Mohd Amree

Rails, Unicorn and Nginx on Slackware

2013-01-06T00:00:00-08:00

Rails, Unicorn and Nginx on Slackware

I was trying to deploy a Rails application using Apache and got into some problems when I tried to configure the app so that Apache will be the one that serves the precompiled assets (javascripts, css, images and others). Since I’m on tight deadline (yeah, I should’ve tested production mode much more earlier), I tried my luck with Nginx and it worked easily without any hassle. So, this is how I did it.

Goals

Precompiled assets will be served by Nginx and not the Rails server itself.
Assets will be served in gzip.

Environments

Slackware v14
Rails v3.2.10
Ruby v1.9.3

Guides

First of all, install nginx from Slackbuild.

Be sure to turn off Apache’s startup script if you have it installed:

chmod -x /etc/rc.d/rc.httpd

This is to ensure no conflict since both of them by default will use port 80.

Put unicorn in your Gemfile and run bundle.

I’m putting my Rails app in /opt/neuro, so, adjust it accordingly.

Create nginx.conf in /opt/neuro/config/nginx.conf:

upstream unicorn {
  server unix:/tmp/unicorn.neuro.sock fail_timeout=0;
}

server {
  listen 80 default deferred;
  server_name neuro.husmnet
  root /opt/neuro/public;

  location ^~ /assets/ {
    root /opt/neuro/public;
    gzip_static on;
    expires max;
    add_header Cache-Control public;
  }

  try_files $uri/index.html $uri @unicorn;

  location @unicorn {
    proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
    proxy_set_header Host $http_host;
    proxy_redirect off;
    proxy_pass http://unicorn;
  }

  error_page 500 502 503 504 /500.html;
  client_max_body_size 4G;
  keepalive_timeout 10;
}

Create unicorn.rb in /opt/neuro/config/unicorn.rb

root = "#{Dir.pwd}"

# Define worker directory for Unicorn
working_directory root

# Location of PID file
pid "#{root}/tmp/pids/unicorn.pid"

# Define Log paths
stderr_path "#{root}/log/unicorn.log"
stdout_path "#{root}/log/unicorn.log"

# Listen on a UNIX data socket
listen "/tmp/unicorn.neuro.sock", :backlog => 64
# houllisten 8080, :tcp_nopush => true

worker_processes 2

# Load rails before forking workers for better worker spawn time
preload_app true

# Restart workes hangin' out for more than 240 secs
timeout 240

Replace `/etc/nginx/nginx.conf` with this content:

# user  root;
worker_processes  1;

error_log  /var/log/nginx/error.log;
#error_log  logs/error.log  notice;
#error_log  logs/error.log  info;

#pid        logs/nginx.pid;

events {
  worker_connections  1024;
}

http {
  include       mime.types;
  default_type  application/octet-stream;

  sendfile        on;

  keepalive_timeout  65;

  gzip  on;

  include /etc/nginx/sites-enabled/*;
}

Create sites-enabled directory in /etc/nginx and create a softlink to the nginx.conf in our app:

$ mkdir /etc/nginx/sites-enabled
$ ln -s /opt/neuro/config/nginx.conf neuro.conf

Before starting it for the first time, let us monitor important logs (open it using different terminals) :

$ tail -f /var/log/nginx/error.log
$ tail -f /opt/neuro/log/unicorn.log

Make sure you’ve precompiled your assets:

$ rake assets:clean
$ rake assets:precompile

Start unicorn:

$ cd /opt/neuro
$ unicorn -c config/unicorn.rb -E production -D

Start nginx:

$ nginx

Congratulation! Make sure there’s no error in your logs. If there’re, you can use these commands to stop nginx and unicorn to start everything back:

$ nginx -s stop
$ killall unicorn

Let’s test the gzip compression using curl:

$ curl -LI --compressed http://neuro.husmnet/

You’ll get something like this (notice the gzip info):

HTTP/1.1 200 OK
Server: nginx/1.2.2
Date: Sat, 05 Jan 2013 21:00:26 GMT
Content-Type: text/html; charset=utf-8
Connection: keep-alive
Status: 200 OK
X-UA-Compatible: IE=Edge,chrome=1
ETag: "a66ac1d43d8f07ecc0737e64dd8a3366"
Cache-Control: max-age=0, private, must-revalidate
Set-Cookie: _neuro_session=BAh7B0kiD3Nlc3Npb25faWQGOgZFRkkiJTQ2YTU2NjE3MDY0Y2RkNzk0Yzk3ODhhNDJlYmQ3ODA3BjsAVEkiEF9jc3JmX3Rva2VuBjsARkkiMVNvOXVTS2I2RXJIcTljbS9WeGRDODZyTUpVZENXL2NPVnJhTGR0V2xydGM9BjsARg%3D%3D--173e9d56a5cc434eb675b6626d90aced1cc17cd6; path=/; HttpOnly
X-Request-Id: 760e233de4980dc75169f2c68a53dc31
X-Runtime: 0.017885
X-Rack-Cache: miss
Content-Encoding: gzip

I admit, this is a very simple config, but it’s good enough for a beginner like me to get started. So, good luck!

