Your migration is about to hit its most exhilarating point. The Migration step proves out all the hard work and planning you’ve done prior. The many hours of hard and sometimes monotonous project planning work, assessments, testing, vendor selection, purchasing, contracting, etc., and other preparations are done, and part four is the final step that stands between you, your team, and a truly good night’s sleep.
This entry will be on the shorter side, as well, since in parts one (“Readiness Assessment and Creation of the Pre-Migration Plan”) through three (“Migration Prep & Solution Design”), we did the majority of the hard work, having discussed the planning details, designed our solutions, ensured security boundaries existed and would be respected, established, refreshed, and validated SLAs, and all the other nuances that got you here. This fourth article will consist mostly of last-minute reminders and tips, and a strict insistence that you stick to the plan you worked so hard to create:
- Stick to the plan! [aka “If it’s not on the plan, don’t do it!]
- Ensure SLA’s are met and be prepared to back out if necessary. Be prepared to make the back out call, and ensure each designee is clear on the criteria that would trigger that call.
- Verify designated communication channels are all functioning and all groups are ready.
- Ready, set … MIGRATE!
- Monitor everything closely for anomalies, and independently validate each migration group
Stick to the plan! [aka ‘If it’s not on the plan, don’t do it!]
It sounds obvious, right? You’ve spent months coming up with it and worked out every detail and nuance of who will do what, where, and when. The temptation, however, to make last minute changes and adjustments can sometimes be so overwhelming that it can lead to rash, and rather poor decisions.
If it isn’t in the plan, it shouldn’t happen during the migration, with few exceptions. There was plenty of time to plan and scope everything that will be taking place. The migration should proceed like a well-oiled machine, and your plan even accounts for certain situations where it might not. There will be plenty of time after the migration to make changes!
Ensure SLA’s Are Met, and be Prepared (yes, really) to Back Out!
Nobody ever wants to use a back out plan. Backing out is a worst-case scenario, and if the back out plan comes into play, something, somewhere went terribly wrong.
Fortunately, back out plans are like insurance policies. You almost never need one unless you don’t have one. The bad news is that when you need one, you really, really need one. It’s much better to be prepared.
Ensure that the designated signee is aware of the go/no-go criteria for each step and the criteria thresholds that trigger the back out plan as well. There should be no question in anyone’s mind during the migration regarding to how to proceed. Your plan spells out very clearly what to do when, what SLAs must be met, and what kinds of failures would be too severe to be tolerated by the business.
Verify Designated Communication Channels and all Groups are Ready
Before the migration kicks off, take the time to verify that the designated communication channels are all open and functional. If you’re using screen sharing or collaboration software, ensure the meetings and conference lines have been created, and that those phone numbers and URLs have been distributed to all who will need to join. Ensure that the migration documentation is accessible to all that will need it, and double check that provisions for accessing those resources during the migration are in place.
Ready, Set … Migrate!
It’s okay to be nervous the morning of migration day. The key here is sticking to the plan. You’ve already done much of the hard work, and your plan details exactly what needs to be done. Just take it one step at a time.
On the off chance something should go wrong, don’t panic. Be honest about it, and don’t try to cover up the mistake. Consult your team and figure out if it triggers the back out, or if it can be worked through.
Monitor for Anomalies and Independently Validate Each Step
The monitoring system will be alerting to the changes it picks up during the migration, and your plan accounts for these. You’ve likely silenced the non-critical alerts, as well as those that would not clear on their own.
Look for any alerts that are unexpected or anomalous. Don’t forget that it’s very possible to have the same kinds of production issues that might occur during any other normal business day. If you haven’t completely shut down for the migration, as is likely with many tech businesses operating 24/7, keep a watchful eye out for unrelated alerts.
In the case of an unexpected alert, delegate to someone who is available and waiting for their turn in the migration. Ensure that party has accepted responsibility, and then return focus to the task at hand. If you are the only party that can handle the given situation, pause and fill in your team. If possible, handle the issue using standard procedure, or consult the team and utilize a procedure deemed appropriate until after the migration. If a temporary fix cannot be avoided, document everything.
Lastly, validate each phase of the work. Each participant must independently validate their steps to ensure the process delivered the expected end results. It’s not a bad idea to double check that monitoring systems have been re-activated and are monitoring the new environment, and that your CMDB’s autodiscovery documentation jobs have been edited to include any newly created and/or removed network segments, if applicable.
You are nearing the very end of your hard-fought migration, and you’ve made it through the hardest night. Congratulations on a migration well done! Go ahead and head home for some well-deserved rest. You’ve earned it.
Tune back in for the fifth and final entry in Device42’s Ultimate Guide to Migration. All that’s left, now is the Post-Migration Party, and of course, “Decommissioning the Old Data Center”. Until then, happy migrating!
Sources / References:
Editor’s Note: This post was originally published in June 2017 and has been updated for accuracy.