The Googleplex of New York City

Upgrading OS Boot Partition (C: Drive) from X (500GB) to Y (1TB).

It is a true success story when a machine has reached a level of operation where its only increase in hardware is the local drive capacity.  When mentioned, upgrading an Operating System drive usually refers to the re-installation of the operating system itself however there are these mission critical servers which keep running and for whatever their reasons it’s not applicable for extra disks to be added.

Server has been running for a couple of years now and it has 2 disks in a RAID1 mirror and the C: drive is reaching near capacity, the plan is to upgrade and utilizing engineering expertise let us upgrade with 0% of server downtime and least amount of modifications, mission critical roles here.

Upgrade path is straight forward and simple.
1) Remove one drive from the mirror set
2) Insert the replacement upgrade drive
3) Rebuild the mirror set
4) Remove the other drive from the mirror set
5) Insert the second replacement upgrade drive
6) Rebuild the mirror set
7) Extend the volume into new space

Recently on ServerFault I was asked this question and never had document reference to share to others, a process which I am performing on a regular now made me think, what else hadn’t I documented to share with others… time for more blogging!

To make this short, I stumbled across an outstanding reference guide play by play which would for sure put me into the driver’s seat and explain it with the pictures and detail to answer exactly what was trying to be accomplished.

Credits go out to Russ, and his technical write-up over at Do It Yourself Home Workshop.
Link here.

Clouds are reshaping the traditional IT roles

Clouds use virtualized pools of computing to centralize resources and create a highly efficient agile and redundant environment.  The advancement in this offering is that it produces a way to lessen the complexity, inflexibility and capital expense of infrastructure and more importantly better control. 
Cloud Computing is transforming the evolving IT competencies adding a wider set of skills across the spectrum with the focus on virtualization, enterprise storage and security.

Troubleshooting the cloud technology in a mission critical application might require that you call upon multiple members of an IT department.  Here you will need your Storage, Networking and Virtualization Engineers and in addition possibly an application or operating system administrator.  

This creates a lot of dependencies and having to depend on all of these resources counters the cost saving benefits of the technology unless IT is able to present a broad array of skill competencies.
The outcome is that bleeding edge best-of-breed engineers will need to build stronger competency across all of these technologies as it will be more beneficial and enable us into the cloud with a higher value.

Read more at CIO here

eBay Chooses Joomla!

Free, Windows File System Replication

In the hosting industry we are often presented with problems needing solutions that demonstrate intelligence and cost effective savings without any compromises.
The confidence in this is demonstrating as an expert we have an understanding of what hardware technology suits a given solution and how that solution works, what are the instructions.

In a marketing seminar I once heard more you spend the less you know the less you know the more you spend.

There are factors in this however where the rules don’t always apply this same way.
A hardware manufacturer solution is backed up by the guarantee and for insurance this is generally a more comforting level for which we can sleep at night.  That does not go to say other methods don’t get the job done, it’s just a basis on the cost factor as to when you want to apply dollars to process.  Whether its investment in people, equipment or a guarantee.

Competitive bidding wars in the hosting industry, contending for the RFP and delivering capability the end results are from a paper perspective decided on what looks best on it, usually it is cost and guarantee.

A common problem and solution is that a customer is looking for redundancy between multiple servers on the file system layer and they are looking for a solution which does not add all of the costs of a enterprise storage system.

The scripts come out and today we talk about Robocopy, a free product and solution built into the Windows Operating System for some time now, we have had a tremendous amount of confidence and success in the commanding function.  Not even just my advocacy for Robocopy but I have multiple commentaries from Microsoft staff employees themselves whom work in the Windows Server department at Microsoft whom also speak out that Robocopy is way more successful than even some of the more complex software replication technologies including those others designed in house at Microsoft.

Robocopy is a command line function run from DOS shell that operates using syntax and variable switches.  You tell it what you want to do, where you are copying (replicating) from and where you will copy to (mirroring).  Here is the command we are going to demonstrate, we will follow with explanation of them.

Robocopy “D:\webserver-source” \\client-web2\webserver-bin /mir /mon:1  /mot:1 /R:3 /W:10 /COPY:DAT /log:c:\tools\robolog.txt /NDL /NP /Z

/MIR = mirror mode, keeps trees in sync which means it will also delete files if source has deleted files.
/MON = monitors the source and runs again when more than # of changes are detected.
/MOT = monitors the source and runs again in X minutes if changes are detected.
/R = number of times to retry the current scheduled attempt, three strikes, let’s play baseball.
/W = number of seconds to wait in between each try.
/COPY = what flags to copy for files D-data A-attributes T-timestamps S-securityNTFSacls O-owner information u-auditing info.
/LOG = output log path and filename, overwrites existing log.
/NDL = no directory listing, don’t log the directory names, we know from our source where we are.
/NP = don’t display progress % of copied.
/Z = copy files in restartable mode.

Create and save this as a batch file!
It helps that you have your web-server source and destination the actual same path and folder names to keep things organized.  The first initial copy will always take the longest.

Set it as a scheduled task.  Have it run 23 hours a day, have the process kill when you are most comfortable checking this in the morning or evening and if all looks fine, re-run and repeat process.

For additional reading and understanding of Robocopy please visit Wikipedia link @
Microsoft Robocopy Technet Reference

Check out My Digital Life for an additional overview of Robocopy syntax and switches.