MyKad Codes Takedown

2012-11-04T00:00:00-07:00

MyKad Codes Takedown

Last month my colleague and I had a meeting with Jabatan Pendaftaran Negara. From that meeting I learned people at JPN are not really happy with the code to read MyKad available publicly on the Internet especially not in the Github due to some reasons which I don’t think should be mentioned here.

Since I’ve posted those codes (from Xenon and myself) on the Github, I had to take them down permanently. IMHO I should respect their decision since it’s not mine to begin with. I love how those codes have helped a lot of people especially students as they do not really have the sufficient funds to buy the expensive SDK. I would love to comment further on this decision, but I don’t think I should :(.

Saya yang menurut perintah,

Mohd Amree

Getting Started with Xen on Slackware

2012-09-12T00:00:00-07:00

Getting Started with Xen on Slackware

Install Slackware

I’ll be doing this on Slackware v13.37 64 bit.

Create 1 partition for the whole disk
Set its type to 8e (Linux LVM)

Run these commands to setup the LVM partitions:

$ pvcreate /dev/sda1
$ vgcreate vg01 /dev/sda1
$ lvcreate -L 20G -n root vg01
$ lvcreate -L 20G -n iso vg01
$ lvcreate -L 8G -n swap vg01
$ mkswap /dev/vg01/swap
$ setup

Mount /dev/vg01/root as /
Mount /dev/vg01/iso as /iso
Install all packages from a, ap, d, f, l, n, t, tcl, x

Create custom initrd after the setup complete:

$ chroot /mnt
$ /usr/share/mkinitrd/mkinitrd_command_generator.sh > init
$ sh init
$ ln -sf /boot/vmlinuz-generic-2.6.37.6 /boot/vmlinuz

Add initrd option in your /etc/lilo.conf:

# /etc/lilo.conf

image = /boot/vmlinuz
  initrd = /boot/initrd.gz # Add this line
  root = /dev/vg01/root
  label = slackware
  read-only  # Partitions should be mounted read-only for checking

Install the new bootloader and reboot:

$ lilo
$ exit
$ reboot

Setup Dom0 (Host)

Install Xen

Install acpica
Install xen

Build Xen Kernel

We’ll be using kernel v3.2.28 so that no patching will be required to enable Xen and for that we’ll get the one from Slackware’s repo. By the time of this writing, Slackware v14 has not been released yet (it’s in the current).

$ wget http://slackware.osuosl.org/slackware-current/source/k/linux-3.2.28.tar.xz
$ tar Jxvf linux-3.2.28.tar.xz
$ mv linux-3.2.28 linux-3.2.28-xen
$ mv linux-3.2.28-xen/ /usr/src/
$ cd /usr/src
$ rm linux
$ ln -s linux-3.2.28-xen linux

Get Slackware’s default kernel config for a start. From there, you must add all Xen’s related kernel configs. You can refer to Xen’s wiki on what kernel config that needs to be enabled.

You can also get my configs from this gist. It’s based on 3.2.28 kernel (you can use Gist’s diff to see what config I’ve added).

Whatever you chose, make sure you’ll change the LOCALVERSION so that you won’t replace your current kernel. Use cat /proc/cpuinfo | grep -i 'processor' | wc -l + 1 to get your cpu core count to be used with make -j.

$ cd linux
$ wget http://slackware.osuosl.org/slackware-current/source/k/config-x86_64/config-generic-3.2.28.x64
$ cp config-generic-3.2.28.x64 .config
$ make oldconfig
$ make menuconfig
$ make -j13 bzImage modules
$ make modules_install
$ cp System.map /boot/System.map-xen-3.2.28
$ cp .config /boot/config-xen-3.2.28
$ cd /boot
$ rm System.map
$ ln -s System.map-xen-3.2.28 System.map

Xen + Dom0 + Initrd

Since LILO doesn’t support “module” directive of Grub, we need to use mbootpack.

