Thursday Jan 31, 2013

Deep Inside Every Sysadmin Is ...


photo courtesy of James Vaughan - all rights reserved

... an Oracle ACE!

The thrills. The glory. The fame. Who can resist?

Turns out sysadmins can.

Last time I checked, the Oracle ACE program had 417 ACES and ACE Directors. Of those 417, only 6 have a specialty in Oracle Solaris or Oracle Linux.

That's simply not enough to defend the homeland! I know there are many more sysadmins and developers proficient in Oracle Solaris and Oracle Linux who can qualify to become Oracle ACES. Toss your silk scarf around your neck and grab your goggles. Then talk to me on the OTN Garage page on Facebook. I'll explain the benefits and help you enroll in pilot training.

Current Oracle ACES With Specialties in Oracle Solaris or Oracle Linux

Suk Kim, ACE Director, Oracle Solaris, Korea

Proficient in Oracle Solaris system tuning, troubleshooting Oracle Solaris security, audit information security, penetration tester incident and response, digital forensics virtualization, and cloud computing. Member of Korea Oracle Solaris User Network, Chairman of Oracle Solaris Tehchnet, Manager of Solaris School, adjunct professor at Ansan University, senior consultant at NoBreak Co., LTD.

Diego Aguirre, ACE, Oracle Solaris, Argentina

Diego Aguirre has been a Solaris Support Specialist since 1998. Over the past several years, he has contributed to the Oracle Solaris Community and has published technical articles for Sun Microsystems and now Oracle. He is the author of

Alexander Eremin, ACE, Oracle Solaris, Russia

Alexander Eremin is a user on Solaris and Linux platforms since 1995. Over the past ten years, he has worked as a Senior Unix Administrator. He is also the creator of the MilaX - Small Live Distribution of OpenSolaris. Alexander is also taking part in the Caiman OpenSolaris project.

Julien Gabel, ACE, Oracle Solaris, France

Julien Gabel is a Multi-platform UNIX systems consultant and administrator in mutualized and virtualized environments. He has architecture and expertise in building Solaris and UNIX experience in large enterprises such as banking and financial services, IT services, Telecoms and multimedia companies.

Raimonds Simanovskis, ACE, Oracle Linux, Latvia

Raimonds Simanovskis in founder of EazyOne which develops business intelligence web application Previously he was working at Tieto Latvia where he was using and promoting new technologies, open source and Agile software development. Raimonds has participated in many Oracle E-Business Suite implementation projects as well as Oracle based software development projects. In recent years he is active Ruby language and Ruby on Rails framework user and contributor. He has created and maintains Oracle database adapter for Ruby on Rails as well as PL/SQL and Ruby integration libraries.

Damian Wojslaw, ACE, Oracle Solaris, Poland

Damian is currently working as systems operator since 1999. Since 2006 he has worked with Solaris and OpenSolaris operating systems and other Sun Microsystems born applications. He blogs regularly on TrochejEN and reposts on Planet OpenSolaris. Damian has translated four OpenSolaris related Guides (ZFS Administrator Guide, OpenSolaris Installation Guide: Basic Installations, DTrace User Guide, Device Driver Tutorial) to Polish.

Defend the homeland!

- Rick

Website Newsletter Facebook Twitter My Personal Blog

Friday Jan 18, 2013

Once Upon a Time in the Kingdom of Serv

If you're the type of person who has no time to read fairy tales, scroll to the very bottom for a link to the article.

Once upon a time there was a very happy Kingdom called Serv. It was ruled by inventors called engineers. Most of the engineers were clever, kind, and handsome. They had beautiful wives who cooked them tasty and nutritious meals.

A few of the engineers, however, had wives with big, hairy, purple moles, who sat around all day watching reruns of Bridezilla while chomping loudly on pork rinds. They never served their engineer husbands any meals and instead, screamed at them to get them another bag of pork rinds. And they hated sysadmins.

Sysadmins were the workers of the Kingdom. They were very playful, and they had big strong hands. They spent their days tossing servers back and forth to each other, or playing hacky sack.

