Normally whenever you need to boot the RHEL machine in rescue mode, you do it by using CD-ROM/DVD/PXE/NFS/FTP but  para virtualization doesn’t support CD-ROM/DVD/PXE to boot it  hence  you need NFS/FTP/http bootable image to boot the machine.

 Hence let’s start from here to boot the DomU from CD Image and get recue mode.

Let suppose you have a CD/DVD iso on the server.

Mount the image  mount –t iso9660 –o loop /path/to/image/file.iso  /mnt

copy the vmlinuz and initrd.img from /mnt/isolinux/xen/  to another location /etc/xen/boot/

Lets suppose you have domain named “domain1“ and need to boot this domain in rescue mode.

Edit/update the domain1 configuration file /etc/xen/domain1 as follows.

•  kernel = “/etc/xen/boot/vmliuz”
• ramdisk = “/etc/xen/boot/initrd.img”
• extra = “rescue”

Comment the bootloader line in your configuration

1. Now create your domain as  # xm create –c domain1
2. It will ask you few questions and then linux image location on the domU disks

Hence you need to create NFS/FTP for the same to provide the image.

Lets start using NFS.

update /etc/exports with the following

/mnt  *(ro,sync)

service portmap start

service nfs start

If you already running NFS there then just  # exportfs –a

 Now start again from step 5 to create the domain.

After creating domain it will ask you the way of booting then please choose NFS and then it will ask you for your IP setting which you know better.

Update the NFS server name : IP address of your Dom0

Shared name: /mnt

Now your domU is booting in rescue mode and you will get a shell a shell.

After fixing the problem, revert back the changes in the domain1 configuration file.

Please let me know if you have any question.

But that was really easy to build and install.

There are few points to remember to make the build and use it.

Hence before starting I would like to share few requirements for this.

1. Install all required RPM packages like rpmbuild…etc

2. Create Directories in /usr/src/redhat/

BUILD/
RPMS/
SOURCES/
SPECS/
SRPMS/

3. Download source code of your required software tar.gz file.

4. Put it in /usr/src/redhat/SOURCES/

5. Copy the *.spec file (one file) from source file  to /usr/src/redhat/SPECS/

Now we are ready to make RPM but need to edit the spec file.

Hence for an example I have taken openssl and defining all steps to build new RPMs.

Edit openssl.spec file in /usr/src/redhat/SPECS/  which have been highlighted there.

%define libmaj 0
%define libmin 9
%define librel 8
%define librev k
Release: 1

%define openssldir /usr/local/openssl

Summary: Secure Sockets Layer and cryptography libraries and tools
Name: openssl
#Version: %{libmaj}.%{libmin}.%{librel}
Version: %{libmaj}.%{libmin}.%{librel}%{librev}
Source0: /usr/src/redhat/SOURCES/%{name}-%{version}.tar.gz
License: GPL
#Copyright: Freely distributable
Group: System Environment/Libraries
Provides: SSL
#URL: http://www.openssl.org/
Packager: Damien Miller
BuildRoot:   /var/tmp/%{name}-%{version}-root

%description

…..

…..

#%files doc
#%defattr(0644,root,root,0755)
#%doc CHANGES CHANGES.SSLeay LICENSE NEWS README
#%doc doc

—–

—–

In the above file there are few changes required.

1.   Comment  the version and copyright field

2.  Comment the URL field else at building time you will get new spec file from there.

3.  Define field provides the location where RPM will install

4. Update source link

5. Update your build root where temporary build can be put at installation time

Now your file is ready in make RPM

cd /usr/src/redhat/SPECS

rpmbuild –ba openssl.spec

You may get the error at the building of RPM so check the dependencies required to build the RPM at the screen or in the file at /var/tmp/.

When you get any error check the screen and that may be because of some dependencies and need to install the RPM for that. The directory structure in /var/tmp will be same as your installation directories.

Now after completion of above steps your RPMs have been created and  are in /usr/src/redhat/RPMS/ and Source RPMS are in /usr/src/redhat/RPMS

The options used in rpmbuild command above will create both source rpm and rpm.

Before installing the RPM on the server, need to pass some check points as mentioned below.

1. On all rpms in /usr/src/redhat/RPMS run the below command.

rpm -qip *.rpm

rpm -qlp *.rpm

The above command provided you list of all the files which will install on your system. Hence need to take backup of all your files which are going to overwrite. This is for precaution and need them if your new RPMs does not support your OS.

Now after taking backups of your all files run the following commands.

rpm -ivh *.rpm

If you get conflicting error then do update -Uvh

Openssl has been updated on your system and need to check

# ssh -V

OpenSSH_3.9p1, OpenSSL 0.9.7a Feb 19 2003

# /usr/local/openssl/bin/openssl version

OpenSSL 0.9.8k 25 Mar 2009

# which openssl

/usr/bin/openssl

# cp /usr/bin/openssl /usr/bin/openssl-0.9.7a

# rm /usr/bin/openssl

rm: remove regular file `/usr/bin/openssl’? y

# ln -s /usr/local/openssl/bin/openssl /usr/bin/openssl

As you installed openssl in /usr/local/openssl and this is not in $PATH hence replace the old openssl witih the new.

The overall process above is a simple RPM build and installation.

You  may have difficulties at building time so start with simpler one and go to big installation.

The main point, I would like to highlight here that if your are building RPM on RHEL 4  32bit then it can be installed on any 32bit server but lower than RHEL4.

RHEL 2.1 has different directory structure so will not work there…

All the best …let me know if you have any question.

This is my first blog on a small topic but really helpful to save the time.

Lets say that you have 50 network printers connected on a particular linux server and now need to migrate all of them to the newly build server.

Do you imagine  how time consuming and hectic task is it ? ..but you can do this and consume only a single minute out of hours.

so the steps are here…

Lets say there are two servers.

server –> a.com (where all printers are already connected)

server –> b.com (where you need to configure the printers)

Moving printers from server “a” –> server “b”

Tasks on server “a”

1. Make a tar of the directory /etc/cups on the server “a” on which printers are already configured
tar –cvf  cups.tar /etc/cups

2. Export the printer conf file (used by printconf-tui) on server “a”
      /usr/sbin/printconf-tui –Xexport > settings.xml

3. copy the tar to the server “b”
scp cups.tar  root@b.com:/tmp

4. Copy the above settings.xml to new server
   scp settings.xml root@b.com:/tmp

Setting on server “b”

1. As a precaution take the backup of cups dir on the new dir and untar the tar file copied.
    # mv /etc/cups /etc/cups.backup
    # cd /etc/
    # tar -cvf /tmp/cups.tar

2. Import the settings
       /usr/sbin/printconf-tui –Ximport < /tmp/settings.xml

2.1 If you have already configures some printers on the new server and need to merge the configuration use
        /usr/sbin/printconf-tui –Ximport –merge < settings.xml

3. Restart cups service
/etc/init.d/cups restart

4. Verify the printers

  lpstat -a

Practical and theoretical have major difference when you work in an industry.

Need Practical guidance on Unix? I will share with you shortly.

 Keep  in touch with me to discuss any query