I could really use some sage words of wisdom from the SQL Server Central community on Availability Groups in SQL Server 2017. I’ve wanted to implement AG for a while now but a coming increase of 200x our current work and going to 24x7 operation with no maintenance window has pretty much forced them into accepting that we need to setup an AG now. I’m reading up as much as I can on this, but I have some questions about some of what I’m finding.
It says that SQL Server 2017 no longer needs a Windows Cluster to setup an AG. Does anyone know of a downside to not using a Windows Cluster?
There is a lot of data syncing going on between an InterSystems Cache database and SQL Server (both in the same datacenter) so a lot of bits are moving around and lots to sync to a secondary. IT is really hot to implement the secondary in Azure instead of another host server in the datacenter. Latency is <1ms VM to VM in the datacenter but from datacenter VM to Azure it’s around 20ms. Both numbers from using ping. Thoughts on a busy server hosting 2TB of databases moving lots of data having a secondary in Azure vs the datacenter? I don’t get a warm and fuzzy feeling with the higher latency but maybe I’m being needlessly paranoid.
Some stats on the current server:
Windows Server 2016, 8 core, 96GB RAM – 90GB to SQL Server (no other apps on the box)
SQL Server 2017 CU20 – 28 databases. Largest 5 databases total 1.8TB, the remaining DBs are smaller taking up the remaining 200GB of space.
The SQL Server disks are from the faster of the 2 SAN segments and there is no second SAN. I know this leads to a single point of failure at the moment, but I don’t think they’re planning on updates to the SAN anytime soon. This might be where a secondary in Azure might make more sense if the latency isn’t too big an issue.
I appreciate any advice anyone has to share. As always, thanks for your input.