The Kingdom was a happy place because the clever, kind, and handsome engineers had long ago invented a wonderful contraption called, as you would expect, a "server." Servers were loved throughout the Serv kingdom and all the surrounding kingdoms. They came in shiny metal boxes and had blinking lights. Best of all, they had straight edges so that sysadmins could toss them back and forth to each other. Sysadmins loved tossing servers back and forth to each other, and at lunch time it was not uncommon for several servers to be in the air at once. But when a sysadmin dropped a server, it usually broke. And when a server broke, it was called a "failure." And a failure always woke up The Boss.

The Boss was a hairy ugly giant with one eye. He did only two things. He slept. And he fired sysadmins for waking him up. Naturally, everybody preferred to keep the boss asleep. Especially sysadmins.

Polite people in the Kingdom never mentioned the word "failure" at dinner parties, not even in a whisper, lest they unwittingly awaken The Boss. But everybody knew that if sysadmins began to appear on their sofas in the middle of the night, somewhere in the Kingdom a failure had occurred.

The wives of the clever, kind, and handsome engineers begged their husbands to do something about the plight of the playful sysadmins. And so the clever, kind, and handsome engineers invented the cluster. A cluster was an enchanted cable that connected groups of servers in a magical way. When one server was dropped by a sysadmin, the cable moved that server's applications to another server so fast that nobody had time to even think of saying "failure," much less say it loud enough to wake The Boss. When the dropped server was fixed, the enchanted cable moved that server's applications back.

And so the Kingdom was full of happy sysadmins tossing servers back and forth during lunch, and sleeping in their very own beds at night.

This turn of events, of course, made the pork rind and Bridezilla wives jealous. During the commercials they screeched at their browbeaten husbands until they invented a curse to get the sysadmins fired again and back on the sofas of the beautiful wives who cooked their engineer husbands tasty and nutritious food.

It was an unspeakable curse, and polite people at dinner parties didn't dare to even whisper its name. When this curse was unleashed upon the Kingdom, all the beautiful metal servers disappeared. Except one. And inside that one server were trapped the spirits of all the other servers. The sysadmins stood around staring at it, wondering of what use their big strong hands were when the servers no longer had bodies.

One by one the sysadmins grew sad and left, and in no time at all, almost all the clever, kind, and handsome engineers had sysadmins sleeping on their sofas again.

The Kingdom was not a happy place.

Until one day, it occurred to the cleverest, kindest, and most handsome of the clever, kind, and handsome engineers to put a spell on the enchanted cable so that it could do the same thing for the spirit servers that it once did for the physical servers.

It was a wonderful invention, and the sysadmins jumped off their sofas to learn how to use it. And to keep the pork rind-chomping, Bridezilla-watching wives of the browbeaten engineers guessing, the enchanted cable could be used in two different ways:

Two Ways to Create a Cluster from Logical Domains

  • Configure logical domains within Oracle Solaris Cluster
  • Configure Oracle Solaris Cluster within Oracle VM Server for SPARC

The first approach is fairly obvious. You can put one or more applications inside each domain and create a cluster from all the domains. When a particular domain goes down, the applications running inside it get moved to a working domain. The domains are controlled individually through Oracle VM Server for SPARC, and the cluster is controlled by Oracle Solaris Cluster.

The second approach is more involved, but it provides significant benefits. It consists of setting up Oracle Solaris Cluster inside the control domain of Oracle VM Server for SPARC. When deployed this way, Oracle Solaris cluster can manage guest domains as "black boxes," which allows a site to isolate the administration of guest domains from each other. With this approach, from within Oracle Solaris Cluster you can:

  • Create guest domains
  • Live- and warm-migrate the guest domains
  • And manage individual applications like you can with the first approach

The second approach is well documented. In fact, Venkat Chennuru, a sysadmin with big strong hands who was elevated to the rank of clever, kind, and handsome engineer, took the trouble to write it down for us. You can find his article on OTN:

How to Configure a Failover Guest Domain in an Oracle Solaris Cluster

Read it, learn how to do it. Because as you know, evil never rests.

