![]() ![]() But if you playing with a serious SQL Server, you will end up doing diskspd.exe testing (built by Microsoft) and running a ton of tests on each volume. The real question is how far apart are we? You can do some basic SAN testing using a program such as CrystalDiskMark. Those two numbers never match – but that’s OK. Here I want to compare how SAN theoretical advertised maximum performance compares to what I am actually seeing. Storage is where most SQL Servers bottleneck on today. SAN Config for SQL Server – I’d make this item #1 on this list (I used some automated SQLAgent jobs for this).ģ. Keeping things in synch – requires more work. ![]() Setting this up the first time – is easy. Just make sure to think about how you will keep SQL security, SQL Agent jobs, linked servers synched between all nodes at all times (and No, AlwaysOn doesn’t help you there). If you get pre-setup correctly – this part is going to be a breeze. SQL Distributed Replay eliminates SQL migration unpredictability I have used the SQL Distributed Replay feature for this – it worked great! It’s a lot of steps and maybe your SQL Server is not that important to go through extra pain and time this will take, but it is – is really worth it! I can probably even send you some Excel files that show a comparison of run1 vs. Such as how CPU cores are presented. In those cases, I suggest grabbing production workload and re-running it on the NEW SQL server with setting A and then with setting B. Some of the settings are not as obvious though. The key here is to NOT to allow other guest OS’es to take SQL Server resources. If you just google “VMware SQL Server best practices”, you will find them. Make sure you talk to VMWare person to get some specific settings turned on for SQL Server. ![]() There are few settings in virtualization that you need to get right for SQL Servers. ![]()
0 Comments
Leave a Reply. |