According to International Data Corporation statistics: more than 53% of small businesses have already moved to the clouds. Companies migrate to cloud hosting to improve the efficiency of using their IT infrastructure and reduce the costs required for its maintenance. But do they always succeed? Unfortunately no.
The most common mistakes during migration
– Absence of a migration plan: you did not plan which systems, in what order and how you will be transferring your infrastructure to the cloud.
Without an accurate plan of migration this action can only make a hole in your pocket. Sometimes really deep and wide one.)
– Absence of an application dependencies scheme: you are transferring applications to the cloud that are directly connected to some external services or even other servers / clouds.
When you perform the migration you expect that your website / application will work the same or even better way on a new server. But how can you make it work if you don’t have a scheme of dependencies. The thing is that besides just regular move of the data & services you have to do some re-configuration of the services involved in the work of your app / website. Without that all you’ll receive after the transfer – huge and volumed crackle which your project will go down with.
– Absence of a recovery plan: you do not have a Disaster Recovery plan with clearly defined actions in case of a failure.
With any force majeure, you risk to lose important / business information. That’s why we do backups as the first thing and consider this action as one of most important for any project.
– Ignorance of security policies: you are transferring your old infrastructure to a brand new environment, but do not change the security configurations. Interesting thing here is that hacking never sleeps.
Such carelessness increases a possibility of unauthorized interference to your system and can cause nothing but pain to your and your system.
– Migration without testing: you take applications, developed on specific environment with specific configs, software versions, dependencies, packages, etc. and try do deploy them in the cloud immediately.
The chance of successful launch in such case is very low. Most likely they will not work correctly or will not run at all. The important thing is you need to check and test if everything works correctly on a new server before switching DNS. Error at this stage may cause unexpected and really harmful downtime.
– Wrong choice of the provider: you think that all companies are the same and entrust the work to the first googled executor. How then can you be sure of the proper quality of the services? And how you can know for sure what kind of server / cloud do you need without testing?
Such mistakes make the process of migration longer, draw out unnecessary resources and sometimes make the migration itself inexpedient. I am describing a good example of how to do it a few lines below.
What steps should you perform to complete the migration successfully?
Most of the migrations are going with the same algorithm:
– Do a fresh backup of the domains and databases: just because there shouldn’t be a problem doesn’t mean that something unexpected isn’t going to happen. The most important thing in each IT project is backup. Do not take any actions with your server until you 100% sure that you have a backup and a rollback plan of the changes that messed up your server.
– Migrate files to the new server: that’s pretty straightforward task even for L1 / L2 sysadmin so there should not be any problems that may cause downtime or something bad, because it is just file transferring.
– Restore backups: that’s simple too 😉
– Check and test if everything is in its proper place: and here we come to a critical and the most important step of the migration, because if you don’t check everything accurately and go to next step – it will cause painful and unexpected downtime.
– Switch DNS to new server: if all of the above steps were performed correctly – then your migration passed successfully.
ITsyndicate Limited have performed dozen of various migrations, including complex infrastructure moves. For some particular cases we use special and custom approaches. To complete the migration successfully – we need to analyze your backend first 🙂
To be 100% sure that everything is set up, configured and working properly on a new infrastructure after the transfer – a good idea would be to give your website or application to a person or a group of people for a few days to test it. This will be live and honest testing that will not only show errors made during transfer but may also display bugs that were not caught before.
Created by Leonid @ ITsyndicate Limited on 15.02.2018. All rights reserved.