- Rick

Website Newsletter Facebook Twitter My Personal Blog

Friday Jan 11, 2013

How to Install Oracle Linux from a USB Stick


If you want to install Oracle Linux from a USB drive, keep in mind that not all hardware supports USB device booting. Also, during the boot process you may have to instruct your BIOS to boot from that specific USB device. Finally, keep in mind that this method of installation is not officially sanctioned by Oracle support. You'll need an Oracle Linux 6.0 or higher system to produce the key. Earlier versions may work, but additional prerequisites may be required. The examples in this article assume a USB key device name of /dev/sdb1. Be sure to verify the device name of your USB key to avoid accident data loss.


  1. The first thing you will need is an ISO image of Oracle Linux. The quickest way to obtain an ISO image is from the Oracle Software Delivery Cloud
  2. You will need a desktop or server system running Oracle Linux in order to prepare your USB drive.
  3. You will also need to download this script to create the bootable USB drive.
  4. Your Oracle Linux system will also need the package syslinux installed. You can install syslinux using yum with the following command:
  5. yum install syslinux

Marking Partition One as Bootable

Once your prerequisites are in order, you need to designate partition one as bootable. Use the parted application, as in this example:

[root@host]# parted /dev/sdb 
GNU Parted 2.1 Using /dev/sdb Welcome to GNU Parted! Type 'help' to view a list of commands.
(parted) toggle 1 boot
(parted) quit
Information: You may need to update /etc/fstab.

The example above uses a USB key labelled /dev/sdb. The parted application will only accept device files without partition numbers. So, if we had selected /dev/sdb1 instead, we would have gotten an error message when we tried to write the changes to disk.

Creating the USB Key

Now you can start creating USB key via the script that you downloaded earlier. The script accepts two paths: first the source ISO file and then the USB key:

[root@host]# sh Install_OL_fromUSBStick_Script --reset-mbr /home/user/OL6.3.iso /dev/sdb1 
Verifying image... line 527: checkisomd5: command not found Are you SURE you want to continue?
Press Enter to continue or ctrl-c to abort
Size of DVD image: 2957
Size of images/install.img: 132
Available space: 31186
Copying DVD image to USB stick
    137834496 100%   10.87MB/s    0:00:12 (xfer#1, to-check=0/1)
sent 137851396 bytes  received 31 bytes  11028114.16 bytes/sec total size is 137834496  speedup is 1.00
sent 37 bytes  received 12 bytes  98.00 bytes/sec total size is 3100217344  speedup is 63269741.71 Updating boot config file Installing boot loader USB stick set up as live image!

Once the script is finished running you have a bootable USB drive that can install Oracle Linux. While booting, pay attention to your BIOS boot screens as they will often provide direction on how to select a specific boot device other than the ones in the standard boot sequence. For some older systems you may need to go directly into the BIOS setup utility to specify the USB device in your boot sequence. Once you have booted successfully off of your USB device and the installer starts installation will proceed just like an installation from regular DVD media.

- Robert Chase

Website Newsletter Facebook Twitter

Thursday Jan 10, 2013

Recent How-To Articles About Oracle Solaris Zones

LEGO Clone Army Collection

How to Put Oracle Solaris Zones on Shared Storage for Easy Cloning

by Jeff Victor

What is ZOSS? Zones on shared storage. Why would you do that? When you configure a zone on shared storage, you can quickly clone it on any server that uses that storage. Jeff explains how.

How to Create Oracle Solaris 11 Zones with Oracle Enterprise Manager Ops Center

You can also create Oracle Solaris 11 zones with Oracle Enterprise Manager Ops Center 12c. When you set up a zone this way, you can add the zone to a server pool and use the zone migration feature of Oracle Solaris 11. In this article, Laura Hartman shows you how to create and configure a non-global zone from the Plan Management View of this handy-dandy tool.

