Share what you know with millions of people
Focus is the best place to turn what you know into remarkable content
0
What are your server migration strategies?
We know we’re going to have to upgrade our servers next year, and I’m curious to know what your server migration strategies are. I want to make this process as painless as possible. How did you go about migrating your servers?
Events
- Dos and Don'ts of Small Business Marketing May 29 @ 11 am PT
- Lead Nurturing 202: The Next Generation May 31 @ 11 am PT
- The Tricks to Paid Media June 6 @ 11 am PT
- Display Advertising for Brand Awareness June 20 @ 11 am PT
Related Activity
3 Answers
Guy, before you embark on this size of a project, I recommend you pull together a Technology Roadmap. What do you want your IT department to look like in 3 to 5 years. Do you want it to look and run much the same as it is? Or do you want it to incorporate new technologies?
Specific to your question, will your IT utilize Cloud Computing in the near future? Will you consolidate sites? Consolidate Servers? Outsource some (or all) IT functionality? Consolidate storage?
Once you have a roadmap you can determine the best strategy for migrating your servers.
Regarding basic migration strategies for a one to one replacement, you can either Platespin the images from the old to the new or you can build new servers with upgraded Operating Systems/app versions and copy the data from the old.
Depending on the amout of data on the servers your data migration strategy may vary. If you have large data stores to copy from local drives it can take days to copy to new servers. If server downtime is an issue you may need to use intermediary storage such as Datamovers.
Your data migration strategy may differ if you have NAS or SAN capabilities in house.
Also, consider the speed of your network and NIC's? if they are still running at 10MB/sec your data copy speeds will be painfully slow.
Hopefully this give you food for thought. Map out the big picture of where you want to go first, then use that as your guide to upgrading your servers.
I also recommend reading the following article about the pitfalls of legacy IT thinking in a modern IT world.
http://www.itanswers4u.com/Wordpress/?p=215
If you have any question feel free to contact me.
Guy,
John gives excellent advice!
One issue that always comes into play is budget. If you got a lot of dough you can do all of these upgrades at once, if not then a step by step plan will be more appropriate. If the latter is the case the real challenge becomes what do I do first and what makes the most sense for the dollars I have to spend now, next step, next quarter, etc.
Also user interruption is another consideration, as John made some mention of. But the easiest way to sour all of your users impression of a technology upgrade is to effect their production. This often means the IT staff working after hours or on weekends to limit that interruption, which again plays out in the budget of the upgrade in terms of overtime, etc.
But John's most impoortant point is to have a plan that has been meetted out and rehearsed until all of the details are covered and you don't get hit with a big OOOPS in the middle of the project!
Hello Guy,Jonh and Don gives excellent advice, and i would like to give you share my latest experience on the matter.
I´m a virtulization it pro in microsoft technologies so this was the technology we decided to use because of the experience of my team.
A client wanted to move their legacy servers into new ones, this servers where 6 yers old, so they were having a lot of storage and performance problems.
After showing various scenarios we decided to go with DELL blade solution, because they have great warranty and support.
Then it was time for the storage, after looking some solutions we suggest a clarion ax/4 by emc, with 1 TB sata hd.
The client had software assurance for all their microsoft apps ans OS, so we were going to migrate to the new platforms, (not a good idea to make it the same time Jonh is right make your roadmap), so the strategy was this:
1.- We take a month to measure the bandwith in use on the network, the switches capacity and protocols, and finally the user profiles to implement network security joining the common groups to different vlans.
We use one vlan for every department because of their needs and the privacy requirements, one production vlan that join all the other client vlans, a backup vlan, and administration vlans, one for managment servers and another one for business servers. (the multi layer switches are really expensive so think deeply if you need something like this)
2.- We install Windows Server 2008 R2 on 11 blade servers (14 total) and took 3 servers for the applications that we decided were not candidates for virtualization.
With SCDPM we took images of those servers and then restore them to the new ones, with less than 4 hours of downtime. (a SAN is great for this type of migration, you can boot from lan with another server and then just restore the image)
3.- Then the other 11 servers where configured as a cluster for hyper-v, because they use sharepoint (if you use file servers i recommend a cluster too).
All the other 20 servers where moved with SCVMM p2v with no downtime, just one or two minutes for the switching between the two of them.
SAN.- The SAN was configures with 2 tb LUNS for the vms, we use iscsi and csv in the hyper-v cluster to allow live migration and other great features.
All the servers are managed with the System Center family products.
Now we are migrating each individual server to a test enviroment for deploying them when we know it´s working on the new OS or app, their are some solution accelerators if you have microsoft platform at http://www.microsoft.com/MOF.
The biggest deal was the active directory, be really careful if you have one, take it step by step, depending on the functionality level go one by one upgrading and testing before you upgrade the forest.
The results for this strategy where great, expensive at first but ROI will pay off in 26 months.
Hope it helps you.
Answer This Question