Get Blogged by JoKi

"The only frontiers are in your mind"
01 | 08 | 2014
Navigation
Syndication
feed-image RSS 2.0
Latest Tweets

Most Read Articles
Article Time Line
About me
Family guy, geek, entrepreneur, software craftsman: Visual FoxPro, C#, SQL Server, MySQL, Linux consultant, conference speaker

Certificates & Awards

Microsoft Certified Professional

Microsoft Specialist - Programming in HTML5 with JavaScript and CSS3 Specialist

Microsoft MVP - Visual Developer for Visual FoxPro 2006 & 2007
Visual Developer for Visual FoxPro 2007
Visual Developer for Visual FoxPro 2006

Get in touch

Sharing is caring

Recent books

Sponsoring
If you like the information on these pages, your support is highly appreciated.
Thank you very much!
Affiliates
Spacer for layout formatting
Accessing your web server via IPv6
Linux
Sunday, 16 March 2014 07:51

Being able to run your systems on IPv6, have automatic address assignment and the ability to resolve host names are the necessary building blocks in your IPv6 network infrastructure. Now, that everything is in place it is about time that we are going to enable another service to respond to IPv6 requests. The following article will guide through the steps on how to enable Apache2 httpd to listen and respond to incoming IPv6 requests.

This is the fourth article in a series on IPv6 configuration:

Piece of advice: This is based on my findings on the internet while reading other people's helpful articles and going through a couple of man-pages on my local system.

Surfing the web - IPv6 style

Enabling IPv6 connections in Apache 2 is fairly simply. But first let's check whether your system has a running instance of Apache2 or not. You can check this like so:

$ service apache2 status
Apache2 is running (pid 2680).

In case that you got a 'service unknown' you have to install Apache to proceed with the following steps:

$ sudo apt-get install apache2

Out of the box, Apache binds to all your available network interfaces and listens to TCP port 80. To check this, run the following command:

$ sudo netstat -lnptu | grep "apache2\W*$"
tcp6       0      0 :::80                   :::*                    LISTEN      28306/apache2

In this case Apache2 is already binding to IPv6 (and implicitly to IPv4). If you only got a tcp output, then your HTTPd is not yet IPv6 enabled.

Check your Listen directive, depending on your system this might be in a different location than the default in Ubuntu.

$ sudo nano /etc/apache2/ports.conf

# If you just change the port or add more ports here, you will likely also
# have to change the VirtualHost statement in
# /etc/apache2/sites-enabled/000-default
# This is also true if you have upgraded from before 2.2.9-3 (i.e. from
# Debian etch). See /usr/share/doc/apache2.2-common/NEWS.Debian.gz and
# README.Debian.gz

NameVirtualHost *:80
Listen 80

<IfModule mod_ssl.c>
    # If you add NameVirtualHost *:443 here, you will also have to change
    # the VirtualHost statement in /etc/apache2/sites-available/default-ssl
    # to <VirtualHost *:443>
    # Server Name Indication for SSL named virtual hosts is currently not
    # supported by MSIE on Windows XP.
    Listen 443
</IfModule>

<IfModule mod_gnutls.c>
    Listen 443
</IfModule>

Just in case that you don't have a ports.conf file, look for it like so:

$ cd /etc/apache2/
$ fgrep -r -i 'listen' ./*

And modify the related file instead of the ports.conf. Which most probably might be either apache2.conf or httpd.conf anyways.

Okay, please bear in mind that Apache can only bind once on the same interface and port. So, eventually, you might be interested to add another port which explicitly listens to IPv6 only. In that case, you would add the following in your configuration file:

Listen 80
Listen [2001:db8:bad:a55::2]:8080

But this is completely optional... Anyways, just to complete all steps, you save the file, and then check the syntax like so:

$ sudo apache2ctl configtest
Syntax OK

Ok, now let's apply the modifications to our running Apache2 instances:

$ sudo service apache2 reload
 * Reloading web server config apache2
   ...done.

$ sudo netstat -lnptu | grep "apache2\W*$"                                                                                              
tcp6       0      0 2001:db8:bad:a55:::8080 :::*                    LISTEN      5922/apache2   
tcp6       0      0 :::80                   :::*                    LISTEN      5922/apache2

There we have two daemons running and listening to different TCP ports.

Now, that the basics are in place, it's time to prepare any website to respond to incoming requests on the IPv6 address. Open up any configuration file you have below your sites-enabled folder.

$ ls -al /etc/apache2/sites-enabled/
...

$ sudo nano /etc/apache2/sites-enabled/000-default

<VirtualHost *:80 [2001:db8:bad:a55::2]:8080>
        ServerAdmin This e-mail address is being protected from spambots. You need JavaScript enabled to view it
        ServerName server.ios.mu
        ServerAlias server

Here, we have to check and modify the VirtualHost directive and enable it to respond to the IPv6 address and port our web server is listening to. Save your changes, run the configuration test and reload Apache2 in order to apply your modifications. After successful steps you can launch your favourite browser and navigate to your IPv6 enabled web server.

Accessing an IPv6 address in the browser
Accessing an IPv6 address in the browser

That looks like a successful surgery to me...

Note: In case that you received a timeout, check whether your client is operating on IPv6, too.

Add a comment
 
Enabling DNS for IPv6 infrastructure
Linux
Wednesday, 12 March 2014 10:34

After successful automatic distribution of IPv6 address information via DHCPv6 in your local network it might be time to start offering some more services. Usually, we would use host names in order to communicate with other machines instead of their bare IPv6 addresses. During the following paragraphs we are going to enable our own DNS name server with IPv6 address resolving.

This is the third article in a series on IPv6 configuration:

Piece of advice: This is based on my findings on the internet while reading other people's helpful articles and going through a couple of man-pages on my local system.

What's your name and your IPv6 address?

$ sudo service bind9 status
 * bind9 is running

If the service is not recognised, you have to install it first on your system. This is done very easy and quickly like so:

$ sudo apt-get install bind9

Once again, there is no specialised package for IPv6. Just the regular application is good to go.

But of course, it is necessary to enable IPv6 binding in the options. Let's fire up a text editor and modify the configuration file.

$ sudo nano /etc/bind/named.conf.options

acl iosnet {
        127.0.0.1;
        192.168.1.0/24;
        ::1/128;
        2001:db8:bad:a55::/64;
};

listen-on { iosnet; };
listen-on-v6 { any; };

allow-query { iosnet; };
allow-transfer { iosnet; };

Most important directive is the listen-on-v6. This will enable your named to bind to your IPv6 addresses specified on your system. Easiest is to specify any as value, and named will bind to all available IPv6 addresses during start. More details and explanations are found in the man-pages of named.conf.

Save the file and restart the named service. As usual, check your log files and correct your configuration in case of any logged error messages. Using the netstat command you can validate whether the service is running and to which IP and IPv6 addresses it is bound to, like so:

$ sudo service bind9 restart

$ sudo netstat -lnptu | grep "named\W*$"
tcp        0      0 192.168.1.2:53        0.0.0.0:*               LISTEN      1734/named     
tcp        0      0 127.0.0.1:53          0.0.0.0:*               LISTEN      1734/named     
tcp6       0      0 :::53                 :::*                    LISTEN      1734/named     
udp        0      0 192.168.1.2:53        0.0.0.0:*                           1734/named     
udp        0      0 127.0.0.1:53          0.0.0.0:*                           1734/named     
udp6       0      0 :::53                 :::*                                1734/named 

 Sweet! Okay, now it's about time to resolve host names and their assigned IPv6 addresses using our own DNS name server.

$ host -t aaaa www.6bone.net 2001:db8:bad:a55::2
Using domain server:
Name: 2001:db8:bad:a55::2
Address: 2001:db8:bad:a55::2#53
Aliases:

www.6bone.net is an alias for 6bone.net.
6bone.net has IPv6 address 2001:5c0:1000:10::2

Alright, our newly configured BIND named is fully operational.

Eventually, you might be more familiar with the dig command. Here is the same kind of IPv6 host name resolve but it will provide more details about that particular host as well as the domain in general.

$ dig @2001:db8:bad:a55::2 www.6bone.net. AAAA

More details on the Berkeley Internet Name Domain (bind) daemon and IPv6 are available in Chapter 22.1 of Peter Bieringer's HOWTO on IPv6.

Setting up your own DNS zone

Now, that we have an operational named in place, it's about time to implement and configure our own host names and IPv6 address resolving. The general approach is to create your own zone database below the bind folder and to add AAAA records for your hosts. In order to achieve this, we have to define the zone first in the configuration file named.conf.local.

