This is a server configuration question for a particular use case. It touches on a variety of subjects so I hope I can state my questions clearly. I've done a lotta reading, but am more confused than ever...

 

We have a PE R750 server (16 - 64GB RDIMM, 3200MT/s, Dual Rank) configured thusly...

1.  BOSS: Not used

2.  Two (2) 1.6TB SSDs in RAID 1 using PERC H755 SAS Front (OS)

3.  Six (6) 6.4TB NMVe SSDs...SW RAID 10 (Hyper-V supporting a mix of Web Application and SQL Server VMs)

 

My goal is to increase storage capacity and performance with the current equipment. I'm thinking of rebuilding the server to...

1.  BOSS: Use for OS

2.  Six (6) 6.4TB NVMe SSDs...SW RAID 50 (supporting just the Hyper-V SQL Server VMs)

3.  Eight (8) 1.92TB SAS SSDs (available from another server upgrade)...PERC11 RAID 50 (supporting Web Application VMs )

 

My questions are...

1.  Since we have the PERC11 "H755 SAS Front", can/should I use that for both RAIDs?

2.  Is it better to continue using the SW RAID for the NVMe drives to dedicate the PERC to the 2nd RAID (1.92 SAS SSDs)?

3.  Is it best to isolate the SQL DB VMs on one RAID and run the Web Application VMs on the other? Is so, which is best for the SQL DB VMs...NVMe drives w/SW RAID or SAS SSDs w/HW RAID?

4.  Is switching from RAID 10 to RAID 50 an effective (recommended) way to increase capacity without sacrificing performance and redundancy?

5.  Can the PERC11 "H755 SAS Front" effectively support RAID 50?

6.  What is meant by "synchronized disks"? I've read where RAID 50 requires "synchronized disks".

7.  What is the optimum "strip element size" for SSDs in this scenario? (I've read where smaller is better with SSDs)

8.  Is caching even necessary when using SSD drives? (I read where caching might lower overall IOPS of SSD arrays)

9.  What are the best Read/Write policies for this configuration? ("No Read Ahead" and "Write Through" ???)

10. Are there considerations I'm overlooking?

 

I truly appreciate any and all comments...even if it's a slap.