I recently got my CSS610-8G-2S+ switch and its running great so far.
Yesterday I connected my servers via SFP+ RJ45 modules and removed their 1G links but after doing that, I could no longer the management website of the switch from the servers. If I move the cable to a normal RJ45 Port it works again. I also checked port isolation, nothing configured there and the “Allow from Ports” setting in the general tab, but all is default there - access to management from every port. A reset to default etc. didn’t change the problem. My CSS610-8G-2S+ is running 2.12rc2.
Can anyone of you, also owning that switch, try if you have the same issue?
Problem: It is not possible comunicate with device CSS610-8G-2S+ through
SFP+ ports (I,II) [Mikrotik S+RJ10 /r2]. Web interface is inaccesible, PING to
device failed, (DHCP lease failed) and static IP on the device do the same.
Standard 1G ports 1-8 works correctly (DHCP/PING/WEB). Switching between
ports 1-8 / SFP+ I,II (vice versa) always working well.
Answer SUP-30842 :
Hello,
Thank you for the report!
We have managed to reproduce the issue locally in our labs and look forward to fixing it on upcoming SwOS versions, unfortunately, I cannot comment on the version availability.
I have this problem too. It’s good to know that Mikrotik Support will be working on it. I have yet seen that many people report a problem with the vlan configuration, I haven’t tried it yet, I will wait for the new SwOS, maybe there will be less nerves.
Can anyone of you, also owning that switch, try if you have the same issue?
I can also confirm this problem. I just plugged in two identical CSS610-8G-2S+ switches, connected them via SPF+. I can access both of their web interfaces through RJ45 port, but not with SFP+ port.
It is not just the management interface. It seems that DCHP client is also affected. On the DHCP server side I see that address was “offered” to the switch, but the switch is not bound to the offered address.
A similar issue has been observed also in our labs and we look forward to fixing it on upcoming SwOS versions, unfortunately, I cannot provide an ETA now.
and
Unfortunately, I cannot share any details regarding software availability. My apologies for the caused inconvenience.
Well, I bought these units for a concrete installation, and now I’m not able to use them. I have a deadline. Since they cannot tell when it will be fixed, I have to use something else.
When I bought them, I did not know that web management and DHCP client won’t work on SPF+ port. It also turned out the VLAN is not working normally.
Bump. Same with CSS610-1Gi-7R-2S+ and RB4011iGS+RM
When connected via SFP+ it works pretty much like unmanaged switch. And only in DHCP mode.
Configurations tested:
All configurations assume default router config with IP changed to 10.0.1.1 and DHCP address space to 10.0.1.101 - 10.0.1.254. Client IP set to 10.0.1.50 manual.
Case 1: Switch configured to “Static only” with IP: 10.0.1.100. Switch connected to Router vis SFP+
switch can be accessed when client plugged into switch
router (10.0.1.1) can’t be accessed from client when client plugged into switch
switch can’t be accessed via IP address when client plugged into router
resources connected to router are NOT accessible from client, connected to switch
resources connected to switch are NOT accessible from client, connected to router
Case 2: Switch configured with “DHCP only” or “DHCP with fallback”. Switch connected to Router vis SFP+
switch can NOT be accessed when client plugged into switch
router can be accessed when client plugged into switch
switch can NOT be accessed when client plugged into router
router can be accessed when client plugged into router
resources connected to router are accessible from client, connected to switch
resources connected to switch are accessible from client, connected to router
Case 3: Any switch configuration. Switch connected to router via ethernet
both admin UIs are accessible when client connected to either box
all resources are available when client connected to either box
Same problem. Took me few hours of wondering what am I doing wrong before I found this.
Agree. Quite annoying as my networks are connecting to this switch through sfp.
In the last 10 Years we spend in MKT devices more than € 250,000. Sometimes we had some problems but never something as obvious as this. Aswell we work with Huawei, Ruckus, Netgear, Mimosa, Silvus… I don’t normally buy switches from MKT and now I know I did it right.
Guys, seriously? This is NEW product. Not really tested by earth and fire. If you’re planning a commercial install based on an untested product - it’s not MKT problem, it’s your risk management. Go check Cisco forums for new products they release. Give folks some slack.
Again, you wanted a cool shiny new thingie? Wanna pay a fraction of what you’d pay for Cisco? Be ready to deal with early adopter issues.
To Mikrotik team - you guys are doing great, keep pushing.
And maybe after you resolve this you could think of replacing “G” with “P” in this product name? Having PoE+ for this switch will truly make it THE instant classic
you missed the point completely
bugs do exist, in any product, new or not new.
what we are criticizing is support’s answer to the request to fix a bug which should be trivial to fix
for these, Cisco used to deliver us custom builds back then when I was a customer.
Totally agree. As you say, sometimes there are bugs. The strange thing is this type of error and the support that MKT provides. It’s so strange that they didn’t test it before launching it. On a managed switch I think this is the first point to check because… Do you want to be able to manage it? LOL.
This is like selling a car that has no wheels. We may classify this problem as a “software bug”. But this is an obvious fault, and the product could never work with the simplest configuration.
