Red interface UP and DOWN state reversed


(Matthieu Gaillet) #1

Hi there,

I think there is a bug in the way some internet connection (red interfaces) are reported by email : obviously some the connexion states fore some interfaces are reversed :

The connexion looks like down all the time, which is fortunately not the case :slight_smile:

Interestingly, the backup connexion’s status is correctly reported.

The network interface is an Intel Corporation 82546GB Gigabit Ethernet Controller (4 in one).

Here is what I found when I look for EDPNT in the logs :

Jun 19 19:05:23 serveur esmith::event[30822]: Event: wan-uplink-update down EDPNT   root 0 0 0 0 0 0 0 0  up 1497891923
Jun 19 19:05:24 serveur esmith::event[30822]:   Provider EDPNT (1) stopped
Jun 19 19:05:53 serveur esmith::event[30955]: Event: wan-uplink-update up EDPNT   root 0 0 0 0 0 0 0 0  down 1497891953
Jun 19 19:05:54 serveur esmith::event[30955]: Provider EDPNT (1) Started

How can I further debug this issue ?

Thanks

Matthieu


(Matthieu Gaillet) #3

Update : even more strange


(Alessio Fattorini) #4

Sorry for the late response. @giacomo @davide_marini any tips?


(Giacomo Sanchietti) #5

It seems strange to me, I can’t reproduce it right now.


(Matthieu Gaillet) #6

Any way I could debug this ?


(Marc) #7

I don’t use multiwan so forgive my ignorance, just trying to understand the problem.

The dates within the email body are correct for the down/up state, but they are delivered/received with such delay that they pile up in the mail inbox in the wrong order?
Through which interface is every mail send (some clues in the mail headers)?

Edit: and /var/log/maillog might be worth looking at.


(Matthieu Gaillet) #8

That may indeed be a good clue @dnutan ! I’ll investigate with that in mind.


(Matthieu Gaillet) #9

You nailed it @dnutan ! The messages are sent in the right order. But for one or another reason, they are received in the reverse order by the distant MTA (this server is not used as a mail server).

–> Not a bug.

Matthieu


(Michael Träumner) #10

Hi Matthieu,
please mark the answer as solution.


(Matthieu Gaillet) #11

I did. Look at my first post, the solution is there.