Swap Lag – Time to Retreive Data
I have a virtual machine, a MythTV server, and an email server running on my computer at home. The compiz desktop environment will lag substantially when I get home from work. I conclude that this is because during the day the desktop interface is not being use therefore it is thrown to the swap. There are several things that can be done to alleviate this lag.
1) The easiest thing to do is turn the swap off completely. This is done with a simple command:
sudo swapoff -a
2) I can logout of the desktop session to the GDM window.
Then after logging in the desktop session will be new in memory and will be responsive accordingly. The problem with this is that I need Thunderbird for its mail filtering capabilities. I can do mail filtering server level with Sieve, but I cannot get it to work right. I don’t know why it wont work, but to get around the problem, and achieve the same functionality, I use Thunderbird which is a graphical application. TO use Thunderbird I need the Gnome desktop fully loaded. EOS (End of Story).
3) To help with the situation you can renice running processes to help you desktop environment get better priority with the processor. Overall I wish I could get Sieve working, then I could just logout of the desktop session when I am done using it. The server would just run in the background using a “nohup” command, but im stuck with needing Thunderbird’s mail filtering. So for now I’ve turned the swap off to maintain system responsiveness. This is working ok because of the 8 gigs the system has. I will try to get Sieve working again.



November 26, 2009 - 4:04 am
Very great site.
The information here is truly valuable.
I will share it with my friends.
Cheers
November 27, 2009 - 2:50 am
Very extraordinary website.
The info here is truly useful.
I will invite my friends.
Cheers
December 2, 2009 - 4:12 am
Very extraordinary website.
The message here is super helpful.
I will refer it to my friends.
Cheers
December 18, 2009 - 6:15 am
Such a good site. I am saving this page.