$ sudo nano /etc/bind/named.conf.local

//
// Do any local configuration here
//
zone "ios.mu" {
        type master;
        file "/etc/bind/zones/db.ios.mu";
};

Here we specify the location of our zone database file. Next, we are going to create it and add our host names, our IP and our IPv6 addresses.

$ sudo nano /etc/bind/zones/db.ios.mu

$ORIGIN .
$TTL 259200     ; 3 days
ios.mu                  IN SOA  ios.mu. hostmaster.ios.mu. (
                                2014031101 ; serial
                                28800      ; refresh (8 hours)
                                7200       ; retry (2 hours)
                                604800     ; expire (1 week)
                                86400      ; minimum (1 day)
                                )
                        NS      server.ios.mu.
$ORIGIN ios.mu.
server                  A       192.168.1.2
server                  AAAA    2001:db8:bad:a55::2
client1                 A       192.168.1.3
client1                 AAAA    2001:db8:bad:a55::3
client2                 A       192.168.1.4
client2                 AAAA    2001:db8:bad:a55::4

With a couple of machines in place, it's time to reload that new configuration.

Note: Each time you are going to change your zone databases you have to modify the serial information, too. Named loads the plain text zone definitions and converts them into an internal, indexed binary format to improve lookup performance. If you forget to change your serial then named will not use the new records from the text file but the indexed ones. Or you have to flush the index and force a reload of the zone.

This can be done easily by either restarting the named:

$ sudo service bind9 restart

or by reloading the configuration file using the name server control utility - rndc:

$ sudo rndc reconfig

Check your log files for any error messages and whether the new zone database has been accepted. Next, we are going to resolve a host name trying to get its IPv6 address like so:

$ host -t aaaa server.ios.mu. 2001:db8:bad:a55::2
Using domain server:
Name: 2001:db8:bad:a55::2
Address: 2001:db8:bad:a55::2#53
Aliases:

server.ios.mu has IPv6 address 2001:db8:bad:a55::2

Looks good.

Alternatively, you could have just ping'd the system as well using the ping6 command instead of the regular ping:

$ ping6 server
PING server(2001:db8:bad:a55::2) 56 data bytes
64 bytes from 2001:db8:bad:a55::2: icmp_seq=1 ttl=64 time=0.615 ms
64 bytes from 2001:db8:bad:a55::2: icmp_seq=2 ttl=64 time=0.407 ms
^C
--- ios1 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.407/0.511/0.615/0.104 ms

That also looks promising to me. How about your configuration?

Next, it might be interesting to extend the range of available services on the network. One essential service would be to have web sites at hand.

Add a comment
 
DHCPv6: Provide IPv6 information in your local network
Linux
Tuesday, 11 March 2014 10:57

Even though IPv6 might not be that important within your local network it might be good to get yourself into shape, and be able to provide some details of your infrastructure automatically to your network clients.

This is the second article in a series on IPv6 configuration:

Piece of advice: This is based on my findings on the internet while reading other people's helpful articles and going through a couple of man-pages on my local system.

IPv6 addresses for everyone (in your network)

Okay, after setting up the configuration of your local system, it might be interesting to enable all your machines in your network to use IPv6. There are two options to solve this kind of requirement... Either you're busy like a bee and you go around to configure each and every system manually, or you're more the lazy and effective type of network administrator and you prefer to work with Dynamic Host Configuration Protocol (DHCP). Obviously, I'm of the second type.

Enabling dynamic IPv6 address assignments can be done with a new or an existing instance of a DHCPd. In case of Ubuntu-based installation this might be isc-dhcp-server. The isc-dhcp-server allows address pooling for IP and IPv6 within the same package, you just have to run to independent daemons for each protocol version. First, check whether isc-dhcp-server is already installed and maybe running your machine like so:

$ service isc-dhcp-server6 status

In case, that the service is unknown, you have to install it like so:

$ sudo apt-get install isc-dhcp-server

Please bear in mind that there is no designated installation package for IPv6.

