Today, Scientific Linux 7.0 is officially released.

In the Scientific Linux 7.0 release notes, two points need attention:

• The default Scientific Linux 7 installation provides automatic updates via the yum-cron package: because this distribution is often installed by graduate students, it has been decided to reduce the security risks for a novice by applying security updates automatically each night.
• There is no supported upgrade path from Scientific Linux 6 to Scientific Linux 7: it is a deliberate choice not to take any risk of leaving the system in an intermediate state.

You can download the Scientific Linux 7.0 distribution from now on.

Today, Dell just released its KVM Virtualization made easy for RHEL 7.
To anybody interested in building a KVM lab, this is a must-read.

In this white paper, Jose De la Rosa from Dell explains all the steps involved in this operation:

• required packages,
• required services,
• networking configuration (with or without bridge),
• VM image location,
• VM creation & cloning,
• basic and advanced VM management.

Happy reading!

The I/O scheduler policy has changed with Red Hat Enterprise Linux 7.

The default I/O Scheduler is now CFQ for SATA drives and Deadline for everything else.
Indeed, for faster storage than SATA drives, Deadline outperforms CFQ, giving a performance increase without any special tuning.

Additional information and instructions are available in the I/O scheduler tutorial.

The Red Hat certification team recently made some adjustments to the RHCSA & RHCE exams objectives for RHEL 7.

Concerning the RHCSA exam, VNC and LUKS related tasks have been removed, when firewall settings can now be set up through iptables or firewalld, which is very sensible. The client LDAP configuration, although no longer directly specified, seems to remain an objective through the client configuration to an authentication service.
This appears to be a slight adjustment with very limited consequences.

Concerning the RHCE exam, access to Samba shares through Kerberos authentication has been added when Rsyslog client & server configurations have been removed.
Two easy tasks have been replaced with a complicated one, emphasizing again the weight of Kerberos.
A RHCE 7 candidate will definitely need serious Kerberos skills! At least, it’s what Red Hat people want.

Two days ago, the Linux Foundation announced a new certification program.

It consists in two exams:

• LFCS (Linux Foundation Certified System administrator),
• LFCE (Linux Foundation Certified Engineer).

Both exams are performance-based exam in English lasting 2 hours focusing on system administration and costing $300 ($50 until 24/8).

Candidates can choose between 3 Linux distributions (CentOS 6.4, OpenSUSE 13.1 and Ubuntu 14.04) before the beginning of the exam.

Compared to the RHCSA6 exam, the new LFCS exam on Centos 6.4 requires additional sudo, backup/restore and Raid configuration skills.

Similarly, compared to the RHCE6 exam, the new LFCE exam on CentOS 6.4 requires additional Apache SSL, Squid and IMAP/IMAPS configuration skills.

Time will tell us if this new Linux Foundation certification program can get some success.

Some applications writing files in the /tmp directory can see huge improvements when memory is used instead of disk.

You will find all the details in the tutorial about configuring /tmp on tmpfs.

If you don’t know the kpartx command, you miss something!

If you regularly play with virtual machines under KVM, you know that sometimes things go wrong.
A change in the /etc/fstab file or in the ssh configuration and you are in trouble, you can no longer access your virtual machine because you hadn’t set any virtual console: you need to reinstall it and it’s not fun!

But there is a solution! I call it emergency procedure.

This solution consists in stopping your virtual machine with the destroy command, then map your virtual machine image file in your physical host environment with the kpartx command. Mount the /boot partition and edit your /boot/grub2/grub.cfg file, adding the console=ttyS0 string at the end of every kernel boot line.
Finally, unmount the /boot partition, unmap your virtual machine image and reboot your virtual machine: from now on, you’ve got your virtual console!

This tip works at least for RHEL 6/CentOS 6 and RHEL 7/CentOS 7.

To get all the details, go to the RHEL 7 virtual console page.

With Systemd, the Init scripts are not there any more. Consequently, the execution of tasks at boot time had to change.

Hopefully, a nice solution has been found: it uses the good old rc.local file.

Now, to execute tasks at boot time, you put them into the /etc/rc.d/rc.local file.

Then, you change the execution permissions on this file:

# chmod u+x /etc/rc.d/rc.local

Besides allowing executions of tasks, this simple operation activates the new rc-local Systemd service for all the further boots (this is like systemctl enable in Systemd language).

Finally, you can start the rc-local service and, this way, test the execution of the rc.local file:

# systemctl start rc-local

SSD optimization is a good example of such a rc-local service.

Just a reminder for RHCEv5: you’ve got still 29 days to get recertified by passing the RHCEv6 only (a 50% discount is available in some countries).