On 2/6/24 13:50, Eelco Chaudron wrote:
> 
> 
> On 6 Feb 2024, at 13:46, Ilya Maximets wrote:
> 
>> On 2/4/24 15:40, Timothy Redaelli wrote:
>>> A log message like this one:
>>>
>>> 2024-01-09T06:45:17.201Z|00071|bfd(handler2)|INFO|ovn-0af536-0: BFD state
>>> change: down->up "Neighbor Signaled Session Down"->"Neighbor Signaled 
>>> Session
>>> Down".
>>>
>>> can be hard to read since '->' usually represents a status change, but
>>> in this case the diagnostic code stays constant. Update the log message to
>>> avoid such ambiguity. The log message for the above event become:
>>>
>>> 2024-01-09T06:45:16.211Z|00026|bfd(handler3)|INFO|ovn-0af536-0: BFD state
>>> change: up->down, previous failure: "Neighbor Signaled Session Down",
>>> current failure: "Neighbor Signaled Session Down".
>>
>> Hi, Timothy.  Thanks for the patch!
>>
>> Though I'm not sure that the new version is fully correct either.
>> Since this message is BFD-specific, we should use BFD-specific language.
>> Diagnostic is not a failure in a general case, so logging them as a failure
>> is not correct, IMO.
>>
>> We may also want to specify that it is a local diagnostic and not a remote
>> one to avoid any further ambiguity.
>>
>> However, if the confusion is caused by the fact that it doesn't change, while
>> the arrow suggests that it did, maybe we can just re-structure the message
>> in a following way:
>>
>> BFD state change:
>>   (bfd.SessionState:   up, bfd.LocalDiag: Neighbor Signaled Session Down) ->
>>   (bfd.SessionState: down, bfd.LocalDiag: Neighbor Signaled Session Down)
>>
>> (no need to split lines, I did that only for ease of reading)
>>
>> Since at least one part will change, the arrow should not be confusing 
>> anymore.
>>
>> Spelling out bfd.Xx names as in the RFC is probably not necessary since we're
>> printing out simple words (state, diagnostic) in other places, so can do that
>> here as well.
> 
> Good idea, I do like the RFC naming of the events. Maybe to save space, we
> could remove the “bfd.” prefix, but keep the SessionState, etc.

I think, if we use the names from RFC, we should use them with a prefix.
The prefix is there to avoid confusion between state machine variables
and BFD packet fields that are very similarly named.

> 
> Maybe a follow-up patch to change this in all BFD-related messages to make 
> them
> more uniform?
> 
> 
>> Thoughts?
>>
>> Best regards, Ilya Maximets.
> 

_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to