Setup a Linux Server or VPS » Dedicated Server

Better Network Performance

admin — Tue, 12 Oct 2010 16:10:28 +0000

As we all know a basic install of any OS is never tweaked out for performance but setup for general use. One of the biggest problems in the Linux (CentOS, Ubuntu, Fedora, I have not tested other OS’s) is the network is never really tweaked up.

However running a few simple commands and updating a single file can produce a massive increase in performance.

** Please note that there are other issues that can cause poor performance **

Example Bad Network Cable, Cheap Switch, Overloaded Router, and others.

Run the following commands from the prompt.

mv /etc/sysctl.conf sysctl.bak
cd /etc
wget http://www.setupalinuxserver.com/files/sysctl.conf.txt
mv sysctl.conf.txt sysctl.conf
/sbin/sysctl -p
/sbin/sysctl -w net.ipv4.route.flush=1

You should see a performance increase right away. Please let me know if this helps you out or not.

Basic Server Setup

admin — Sun, 25 Jul 2010 04:05:16 +0000

No matter how new your server or vps is, it is outdated and running old software. While I am not the best, nor do I know everything this is a base of what I do when I get a new box online.

Again assuming you can SSH into the box.

yum -y install nano
yum -y install yum-priorities
wget http://apt.sw.be/redhat/el5/en/i386/RPMS.dag/rpmforge-release-0.3.6-1.el5.rf.i386.rpm
rpm –import http://dag.wieers.com/rpm/packages/RPM-GPG-KEY.dag.txt
rpm -K rpmforge-release-0.3.6-1.el5.rf.*.rpm
rpm -i rpmforge-release-0.3.6-1.el5.rf.*.rpm
wget http://dag.linux.iastate.edu/dag/RPM-GPG-KEY.dag.txt
rpm –import RPM-GPG-KEY.dag.txt

nano /etc/yum.repos.d/dag.repo

[dag]
name=Dag RPM Repository for Red Hat Enterprise Linux
baseurl=http://dag.linux.iastate.edu/dag/redhat/el5/en/$basearch/dag
gpgcheck=1
enabled=1

Save and Exit

yum -y install perl-libwww-perl
yum -y update

chkconfig cups off
chkconfig pcmcia off
chkconfig kudzu off
chkconfig isdn off
chkconfig xfs off
chkconfig atd off
chkconfig nfslock off
chkconfig canna off
chkconfig FreeWnn off
chkconfig cups-config-daemon off
chkconfig iiim off
chkconfig mDNSResponder off
chkconfig nifd off
chkconfig rpcidmapd off
chkconfig bluetooth off
chkconfig anacron off
chkconfig gpm off
chkconfig saslauthd off
chkconfig avahi-daemon off
chkconfig avahi-dnsconfd off
chkconfig hidd off
chkconfig pcscd off
chkconfig sbadm off
chkconfig ossec off
chkconfig acpid off
chkconfig dhcpd off
chkconfig firstboot off

mkdir /root/source
cd /root/source
wget http://www.configserver.com/free/csf.tgz
tar -xzf csf.tgz
cd csf
sh install.sh

nano /etc/csf/csf.conf

Change TESTING=”1″ to TESTING=”0″
AUTO_UPDATES=”0″ to AUTO_UPDATES=”1″

TCP_IN = “20,21,22,25,26,37,43,53,80,106,110,113,143,443,465,587,873,990,993,995,1129,2077,2078,2082,2083,2086,2087,2095,2096,3306,5224,5432,7777,7778,8443,8880,8888,8889,9080,10000,30000:35000″
TCP_OUT = “20,21,22,25,26,37,43,53,80,106,110,113,143,443,465,587,873,990,993,995,1129,2077,2078,2082,2083,2086,2087,2095,2096,3306,5224,5432,7777,7778,8443,8880,8888,8889,9080,10000″
UDP_IN = “20,21,53,953″
UDP_OUT = “20,21,53,113,123,953,33434:33523″

service csf restart
service lfd restart

reboot

. Add Passive Port range 30000-350000 to your Pureftp or Proftp configuration file
(i) Pureftpd
open /etc/pure-ftpd.conf, and this line
PassivePortRange 30000 35000
(ii) ProFTP
Open /etc/proftpd.conf, and add this line
PassivePorts 30000 35000

Now there are many other things you can do, root kit hunters and log rotators and more. However depending on your needs this alone will get you off to a good start.

Switching from Apache to Lighttpd

admin — Sat, 24 Jul 2010 00:26:19 +0000

One of my side projects is running a free image hosting site. FreeImageHostingOn.US with this site I have been pushing out millions of hits a day and as you can imagine it puts a load on many shared hosting providers. So the other day I grabbed a VPS from Ruby Ring Technologies and I installed the CentOS/LXAdmin image. Out of the box this VPS screamed. However I know I could get more performance out of it. The following six lines of code did more then double the thruput and halved the ram usage.

Add the following lines to the lighttpd.conf file.

server.max-keep-alive-requests = 4
server.max-keep-alive-idle = 4
server.event-handler = “linux-sysepoll”
server.network-backend = “linux-sendfile”
server.max-fds = 8192
server.stat-cache-engine = “simple”

As you can see from the following image, things are moving along nicely.

The current load on the server is 0.0 and the ram used it 66MB’s for live stats check out this link. System Usage.