For those who are having problems installing curl-loader under Centos platforms please read this guide.

This guide applies to Centos 5.5 and 5.6 versions, 64-bit platforms.

Install the following packages:

yum install libevent libevent-devel

Download curl-loader from: http://sourceforge.net/projects/curl-loader/

replace in Makefile
OPENSSLDIR=$(shell $(CURDIR)/openssldir.sh)
to
OPENSSLDIR=/usr/lib64/openssl

Run with root user:

./configure
make
make install

If you are getting the following error:

curl-loader: error while loading shared libraries: libcares.so.2: cannot open shared object file: No such file or directory

you are missing c-ares package..you can try to install from Centos yum repository:

yum install c-ares

… Read the rest of this entry »

If you are getting the following error when you are installing memcached from the source:

checking build system type… Invalid configuration `x86_64-unknown-linux-’: machine `x86_64-unknown-linux’ not recognized
configure: error: /bin/sh ./config.sub x86_64-unknown-linux- failed

the problem is that you probably do not have a C compiler available on the system. Please check the server config and make sure you install C compiler with all the required rpms.

Before you try to install LUA on Centos Linux, make sure you have already installed the following development packages:

ncurses-devel
readline-devel

or you may experience the following errors while running “make linux”

/usr/bin/ld: cannot find -lncurses

gcc -O2 -Wall -DLUA_USE_LINUX -c -o lua.o lua.c
In file included from lua.h:16,
from lua.c:15:
luaconf.h:275:31: error: readline/readline.h: No such file or directory
luaconf.h:276:30: error: readline/history.h: No such file or directory
lua.c: In function ā€˜pushlineā€™:
lua.c:182: warning: implicit declaration of function ā€˜readlineā€™
lua.c:182: warning: assignment makes pointer from integer without a cast
lua.c: In function

… Read the rest of this entry »

There are only two Oracle certified operating systems that support hosting Oracle DB:

• SLES (Novell Suse Enterprise)
• Redhat (Enterprise Linux)

I have been using both OSes to hosting Oracle DB and so far it’s been working great. My personal choice would be Redhat, but I know you can get better pricing and features (from the licensing point of view) with Novell… if you know how to deal with these guys.

If you are trying to install Oracle database product(s) on Suse Novell Linux 10 SP 3 you may experience problem with missing compat rpm package. Novell SLES actually provides all the necessary files via libstdc++33 howerver Oracle installation doesn’t really recognize it and reports missing compat libstdc++ package.

If you have installed you may remove the following packages running the following rpm command:

rpm -e libstdc++33
rpm -e libstdc++33-32bit

A good alternative is to find a compat libstdc++ package from SLES 10 SP 2 and it works perfectly. The compact-libstdc++ package you need is as follows:

compat-libstdc++-5.0.7-22.2.x86_64.rpm

Download it… Read the rest of this entry »

By default Centos, RedHat and Fedora starts multiple daemons that are actually not needed, for example:

nfs
nfslock
cups
gpm

and I am suggesting you disable these services asap. You can turn it off via chkconfig command or even remove from the execution at init levels as well.

chkconfig –del cups

chkconfig –level 3 cups off

The above line with –del will simply remove it

the second line with –level 3 option will turn off in init #3

You can easily monitor your Linux server load using built-in bash commands like

w

free

top

However, many folks prefer sysstat package (can be downloaded here, or from your Linux distribution site) package that provides even further performance monitoring utilities:

iostat

sar

vmstat

——

Login into your Linux server shell and execute the above commands. iostat, sar, vmstat can be run like

vmstat 1 10

where 1 is the time between monitoring load, 10 is the count  how many times the data need to be read.

We got an email last night from a fellow reader asking what tcp_fin_timeout and tcp_keepalive_time we are usually using from small to medium blogs.

At this time, we keep the following settings:

net.ipv4.tcp_fin_timeout = 60
net.ipv4.tcp_keepalive_time = 60

Getting the following bug…. bad luck :(
Running Centos, 32 bit

BUG: warning at drivers/net/e1000e/ich8lan.c:361/e1000_acquire_swflag_ich8lan() (Not tainted)
[] e1000_acquire_swflag_ich8lan+0x6e/0×86 [e1000e]
[] e1000e_read_kmrn_reg+0×10/0x4e [e1000e]
[] e1000e_gig_downshift_workaround_ich8lan+0×23/0×66 [e1000e]
[] e1000_intr_msi+0×46/0xd4 [e1000e]
[] handle_IRQ_event+0×23/0×49
[] __do_IRQ+0×84/0xd6
[] do_IRQ+0×93/0xae
[] common_interrupt+0x1a/0×20
[] e1000e_mc_addr_list_update_generic+0×95/0×148 [e1000e]
[] e1000_mc_addr_list_update+0xe/0×12 [e1000e]
[] e1000_set_multi+0xcd/0xd5 [e1000e]
[] __dev_mc_upload+0x1d/0x1e
[] dev_mc_upload+0×28/0x3b
[] dev_open+0×44/0×62
[] dev_change_flags+0×47/0xe4
[] devinet_ioctl+0x24e/0×542
[] sock_ioctl+0×191/0x1b3
[] sock_ioctl+0×0/0x1b3
[] do_ioctl+0x1c/0x5d
[] vfs_ioctl+0x24a/0x25c
[] sys_ioctl+0×48/0x5f
[] syscall_call+0×7/0xb
=======================
pci0000:00: eth0: Link is Up 100 Mbps Full Duplex, Flow Control: RX
pci0000:00: eth0: 10/100 speed: disabling TSO

We are fine tuning sysctl network stack on Linux OS and  change keepalive intervals for much better “response” times and load. These settings are lowered on our all managed performance servers that do host large sites – Drupal, WordPress and custom LAMP services.

Enter the following lines in /etc/sysctl.conf

net.ipv4.tcp_keepalive_probes = 2
net.ipv4.tcp_keepalive_intvl = 30

And execute “sysctl -p” to set these settings in real-time.

By default it’s for Linux (Fedora, Redhat, Centos):

net.ipv4.tcp_keepalive_probes=9
net.ipv4.tcp_keepalive_intvl = 75

Enjoy!