How to be notified if an interface is down?

torontob

Hi Everyone,

Is there a tool already built-in to notify me if an interface goes down? I want to not only be informed but I also want to know if attempts are made to reconnect again.

1- If the PPoE interface goes down, I want an e-mail message to be put in queue to be sent to me once the connection is back up.
2- If PPoE is down, I want the interface to continuously and repeatedly query the DSL provider for an IP and connection. I think that happens by default but for my piece of mind I want to confirm this if it happens.
3- At a later stage, what I want to do is to actually connect to an Asterisk server which is connected to the pfsense router to place a call through the analogue lines to inform me of the down interface. I guess once I have #1 and #2 confirmed and figured out then #3 can be easily implemented.

Thanks,
Bruce

jimp

#2 happens automatically, it reconnects as many times as it takes. It wouldn't be wise to generate an e-mail on each attempt.

As for #1, SMTP (and Growl) notifications are in 2.0, but I don't think we have an 'interface down' notification, though it could be done.

Not sure how you'd pull off #3, but I doubt anything that relied on Asterisk would be in the base system.

torontob

Hello,

Where should I look for the code to detect an interface down. I abandoning other ideas and will be just sending a request to Asterisk server to make a call rather than e-mail alerts. Asterisk server is on the same network so it's very unlikely that LAN port will go down.

Any pointers are greatly appreciated.

Thanks,
Bruce

jimp

If you are looking for a hard interface failure (e.g. the port is unplugged or CPE powered off) look into the link events in /etc/devd.conf (read the FreeBSD man page)

As for a "soft" failure, such as a high delay or connectivity loss (no pings returned) you might be able to get that info from /tmp/apinger.status

Or just route the system logs to another server (the system log settings page can do this) and have something that watches/parses the log and reports on failures there.