Okay, next you have to create a separate configuration file for IPv6 address pooling and network parameters called /etc/dhcp/dhcpd6.conf. This file is not automatically provided by the package, compared to IPv4. Again, use your favourite editor and put the following lines:

$ sudo nano /etc/dhcp/dhcpd6.conf

authoritative;
default-lease-time 14400;
max-lease-time 86400;
log-facility local7;
subnet6 2001:db8:bad:a55::/64 {
    option dhcp6.name-servers 2001:4860:4860::8888, 2001:4860:4860::8844;
    option dhcp6.domain-search "ios.mu";
    range6 2001:db8:bad:a55::100 2001:db8:bad:a55::199;
    range6 2001:db8:bad:a55::/64 temporary;
}

Next, save the file and start the daemon as a foreground process to see whether it is going to listen to requests or not, like so:

$ sudo /usr/sbin/dhcpd -6 -d -cf /etc/dhcp/dhcpd6.conf eth0

The parameters are explained quickly as -6 we want to run as a DHCPv6 server, -d we are sending log messages to the standard error descriptor (so you should monitor your /var/log/syslog file, too), and we explicitely want to use our newly created configuration file (-cf). You might also use the command switch -t to test the configuration file prior to running the server.

In my case, I ended up with a couple of complaints by the server, especially reporting that the necessary lease file wouldn't exist. So, ensure that the lease file for your IPv6 address assignments is present:

$ sudo touch /var/lib/dhcp/dhcpd6.leases
$ sudo chown dhcpd:dhcpd /var/lib/dhcp/dhcpd6.leases

Now, you should be good to go. Stop your foreground process and try to run the DHCPv6 server as a service on your system:

$ sudo service isc-dhcp-server6 start
isc-dhcp-server6 start/running, process 15883

Check your log file /var/log/syslog for any kind of problems. Refer to the man-pages of isc-dhcp-server and you might check out Chapter 22.6 of Peter Bieringer's IPv6 Howto. The instructions regarding DHCPv6 on the Ubuntu Wiki are not as complete as expected and it might not be as helpful as this article or Peter's HOWTO. But see for yourself.

Does the client get an IPv6 address?

Running a DHCPv6 server on your local network surely comes in handy but it has to work properly. The following paragraphs describe briefly how to check the IPv6 configuration of your clients,

Linux - ifconfig or ip command

First, you have enable IPv6 on your Linux by specifying the necessary directives in the /etc/network/interfaces file, like so:

$ sudo nano /etc/network/interfaces

iface eth1 inet6 dhcp

Note: Your network device might be eth0 - please don't just copy my configuration lines.

Then, either restart your network subsystem, or enable the device manually using the dhclient command with IPv6 switch, like so:

$ sudo dhclient -6

You would either use the ifconfig or (if installed) the ip command to check the configuration of your network device like so:

$ sudo ifconfig eth1
eth1      Link encap:Ethernet  HWaddr 00:1d:09:5d:8d:98 
          inet addr:192.168.160.147  Bcast:192.168.160.255  Mask:255.255.255.0
          inet6 addr: 2001:db8:bad:a55::193/64 Scope:Global
          inet6 addr: fe80::21d:9ff:fe5d:8d98/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

Looks good, the client has an IPv6 assignment. Now, let's see whether DNS information has been provided, too.

$ less /etc/resolv.conf

# Dynamic resolv.conf(5) file for glibc resolver(3) generated by resolvconf(8)
#     DO NOT EDIT THIS FILE BY HAND -- YOUR CHANGES WILL BE OVERWRITTEN
nameserver 2001:4860:4860::8888
nameserver 2001:4860:4860::8844
nameserver 192.168.1.2
nameserver 127.0.1.1

search ios.mu

Nicely done.

Windows - netsh

Per description on TechNet the netsh is defined as following:

"Netsh is a command-line scripting utility that allows you to, either locally or remotely, display or modify the network configuration of a computer that is currently running. Netsh also provides a scripting feature that allows you to run a group of commands in batch mode against a specified computer. Netsh can also save a configuration script in a text file for archival purposes or to help you configure other servers."

And even though TechNet states that it applies to Windows Server (only), it is also available on Windows client operating systems, like Vista, Windows 7 and Windows 8.

In order to get or even set information related to IPv6 protocol, we have to switch the netsh interface context prior to our queries. Open a command prompt in Windows and run the following statements:

C:\Users\joki>netsh
netsh>interface ipv6
netsh interface ipv6>show interfaces
Show IPv6 network interfaces using netsh command on Windows

Select the device index from the Idx column to get more details about the IPv6 address and DNS server information (here: I'm going to use my WiFi device with device index 11), like so:

netsh interface ipv6>show address 11
Show IPv6 address information using netsh command on Windows

Okay, address information has been provided. Now, let's check the details about DNS and resolving host names:

netsh interface ipv6> show dnsservers 11
Show IPv6 DNS server configuration using netsh command on Windows

Okay, that looks good already. Our Windows client has a valid IPv6 address lease with lifetime information and details about the configured DNS servers.

