Moving a rack server quick reference guide

I put together this quick reference guide about moving a rack server for my brother who’s working on a data center exit project (migrating servers from one data center to another) as a point of reference, do change it to suit your needs, I have made some assumptions which might not apply to all servers. Where I’ve said ILO, it’s the same concept for Dell’s iDRAC or IBM’s IMM.

At this point I can sense many colleagues, industry experts and vendors saying but the ROI, the cost benefit analysis of physically moving servers just isn’t there, you could p2v them, or do a new for old swap and do it that way. Absolutely but let us appreciate that not everyone is in the same space in their IT lifecycle, that not everyone would necessarily have forward planned such activities nor have relevant buy in, that in some cases, moving it and then addressing at a later stage might work. That olden days issue that I have budget set aside to move, but none to buy new hardware to build out a new virtual farm or replace like for like hardware.

Information required:

Server name

Serial number

Network teaming – yes/no

Network connection(s) and IP – MAC(s), IP/Subnet/Gateway per connection

ILO(iDRAC/IMM) – yes/no

ILO/iDRAC/IMM – ip configuration

SAN storage – yes/no

WorldWideNames – referred to as WWN (the equivalent of the MAC for the SAN fibre cards)

Cabinet location

Size in U

Checks required on power down (Windows but concepts for other operating systems apply)

Ping ILO and check that you can log in.

Save and clear IML

Log on to event logs and check for any hardware or software errors

Run command prompt (in Administrator mode if UAC turned on) ipconfig /full >ipbfmove.txt

Check disk space to avoid any issues

Shut server down

On power up:

Switch server on

Log into ILO config screen and configure new IP details (if applicable), check the port speed and duplex match expected, apply changes and save

Boot into the operating system and apply new IP details (if applicable)

Ping default gateway (for some isolated/customer facing networks this might not work)

Ping a server on the same network to check connectivity

Check that you can see all the drives, that the event logs are clear of errors and all services set to automatic are started (as normal)

Hand server over to project manager for application sign off

Components that I would have to hand during a physical lift and shift data center move project:

Hard disk drives

Power supplies

Memory

Either a system board or servers of the same vintage/range that we can use either as spare swap servers or to obtain parts

This site uses Akismet to reduce spam. Learn how your comment data is processed.