1

I'm working on a failover plan for my (at the moment, single Azure region) webservers. My azure resources are due to legacy applications quite oldschool, Windows server 2016 VM's with SSD disks attached and some VM's with SQL SERVER (2016) installed.

In case that the failover is needed I would simply switch the DNS records to point to the loadbalancer in the failover datacenter as my website is only accessible via a domainname.

I want to create a failover configuration in another datacenter that uses Azure Stack (to avoid crossing country borders because Azure only has 1 region in my country).

My idea is to start of easy and just recreate a my server setup in the failover-datacenter with ssd disks and sql server installed. However, I need to synchronize the data on these disks, and synchronize the database data.

Because the disks are not network shares or storageaccounts, the only way I see is to sync via FTP (winSCP), is this correct? Wondering if you have any ideas on this subject?

user2713516
  • 155
  • 1
  • 1
  • 12
  • RTFM? AOAG (always on availability groups) are standard in SQL, S2D failover cluster can do pure disc replication, as can DPM for the websites. – TomTom Apr 23 '18 at 08:14
  • 1
    SQL Standard AlwaysOn AGs support is quite limited (one DB, no read replicas etc), S2D won't work between data centers (latency!) so Storage Replica is the only viable (and unmentioned) MSFT-only option. P.S. It's DISK and not DISC. – BaronSamedi1958 May 16 '18 at 17:39

1 Answers1

1

At the moment, if you're planning to use Azure Stack then you will need to replicate at the Windows level with things like Storage Replica. AlwaysOn AGs is another viable option for DB-level.

In the future, once Azure Site Recovery is available in Azure Stack then you would be able to use that to failover at the VM level from Azure to Stack.

BaronSamedi1958
  • 13,676
  • 1
  • 21
  • 53
Sam Cogan
  • 38,736
  • 6
  • 78
  • 114