Talking about DNS server...
Your clients should be able to connect to your network servers via IPv6 using hostnames instead of IPv6 addresses. Please read on about how to enable a local named with IPv6.

Add a comment
 
Configure IPv6 on your Linux system (Ubuntu)
User Rating:★☆☆☆☆ / 1
PoorBest 
Linux
Monday, 10 March 2014 12:12

After the presentation on IPv6 at the first event of the Emtel Knowledge Series and some recent discussion on social media networks with other geeks and Linux interested IT people here in Mauritius, I thought that I should give it a try (finally) and tweak my local network infrastructure. Honestly, I have been to busy with contractual project work and it never really occurred to me to set up IPv6 in my LAN. Well, the following paragraphs are going to shed some light on those aspects of modern computer and network technology.

This is the first article in a series on IPv6 configuration:

Piece of advice: This is based on my findings on the internet while reading other people's helpful articles and going through a couple of man-pages on my local system.

Let's embrace IPv6

The basic configuration on Linux is actually very simple as the kernel, operating system, and user-space programs support that protocol natively. If your system is ready to go for IP (aka: IPv4), then you are good to go for anything else. At least, I didn't have to install any additional packages on my system(s). We are going to assign a static IPv6 address to the system. Hence, we have to modify the definition of interfaces and check whether we have an inet6 entry specified. Open your favourite text editor and check the following entries (it should be at least similar to this):

$ sudo nano /etc/network/interfaces

auto eth0
# IPv4 configuration
iface eth0 inet static
  address 192.168.1.2
  network 192.168.1.0
  netmask 255.255.255.0
  broadcast 192.168.1.255

# IPv6 configuration
iface eth0 inet6 static
  pre-up modprobe ipv6
  address 2001:db8:bad:a55::2
  netmask 64

Of course, you might have to adjust your interface device (eth0) or you might be interested to have multiple directives for additional devices (eth1, eth2, etc.). The auto instruction takes care that your device is enabled and configured during the booting phase. The use of the pre-up directive depends on your kernel configuration but in most scenarios this might be an optional line. Anyways, it doesn't hurt to have it enabled after all - just to be on the safe side.

Next, either restart your network subsystem like so:

$ sudo service networking restart

Or you might prefer to do it manually with identical parameters, like so:

$ sudo ifconfig eth0 inet6 add 2001:db8:bad:a55::2/64

In case that you're logged in remotely into your PC (ie. via ssh), it is highly advised to opt for the second choice and add the device manually.

You can check your configuration afterwards with one of the following commands (depends on whether it is installed):

$ sudo ifconfig eth0
eth0      Link encap:Ethernet  HWaddr 00:21:5a:50:d7:94 
          inet addr:192.168.160.2  Bcast:192.168.160.255  Mask:255.255.255.0
          inet6 addr: fe80::221:5aff:fe50:d794/64 Scope:Link
          inet6 addr: 2001:db8:bad:a55::2/64 Scope:Global
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1

$ sudo ip -6 address show eth0
3: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qlen 1000
    inet6 2001:db8:bad:a55::2/64 scope global
       valid_lft forever preferred_lft forever
    inet6 fe80::221:5aff:fe50:d794/64 scope link
       valid_lft forever preferred_lft forever

In both cases, it confirms that our network device has been assigned a valid IPv6 address.

That's it in general for your setup on one system. But of course, you might be interested to enable more services for IPv6, especially if you're already running a couple of them in your IP network. More details are available on the official Ubuntu Wiki.

Continue to configure your network to provide IPv6 address information automatically in your local infrastructure.

Add a comment
 
How to extend a virtual disk and its logical volume(s) in CentOS
User Rating:★★☆☆☆ / 4
PoorBest 
Linux
Tuesday, 04 March 2014 12:02

Lately, I ran into the situation that one of my services which is hosted in a virtual machine stopped working. A quick check revealed that the hard disk ran out of disk space and it was about time to increase the available storage. Don't laugh but the system is running on CentOS 5.x with a mere 13 GB virtual disk - well, since years already. And one golden rule learned from experience: Never touch a running system - kept me away from any modifications. Well, there's always a time that change has more benefits than not touching the system...

