All:
Having severe issues getting a simple 1 gig SM link up with a Ciena. I have an ATT circuit on site, they’ve swapped optics as well as the actual Ciena device itself. Circuit is confirmed good when plugged into a Cisco router (ISR4331) with the same optics used in the Mikrotik. (using official MT optic, FS.com and Cisco) Circuit was previously hard coded in Ciena for 1gig Full, had them change to autonegotiate - no difference. Changing the MT side either way makes no positive impact on circuit availability.
Symptoms:
Links flap when changing the SFP out of the Ciena - once every 2 seconds or sooner. Eventually the Ciena gives up in negotiating and stops sending light (-40db Rx power)
SFP Page: When the link comes up, shortly after it goes down - and most of the page (Wavelength, Temp, Supply V, Tx Bias, Tx Power, Rx Power) blank out in addition to the Connector Type and SM Link Length.
It’s almost like the SFP is rebooting, yet the Vendor name, part number, rev, serial and manufacturing date all remain as valid items on the page.
What I see in the MT logs:
08:18:45 interface,info combo1 link up (speed 1G, full duplex)
08:18:46 interface,info combo1 link down
08:18:47 interface,info combo1 link up (speed 1G, full duplex)
08:18:54 interface,info combo1 link down
08:18:54 interface,info combo1 link up (speed 1G, full duplex)
08:18:56 interface,info combo1 link down
08:18:56 interface,info combo1 link up (speed 1G, full duplex)
08:18:58 interface,info combo1 link down
08:18:58 interface,info combo1 link up (speed 1G, full duplex)
08:18:59 interface,info combo1 link down
08:19:00 interface,info combo1 link up (speed 1G, full duplex)
08:19:01 interface,info combo1 link down
\
- Has anyone else experienced this with the specific CRS312-4C+8XG device or otherwise?
- Any further debugging I can do to see why this is going sideways and potentially remediate?
I’ve been on several software releases in 7.x on this device and the upgrades don’t seem to impact this performance.