Install mbootpack

You can use /usr/share/mkinitrd/mkinitrd_command_generator.sh for mkinitrd recommendation. Create an initrd for Xen’s kernel:

$ depmod 3.2.28-xen
$ mkinitrd -c -k 3.2.28-xen -f ext4 -r /dev/vg01/root -m mptbase:mptscsih:mptsas:usbhid:ehci-hcd:uhci-hcd:jbd2:mbcache:ext4 -L -u -o /boot/initrd-xen.gz

Create the boot image using mbootpack:

$ cd /boot
$ gzip -d -c /boot/xen-4.1.2.gz > /boot/xen-4.1.2
$ gzip -d -c /boot/initrd-xen.gz > /boot/initrd-xen
$ mbootpack -o /boot/vmlinuz-xen-3.2.28 -m /usr/src/linux-3.2.28-xen/vmlinux -m /boot/initrd-xen /boot/xen-4.1.2
$ ln -s vmlinuz-xen-3.2.28 vmlinuz-xen

With mbootpack, we don’t have to specify initrd option. Add these new configs into your /etc/lilo.conf:

# /etc/lilo.conf

image = /boot/vmlinuz-xen
  root = /dev/vg01/root
  label = slackware-xen
  append="-- nomodeset"
  read-only

Put these configs on the top of your lilo.conf:

default = slackware-xen
timeout = 30

Run lilo and make sure there’s no error from the output:

$ lilo
Warning: LBA32 addressing assumed
Added slackware
Added slackware-xen *
One warning was issued.

Put these settings in your /etc/rc.d/rc.local:

# /etc/rc.d/rc.local

if [ -d /proc/xen ]; then
  if [ -x /etc/rc.d/rc.xencommons ]; then
    echo "Starting XEN commons:  /etc/rc.d/rc.xencommons"
    /etc/rc.d/rc.xencommons start
  fi
  if [ -x /etc/rc.d/rc.xendomains ]; then
    echo "Starting XEN domains:  /etc/rc.d/rc.xendomains"
    /etc/rc.d/rc.xendomains start
  fi
fi

And these commands in your /etc/rc.d/rc.local_shutdown:

# /etc/rc.d/rc.local_shutdown

if [ -d /proc/xen ]; then
  if [ -x /etc/rc.d/rc.xendomains ]; then
    echo "Stopping XEN domains:  /etc/rc.d/rc.xendomains"
    /etc/rc.d/rc.xendomains stop
  fi
  if [ -x /etc/rc.d/rc.xencommons ]; then
    echo "Stopping XEN commons:  /etc/rc.d/rc.xencommons"
    /etc/rc.d/rc.xencommons stop
  fi
fi

After the reboot, you should get something like this:

$ uname -r
3.2.28-xen

$ cat /proc/xen/capabilities
control_d

$ cat /sys/hypervisor/properties/capabilities
xen-3.0-x86_64 xen-3.0-x86_32p hvm-3.0-x86_32 hvm-3.0-x86_32p hvm-3.0-x86_64

$ xl list
Name                                        ID   Mem VCPUs  State Time(s)
Domain-0                                     0 15243    12     r-----       9.0

Disable autosave and restore of DomU:

# /etc/default/xendomains

XENDOMAINS_RESTORE=false
XENDOMAINS_SAVE=""

Setup DomU (Guest)

I’m going to install a Slackware v13.37 64 bit as the first guest. I’ll be using an ISO image that I’ve copied into my /iso.

$ cd /iso
$ wget http://mirrors.xmission.com/slackware/slackware64-13.37-iso/slackware64-13.37-install-dvd.iso

For a start, we need to create a config file for the Slackware installation:

$ mkdir ~/machines
$ cd ~/machines
$ vim slackware-install

Put these initial configs in your /root/machines/slackware-install:

kernel = 'hvmloader'
builder='hvm'
memory = 1024
name = "slackware"
vif = [ 'mac=00:16:3E:AD:81:AE, bridge=virbr0, model=e1000' ]
dhcp = "dhcp"
disk = ['phy:/dev/vg01/slackware,hda,w', 'file:/iso/slackware64-13.37-install-dvd.iso,hdc:cdrom,r']
boot='dc'
sdl=0
opengl=1
vnc=1
vncpasswd=''
serial='pty'

As you can see, there’re various of thing thats we need to take care of before we start the installation. First of all the networking.