Running out of disk space

That is (now: was) the situation before the following steps.

# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
                       12G   12G    0G 100% /
/dev/hdc1              99M   36M   59M  38% /boot
tmpfs                 252M     0  252M   0% /dev/shm

And I can tell you that even a Linux OS doesn't like the lack of space on the system drive at all.

Backup your data and drives

Although, I think it is not necessary to stress the fact that you should (read: have to) backup your data and drives always. Just to be on the safe side, and to revert back to the original state easily and quickly.

Preparing the virtual disk

First, shut down your virtual machine. Even though it is possible to run almost all of the following in a running environment I doubt that for the casual (home) administrator reading this article it might not be necessary to risk your data. Okay next, it would be necessary to expand the size of virtual hard disk, and then afterwards extend the logical volume with the newly created partition.

Converting the disk format from VMDK to VDI

Well, the system has kind of a past and it was originally created on VMware Server 1.0, then upgraded to VMware Server 2.0, and some years back I switched over to SUN, eh Oracle, VirtualBox. So, the original virtual disk was still in VMware's format - vmdk. VirtualBox is not able to expand that type of disk drive (yet):

$ vboxmanage modifyhd virtual-disk1.vmdk --resize 40960
0%...
Progress state: VBOX_E_NOT_SUPPORTED
VBoxManage: error: Resize hard disk operation for this format is not implemented yet!

In order to complete this initial task it was necessary to convert the format from VMDK to VDI. This is done by cloning the disk into the new format:

$ vboxmanage clonehd virtual-disk1.vmdk virtual-disk1.vdi --format VDI
0%...10%...20%...30%...40%...50%...60%...70%...80%...90%...100%
Clone hard disk created in format 'vdi'. UUID: 37ef6965-0000-4159-861a-d1c64d9c060f

Depending on the actual disk size and your system performance this might take some time. Remain patient and let the system do the job. Meanwhile, you might check your mails or post your intentions on various social media networks. Just kidding! ;-)

Note: Additionally to your backup, it might be interesting to keep an archive of the original VMDK file. Just in case...

Resizing the virtual drive

Now, that we have our drive in VDI format we are able to expand the disk size. Let's try our previous statement again, but this time with the newly created VDI disk:

$ vboxmanage modifyhd virtual-disk1.vdi --resize 40960
0%...10%...20%...30%...40%...50%...60%...70%...80%...90%...100%

After successful expansion it is about time to check our disk modifications in the settings of the virtual machine. Launch the Oracle VM VirtualBox Manager, select your virtual machine and either press the Settings button in the toolbar or choose from the menu Machine -- Settings (Ctrl+S). Next, select the Storage entry from the side pane and then the virtual hard disk that we just modified. You should see something similar to the following screenshot.

Settings of virtual disk size and proper assignment
Oracle VM VirtualBox: Settings of virtual disk size and proper assignment

In case that you had to convert an VMDK drive to VDI, please select the newly cloned drive from the dropdown list in the Attributes area. You should have the .VDI disk attached to your virtual machine.

Pro tip: In case that you want to keep the original disk type in VMDK format, do a complete round-trip and clone the extended drive back to VMDK format, like so:

$ vboxmanage clonehd virtual-disk1.vdi virtual-disk1.vmdk --format VMDK
0%...10%...20%...30%...40%...50%...60%...70%...80%...90%...100%
Clone hard disk created in format 'vmdk'. UUID: 37ef6965-0000-4159-861a-d1c64d9c060f

Using VMware Player

The above mentioned conversion is not necessary in case that you have an installation of VMware Player or even VMware Workstation at hand. In my case, I didn't. Anyway, VMware Player gives you the ability to expand the disk capacity through the UI directly. Open the Virtual Machine Settings, then select the Hard Disk from the list of devices and below the Disk information you'll have a list of utilities in the dropdown list. Choose "Expand..." in order to change the disk's capacity.

Virtual Machine Settings and ability to expand the disk capacity
VMware Player: Virtual Machine Settings and ability to expand the disk capacity

That's all for the 'physical' expansion of our hard drive. Now, start your virtual machine as we are going to take of the software part.

