Highlighting pitfalls to be aware of and the steps to take to ensure all goes smoothly.
Hello and welcome back to the monthly Digistor Blog – today we are reviewing some key considerations we work through with our clients to facilitate a stress-free and successful installation for all parties involved.
It’s our job to make sure the process is as smooth as possible, leaving our clients feeling satisfied with our service and eager to use the product they’ve purchased.
With any Installation, Preparation is Key – In order to avoid potential roadblocks and issues, we endeavor to be as prepared as we possibly can – hopefully this blog provides an insight into some of the thinking, planning and preparations that we put into these projects – each one will have its own unique challenges.
Here’s a few things we look out for and try have lined up before the install date:
- - Having the sales orders on hand for confirmation of components while on site.
- - Having the delivery receipts so we know when items were received and who signed for them.
- - Having the networking and cabling diagrams so we know exactly how everything fits together.
- - Ensuring we will have easy access to the server room, rack and client machines – many larger sites have very restricted access to their data rooms and labs and access can take several days to arrange.
- - Parking – is parking required for extended hours? – some areas only have 1 hour parking and we don’t want to be spending 15 minutes every hour to move our car – we ask the client if they have a spot in their parking area, if not then we find alternative transport.
- - Is the equipment still where it was delivered? – sometimes gear will be moved to a store room – before we go on-site, we make sure the client knows its location and ensure we don’t have to move it around a large building or campus.
- - If possible, we do an on-site visit to confirm these aspects.
- - Always check the rack size and ensure the hardware will fit in there with enough space front and back to install safely and easily – I’ve had rooms where we’ve needed to push a half populated rack across the floor so we could install the servers on the rails… and then had to push an even heavier rack back to location.
- - Some old racks are not very deep - you can’t safely balance a 1m deep server in a 300cm deep rack, let alone put the rails in.
- - Is there sufficient power (most preferably on UPS) and cooling in the allocated room for the extra hardware? – We always provide the client with the BTU and Power Consumption requirements well in advance.
- - Does the rack have enough spare RU to accommodate the new kit?
- - If the kit is heavy, always try to place it as low as possible in the rack.
- - We always bring our own tools to an installation.
- - We always pre-confirm all IP’s and hostnames – these can take a long time to provision in larger institutions.
- - If any specific ports are required for communications between systems, we let the client know and ensure they are open – some more secure setups will have all sorts of ports blocked and this could affect some of the intended functionality of the hardware.
- - Check if Jumbo Frames is setup – for media workflows and 10G (Ethernet) links we ideally like to set jumbo frames to facilitate the traffic, but if this can’t be guaranteed, we will likely run with 1500.
- - Are there any VLANs to consider?
- - We try our best to get an internet connection for the technician, though with phone technology these days we can tether to that in most situations – this is to allow access to information as required to help the install proceed efficiently.
- - We endeavor to ensure the new hardware is externally accessible by either VPN or other tools – it’s easier to fix unexpected problems quicker if the vendor can have access to their hardware that in being installed.
- - We check there are enough physical ports of the right type and speed for the hardware being added. Bonding will obviously require more ports, but this can often be overlooked by the customer.
- - We try to get a good overview of the client machines that may be connecting to the system – some protocols play better than others in mixed client environments and must be considered.
- - Are the client machines a mixed bag of operating systems and versions? – each release has its own intricacies, older SOE’s will have extra content added over time and often give erratic results if we are trying to get the same speed out of all of them. A new/vanilla SOE gives us the best chance to see consistent results across all machines.
- - Being aware of the applications the client expects to have leveraging the new system(s) is always an advantage – this includes AntiVirus software which can heavily impact the experience of quickly changing data.
OK, now we should have enough information handy to be as ready as we can for install day… let’s proceed to the actual installation.
- - First thing is racking in the hardware, usually starting from the bottom of the rack and working upwards – maybe putting the switch and patching at the top of the rack to allow for future expansion.
- - Carefully label and tidy all your cabling – it makes a good impression when things look neat and the next engineer can easily follow where everything goes. If the client has already racked the kit in for us, we triple check all the connections.
- - If networking has not been applied in the workshop, it is configured and tested now.
- - Now that networking is up, with any luck we can get out of that noisy, cold server room and continue configuration from the comfort of a desk with a swivel chair.
- - Set NTP, email alerts, UPS shutdown, etc as required.
- - Perform some initial testing of speed and functionality and note it down.
- - Go to the client machines and begin testing connectivity of each, doing a speed test on each to confirm consistency – if anything is out of whack, consider the variables at play (old machine vs new machine, ram, network speed lower than anticipated / bad patch cable, etc)
Hopefully everything has gone smoothly to this point and we can hand over to the client.
Handover is paramount to enabling the client to get straight on their new kit in the most efficient manner. As part of our handover, we perform an acceptance test (ATP) which basically takes the client through the specs and functions of the system and provides a signed agreement that what was promised has been delivered. This includes the initial speeds between clients and servers so we have a known benchmark.
Besides general agreement of services delivered, it’s always good to provide some basic training on the system – first showing them while explaining the logic, and then getting the client to do it themselves, for example:
- - Showing the client how to use the admin tools (often via a friendly web UI) to get an overview of the system and quickly spot if anything is wrong – if we have email alerts setup, we’re probably already aware, but at least they can have an initial look around before calling.
- - Showing the client how to connect a new machine to the system – every system is different, SANs in particular usually require a driver installation and/or some specific networking or license – it’s easy when you know how, not so much when you don’t.
- - If they have any workflows, we like to discuss and walk through some scenarios with them – best to get the client to choose a scenario that is relevant to them, they can start using that immediately and build good habits from the beginning.
A further benefit of the handover is it gives the client an opportunity to ask more relevant questions now that they see their new system in action – answer them and take a note of these questions so they can be included in the build documentation.
The more we enable the client, the greater their ability to understand and support their own system – everyone has their own way of learning, and there will always be more questions in the first week or two. It is important they don’t feel totally helpless as soon as we leave the building – we want the client to be rightfully excited with their new purchase. At the same time, we’ve provided a lot of new information during the handover, and much of it won’t have sunk in – we are ready to support them with priority.
Once back in the office and with everything is still fresh in our heads, we get onto the build documentation – the client is still learning their system, so it’s good to provide these as soon as possible.
I have found that all people differ in their preferred way to digest new information – some like just the basics to get started (usually creative types), while others prefer a complete rundown with wiring diagrams and the like (usually IT/engineer types). If possible, we build the comprehensive documentation, then cherry-pick a quick-start guide most relevant to the requirements of the client, giving them both documents.
While a lot of this may seem obvious, it’s easy to overlook something that seems innocuous at the time. To that end, I hope this article is of assistance to others.
If you are interested in this or other Digistor services please contact the Digistor sales team for more details.
See you again in a few months!