[mrtg] Problem with Forks & Non-duplication (hanging around)

[Grossman, Benjamin]

For some reason that I am having trouble understanding, if I use Forks in my CFG file, the logic that prevents a second MRTG from starting up if the first hasn't finish doesn't work.

Case A: no Forks statement. Cron set to start CFG file every 5 minutes. Job usually takes 12 minutes. Attempts to start at +5 and +10 minutes are blocked, generating "hanging around" messages.


Case B: Forks: 4. Cron set to start CFG file every 5 minutes. Don't know how long job should take; usually end up with 3 or more jobs running at the same time.

Can someone explain to me why this is, and if there is something I can to do fix it? I need to get poll cycles down under 5 minutes, and I was planning to use forks for that. The server has lots of resources available, but some of my targets take a long time to fail to respond, so the overall time for a cycle is fairly large. Manually breaking the CFG up into units is a lot of extra bother I would like to avoid as this is a dynamically built file based on what devices are on the network today.

MRTG 2.15.2 
Solaris 10/ Sun T2000 server 

Benjamin k Grossman 
Network Tools Sachem 
King County OIRM 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/mrtg/attachments/20071220/e83c9b84/attachment.html