Sorting genuine Cisco optics out of a used batch: which checks actually mean something?
A site decommission left me with a drawer of pulled Cisco optics, a mix of GBICs, 1G SFPs and 10G SFP+, and I have to decide which of them go back into production and which go in the bin. The GBICs I have put aside for now, since nothing on the bench takes them, so everything below is about the SFP and SFP+ pile. Provenance is unknown for about half: some still have their bags, some do not, and the labels are visibly not all printed the same way.
- Catalyst 9300 on the bench for testing
- mixed pulls, including several SFP-10G-LR-S and a handful of GLC-LH-SMD
- a few modules whose serial numbers look nothing like the ones on parts I know are genuine
What I run on each one before deciding:
show idprom interface Te1/0/1 detail
show interfaces Te1/0/1 transceiver detail
show interfaces transceiver supported-listSo far I have really only sorted by "does it link on the bench", which is not much of a filter, since a re-coded module links perfectly happily. I compared serials against the bags that survived and found two that match nothing.
What in the idprom output is actually worth reading, and is there any field that gives away a re-coded or counterfeit module rather than just one the platform does not have on its list?
Comments 2
Read the serial properly, it is structured: three letters for the factory, four digits for the date code, then four alphanumerics for the unique ID. Anything that does not fit that shape goes straight into the suspect pile. Then look at the vendor OUI in the same output, because it has to map to a real manufacturer, and on the bad ones it frequently maps to nothing at all.
After that it is physical. The serial on the module has to match the serial on the bag and on the box, which is exactly the check you have lost on the unbagged half, so I would treat those as unknown by default. Genuine labels are a light yellow-orange with the part number laser-etched rather than ink-stamped, and once you hold a known-good module next to a copy the difference is easier to see than to describe.
None of this is proof on its own. A decent copy passes the serial format check without trouble, so treat it as a filter that removes the obvious rubbish rather than a verdict, and expect whatever survives to still fail at a noticeably higher rate than new stock.
Two separate questions, as you say, and
show interfaces transceiver supported-listonly settles the second one. Off the list means this box may refuse to drive the module, and nothing beyond that.For the other one I trust DOM more than the idprom.
show interfaces Te1/0/1 transceiver detailprints every value beside its thresholds: supply voltage wants to sit in the 3.13 to 3.46 V band, bias current 18 to 70 mA before it warns, and Rx under Low Warn is the far end or the cable talking, not the module in your hand. Those fields live on the SFF-8472 diagnostics page, and a lazily cloned EEPROM tends to show it right there, either garbage numbers or empty fields.Leave
show loggingrunning while you swap them.%TRANSCEIVER-3-NOT_SUPPORTED, or a media type printed as unknown, means the switch never really took the module. Different command if one of these lands in an ISR 4000 later, mind:show hw-module subslot 0/0 transceiver 0 status.