We have two identical Brocade 300 SAN switches ISL'd together on ports 0 and 15 over LongWave 30km SFP transceivers. Running FOS 7.0.0c
Issuing a "porterrshow" on switch1 lists a very high number of enc_out errors on PORT15.(>500 errors per second)
Issueing the same command on the partner, switch2, shows no errors.
A "portshow 15" command on switch1 indicates that most all errors are of type "Invalid_word".
The same command on switch2 shows no errors.
A "portstatsshow 15" command on switch1 reports a large amount of "er_enc_out" and "er_bad_os" errors.
Additionally, issuing a "sfpshow 15" command on switch 1 shows an Rx Power value of -20 dBm and a Tx Power of 2.2 dBm.
An "sfpshow 15" command on the partner switch2 shows Rx Power value of -11 dBm and a Tx power of 2.2 dBm.
We're at a loss as to what the issue is at this point. All fiber jumpers have been replaced/cleaned, and our network team has re-routed the ISL connections on different fiber strands and the errors remain.
Can someone help to interpret these output values for the errors as well as provide input on the variation of the Rx/Tx power values between the 2 switches?
Hi barrickd. I have very basic information regarding your request.
For the first, all i can give you is the definition. "Invalid_word"-- Encoding errors inside of frames
Regarding "er_enc_out" and "er_bad_os", That counter (er_bad_os) is invalid. Ignore it. Also, you frequently see EncOut (Encoding Error Outside of frame) errors when there are BadOS errors. You can ignore the EncOut errors when there are BadOS errors as well; however, EncOut errors when there are no BadOS errors are geniuine and should be debugged.
But for more information, you would need to contact support.
Every Fibre Channel connection (an ISL in this case) is made up of two separated fibers (one for Tx and the other one for Rx).
We can see the following from your post:
Please, review again your infrastructure since there is an element that is attenuating the signal by 9dBm.