Taking Over an Undocumented IT Estate: Where to Start
Hardware you inherit rather than buy is the hardest kind to manage, because nobody knows what is there or why. This covers building a picture of an estate you did not create.
Start with what is powered on
Before opening anything, find out what exists.
Management controllers are the fastest route. iLO, iDRAC and IMM report model, serial, configuration, firmware and health without touching the machine, and they work over the network. If you can reach them, you have most of an inventory already. Our management guide covers access.
The problem is credentials. Inherited estates frequently have management controllers nobody can log into. Where that is the case, the controller can usually be reset from the machine itself, which needs physical access and a reboot.
Network discovery finds devices that respond but that nobody documented — switches, storage, appliances, and the machine in a cupboard that turns out to be running something important.
Record the things that are hard to find later
Six fields per machine, and these are the ones that matter when something fails.
Model and generation. Our naming guide covers identifying it.
Serial and product number. Needed for any support or entitlement question. Our serial lookup guide covers finding them.
Configuration — processors, memory total and module type, drives, controller.
Firmware levels across BIOS, controller and adapters.
What it runs, and who cares if it stops.
Where it physically is. Rack, unit position, which site. Obvious until someone needs to find it at 2am.
The questions that matter more than the inventory
Three, and they change what you do next.
What is unsupported? Platforms past manufacturer support, operating systems past updates, firmware that can no longer be downloaded. Our entitlement guide covers checking.
What has no spare and no redundancy? A single machine running something important, with no spare parts and no second copy, is the risk in the estate regardless of its age.
What would you not be able to rebuild? A machine nobody knows how to reconfigure is worse than an old machine. Documentation gaps are frequently a larger risk than hardware age.
Finding what nobody mentioned
Four places things hide.
Machines with no monitoring. If it is not monitored, nobody hears about it until it stops.
Backup jobs pointing at things. The backup system frequently knows about systems the documentation does not.
DNS and DHCP records for hosts nobody can account for.
Cabling. Follow the cables out of the rack. Things turn up. Our network documentation guide covers doing this systematically.
What to do first
Four actions, in this order.
Get management controller access on everything. Without it you are blind and every incident needs someone on site.
Set up monitoring and route the alerts somewhere a person reads. An estate with no alerting is an estate where every problem is discovered by a user. Our monitoring guide covers what to watch.
Verify backups restore. Not that they run — that they restore. Our backup guide covers the principle.
Identify the single points of failure and decide deliberately whether each is acceptable. Some will be; the point is deciding rather than discovering.
Then, and only then, plan hardware
With a picture in hand, three decisions become answerable.
What needs spares. Our spares guide covers sizing a holding by consequence.
What needs replacing, on risk rather than age. Our lifecycle guide covers the sequence.
What can be consolidated or retired. Inherited estates almost always contain machines doing nothing that anyone can name. Our consolidation guide covers this, and our selling guide covers recovering value from what comes out.
Frequently asked questions
Where do I start with an estate I inherited?
Management controllers. They report model, serial, configuration, firmware and health over the network without touching the machines. Combine with network discovery to find devices nobody documented.
What if nobody has the management controller credentials?
The controller can usually be reset from the machine itself, which needs physical access and a reboot. Getting access on everything is the first action worth taking, because without it every incident needs someone on site.
What should I record for each machine?
Model and generation, serial and product number, configuration, firmware levels, what it runs and who cares if it stops, and where it physically is.
How do I find systems nobody mentioned?
Backup jobs frequently know about systems the documentation does not. Also DNS and DHCP records for unaccounted hosts, machines with no monitoring, and following the cables out of the rack.
What is the biggest risk in an inherited estate?
Frequently documentation rather than hardware age. A machine nobody knows how to rebuild is worse than an old machine, and a single system running something important with no spare and no redundancy is the real exposure.
Should I replace old machines first?
Replace on risk rather than age. Get management access, monitoring and verified backups first, then identify single points of failure. Some old machines are fine; some new ones are the exposure.
Send us a list of what you have found and we will tell you which platforms still have parts available and which do not.
