super-flat” may lack the twisting of pairs (hopefully only very short patch cables are made with this)
Twisting pairs, at different twist rates, is actually a very smart way to reduce crosstalk between pairs, still Cat 6+ round cables have anyway a cross-shaped insulator between the pairs (good) and the outer insulation is way thicker than what the flat cables have.
Flat cables also have the benefit of your wife not asking, ‘Why can’t you just use Wi-Fi like a normal person?’ As they can be hidden nicely, which was the whole reason with ‘going flat’
I think it depends on the exact model/version/release of wife, I don’t believe most would consider a black cable on a whitish wall as “hiding”, let alone “nicely”.
)
it has two ether4 ports 