Generate Random MAC

Use this command (thanks to Unixtitan):

perl -e 'printf "00:16:3E:%02X:%02X:%02X\n", rand 0xFF, rand 0xFF, rand 0xFF'

Setup Networking Bridging

There are a few methods for a DomU to access the network. For this guide, I’ll use network bridging.

Put these configs (before starting anything related to Xen) into your /etc/rc.d/rc.local to enable network bridge:

# /etc/rc.d/rc.local

ifconfig eth0 up
ifconfig eth0 0.0.0.0

brctl addbr virbr0
brctl addif virbr0 eth0
ifconfig virbr0 192.168.1.10 netmask 255.255.255.0
route add default gw 192.168.1.1

Clear out any values set in your /etc/rc.d/rc.inet1.conf. Usually the value of IPADDR[0], NETMASK[0] and GATEWAY.

# /etc/rc.d/rc.inet1.conf

# Config information for eth0:
IPADDR[0]=""
NETMASK[0]=""
USE_DHCP[0]=""
DHCP_HOSTNAME[0]=""

# Default gateway IP address:
GATEWAY=""

This configs assume your main networking interface is eth0. Change other settings based on your network environment. Reboot your Dom0.

You’ll get something that looks like this after the reboot (if you got it right).

$ ifconfig

eth0      Link encap:Ethernet  HWaddr bc:30:5b:db:e2:99
          inet6 addr: fe80::be30:5bff:fedb:e299/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:163836 errors:0 dropped:0 overruns:0 frame:0
          TX packets:337 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:18347806 (17.4 MiB)  TX bytes:30535 (29.8 KiB)
          Interrupt:36 Memory:d6000000-d6012800

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:4 errors:0 dropped:0 overruns:0 frame:0
          TX packets:4 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:336 (336.0 B)  TX bytes:336 (336.0 B)

virbr0    Link encap:Ethernet  HWaddr bc:30:5b:db:e2:99
          inet addr:192.168.1.10  Bcast:192.168.1.255  Mask:255.255.255.0
          inet6 addr: fe80::be30:5bff:fedb:e299/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:147426 errors:0 dropped:506 overruns:0 frame:0
          TX packets:129 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:12417385 (11.8 MiB)  TX bytes:16779 (16.3 KiB)

Enable X11 Forwarding in Dom0

Install vncviewer

Uncomment or create these options in your /etc/ssh/sshd_config:

# /etc/ssh/sshd_config

X11Forwarding yes

Restart your ssh server:

$ /etc/rc.d/rc.sshd restart
$ exit

Reconnect to Dom0 using -Y option:

$ ssh -Y root@192.168.1.10
$ xclock

If you can see a clock ticking, that means the X11 forward is working.

Prepare space for DomU

I’m going to store my DomU in the a new logical volume.

$ lvcreate -L 15G -n slackware vg01

Install Slackware in DomU

$ cd machines
$ xl create slackware

If everything is good, you should see slackware when you run xl list.

$ xl list
Name                                        ID   Mem VCPUs  State Time(s)
Domain-0                                     0 13988    12     r-----      40.0
slackware                                    2  1019     1     -b----       4.6

Install Slackware through VNC:

$ xl vncviewer slackware

Just install as if you’re installing on a normal server. halt after you’ve finished your installation.

Start Slackware in DomU

$ cd ~/machines
$ cp slackware-install slackware
$ vim slackware

Change these two lines in your config file to make sure it’ll boot from the hardisk:

# /root/machines/slackware

disk = ['phy:/dev/vg01/slackware,hda,w']
boot='c'

Start the guest:

$ xl create slackware

Enable Console in DomU

It’s pretty annoying to use vnc everytime we need to access our DomU directly. It’s possible to connect to our DomU using console connection:

Uncomment these lines:

# /etc/inittab
s1:12345:respawn:/sbin/agetty -L ttyS0 9600 vt100

# /etc/securetty
ttyS0

$ reboot

For Ubuntu’s DomU, please check out https://help.ubuntu.com/community/SerialConsoleHowto.

You can now connect to your Slackware DomU using:

$ xl console slackware

Additional Info

Disable auto save and restore of domUs on host reboot:

$ vim /etc/default/xendomains

XENDOMAINS_RESTORE=false
XENDOMAINS_SAVE=""

In order to run xl shutdown guest so that your DomU will shutdown gracefully, you need to use PVHVM drivers. You can refer here for the related kernel configs that you need to enable in your kernel.

There are still many things that I’d like to add, but I think it’s good enough for now. After all, this post is about getting started, not a complete guide. I’ll update this post if I found anything worth of sharing.