How to Create a Load-Balanced Application Across Two Oracle Solaris Zones

by Laura Hartman

Install Apache Tomcat on two Oracle Solaris zones. Connect them across a VPN. And let the Integrated Load Balancer in Oracle Solaris 11 manage traffic. Presto: high(er) availability in a single server.

- Rick

Website Newsletter Facebook Twitter My Personal Blog

Wednesday Jan 09, 2013

How to Treat an NFS File As a Block Storage Device


Wim actually beat me in blogging about this feature while I was on vacation, but I'd like to add a little more background about dm-nfs, which I gathered from our kernel developers:

What is dm-nfs?

The dm-nfs kernel module provides a device-mapper target that allows you to treat an NFS file as a block device. It provides loopback-style emulation of a block device using a regular file as backing storage. The backing file resides on a remote system and is accessed via the NFS protocol.

The general idea is to have a more-efficient-than-loop access to files on NFS. The device mapper module directly converts requests to the dm device into NFS RPC calls.

dm-nfs is used transparently by Oracle VM's Dom0 when mounting NFS-backed virtual disks. It essentially allows for asynchronous and direct I/O to an NFS-backed block device, which is a lot faster than normal NFS for virtual disks. The Xen block hotplug script has been modified on OVM to look for files which are on NFS filesystems. If the file is on NFS, OVM uses dm-nfs automatically, otherwise it falls back to using the regular (but slower) loop mount method.

The original dm-nfs module was written by Chuck Lever. It has been supported and used by Oracle VM since version 2.2 and is also included in the Unbreakable Enterprise Kernel for Oracle Linux.

Why this feature matters

This feature creates virtual disk devices (LUNs) where the data is stored in an NFS file instead of on local storage. Managed networked storage has many benefits over keeping virtual devices on a disk local to the physical host.

A sample use case is the fast migration of guest VMs for load balancing or if a physical host requires maintenance. This functionality is also possible using iSCSI LUNs, but the advantage of dm-nfs is that you can manage new virtual drives on a local host system, rather than requiring a storage administrator to initialize new LUNs on the storage subsystem. Host administrators can handle their own virtual disk provisioning.

For durability and performance, dm-nfs uses asynchronous and direct I/O so all I/O operations are performed efficiently and coherently. Guest disk data is not double cached on the underlying host. If the underlying host crashes, there's a lower probability of data corruption. If the guest is frozen, a clean backup can be taken of the virtual disk, as you can be certain that its data has been fully written out.

How to use it

You use dm-nfs by first loading the kernel module, then using dmsetup to create a device mapper device on your file. The syntax is very similar to the dm-linear module.

The following sample code demonstrates how to use dmsetup to create a mapped device (/dev/mapper/$dm_nfsdev) for the file $filename that is accessible on a mounted NFS file system:

nblks=`stat -c '%s' $filename`
echo -n "0 $nblks nfs $filename 0" | dmsetup create $dm_nfsdev

Now you can mount /dev/mapper/$dm_nfsdev like any other filesystem image.

- Lenz Grimmer (Oracle Linux Blog)

Website Newsletter Facebook Twitter

Tuesday Jan 08, 2013

How to Upgrade an Oracle Solaris 11 Repository with 11.1 Packages

These instructions assume you already have a local Oracle Solaris 11 11/11 IPS package repository, and you want to update that local repository by adding Oracle Solaris 11.1 packages.

You can either use the pkgrecv command or you can download a repository image file.

Using pkgrecv

Using pkgrecv takes more time and requires your system to be connected to the Internet for a much longer time. If you use pkgrecv, specify as the source, and be sure to specify the -m all-versions option. See the pkgrecv(1) man page for more information.

Using a Repository Image File

If you prefer to use a repository image file, first download the image file and then copy the contents to your existing repository. You can get the Oracle Solaris 11.1 repository image files from OTN or from eDelivery. Then follow these instructions.

Step 0. Assume the root role and snapshot your local Oracle Solaris 11 11/11 repository.

In this example, your local Oracle Solaris 11 11/11 repository is located at /export/repo/Solaris11.

