Editor's note: This is the sixth and final part in a series of articles on Unix-to-Linux migrations on SearchEnterpriseLinux.com.
After you've learned about the hurdles for the actual Unix-to-Linux migration, what about training? What about backup and high availability? How do you monitor performance and tune your systems? How do you ensure adequate skill sets and certifications for your staff?
Systems management across Unix and Linux
Don’t let anyone convince you that Unix and Linux are mostly the same. There are many differences –some subtle and some not so subtle.
The three popular Unix flavors that dominate the market – AIX, HP-UX and Solaris – are different in so many ways. These differences include the kernel, performance monitoring and tuning commands, networking, configuration, virtualization and other processes.
With AIX, you have a dominant GUI, or SMIT, short for the system management interface tool. The other flavors of Unix have similar front ends as do the Linux distributions.
Although you can use the command line for most AIX tasks, it’s actually recommended that some
tasks be performed by SMIT, or else they won’t save on a reboot. Networking commands such as
ifconfig are a good example of this. On Linux, you can use
it will save on a reboot.
With AIX, you use certain commands to change kernel processes such as
schedo. With Solaris, you use
Linux, you use
If you want to view your run-level, with AIX and Solaris you type in
who –r. With
Linux, you use the
runlevel command. When you want to check your swap space, with AIX
you use the
lsps command, and with Linux you use
swapon – s or
When looking for filesystem information, with Solaris you use
Linux, you use
fstab, and with AIX it’s
These are just the basic commands. When you’re working with logical volumes and filesystems, things can get that much more complex. Remember that you need training to be effective. Many vendors today offer custom classes specifically for Linux administrators who have come from Unix environments. Look into taking specific classes from your hardware vendor. IBM offers a class called Linux Jumpstart for Unix professionals.
There are also several advantages of staying on the same hardware platform when considering a Unix-to-Linux migration. Let’s look at IBM’s Power Systems, which can run either AIX or Linux distributions from Red Hat or Novell SUSE. Because the virtualization engine–PowerVM–on IBM Power Systems is the same regardless of whether you run AIX or Linux partitions on your server, managing virtualization and logical partitioning is the same across operating systems. Only having to understand one type of virtualization technology across several OS platforms is a huge plus.
Another plus to staying on one hardware platform is that you don’t have to worry about endianess– a concept that can cause big problems when moving to different platforms. And you don’t have to learn another architecture or retrain your hardware staff, assuming you have a large enough environment to have one.
Backup and High-Availability Tactics
How do you back up your systems now? In most large enterprise IT shops, you would probably use a third-party product like NetBackup or TSM. This kind of software offers both Unix and Linux clients.
If you’re using a generic Unix tool to back up the systems portion of your data, you will need
to look at equivalent tools for your Linux distribution because your Unix-specific tools will not
work. I’ve even seen problems using generic Unix/Linux tools such as
cpio between the different platforms, so be careful.
As far as high availability is concerned, you may be in a position to use similar tools. For example, if you’re using Veritas for High Availability on Solaris, Veritas also comes with a Linux client. Same thing with IBM’s PowerHA, formerly referred to as HACMP, which also has a Linux version in addition to the AIX version. The best bet is to use the tools that you have the most experience with, if at all possible.
Monitoring and Performance Measurement Tactics
Performance monitoring is also different with Unix and Linux. The following table shows some of the differences among the performance tools for Unix and Linux as well as their purposes. Again, training is key.
Unix/Linux Performance Tools Matrix
|HP-UX||GlancePlus, MeasureWare/Perfview, Caliper||
|RHEL||sysstat, systemTap, oprofile.||
|Solaris||SE Toolkit, sysperfstat||
Ensuring Adequate Skill Sets or Certifications
Linux certifications may have come a long way in recent years, but Red Hat still sets the bar with its Red Hat Certified Engineer, or RHCE, available since 1999. Red Hat now offers the RHCT (technician) and the RHCA (architect). With the RHCE certification, you need to pass a full-day hands-on lab consisting of a written test, server install and network lab.
The Linux Professionals Institute has the LPI certification, which is designed to be distribution-neutral following the Linux standard base and other related conventions. It is three-tiered: Level 1 is for junior administrators, Level 2 is for intermediates, and Level 3 is for more advanced engineers and administrators.
CompTIA Linux+ is another vendor-neutral certification, though it is not as prestigious. Novell has three certifications: the CLP for entry level, the CLE engineer classification, and its newest certification called the Novell Certified Linux Administrator, or CLA, which certifies that the recipient knows how to perform day-to-day administration of installed SUSE Linux Enterprise Server networks.
About the author:
Ken Milberg is the President and Managing consultant of PowerTCO (formerly known as Unix-Linux Solutions), a NY-based IBM Business Partner. He is a technology writer and site expert for techtarget.com and provides Linux technical information and support at SearchEnterpriseLinux.com.
This was first published in July 2010