http://i.imgur.com/htMcH.png
Above is a link to a screen cap of the consistantly top offenders of my system. Even after all updates have been applied, and a complete reboot, these three services are almost immediately taking 36% polkitd, 36% dbus-daemon, and 12% NetworkManager with vary little fluxuation from the screen cap. Not only is this keeping my poor CPU pegged, but it's likely the cause for the abysmal battery life I have even after upgrading to an SSD.
Sometimes Polkitd will consume over 1GB of RAM! After the computer has been left on for a few days. This is confirmed on two computers of different ages of F17.
Is this a bug I need to report/add to? Does anyone have a suggestion for investigating why these three apps are almost locked in CPU usage.
I have done systemctl restart NetworkManager.service and when the service reloads, it's back up to 12% immediately. I think it's realy strange that these three services are consistantly at the same usage level. 4x CPU system
Some system info:
tmpfs /tmp
cpu throttling set to concervative
tuned set to laptop-battery-powersave
brightness turned all down
The curious thing is that in a test user, I don't have the same problem, so I can at least confirm that its tied to something I'm doing in my user profile. The only problem is I'm not sure where to start looking since these services explode on log-in.
Found the problem, it's a script I run that is supposed to remount my gvfs share when the system reconnects to my local network. Looks like my quick hack caused un intended consequences. any ideas how I can optimize this to only make remount requests when network state and folder detection state changes?
.
Code:
i=0
while [ $i -eq 0 ]; do
CONSTATE=`nmcli -t -f STATE nm status`
FOLDERSTATE=`gvfs-info /run/user/ben/gvfs/mount/ 2>&1|grep "type: directory"`
if [[ $CONSTATE = "connected" ]]; then
if [[ $FOLDERSTATE != "type: directory" ]]; then
gvfs-mount "smb://NEXUSNEBULA.NET;ben@file.nexusnebula.net/ben"
fi
fi
done
Linear Mode