Good luck!

More references:

Corosync and Pacemaker in Slackware

2012-05-14T00:00:00-07:00

Corosync and Pacemaker in Slackware

This will be multi part post about high availability solution for Slackware. My first post will be about Corosync and Pacemaker.

You need to combine Corosync and Pacemaker with other distributed storage system such as DRBD/OCFS2/GFS. I’ll talk about these stacks in another post.

Goal:

A MySQL server will always be available at the same IP even though it’s actually down (another server will take over automatically without the needs for manual intervention).

Environments:

Slackware v13.37

Two nodes will be used:

Node 1:
192.168.1.101

Node 2:
192.168.1.102

Cluster/Main/Failover IP:
192.168.1.100

The MySQL data is not syncronized, this post is just about Corosync and Pacemaker.

Guides:

Download and install these packages (by this order) in both nodes:
- http://slackbuilds.org/repository/13.37/libraries/libnet/
- http://slackbuilds.org/repository/13.37/libraries/libesmtp/
- http://slackbuilds.org/repository/13.37/system/clusterglue/
- http://slackbuilds.org/repository/13.37/system/clusterresourceagents/
- http://slackbuilds.org/repository/13.37/system/corosync/
- http://slackbuilds.org/repository/13.37/system/pacemaker/
I strongly suggest you build these packages one by one just to be sure there are no missing dependencies. BTW, some script adjustments are needed for Cluster Resource Agents but I’m sure you guys can handle it ;-)
It would be easier for the next steps if password-less login with OpenSSH is enabled. In your Node 1:
```
ssh-keygen -t rsa
ssh-copy-id -i ~/.ssh/id_rsa.pub root@192.168.1.102
```
Generate an authentication key for Corosync:
```
corosync-keygen
```
If you’re connecting remotely, pressing your keyboard won’t do any good. The fastest way would be typing directly into the server. The other way is running find . on your / directory (press Control + C when the key has been generated).

Copy the new generated authentication key to Node 2:

scp /etc/corosync/authkey 192.168.1.102:/etc/corosync

Copy the default corosync configuration file:

cp /etc/corosync/corosync.conf.example /etc/corosync/corosync.conf

Replace bindnetaddr and logfile (optional):

bindnetaddr: 192.168.1.0
logfile: /var/log/corosync

You can checkout the reference about those values. From corosync:

If the local interface was 10.12.12.93 and the netmask was 255.0.0.0, Totem would execute the logical operation 10.12.12.93 & 255.0.0.0 and produce the value 10.0.0.0. This value would be compared against bindnetaddr and bind Totem to the NIC that matches.
This can cause confusion if netmask or bindnetaddr are not set properly.
In the example above, if bindnetaddr is 10.12.12.0, the network interface will never be matched. If bindnetaddr is 10.0.0.0 the interface will be matched.

Copy corosync.conf to Node 2:

scp /etc/corosync/corosync.conf 192.168.1.102:/etc/corosync

Create pacemaker file so that Corosync will automatically load Pacemaker when it’s started:

touch /etc/corosync/service.d/pacemaker

Put these configs in that file:

service {
  # Load the Pacemaker Cluster Resource Manager
  name: pacemaker
  ver:  0
}

Copy the pacemaker file to Node 2:

scp /etc/corosync/service.d/pacemaker 192.168.1.102:/etc/corosync/service.d/

Start your Corosync and let the magic begins:
```
/etc/rc.d/rc.corosync start
```

Check your log for any error:

tail -f /var/log/corosync

Check your process list:

ps auxf

Corosync should also load other processes automatically:

root      2008  0.5  3.4  52668  3964 ?        Ssl  13:55   0:00 corosync
root      2015  0.0  1.9  12140  2248 ?        S    13:55   0:00  \_ /usr/lib/heartbeat/stonithd
226       2016  0.3  3.3  13004  3796 ?        S    13:55   0:00  \_ /usr/lib/heartbeat/cib
root      2017  0.0  1.6   6812  1848 ?        S    13:55   0:00  \_ /usr/lib/heartbeat/lrmd
226       2018  0.1  2.2  12404  2540 ?        S    13:55   0:00  \_ /usr/lib/heartbeat/attrd
226       2019  0.0  1.7   8664  2032 ?        S    13:55   0:00  \_ /usr/lib/heartbeat/pengine
226       2020  0.1  2.5  12528  2904 ?        S    13:55   0:00  \_ /usr/lib/heartbeat/crmd

Monitor your cluster using Pacemaker tools:

crm status

It should be something like this:

============
Last updated: Sun May 13 13:57:43 2012
Stack: openais
Current DC: node1 - partition with quorum
Version: 1.1.1-b9b672590e79770afb63b9b455400d92fb6b5d9e
2 Nodes configured, 2 expected votes
0 Resources configured.
============

Online: [ node1 node2 ]

Give them some time to online if they’re offline.