Preparing the partition

The system in our virtual machine is still out of space and we are going to change this now. Please follow the next steps closely. Accidentally, I trapped myself and had to revert a couple of steps because I didn't pay enough attention to the details. I'll come back to that at the end of this article. Okay, we enlarged the physical drive, and now we have to take care of that unallocated space. Let's check the partition table first like so:

$ sudo fdisk -l

Disk /dev/hdc: 42.9 GB, 42949672960 bytes
255 heads, 63 sectors/track, 5221 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hdc1   *           1          13      104391   83  Linux
/dev/hdc2              14        1827    14570955   8e  Linux LVM

This will give us an overview of the existing partitions. The new disk size has been accepted already and it's now time to create a new partition.

Creating a new partition

This article is not about how to use fdisk on the command line. There are other tutorials with more detailed information about the particular steps, and you might have a look at the manual page of fdisk directly. To create a new partition we launch fdisk on our hard drive

# fdisk /dev/hdc

Then we create a new full-size partition at the end of the disk drive with the following sequence of keystrokes:

n - new partition
p - primary partition
3 - next partition number
[Enter] - confirm first cylinder
[Enter] - confirm last cylinder
t - change partition's type or system id
3 - choose newly created partition id
8e - specify hex code of system type (Linux LVM)
w - write partition table to disk

Of course, this might vary depending on your existing partition table. Please, adapt your inputs accordingly.

After completing all steps your hard drive might have a similar output like this one:

# fdisk -l

Disk /dev/hdc: 42.9 GB, 42949672960 bytes
255 heads, 63 sectors/track, 5221 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/hdc1   *           1          13      104391   83  Linux
/dev/hdc2              14        1827    14570955   8e  Linux LVM
/dev/hdc3            1828        5221    27262305   8e  Linux LVM

Alternative: Extend an existing partition

During my research in preparation of this 'surgery' I also came across a couple of resources online where it is stated that it is also possible to enlarge an existing partition to the full extend of the new capacity of the disk drive. The general recommendation is to use a more sophisticated partitioning tool like the freely available GParted. While using a Windows operating system, I clearly would recommend this approach, too. Mainly because it's your system drive that runs out of space and Windows is hardly capable to span over multiple drives (at least in casual environments like home and office desktops). I already did this in the past to enlarge a Windows Server 2008 R2 system, and it went as smooth as expected.

You can download the GParted LiveCD ISO here.

Expanding the logical volume

Once you understood the necessary steps to extend an existing logical volume it's fairly easy to reproduce. The general concept is well documented in Chapter 4 of the LVM Administrator's Guide for CentOS, and is based on the following procedure. First, you create a physical volume (PV) based on your new partition which is then extended into an existing volume group (VG), and then finally extended into a logical volume (LV) within a volume group. After resizing the logical volume you are ready to go and the new disk space is available in your system.

Creating the physical volume

# pvcreate /dev/hdc3
  Physical volume "/dev/hdc3" successfully created

This will prepare the partition for use in a volume group.

Note: Depending on whether you are going to extend your system with an additional partition or a whole drive, you might even skip the partitioning and create a physical volume using the whole hard disk.

Extending the volume group

Prior to any extension is not too bad to get an overview of the existing volume groups. This can be done like so:

# vgdisplay
  --- Volume group ---
  VG Name               VolGroup00
  System ID            
  ...

With the information of the available name of the volume group, we are now able to extend it with our physical partition:

# vgextend VolGroup00 /dev/hdc3
  Volume group "VolGroup00" successfully extended

Done. Next step is to assign the available space to the logical volume that we would like to expand.

Extending the logical volume, finally

Same procedure here, it's advised to have an overview of the existing logical volume first:

# lvdisplay
  --- Logical volume ---
  LV Name                /dev/VolGroup00/LogVol00
  VG Name                VolGroup00
...
 
  --- Logical volume ---
  LV Name                /dev/VolGroup00/LogVol01
  VG Name                VolGroup00
...

In my case, I'm going to extend the first logical volume with the new physical volume.

# lvextend /dev/VolGroup00/LogVol00 /dev/hdc3
  Extending logical volume root to 37.94 GB
  Logical volume LogVol00 successfully resized

Almost done. After expanding the volume group and the logical volume we only have to populate the information about the new total size of our logical volume.

# resize2fs /dev/VolGroup00/LogVol00

That's it!
Just to complete the whole process check the available disk space:

# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00
                       37G   12G   24G  32% /
/dev/hdc1              99M   36M   59M  38% /boot
tmpfs                 252M     0  252M   0% /dev/shm

Okay, my system has now more than triple the original disk size. Let's see how many years it is going to run before I might have to add another partition to the logical volume.

