Google

Saturday, February 16, 2008

Welcome!

This is my first post of this blog. I have lived nearly half-a century on this planet and have committed a number of mistakes. I thought it would be great to share those, so that others can learn from My mistakes, so that they need not repeat, but creatively make new mistakes!

Today's is not exactly a mistake, but can turn out to be:

We wanted to buy some blade servers and after going through lot of claims and counter claims in various web pages and literature, by a slim advantage of price, we finalized IBM Blade Center.

Even before we commissioned, we decided to check the claim of full redundancy of power. (Please remember: This is the area where bloody battles were fought between HP & IBM!).

The claim of ALL major players is that they would be able to power their blade setup (with fully configured blades, occupying all the slots) with half of their power supplies being shut-off.

In the present case, the chassis was fully occupied, each blade was configured with max processors, but only part capacity of RAM and max hard disks. When we set the power options to "Fully redundant", the blades simply did not power up! Only if we allowed "Redundancy with potential impact on performance" or "No redundancy" options, in the management console, did the blades power on. This means that, in the event of one power supply failing in a zone (This system is divided into 2 zones and each zone is supplied with 2 power supplies) the other supply

For those, who are planning a setup with heavy reliance on x86, Blades & Virtualization, this is a fact which should be given careful consideration!


Disclaimer: This blog is purely my personal opinion and based on my own judgement. You may use the contents freely, but at your own risk. No warranties or guarantees, explicit or implicit. Period!

No comments: