VIO:

ioslevel –> VIO version

oem_platform_level –> AIX version of VIO server

HMC:

lshmc -V –> list the version of your HMC

First off, you will need to get the following dependencies.  I was able to get most of them from the IBM Linux Toolbox here, and the rest I got from the great people at Perzl.org

Dependencies:

bison-1.875-3.aix5.1.ppc.rpm
coreutils-5.2.1-2.aix5.1.ppc.rpm
expect-5.45-1.aix5.1.ppc.rpm
flex-2.5.4a-6.aix4.3.ppc.rpm
gcc-4.2.0-3.aix6.1.ppc.rpm
gcc-cplusplus-4.2.0-3.aix6.1.ppc.rpm
gdbm-1.8.3-5.aix5.2.ppc.rpm
gettext-0.10.40-8.aix5.2.ppc.rpm
git-git-5bc2dc2
git-git-v1.7.10.2-569-g5bc2dc2.zip
libXft-2.2.0-1.aix5.1.ppc.rpm
libgcc-4.2.0-3.aix6.1.ppc.rpm
libstdcplusplus-4.2.0-3.aix5.3.ppc.rpm
libstdcplusplus-devel-4.2.0-3.aix6.1.ppc.rpm
make-3.80-1.aix5.1.ppc.rpm
popt-1.7-2.aix5.1.ppc.rpm
python-2.6.2-1.aix5.3.ppc.rpm
python-devel-2.6.2-1.aix5.3.ppc.rpm
readline-4.3-2.aix5.1.ppc.rpm
readline-devel-4.3-2.aix5.1.ppc.rpm
unzip-5.51-1.aix5.1.ppc.rpm
zlib-1.2.3-4.aix5.2.ppc.rpm
zlib-devel-1.2.3-4.aix5.2.ppc.rpm

My suggestion, just put them all in a folder and run “rpm -i -v *” in the directory.

Now that all dependencies are met (hopefully), we can go ahead and grab the GIT source from here.  I simply downloaded this, SCP-ed it to my machine and ran “unzip *.zip” on the file.

cd into the git directory and “vi Makefile” in order to change a few things in the make instructions.  Here are the changes I made:

CC = /opt/freeware/bin/gcc
AR = ar
RM = rm -f
DIFF = diff
TAR = /usr/bin/tar
FIND = find
INSTALL = /usr/linux/bin/install

Please don’t blindly follow these changes.  Change these values to the corresponding directory.  You can get this information by running “whereis” or “which” on each utility.

The final step is the command.  Since I don’t have X installed and don’t want any GUI options, I modified my install command to eleviate some of the extra dependencies that I didn’t want to install.

“gmake NO_CURL=1 NO_FNMATCH_CASEFOLD=1 NO_MSGFMT=1 NO_TCLTK=1 install prefix=/opt/freeware”

This turns off things like curl, fixes my issues with fnmatch and turns off tc.  It also installs GIT to /opt/freeware.

Hope you all enjoy and things go smoothly.  If not, post any problems or questions here.

Myths:

Certifications are Vendor-centric

• While this is true, the negative outlook on this truth is unwarranted.  Most technologies are developed by specific vendors, and certain vendors are the best to oversee certification over the technology.  This means that while a certification may be vendor specific, that vendor’s certification will prove a highly technical mastery of the subject at hand.

Certification’s Life Cycle Is Short

• This is a benefit of the changing times and increasingly powerful IT innovations.  The ability to prove mastery over a current, cutting edge technology has many benefits and employers do see a difference between those certified in archaic technologies and those keeping up with innovation and trends

Certifications Are Not Real-World Oriented

• This criticism has holes in it, being that real world changes and is different from place to place.  Holding certifications in multiple technologies at different levels shows your ability to understand multiple aspects of a technology.  Many certifications also provide real world simulations or environments as a part of the process, and therefore this myth loses truth between vendors and technologies

Certifications Have Been Devalued

• It has been proven that those holding certifications have higher than average salaries, and that cheating is becoming increasingly difficult if not impossible for many certifications.

No Oversight Body

• The Information Technology Certification Council (ITCC) was formed to oversee the certification process and is comprised of many IT leaders such as IBM, HP and Microsoft

Degree vs. Certification vs. Experience

• Certifications prove a deeper understanding and knowledge of specific technologies that degrees cannot match.  Also, certifications speak volumes to those holding them.  Certifications are often seen as tie breakers when other credentials are equal

HR People Are Not In Touch with the Real World

• Certifications stand as a clear delimiter of in depth knowledge over a subject and therefore are an obvious benefit to HR execs.

Budget Cuts

• Studies have shown that companies supporting tuition and certification reimbursement attract higher attention and therefore these remain to be values of top competitor companies.

Glut of Certified People

• A gap exists between the required knowledge and the available skilled professionals.  Certifications help bridge that gap by providing the market with skilled employees.

No One Knows Which Certs Matter

• Certifications add a widely accepted benefit to any professional.  An employer needn’t know the specifics of the certification to understand the benefit of hiring certified professionals.

Overall I found this write up to be beneficial and to provide further justification for seeking certifications whether you are currently employed with an organization that provides reimbursement, or if you are on the job market and looking for an edge over your competition

• Multiple choice
• Video recording
• Computer based
• Timed
• Given a scrap of paper for notes
• Cannot bring cell phone’s into room

If you go to the IBM website you will see that most tests are broken down into categories.  These breakdowns are very accurate as you also get a print out at the end of how you did in each category. If you aren’t strong in a subject, LEARN IT.  It won’t be like in highschool where the teacher says something might be on the exam and it isn’t.  Each category listed is in fact on the exam, with weighting how it is outlined.

1. Because you have a well monitored system, you will get an email alert that a threshold has been reached and the file system is filling up
2. Users are screaming because their applications are breaking and files aren’t being saved…

Regardless of which is happening to you, first thing to do is figure out WHY the file systems are filling up.

Nine times out of ten you’ll see that users are putting files where they shouldn’t.  Storing files or logs in the /home directory, saving stuff to /tmp and not deleting it, these are things you’ll see everyday.  Here’s how to figure out what’s going on.

$ df -m -> display the free space, in megabytes, on all file systems

$ who -> always good to see who is on the lpar (if the problem just started, you’ll see a list of possible culprits)

$ last -> see description for “who”

$ du -m | sort -nr | more -> this is my favourite.  This will tell you the disk usage in megabytes, sorting it by the largest files. Go to the file system that is receiving errors, or that is full as per “df”, and complete this command.  You will see right away what files are taking up the most space.  Often times you will be able to pick out things like log files for runaway processes, installation files that have not been deleted, misplaced data files…

I hope these tips help you in the future to determine where your space went when a file system fills up

**NOTE: this is made much simpler by disabling root SSH capabilities.  If you have not done this, please not that it is recommended to disallow SSH directly for root user as this will mask who has been logging into the system (as well as cause other security holes)

$ who -> This will tell you who is currently logged into the system, from where and when they logged on

$ last -> This will give you a list of previous logins, the source and time, both in and out

$ who /etc/security/failedlogin -> List all of the failed logins with the same details as above

$ history -1000 -> If you use shared user ids, or you believe someone using root did something, simply check the history using “history -”number_of_lines”

I will include a script to view the history of users.  Check back soon.

I hope this problem determination guide helps you keep track of users logged in and tracing back who did what

1. Identify resources used by Cluster Aware AIX
2. Configure dump devices and analyze output
3. Determine elements necessary to reduce single points of failure
4. Understand geographical logical volume manager (GLVM)
5. Maintain hardware (CEC/Blade Hardware), deferred or concurrent firmware, AMM firmware (for Power Blades), whether an adapter replacement or new installation.

Storage Management (21%)

1. Create and manage filesystems
2. Create and manage logical volumes
3. Create and manage volume groups
4. Manage physical and virtual devices
5. Manage storage devices (traditional disk, Solid State Drives, and tape) including redundancy

System and Network Security (4%)

1. Configure Role Based Access Control
2. Configure and Manage remote access

Partition Management (11%)

1. Configure and manage Logical Partitions (LPARs), including DLPAR operations
2. Create and manage Workload Partitions (WPAR), including Versioned WPARs and planning for Live Application Mobility
3. Understand HMC, SDMC and IVM interfaces
4. Create and Manage VIO Server partitions (including working with virtual repositories)
5. Understand and explain LPAR and WPAR migration and mobility at a basic level

Performance Management and Tuning

1. Use performance monitoring tools, and plan for future growth
2. Analyze output from performance monitoring tools
3. Configure system tunables to support optimal application performance

