Gateways monitoring still screwed up (2.2.2, apinger issue)
-
Hi guys.. I'm running pfsense for almost an year in a Multi-WAN (TWO Internet gateways plus 8-10 other gateways) virtualized enviroment.
In the beginning we were using 2.1.4 with XEN, then we used 2.1.5 with ESXi and now 2.2.1 and 2.2.2.ALL these combos were giving ridiculously false % loss and RTT ping values on status_gateways.php and this highly compromises our multi-wan setup because we need to automatically change gateways on high ping or packet losses!
I have googled about it since 2.1.4 days and I've seen lots of users complaining about apinger becoming crazy and afaik there's still no real fix for it..
The only workaround I've seen is restarting apinger for every X minutes, which really works, but in my opinion it's so lame I wanna puke.
Example:
Name Gateway Monitor RTT Loss
GWEBT 201.72.x.1 8.8.4.4 0.4ms 25%This exact same "8.8.4.4", while PINGed from within pfsense's SHELL itself, is giving ~61ms and 0% packet loss. EVEN with "ping -f" it would never ever give 25% loss. And even if google had a direct link with us, it would never ever give less than 1ms pongs!
And wrong results like these persists FOREVER if I don't restart apinger. You can imagine how screwed up our gateways graphs are!!
Question: how hard could it be to code an app to PING an IP and show an accurate result?
-
Yaaaaaawn. Tell us when it works instead.
-
Question: how hard could it be to code an app to PING an IP and show an accurate result?
It's an OpenSource project - you can complete the license agreement and get access to the pfsense-tools repo. Look through the apinger source code and work out what goes wrong with the running average latency and packet loss calculation. Many people will be very grateful.