fsck is used to check and optionally repair one or more Linux file systems. File system can be a device name (e.g. /dev/sda2), a mount point (e.g. /, /usr,… ), or an ext2 label or UUID specifier. By default, the fsck will try to handle filesystems on different physical disk drives in parallel to reduce the total amount of time needed to check all of the filesystems.

I hope you won’t need this instructions but in case you have some problems …

First, keep in mind that running fsck on a mounted file system can result in data corruption so use this commands at your own risk. I’m not responsible for any damage you may cause.

To avoid this, you can:

• Change the running state of the system to single user mode and unmount the file system

(Note: Skip this option if you need to check / part.)

• Boot the computer into Rescue Mode using the installation CD

1) Single user mode

Change the run level to 1 and unmont the file system with (I suppose the /home was affected with problems)

# init 1
# umount /home

Then run fsck with

# fsck /dev/sda2

2) Rescue mode with install CD

Insert the Installation CD into the drive and reboot your system with

# reboot

Wait a while and after booting from the Installation CD type in command prompt:

# linux rescue nomount

Note: nomount directive will skip mounting so you can safely run fsck on your filesystem

Then enter

# fsck -yvf /dev/sda2

where /dev/sda2 is problematic partition.
Note: -yvf means
-y attempt to fix any detected filesystem corruption automatically (answer all questions with “yes”)
-f force a check even if reported in a clean state
-v verbose output, including all file system-specific commands that are executed (you will see all errors and recovery process)

LVM Partitions
Situation with LVM partitions is a little different. To run fsck on lvm partitions first you need to find the pv’s, vg’s, lv’s and then to activate them.

# lvm pvscan
# lvm vgscan
# lvm lvchange -ay VolGroup00
# lvm lvscan
 
# fsck -yfv /dev/VolGroup00/LogVol00

fsck can be scheduled on boot with

# shutdown -Fr now

If the Linux server time and date is wrong and you need to set it to new values from the shell prompt, you should use date command. You must login as root user to use date command. Also from command prompt you can check and set system clock (hwclock command).

Use the following syntax to set new data and time:

# date -s "16 MAR 2010 09:36:00"

Also you can use next syntax:

date set="16 MAR 2010 09:36:00"

After you set date and time, it is good idea to check your system clock with

# hwclock --show

If there is some difference you can sync system date and time with

# hwclock --systohc

If you want to test server performance, you can think about SysBench. SysBench is a modular, cross-platform and multi-threaded benchmark tool for evaluating OS parameters that are important for a system running a database under intensive load. The idea of this benchmark suite is to quickly get an impression about system performance without setting up complex database benchmarks or even without installing a database at all.

Current features allow to test the following system parameters:

* file I/O performance
* scheduler performance
* memory allocation and transfer speed
* POSIX threads implementation performance
* database server performance (OLTP benchmark)
(Primarily written for MySQL server benchmarking, SysBench will be further extended to support multiple database backends, distributed benchmarks and third-party plug-in modules)

I couldn’t find CentOS RPM so here are few tips how to install it manually.

Download Sysbench (current version is 0.4.12)

# wget http://garr.dl.sourceforge.net/sourceforge/sysbench/sysbench-0.4.12.tar.gz

Then unpack it and install with

# tar -xvzf sysbench-0.4.12.tar.gz
# cd sysbench-0.4.12
# libtoolize --force --copy 
# ./autogen.sh
# ./configure
# make
# make install

To test CPU performance you can try

# sysbench --test=cpu --cpu-max-prime=20000 run

For MySQL test, you’ll need to prepare database for testing with

# sysbench --test=oltp --mysql-table-engine=innodb --oltp-table-size=500000 --mysql-user=test_database --mysql-password=test_database_password --mysql-socket=/var/lib/mysql/mysql.sock prepare

(replace test_database with valid username and test_database_password with valid password)

This command will create sample table inside test_database and it will have 500 000 rows (InnoDB engine).

sysbench 0.4.12:  multi-threaded system evaluation benchmark
 
No DB drivers specified, using mysql
Creating table 'test-database'...
Creating 500000 records in table 'test-database'...

Now to start read test

# sysbench --num-threads=16 --max-requests=100000 --test=oltp --oltp-table-size=500000 --mysql-socket=/var/lib/mysql/mysql.sock --oltp-read-only --mysql-user=test_database --mysql-password=test_database_password run

For read-write test you can try

# sysbench --num-threads=16 --max-requests=10000 --test=oltp --oltp-table-size=500000 --mysql-socket=/var/lib/mysql/mysql.sock --oltp-test-mode=complex --mysql-user=test_database --mysql-password=test_database_password run

More info about specific parameters can be found in official docs (http://sysbench.sourceforge.net/docs/)