6 Solvable Data Centre Management Mistakes

Building and maintaining a successful data centre environment to house your organisation’s vital infrastructure is no easy task, but by avoiding these common data centre management mistakes, you can help to ensure things continue to run smoothly and deliver the desired performance for your business.

Mistake 1: Cardboard

Cardboard and data centres don’t mix. Full stop. Not only does cardboard present a significant fire hazard, it also sheds particles which can cause significant damage to fans. Blocking vital airflows that work to keep equipment cool. At best, these blockages lead to greater energy usage, at worst, they can cause complete equipment failure.
“But wait!” I hear you cry, “all my new equipment arrives fully clad in this, potentially fatal, material!” That’s true. Whilst cardboard poses a hazard within the data centre, it still remains the best packaging solution for most suppliers shipping equipment to the data centre.

  • Solution
    All packaging needs to be removed before equipment is transferred into the data centre. This can be a time consuming process, so Hardware.com has introduced their pre-delivery logistics services to help. We can remove all combustible packaging from your equipment. Repackaging it into re-usable plastic crates to allow for risk-free delivery directly into the data hall, so you don’t have to worry about any potential damage.

Mistake 2: Documentation

A surprisingly common data centre mistake is not keeping thorough and up-to-date documentation of the network. One of the pain points that is often faced over time is a change in staff and therefore a lack of knowledge as to how and why the network is set up as it is. This not only makes it difficult to resolve any problems that occur, it also hinders the planning of future developments. Thorough documentation can help to plug this knowledge gap, providing any newcomers, or temporary engineers, with all the information they need to fulfil their jobs.

  • Solution
    Clear documentation should be created to record the systems that are in place, the equipment that’s been used, how it’s been wired together and how it’s been configured. This should include a visual diagram of the network, cable mapping, an equipment inventory, software versions, IP addresses, server configurations, backup/disaster recovery procedures and administrative accounts and passwords. Due to the sensitivity of some of this information, the documentation should be available in digital and hard-copy format and stored in a secure place. It is also worth keeping a separate copy in an offsite location for disaster recovery purposes.If documentation does not currently exist or is largely out of date, this task can be a daunting prospect. Hardware.com offers full audit services to help you gain a thorough understanding of your existing infrastructure which can include full documentation. We are also able to offer documentation services alongside new installations.

Mistake 3: Data Centre Security

Colocation facilities generally provide high levels of security including hand-print recognition, CCTV, and on-site personnel but, even with these measures in place, security is not a guarantee, with people often representing the greatest point of vulnerability.

  • Solution
    Businesses still need to play an active role in security management. Who has been given access to the cages and why? If individuals have been granted temporary access for a particular purpose, has this been revoked once the task in question has been completed? Equally, has the access of any company leavers been revoked? Keeping on top of who has access and who needs access will help to ensure all bases are covered with regards to security.There are technologies available to assist with this. Usually in a colocation facility, once access has been granted to a cage, all individual racks are accessible with just one key. Smart card readers provide a strong solution, demanding a specifically programmed smart card to gain access to each individual rack. These smart cards can be pre-programmed with rack numbers and a specific time frame during which access will be permitted. Any engineer requiring access can then be allocated a smart card for the specific work they need to carry out. Reports can also be generated detailing which racks have been accessed and at what time.

Mistake 4: Inadequate UPS power

In the event of power outage, systems are typically backed up by UPS systems which keep equipment running until a generator kicks in, or systems can be safely shut down. The misconception here is that a UPS will support your equipment until mains power returns. This is simply not true.

  • Solution
    It is important to calculate the run-time in the event of failure and ensure your chosen UPS has the battery power to support this. How long does equipment need to be powered for until a generator kicks in or an engineer can be contacted? What is the draw on power in this time? How many UPS’s are required to deliver this level of power? Backup power should always be tested and UPSs continually maintained – the average battery life of a UPS is around 18 months.An environmental audit can assist by assessing power usage along with UPS and battery health to ensure the necessary power backup is in place.

Mistake 5: Backup Management Connectivity

More and more organisations are housing their equipment in remote data centres, but what happens if their internet connection fails? How will their equipment, and the power supporting it, be managed in this scenario? Without some form of backup connectivity and remote access, companies could face significant periods of costly downtime.

  • Solution
    A 3G or 4G connection option can provide uninterrupted availability to continue managing the equipment. Managed PDUs also mean power can be turned off and on remotely to reset systems in the event of failure. Although an investment, the cost is small compared with the potential cost of downtime.

Mistake 6: Cabling

Networks will always grow and, as they grow, the associated cabling can become a bit haphazard. Making it difficult to identify the source of any errors. And turning, what should be a straightforward task, into a time-consuming nightmare. An unnecessary level of complexity is also added to the deployment of any further additional equipment into the network. Any company housing equipment in a shared data centre should additionally be aware of other external eyes on their racks. Messy cabling may, unwittingly, be presenting a negative image of the company and their ability to deliver high standard services.

Solution

  • Cables should be bundled into relevant groups to keep them tidy and then labelled and documented. Presenting something of a roadmap to any engineer faced with making changes. It’s also prudent to allow plenty of additional cable for any future additions. Hardware.com provide these services as routine alongside any deployments or data centre moves. We can also provide a full cable management service in which all patch cables within a server room are audited and documented before being recabled and labelled neatly during an agreed period of downtime.