Put some main configurations to your cluster:
```
crm
configure
property stonith-enabled=false
property no-quorum-policy=ignore
commit
quit
```
If you’re getting some errors such as ERROR: cib-bootstrap-options: attribute last-lrm-refresh does not exist, just proceed. It maybe a bug.

We had to disable stonith since we just want our Pacemaker to be running. However, in real production environment, you really need to configure stonith, you can read more about it here.

We also need to ignore quorum policy since we’re only using 2 nodes and you can read more about it here.

You can see your new configuration by running:
```
	crm configure show
```
Which will output:
```
	node node1
	node node2
	property $id="cib-bootstrap-options" \
		dc-version="1.1.1-b9b672590e79770afb63b9b455400d92fb6b5d9e" \
		cluster-infrastructure="openais" \
		expected-quorum-votes="2" \
		stonith-enabled="false" \
		last-lrm-refresh="1336919205" \
		no-quorum-policy="ignore"
```
If you accidentally put some wrong configurations and don’t know how to edit it, you can use crm configure edit to change your configurations directly but this method is highly not recommended since it’s error-prone.

It’s time to configure our main/failover/cluster IP (our client will use this IP, not the nodes IP):

crm
configure
primitive ip ocf:heartbeat:IPaddr params ip="192.168.1.100" op monitor interval=10s
commit

If everyting goes well, you should be able to ping the cluster IP (192.168.1.100) and crm status should yield this result:

============
Last updated: Sun May 13 14:28:19 2012
Stack: openais
Current DC: node1 - partition with quorum
Version: 1.1.1-b9b672590e79770afb63b9b455400d92fb6b5d9e
2 Nodes configured, 2 expected votes
1 Resources configured.
============

Online: [ node1 node2 ]

ip     (ocf::heartbeat:IPaddr):        Started node1

We’ll now setup MySQL monitoring with Pacemaker. But before that, make sure you:

Installed MySQL in both of the nodes.

Able to connect to your MySQL from other than localhost:
```
	mysql -u root -p -h 192.168.1.101
	mysql -u root -p -h 192.168.1.102
```
You can use this command to allow any host to connect to your MySQL:
```
	GRANT ALL PRIVILEGES ON *.* TO 'root'@'%' IDENTIFIED BY 'password' WITH GRANT OPTION;
	FLUSH PRIVILEGES;
```
Created a database in Node 1 and Node 2. For an example, a databased named node1 in Node 1 and node2 in Node 2. This is just for verification.

Add this resource:

crm
configure
primitive mysql ocf:heartbeat:mysql \
params binary="/usr/bin/mysqld_safe" config="/etc/my.cnf" user="mysql" pid="/var/run/mysql/mysql.pid" datadir="/var/lib/mysql" socket="/var/run/mysql/mysql.sock" \
op monitor interval="30s" timeout="30s" \
op start interval="0" timeout="120" \
op stop interval="0" timeout="120"
commit
quit

The parameter above is purely based on the standard Slackware’s MySQL package. So make sure you’ve created /etc/my.cnf which is not available by default. Just copy from the default file:

cp /etc/my-small.cnf /etc/my.cnf

Your latest crm status would show something like this:
```
============
Last updated: Mon May 14 01:13:23 2012
Stack: openais
Current DC: node1 - partition with quorum
Version: 1.1.1-b9b672590e79770afb63b9b455400d92fb6b5d9e
2 Nodes configured, 2 expected votes
2 Resources configured.
============

Online: [ node1 node2 ]

 ip	(ocf::heartbeat:IPaddr):	Started node1
 mysql	(ocf::heartbeat:mysql):	Started node2
```
As you can see, mysql has been started on Node 2. Actually it doesn’t matter in which node it will start first (for this tutorial, not for the production server), what important is that if one of the nodes is down, the other node should start its MySQL automatically. You can test this situation by running these commands in your Node 2 to simulate a node failure:
```
crm
node
standby
quit
```
crm status would show something like this (give Node 1 some time before it starts its MySQL):
```
============
Last updated: Mon May 14 01:21:12 2012
Stack: openais
Current DC: node1 - partition with quorum
Version: 1.1.1-b9b672590e79770afb63b9b455400d92fb6b5d9e
2 Nodes configured, 2 expected votes
2 Resources configured.
============

Node node2: standby
Online: [ node1 ]

 ip	(ocf::heartbeat:IPaddr):	Started node1
 mysql	(ocf::heartbeat:mysql):	Started node1
```
Right now, your client can use the cluster IP (192.168.1.100) to connect to your MySQL. The client won’t realize which node it connected to. In this case, he/she will connect to Node 2 if both of them (the nodes) are online. If Node 2 is offline, 192.168.1.100 will automatically connect the client to MySQL in 192.168.1.101. If Node 1 is offline, 192.168.1.100 will automatically uses MySQL in Node 2 which is in 192.168.1.102.