Summary

Once you understood the necessary steps on how to expand a logical volume it's actually very simple to reproduce. You start with the physical preparation of the hard disk, either by adding a new drive or by expanding an existing virtual drive. Then you create a partition of type 8e - Linux LVM - for that new unallocated drive space. And to complete the procedure you have to walk through the LVM handling by creating the physical volume, extending the volume group and finally extending and resizing the logical volume in that volume group.

Even in case of an error it's relatively simple to track down the root cause and take care of it.

Beware of the details... Troubleshooting

As mentioned earlier I got stuck in the process because of two issues.

First, my Linux operating system, here CentOS 5.3, didn't get the new partition table automatically. So when I run the following command I got the response that the specified device isn't available.

# pvcreate /dev/hdc3
  Device /dev/hdc3 not found (or ignored by filtering).

A quick look on the internet revealed that this could happen and the modified partition table can be updated manually like so:

# partprobe -s
/dev/hdc: msdos partitions 1 2 3

# partx -a /dev/hdc

Or to keep things simple, just reboot the system. But that would be too easy and a waste of time...

And second, I mistyped the name of the logical volume - LogVol01 instead of LogVol00 - and added the new partition to the swap area. Well, following are some details on how to detect that kind of issue and how to remove a physical drive from a logical volume and volume group.

I came to an abrupt stop on the very last step, resizing the file system with the newly allocated volume:

# resize2fs /dev/VolGroup00/LogVol01

resize2fs: bad magic number in super-block while trying to open /dev/VolGroup00/LogVol01
couldn’t find valid filesystem superblock

That's kind of bummer. Especially, so close to the finish line. Well, it turned out that volume group had two logical volumes LogVol00 and LogVol01. And the second one is actually used for the swap area. Which easily explains why there is absolutely no valid filesystem superblock to be found - no matter how hard I would try it. First, I thought that it might have been a problem with the filesystem type, as I am using ext2, ext3, ext4, and xfs interchangeable. But a quick check confirmed that I'm using the right command. In case of xfs you might have to work with xfs_grow command instead of resize2fs.

Then I checked the overview output of my logical volumes again and I was kind of lucky that I discovered that I increased the wrong volume:

# lvdisplay
  --- Logical volume ---
  LV Name                /dev/VolGroup00/LogVol01
  VG Name                VolGroup00
...
  LV Size                24.83 GB
  Current LE             891
  Segments               2

The original output stated a LV Size of 1.91 GB. Due to the nature of LVM I couldn't remove the physical hard drive because it is reported to be in use. Therefore, it is first necessary to decrease the volume size to its original value like so:

# lvreduce -L 1.9G /dev/VolGroup00/LogVol01
  Rounding up size to full physical extent 1.91 GB

Eventually, you might have to move all allocated space off the physical volume to other free physical volumes in the same volume group like so:

# pvmove /dev/hdc3

Then it's possible to remove (or better said to reduce) the drive from the volume group like so:

# vgreduce VolGroup00 /dev/hdc3

And then to start over again by adding the fresh physical volume back to the volume group again and then assigning it to the proper logical volume, like so:

# pvcreate /dev/hdc3
# vgextend VolGroup00 /dev/hdc3
# lvextend /dev/VolGroup00/LogVol00 /dev/hdc3
# resize2fs /dev/VolGroup00/LogVol00

Now, everything went smooth and my last system check of the logical volumes looks like so:

# lvdisplay
  --- Logical volume ---
  LV Name                /dev/VolGroup00/LogVol00
  VG Name                VolGroup00
...
  LV Size                37.94 GB
  Current LE             1214
  Segments               2
  
  --- Logical volume ---
  LV Name                /dev/VolGroup00/LogVol01
  VG Name                VolGroup00
...
  LV Size                1.91 GB
  Current LE             61
  Segments               1

As usual, follow administrator's rule number 1: Don't Panic!

Have a look at the history of your inputs, here your commands on the console, and check them entirely with the outputs of the various display commands - pvdisplay to show information about physical volumes, vgdisplay for details about volume groups, and finally lvdisplay to get an overview of your logical volumes.

Good luck with your mission and leave your comments below.

Add a comment
 
« Start : Prev : 1 : 2 : 3 : 4 : 5 : Next : End »

Page 1 of 5
Spacer for layout formatting