When opt1 goes offline (non default gateway) gui almost not usable

grandrivers · Mar 23, 2016, 4:57 PM

when opt1 goes offline (non default gateway) gui almost not usable
disabling smtp notifications fixes it

cmb · Mar 25, 2016, 12:27 AM

Is your mail server not reachable when that WAN is down?

grandrivers · Mar 25, 2016, 12:35 AM

its gmail so should be

grandrivers · Mar 25, 2016, 12:36 AM

but still end up with 100 of emails so maybe thats more of the issue?

cmb · Mar 25, 2016, 2:33 AM

What exactly are the contents of the hundreds you end up with? With a single WAN down it should deduplicate and notify once.

grandrivers · Mar 25, 2016, 12:30 PM

yes lots of duplicates, i hadn't seen any changes 2.2.x-2.3 regarding deduplication did i missing it or am i extreme edge case?

luckman212 · Mar 25, 2016, 12:45 PM

I've seen these "mail bombs" too. There's even a ticket for it- https://redmine.pfsense.org/issues/4031 (which now says target version is 2.4)

How does the dedupe work? Does it just hash the outgoing messages and not send again if the new hash matches the previously sent one within a certain amount of elapsed time? I could see how this results in a bomb because sometimes you get alerts in rapid fire like

MONITOR: WAN1 is down
MONITOR: WAN2 is down
MONITOR: WAN1 is down
MONITOR: WAN2 is down
…ad infinitum...

If that's how it works, I would suggest it would be great to have a knob we could tweak that says "Send no more than X number of alerts within Y minutes". And if this was set to e.g. X=2, Y=15 then regardless of the contents of the emails (even if it was "the CPU is on fire!") the unit would not send out more than 2 alerts in a 15 minute window. Chances are the 1st critical alert is enough to prompt you to try to remediate the situation anyhow.