Switching

So we have 2 QFX-5100-48s switches. 

We plugged a certified compatible 10GB optic into a port 46 on switch 1. It didnt light up. We figure, bad optic. We plug another known good optic into the same port. Doesnt work. Take this same known good optic, plug it into the port next door, works. Take it out, plug the first optic into the new port, doesnt work. Plug the known good one into this same new port, doesnt work. 

Its almost like this first optic breaks ports ? This cant be true right ? I reboot the cluster, the condition persists. I have a case open. 

Any thoughts ? Thanks. 

finding this is show log messages : 

Jul 23 00:00:02 qfx-hpc-s01 fpc1 FAILED to read the QSFP IDEEPROM for port: 46 start addr: 0 - length: 255
Jul 23 00:00:02 qfx-hpc-s01 fpc1 sfp-1/0/46(46), SFP 1: in sfp_read_range, failed to read SFP EEPROM address 0

It does seem like the bad optic is damaging something.

With the bad optic in a bad port can you run the hardware list to see if it is seen at all. Substitute your fpc and pic slot numbers this should return the make model of all the optics for that pic currently inserted.

show chassis pic fpc-slot 0 pic-slot 0

With the bad optic in a bad port it just acted just as if there was no optic in the slot. 

I tried formatting and starting from scratch. Still same issue. I ended up RMA'ing the boxes. New boxes work fine. I dont dare plug the bad optic in. The whole thing is just weird. Never heard of an optic damaging a slot. 

Thanks for the update.  This does seem pretty extreme that a bad optic can kill the port hardware.  Do you have access to support from the optic vendor to see what the root cause may be?