I've been trying to diagnose why my server has a constant idle time of 90% even when nothing is running. After finally discovering what "hi" means in 'top' (it means hardware interrupts) I find that this percentage always averages around 7-10%. How can I find out what is causing this constant load of interrupts? I have a Dell 1850 3.0Ghz with on board RAID and 2GB RAM. Anyone else experiencing this? -Matthew
Matthew Boehm wrote:> I've been trying to diagnose why my server has a constant idle time of 90% > even when nothing is running. > > After finally discovering what "hi" means in 'top' (it means hardware > interrupts) I find that this percentage always averages around 7-10%. > > How can I find out what is causing this constant load of interrupts? >cat /proc/interrupts One of your devices will have a *very* large number of interrupts I expect. Tony
> I've been trying to diagnose why my server has a constant idle time of 90% > even when nothing is running. > > After finally discovering what "hi" means in 'top' (it means hardware > interrupts) I find that this percentage always averages around 7-10%. > > How can I find out what is causing this constant load of interrupts? > > I have a Dell 1850 3.0Ghz with on board RAID and 2GB RAM. > > Anyone else experiencing this?Do a ps ax |grep mpg and see if you have more then two mpg's running. If you do, stop asterisk, kill all remaining mpg's, and restart asterisk.
OK. Perhaps I was not clear. Please read my original post again:>>>> I've been trying to diagnose why my server has a constant idle time of 90% >>>> even when nothing is running.I did not say "90% usage" I said "90% idle". Meaning I have a constant CPU usage of 10%. This 10% measurement comes from the "hardware interrupt" (hi) from within "top": Cpu(s): 3.8% us, 2.1% sy, 0.0% ni, 85.5% id, 0.2% wa, 8.0% hi, 0.4% si Even when all other percentages are at 0%, hi remains around 10%. How can I figure out what is causing all these interrupts? -Matthew