[root@patrick ~]# ls -l anaconda-ks.cfg
-rw-r--r--  1 root root 1483 Aug  6  2005 anaconda-ks.cfg

So the first time I installed Linux on that server is 6 Aug 2005, if I haven’t reinstalled the OS, then that date will be also the date for the hardware is being use for the first time also. Now after 5 happy years running 24/7. Isn’t about time to replace the hardware with the new and faster one?

Related posts:

1. Check Linux Foldersize I don’t know if there are any utilities out there...
2. Add Swap on Linux From: FelipeCruz To add swap partition use command # mkswap...
3. Install jabberd on CentOS Create your own Instant Messaging Server!! The company need it...

Jun 19 15:51:56 dbdemo smartd[2117]: Device: /dev/hdb, 37997 Currently unreadable (pending) sectors
Jun 19 15:51:56 dbdemo smartd[2117]: Device: /dev/hdb, 37997 Offline uncorrectable sectors
Jun 19 16:21:57 dbdemo smartd[2117]: Device: /dev/hdb, 38010 Currently unreadable (pending) sectors
Jun 19 16:21:57 dbdemo smartd[2117]: Device: /dev/hdb, 38010 Offline uncorrectable sectors
Jun 19 16:23:47 dbdemo kernel: hdb: task_in_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }

Luckily, uncle Google very nice to me and show me article on Howtoforge.com, following the article I am confident to replace failed harddisk. Below is my Raid configuration:

Array: /dev/md0; Member: /dev/hda1 /dev/hdb1; Raid: 1; Filesystem: ext3fs; mountpoint: /boot
Array: /dev/md1; Member: /dev/hda2 /dev/hdb2; Raid: 0; Filesystem: swap; mountpoint: swap
Array: /dev/md2; Member: /dev/hda3 /dev/hdb3; Raid: 1; Filesystem: ext3fs; mountpoint: /

Summary of Howtoforge’s article is below:

# mdadm --manage /dev/md0 --fail /dev/hdb1
# mdadm --manage /dev/md1 --fail /dev/hdb2
# mdadm --manage /dev/md2 --fail /dev/hdb3
# mdadm --manage /dev/md0 --remove /dev/hdb1
# mdadm --manage /dev/md1 --remove /dev/hdb2
# mdadm --manage /dev/md2 --remove /dev/hdb3

I got problem when removing /dev/hdb2 from array /dev/md1, since its Raid0 and used for swap. I have tried to disabled swap, but whatever I do, I get error “Resource busy..” What the heck I shutdown the server since it’s only swap partition.

Replace the broken HDD, turn on the server and type below command:

# sfdisk -d /dev/hda | sfdisk /dev/hdb
# fdisk -l
# mdadm --manage /dev/md0 --add /dev/hdb1
# mdadm --manage /dev/md2 --add /dev/hdb3

And for /dev/md1 (the swap partition), I am following Jackie’s article on NetworkNewz:

# mdadm --create --verbose /dev/md1 --level=0 --raid-devices=2 /dev/hdb1 /dev/hda1
# mdadm --detail --scan > /etc/mdamd.conf
# mkswap /dev/md1
# swapon /dev/md1

Done, now it’s good as new

Related posts:

1. Add Swap on Linux From: FelipeCruz To add swap partition use command # mkswap...

To add swap partition use command

# mkswap /dev/hdb2
# swapon /dev/hdb2
# vi /etc/fstab
# /dev/hdb2   swap   swap   defaults   0 0

With assuming you already create /dev/hdb2 partition before using fdisk and chose swap filesystem.

To add a swap file:

# dd if=/dev/zero of=/swapfile bs=1024 count=65536
# mkswap /swapfile
# swapon /swapfile
# vi /etc/fstab
#  /swapfile   swap   swap    defaults        0 0

Check swap status:

# cat /proc/swaps
# free

Related posts:

1. Check Linux Foldersize I don’t know if there are any utilities out there...
2. hdd partition for cpanel server Simple Harddisk partition, assume only for vps and have or...

After googling a while I found the solution on Always Hosting Forum. And the only way to resolve it permanently is by running fsck utility on linux single mode before mounting the harddrive. Yes there is a temporarily solution by restarting the server, but when the system is about to write into failed inode it will freeze, daemons stop responding, connections rejected, etc to make sure it wont cause more broken inode will lead into data corruption.
Thanks God, after running fsck, all data still intact, I haven’t found corrupted/missing data (yet). Important lesson is when your system is hang, please try using soft-restart (init 6, shutdown -r) and not just pressing the Power button directly (hard-restart) which can cause data instability.

It’s a very hard lesson.

@weekly find /home2/backup -mtime +7 -type d -depth -exec rm -r {} \;

Related posts:

1. Check Linux Foldersize I don’t know if there are any utilities out there...

# ./foldersize.sh /var/lib/mysql

To show how much each folder/database have. So here foldersize.sh script

#!/bin/bash
ls -1d $1/* > /root/folderlist.txt
while read FOLDER
do
      du -hs $FOLDER
done < /root/folderlist.txt

Kusma!!!, Now I know how to make old mysql front ver 2.5 which created on 2002 and using libmysql from mysql 3.23 accessing mysql5 server. I just realized this when Andry complaining and whining can’t access Ivory DB on new server which also mysql5 using his old mysql-front. He said, he can use mysql-front 2.5 accessing Mysql server 5 on his localhost and BOX server.

Now, if I can make sure old mysql-front program accessing mysql server 5 this will make major leap!!!
old zeos component (which also based on libmysql from mysql 3.23) for delphi 5 now can access Mysql 5 server. God know what happen when mysql5 take control, replacing old and dull mysql4. We can make master-master replication, clustering mysql server or make mysql serve faster by running mysql proxy. But it depend on how much the company will spend money for investing on hardware (and pay me of course to execute it)

Kusma, do you want to know how to enable Mysql server 5 to acknowledge old mysql-front program to access it? if so comment to this article please

Related posts:

1. Running MySQL 4 and MySQL 5 Concurrently From Howtoforge.com’s article: Running MySQL 4 And MySQL 5 Concurrently...