Good Day
I am having some problems in the last month where I am monitoring the “ping” probe on my devices on a large Wireless Network. The Dude sometimes thinks they are unreachable, but If I do a ping from “local” I can reach it.
The ping from “server” and from “local” I have done from the server itself. I have also run a ping from Command Prompt and it works as well.
I have tried to do an sqlite3 “vacuum” of the database but this has not helped at all.
To temporarily resolve the problem I have to restart the dude service/server and reprobe all “down” devices.
Please see the screenshot below for my example of a site which is showing as “down” but is infact up
Server Information
Virtualbox Virtual Machine running
Windows XP Pro
512MB RAM
2Ghz Xeon CPU (Shared)
I have seen a similar threads about this but all relating to UAC in Vista or Windows 7 causing this problem.
Any ideas?
If it is uac just type uac in the search box and move the slider all the way down, also make sure you are running dude as admin. If on the other hand it is not UAC you might have to increase your polling from 30 seconds to say 60 seconds. This was the only way I could reduce false positives.
Thanks lebowski, I will try change my polling settings and let you know how that works out.
As I am running on Windows XP I don’t believe the problem is UAC.
You might want to look at “negative cache time” and hopefully in the next version they will add a globally configurable negative cache time.
Thanks lebowski, the up/down representation has been perfectly accurate.
The only thing now is it takes a long time before a device shows that its down.
What are some good polling settings to most timeously represent a site down without showing a false positive?
I’d stick to 1 minute with 2 or 3 failed attempts. If you only have 200 or so probes you might get away with 30 seconds. Also note that with out negative cache time below retry interval a single missed poll will cause an outage.
If you want to be proactive you might want to make a new “stat view” panel, add a left column and a bottom row. Place the map in the middle the services on the right and the action log on the bottom. Sort the action log by time and the services by problem, now you can see events happening when they happen, a service will show unstable before it is down and log file messages will scroll along the bottom.
HTH,
Lebowski
Just to be sure I’ve understood you correctly, please see below the settings I have adjusted to from “default”
Yep looks good, It will take 4 missed “pings” to have a device go from “unstable to down” the first ping and 3 retries. If you have false positives move to 1 minute and retries of 2.
The check in email notification is what you were missing?
It has been working beautifully 
Perfectly accurate
I am having a similar issue except it is with the dude server itself. It used to poll fine but about 3 months or so ago it stopped polling and will not reprobe. Only the dude server shows down but I can ping it locally I just cannot get it to show up in the dude network map. Any suggestions?