[c-nsp] SFP high power alarm

[Wayne Tucker]

On Tue, Aug 21, 2012 at 7:47 AM, Peter Rathlev <peter at rathlev.dk> wrote:
> On Tue, 2012-08-21 at 06:13 -0700, Wayne Tucker wrote:
>> On Tue, Aug 21, 2012 at 3:13 AM, marc williams <marcuk at me.com> wrote:
>> > %SFF8472-5-THRESHOLD_VIOLATION: Te4/1/2: Tx power high alarm; Operating
>> > value:   0.6 dBm, Threshold value:   0.0 dBm
>>
>> Too much signal can cause receiver saturation, which in turn leads to
>> errors on the interface.  If you aren't seeing errors then you're
>> probably OK, but I always prefer to attenuate the signal so the errors
>> go away.
>
> Maybe I misunderstand the DOM parameter, but would attenuation have an
> influence on what the TX power is according to the transceiver itself?

Good point - I missed that detail.


> To me it sounds like a faulty transceiver, though an optical power meter
> could qualify that a bit.
>
> Not really related but we have previously had a few transceivers logging
> "Voltage low alarm; Operating value: 0.00V". We ignore these since they
> work fine and since 0.00V sounds more like a measuring bug than actual
> low voltage.

The OP mentioned that it's "Cisco compatible" - maybe it's not 100%
compatible or was programmed for a different type of device (I've seen
quirks when SFPs get mixed up, though they're usually more extreme).

:w