Are you swapped? Increase the performance of Linux machine.

With the ever increasing cost of the Hardware, the amount of physical RAM available on the system is increasing day by day. For example, couple of years back, I had a system which was very high end Desktop with 256MB RAM and today I have a 2GB RAM Desktop. So, whats the point.

The Linux systems (right word should be kernel) are desiged to use both RAM and swap partition. Swap partition is a partition on Hard disk and is used mostly like RAM. Problem is that HDD access is always slower than RAM access and hence inherently, the system will work little slower even if you have enough RAM not to use swap. The term \”swappiness\” is used to determine how the kernel should try to seam-balance between the use of RAM and swap. By default, most of the distro\’s have a swappiness of 60. A higher value of swappiness means that the RAM will be swapped out faster.

There are two ways to look at the swappiness:

1) If the user has a higher swappiness then the used memory will be swapped faster to the swap and thus free\’ing the RAM for other useful purposes.

2) A lower value of swappiness would mean that bloaty applications will not be swapped and thus when the user returns to the application, the application would load faster or rather look faster as there will be no swapping from the swap to the RAM.

I personally keep the swappiness to a value of 100 in the Desktop. But then, anyway I kill firefox as soon as I am done and restart when required.

If you see that the RAM is underutilized or feel that the system performance is not that good then you can tickle with this setting and set it to 10-15. How to do it :

Login as roo (\”su -\”)

echo 15 > /proc/sys/vm/swappiness

This will take effect immediately, but if you want to change this value permanently then you can do this(again as root):

echo \”sys.vm.swappiness = 10\” > /etc/sysctl.conf

So play with your swappiness 🙂

\"Enhanced
Enhanced by Zemanta

Logwatch for Linux Systems – Scheduled email.

On my personal desktop at home, I like to see the statistics at least once a day, for what was installed, what was run with sudo and other such details like kernel errors.

Running this monotonously every day is quite boring, so comes to rescue is logwatch. I have Fedora installation so I will talk about the location with respect to that so for your distribution it might be a little different.

First install logwatch using 🙂 yum install logwatch

and you are done.

If you need to do more configuration and want to see something different in the mail that is sent( BTW sendmail is assumed to be default mail client), the keep reading.
First open the file /etc/logwatch/conf/logwatch.conf
The contents for my distro is just one line 🙂
# Local configuration options go here (defaults are in /usr/share/logwatch/default.conf/logwatch.conf)
So, open the file /usr/share/logwatch/default.conf/logwatch.confbr
I will highlight the important lines in the file here:
LogDir = /var/log — This is the directory for all the log files
MailTo = root — Whom should the mail be sent to
Print = If this is set to true then there will be no mail sent and the output will be displayed on the stdout.
Detail = The level of details you want to see in the mail or the output on screen.

Thats pretty much it.. If you want to further modify the details in the mail you can configure the services. For more advanced usage you can even go to /usr/share/logwatch/scripts/services and configure the individual scripts.

\"Reblog