To reonline Node 2, just use these commands in your Node 2:
```
crm
node
online
quit
```

However, usually you want to control which MySQL will be up first, either in Node 1 or in Node 2. To make this happen, you need to use colocation:

crm
configure
colocation ip-mysql inf: ip mysql
commit
quit

crm status would show something like this:

============
Last updated: Mon May 14 01:26:41 2012
Stack: openais
Current DC: node1 - partition with quorum
Version: 1.1.1-b9b672590e79770afb63b9b455400d92fb6b5d9e
2 Nodes configured, 2 expected votes
2 Resources configured.
============

Online: [ node1 node2 ]

 ip	(ocf::heartbeat:IPaddr):	Started node1
 mysql	(ocf::heartbeat:mysql):	Started node1

That means, your mysql has been started on Node 1. So, everytime corosync is started on both of the nodes, mysql will be started on Node 1 due to the colocation configuration.

Try turning off Node 1 or Node 2 and see how MySQL switches side from both of the nodes.

I think that’s it, next tutorial should be mainly about DRBD. Good luck!

Using GRUB instead of LILO in Slackware

2012-05-11T00:00:00-07:00

Using GRUB instead of LILO in Slackware

These are the steps I used to replace LILO with GRUB as the boot manager after upgrading my kernel to v3.

Download directly and install:

wget http://slackware.osuosl.org/slackware-13.37/extra/grub/grub-0.97-i486-9.txz
installpkg grub-0.97-i486-9.txz

Then jump to Step 3 and make sure you’ll skip Step 4.

If you haven’t set any mirror for your slackpkg, you need to edit your mirror list:
```
vim /etc/slackpkg/mirrors
```
Add a mirror, for an example, I use the one from Oregon State University:
```
http://slackware.osuosl.org/slackware-13.37/
```
Do not forget the trailing /.
Run the update:
```
slackpkg update
```
Remove LILO
```
removepkg lilo
```
Install GRUB
```
slackpkg install grub
```

Run configuration editor for GRUB (just like liloconfig):

grubconfig

This is my working config from /boot/grub/menu.lst

# GRUB configuration file '/boot/grub/menu.lst'.
# generated by 'grubconfig'.  Fri May 11 02:21:11 2012
#
# The backup copy of the MBR for drive '/dev/sda' is
# here '/boot/grub/mbr.sda.1594'.  You can restore it like this.
# dd if=mbr.sda.1594 of=/dev/sda bs=512 count=1
#
# Start GRUB global section
#timeout 30
#color light-gray/blue black/light-gray
# End GRUB global section
# Linux bootable partition config begins
  title slackware on (/dev/sda1)
  root (hd0,0)
  kernel /boot/vmlinuz root=/dev/sda1 ro vga=normal 
  initrd /boot/initrd-3.2.13-smp
# Linux bootable partition config ends
title --- For help press 'c', type: 'help'
root (hd0)
title --- For usage examples, type: 'cat /boot/grub/grub.txt'
root (hd0)

That’s it :)

References:

Upgrading to Kernel v3 for Slackware v13.37

2012-05-11T00:00:00-07:00

Upgrading to Kernel v3 for Slackware v13.37

Get all the necessary packages:

wget http://slackware.osuosl.org/slackware-current/slackware/a/kernel-generic-smp-3.2.13_smp-i686-1.txz
wget http://slackware.osuosl.org/slackware-current/slackware/a/kernel-modules-smp-3.2.13_smp-i686-1.txz
wget http://slackware.osuosl.org/slackware-current/slackware/a/lilo-23.2-i486-1.txz

