More often than not when I visit customers disaster recovery is still the elephant in the room.
“Why is there never a big red button” Come on I’ve got to fit in a Doctor Who reference somewhere! But perhaps the dear War Doctor is right, perhaps there should be.
I’ll come to the title of this blog post shortly but lets talk about common disaster recovery situations I see. IT departments often have a great percentage of their infrastructure now virtualized which is great as it technically provides resilience and agility right out of the box. There are still some services that are not virtualized in many shops and this can be for many reasons, too difficult, fear of breaking it, or just have not got round to it yet. Its time to get round to it and if there is a licensing issue or compatibility problem then get onto the vendor or if they don’t want to help then migrate to one who will. There is no need to be tied to vendors of the past there are plenty out there keenly looking for business and will help you migrate from cranky old NT4 based application to a modern and actively developed one.
Many companies also have Citrix products either commonly for internal thin client use or remote access and if your really lucky sometimes both. A modestly backed up or replicated/co-located Citrix or VMware Horizon View can be provisioned within a DR event with very little effort. NetScaler technologies such as GSLB can help with that tremendously.
Profile management is also another common hot potato when it comes to SBC or VDI installations, roaming profiles, Citrix profile management, persona management or other products can be used such as AppSense.
Profile management is generally deployed with the premise of serving profiles to a local unmoving service as Microsoft that still do not support technologies such as DFS-R for site to site roaming staff members. Disaster recovery for profiles often relies on backup tapes or offloaded backup storage and occasionally DFS-R in an active passive configuration. This needs to be considered as often its not just cat pictures or holidays to spain that are stored within the profiles even after all these years company data is stored in the most unusual places within profiles and is critical for staff being able to work in the event of a disaster. This can be eased by folder redirection among other technologies which will simplify designing DR for user profiles.
Telephony is also not considered that much either when it comes to DR, Cloud VOIP and co-located services are great ways to maximize the ability to recover quickly in the event of some pressing the big red “DO DR” button.
So with the above, infrastructure, VDI/SBC etc in mind how much has your disaster recovery plans been tested, are you sure everything is going to come back up as it should elsewhere if you have that facility at all.
Technologies such as Veeam and Zerto allow you to create and test very comprehensive disaster recovery runbooks. However often even if these technologies have been configured they have not been tested fully. The test plans should be actually carried out and ideally your business should be able to run within the DR environment without even realising it. Regularity too is very important when testing disaster recovery as even minor changes within the infrastructure can cause problems when executing a disaster recovery event.
You can utilize the cloud to host all the critical infrastructure you need to keep your business running evening if the worst should happen, indeed XenDesktop even has a native cloud provider connector which is a great way of provisioning SBC/VDI sessions for just seasonal staffing requirements or DR planning.
DR has traditionally involved writing documents , lots and lots of documents on procedures and settings and boot orders and configuration changes required to bring the environment up within a new location. That’s obviously not including the amount of time its taken to restore the environment from whichever medium you have chosen to store it on (and they also have the good fortune to work when required) . Well that was tradition and if a disaster recovery plan extends more than a double sided A4 sheet of paper it is NOT the DR you should be deploying.
With technologies such as Veeam and Zerto, the possibilities of cloud and the relative high speed of site to site links there is no reason that DR should be complicated. Cost too should be little reason to not do it, cloud infrastructure is cheap enough, as well as hardware such as storage is as cheap as chips, servers for hypervisors are now commoditized and can be crammed to the rafters with CPU and RAM like never before.
Staff also are now a highly mobile workforce nearly everyone has a tablet and smartphone and are not the helpdesk calling luddites they used to be. Plan and utilize the mobility your staff offer!
Today’s infrastructure is all about automation and so it should be from simply deploying a single Windows OS instance with MDT to rebuilding and failing over an entire environment from a backup or replicated instance there should be little difference in the amount of mouse clicks required to perform either operation.
Nobody wants this anymore when the single red button approach is not only obtainable but easier and cheaper than its ever been before.
So back to the title of this post why should you do DR like a coffee shop , I’m going to show a bit of brand loyalty here. If I’m off out to lunch and wondering down the high street I’ll go to a Costa as my first choice. There are plenty of local coffee and sandwich shops to choose from but my choice is a repeatable template, I know what I want, I know what to do when I get in the shop and I know the layout even before I’ve walked into Costa.
This repeatability has lead to the success of all the similar coffee chains everyone has their preference and this is mine.
DR should be like a coffee shop you should know exactly what to do in the event, it should be simple to do it and the results should be dependable.
Automation has taken hold in a big way and there is no excuse not to have a fantastic and extremely simple disaster recovery plan, with which any business can continue to function with the least amount of downtime and most importantly the least amount of stress to the IT department.
Next time you plan for DR think about the coffee your drinking!
Author: Dale Scriven