# zfs snapshot rpool/export/repo/Solaris11@11-1111

Step 1. Download Oracle Solaris 11 from OTN.

Go to the download page. Read the "OTN License Agreement for Oracle Solaris" and click the Accept License Agreement button.

Step 2. Click the "MD5 checksum" link to display the checksums for the files you want to download.

Step 3. Scroll down to "Oracle Solaris 11.1 Repository Image."

Step 4. Click "Download Part A SPARC, x86" to get the file sol-11_1-repo-full.iso-a.

Verify the checksum like this:

# digest -a md5 sol-11_1-repo-full.iso-a

Step 5. Click "Download Part B SPARC, x86" to get the file sol-11_1-repo-full.iso-b.

Verify the checksum.

# digest -a md5 sol-11_1-repo-full.iso-b

Step 6. Click "Download Oracle Solaris 11 Pre-Upgrade Repository Image SPARC, x86 to get the file sol-11_1-upgrade-repo.iso.

Verify the checksum.

# digest -a md5 sol-11_1-upgrade-repo.iso

Step 7. Create one image file from the Part A and Part B files.

Verify the checksum of the resulting image file.

# cat sol-11_1-repo-full.iso-a sol-11_1-repo-full.iso-b > sol-11_1-repo-full.iso

Step 8. Merge the Oracle Solaris 11.1 packages from the repository image file into your local Oracle Solaris 11 11/11 repository.

# mount -F hsfs sol-11_1-repo-full.iso /mnt
# rsync -aP /mnt/repo/ /export/repo/Solaris11
# umount /mnt

Step 9. Merge packages from the Oracle Solaris 11 pre-upgrade repository image into your local Oracle Solaris 11 11/11 repository.

# mount -F hsfs sol-11_1-upgrade-repo.iso /mnt
# rsync -aP /mnt/repo/ /export/repo/Solaris11
# umount /mnt

Step 10. Catalog new packages.

# pkgrepo refresh -s /export/repo/Solaris11
# pkg refresh solaris

You can use the pkgrepo info and pkgrepo get commands to check the properties set on the updated repository.

Step 11. Snapshot your updated repository.

# zfs snapshot rpool/export/repo/Solaris11@11.1

Step 12. Check that your Solaris publisher origin is set to your local repository.

- Alta Elstad

Website Newsletter Facebook Twitter

Thursday Jan 03, 2013

Extreme Sports and the SPARC T5 Chip

In my day, you were extreme if you surfed. Nobody had a leash. If a wave ripped the board out of your hands, you had to swim all the way back to the beach. In big surf, that could take 15 minutes. And then you had to paddle out again.

Today, if you're not juggling rusty chainsaws while riding a BMX bike off the top of the Eiffel Tower with half a dozen angry chimpanzees trying to rip off your helmet and goggles, you're not considered extreme.

Which is exactly why the SPARC T5 chip has 1024 functional CPUs. None of Oracle's SPARC engineers wanted to find himself at a cocktail party having to confess in an embarrassed chortle to a salon full of top hats and sequins that he had designed a mere 512 CPU chip. Imagine the chagrin!

Interview: Deep Dive into the SPARC T5 Chip

So Oracle's SPARC engineers worked wicked hard to scale the T5 to eight sockets. Since each socket has 16 cores, that gives you 128 cores, total. Since each core can support eight individual threads (or strands, if you're not the sartorial type), you wind up with a total of 1024 functional CPUs.

As you know, processing power without bandwidth is kinda like a mega motor with a nano fuel tank. Doesn't get you too far. So the T5 also has memory bandwidth to match its processing power. And lots of other capabilities that you an read about in this:

Interview: Deep Dive into the SPARC T5 Chip

- Rick

P.S., If you want to read about surfing in the old days, check out Chapter 1 of Tocayos, a novel I've been writing and posting online in my spare time.
Website Newsletter Facebook Twitter My Personal Blog

Rick Ramsey
Kemer Thomson
and members of the OTN community


« January 2013 »
Blogs We Like