Change your run level:
```
telinit 1
```

Install the kernel’s package:

installpkg kernel-generic-smp-3.2.13_smp-i686-1.txz
kernel-modules-smp-3.2.13_smp-i686-1.txz

Create initial ramdisk images for preloading modules:

cd /boot
mkinitrd -c -k 3.2.2-smp -m ext4 -f ext4 -r /dev/sda1 -o initrd-3.2.2-smp

Edit your lilo.conf:

# Linux bootable partition config begins 
image = /boot/vmlinuz-generic-smp-3.2.2-smp 
initrd = /boot/initrd-3.2.2-smp 
  root = /dev/sda1 
  label = slackware 
  read-only # Partitions should be mounted read-only for checking 
# Linux bootable partition config ends

Reload lilo:
```
lilo -v
```
You’ll get:
```
Fatal: Setup length exceeds 31 maximum; kernel setup will overwrite boot loader
```
“It’s because the kernel has grown so large. You can fix it by installing a fixed lilo from slackware-current, too” - from LQ.
```
upgradepkg /root/lilo-23.2-i486-1.txz
lilo -v
```
Reboot.

Check your new kernel information:

uname -a

You should get something like this:

Linux slackware 3.2.13-smp #1 SMP Fri Mar 23 23:21:11 CDT 2012 i686 Intel(R) Core(TM)2 Duo CPU     P8800  @ 2.66GHz GenuineIntel GNU/Linux

That’s it.

Redmine with Git and Subversion in HUSM

2012-04-11T00:00:00-07:00

Redmine with Git and Subversion in HUSM

It’s official, almost every piece of our codes in Hospital Universiti Sains Malaysia (HUSM) has been versioned in either Git or Subversion (mostly the latter). We’ve also linked all of the repositories into Redmine for easier project management. It’s not really a walk in the park for us since this is the first time we tried to do it this big.

We’re hoping we can document our workloads using issues feature in Redmine. Maybe we’re a few years late (most of our projects are already matured where no big features added every week), but IMHO this is the perfect time since we’re in the process of upgrading our core application in the hospital which is Lifeline.

So, here are some fun facts about our Redmine + (Git/Subversion):

We’re using the latest Redmine v1.3.2 (through its Git’s distribution).
It’s being served through Apache on our Slackware v13.37 box.
The Ruby and the gems are being managed with RVM for easier installation.
It’s using an old desktop PC model Acer M460 (2.4 GHz).
With advance Git and Subversion integrations, the repositories authentication are using the usernames and passwords from the Redmine itself.
Our Git repositories are being served using Smart HTTP Git protocol.
There are about 56 repositories in it which comprises different type of code languages such as Java, C++, Ruby, HTML, JavaScript, CSS and Pascal.
For backups, we’re using RAID 1 for the hardisks. rsync and mysqldump are being ran three times a day. There’s no offsite backup yet.
There’re about 40++ projects in the Redmine right now (still increasing).
The repositories are hosting codes from different programming languages such as Pascal, Java, C++, Ruby and PHP. So, we need to be extra careful about the .gitigore and svn:ignore configurations.
Tutorials has been written for newcomers on how to use version control for Netbeans, Delphi and also Qt Creator. For the time being, everyone is encourage to use their IDE to do all the commits and not doing it through command line unless they know what they are doing.
Most developers chose Subversion due to its low lurning curve.

In the next few days (or weeks), I’m going to post the configuration that I used to make the intergration between Git/Subversion with Redmine. It seems that the wiki from Redmine itself is not complete enough and I had to dig deeper into other websites. So, hopefully it will help anyone who’s lost just like me before.

By the way, you can view a screenshot from the Redmine’s projects page here.

See Full Referrer URL in Google Analytics

2012-01-21T00:00:00-08:00

See Full Referrer URL in Google Analytics

Most of the tutorial on the web focuses on the old version of Google Analytics. So, after poking around Google Analytics, I found out that there are two ways to see it (I stopped looking once I found the second method, so, there might be more).

Go to Custom Reporting tab.
Click Edit.
If you’ve already added a tab, just click “+add report tab” for this purpose.
Put the name of this report.
Choose Flat Table as the type.
For the Dimensions, search for Full Referrer.
For the Metrics, use Unique Visitors (or anything you see fits).
Click Save.

That’s it. You can refer to the screenshot below if you’re feeling lost:

Good luck!

A Not So Simple Introduction to Git

2012-01-07T00:00:00-08:00

A Not So Simple Introduction to Git

Found another presentation from Speaker Deck. I find it useful with its examples. Check it out.