Network Management (11%)

1. Configure network devices (including Etherchannel, IPv4, and IPv6)
2. Troubleshoot network issues
3. Configure TCP/IP, with and without VLAN support

System Management (18%)

1. Create, maintain and modify user accounts
2. Manage services and subsystems (using chtcp, etc)
3. Configure Electronic Service Agent
4. Configure the system and device attributes
5. Use AIX system management tools (for example: DSM, Director, NIM)
6. Install, apply, commit, or reject software
7. Create and manage paging space
8. Use Cron and At functions

Install and Manage AIX (11%)

1. Understand and manage AIX instance startup
2. Backup and restore AIX
3. Migrate WPARs from AIX 6 to AIX 7
4. Install AIX and use NIM environments
5. Install Versioned WPARs

General administrative tasks (7%)

1. Create and use ksh and Perl scripts at a basic level
2. Use AIX commands such as TAR, CPIO, DD, RPM, SAVEVGSTRUC, and explain their use

By consolidating your LPAR environment, you can create more easily manageable servers for environments with multiple development, sandbox and testing deployments.  Often times with the increase in virtualization, we treat LPARs as though they are disposable machines (which technically they are).  By creating a WPAR, you increase the ease of set up, tie up less physical resources, and keep your environment breadth steady.  You will also only need to use one OS image, rather than maintaining and purchasing multiple AIX licences.

The downside to WPARs is the single point of failure.  Performing general maintenance on an LPAR which results in an outage will create outages for all running WPARs on the LPAR.  This means that upgrades etc. on an LPAR become the equivalent of those on a frame.  If it goes down, you might be bringing down several LPARs at a time.

Two types of LPARs exist: System and Application. System WPARs are the most similar to a fully functioning LPAR.  It has inetd and cron daemons and consists of file systems etc.  Application WPARs are WPARs created solely for a single application.  Once the application completes and closes, the WPAR disappears.  Application WPARs don’t run daemons and do not contain file systems.

Some commands you might want to familiarize yourself with are:

mkwpar -n wpar_name -> create a new WPAR with the specified name

lswpar -> list existing WPARs and distinguish those that are active from those that are not

startwpar wpar_name -> start the specified WPAR

clogin wpar_name -> login to the WPAR (will login to an AIX environment)

savewpar -> create backup of WPAR

restwpar -> restore the WPAR

** Creating File Systems -> You will have to create file systems from the global environment under the /wpar file system and then mount the file system in the WPAR

**Creating Users -> Users can be managed within the WPAR and those created in the WPAR will be restricted to the WPAR (will not affect the global environment)

I hope this explanation of WPARs and their place in your environment has helped, and that the included commands will help get you started.

We all know that vi (or any other text editor) can easily edit a text file such as the inittab.  We also know that nobody is perfect, and that mistakes do happen (escape characters, wrong formatting etc.)  The problem is, when editing the inittab, if you make an error in the wrong place, you can render the system unbootable.  This is because the inittab is an important part of the boot sequence for an lpar, and once the boot sequence get mucked up, bad things start to happen.

This guide will go over the correct ways to edit the inittab without the risk of breaking the lpar altogether. (The only way to recover from some inittab errors are to network boot and fix from there…)

lsitab -a -> list contents of the inittab

lsitab identifier -> find specified identifier in inittab and display entry

chitab identifier:run_level:action:command -> change the run level, action and command of the specified identifier

mkitab identifier:run_level:action:command -> create new entry in the inittab

mkitab -i search_term identifier:run_level:action:command -> create new entry in the inittab after specified identifier

rmitab identifier -> remove specified entry from the inittab

Often times in a script I’ll want to strip output based on a delimiter or columns.  I found myself constantly fighting with IFS and trying to make sense of delimiters that don’t follow a common pattern.  My two ways of extracting columns from a file/variable are as follows:

echo $variable | cut -f 2 -d “:” –> (-f to select column, -d to select delimiter)

echo $variable | awk ‘{ print $2 }’

Believe it or not, for someone that started off with no formal shell training or anything, this took me awhile to figure out.  Cut has its ups and downs but I will say this, awk is without a doubt worth learning.  In its entirety.

Anyways, this was just a quick update on how to extract columns from data using cut and awk