SAP HANA VMs and storage layout
HANA runs in memory, but it still depends on fast, consistent storage for logs and savepoints. The exam checks that you match each HANA volume to the right disk and cache setting.
Certified VMs
SAP supports HANA production only on SAP HANA certified VMs. For large databases use M-series (Mv2, Mv3), which has several terabytes of RAM and supports Write Accelerator. Smaller systems can use certified E-series. HANA Large Instances (bare metal) are being retired, and M-series covers their scale.
Volumes and disks
| Volume | Content | Premium SSD guidance |
|---|---|---|
| /hana/data | Savepoints | Stripe several disks, no caching or ReadOnly |
| /hana/log | Redo log | Stripe disks with Write Accelerator (M-series), no caching |
| /hana/shared | Binaries | One disk, or NFS |
| /usr/sap | SAP executables | One small disk |
| /hana/backup | Local backup | Standard SSD or HDD |
Do not place data on the temporary disk. It is lost on redeploy.
Disk types
- Premium SSD v2 and Ultra Disk: you set IOPS and throughput independently of size. Ultra suits log and data for the most demanding systems, and Write Accelerator is not needed with it.
- Premium SSD: needs striping to reach the target IOPS and throughput.
- Standard SSD/HDD: only for backup or non-production.
Striping uses LVM (or mdadm) on Linux. Follow the Azure HANA storage guidance for stripe sizes.
Azure NetApp Files
Azure NetApp Files (ANF) provides NFS v4.1 volumes managed by Azure. It is certified for /hana/data, /hana/log and /hana/shared and avoids striping. Choose the Ultra or Premium service level; the capacity pool and volume quota decide throughput. For lowest latency, use an application volume group for SAP HANA, which places the volumes close to the VM.
Sizing steps
- Pick the VM from the SAP certified list for the memory you need.
- Check the VM limits for disk IOPS and network throughput.
- Size the disks so the stripe meets HANA’s KPIs, which you verify with HCMT (HANA Hardware and Cloud Measurement Tools).
Exam traps
- Write Accelerator is for M-series with Premium SSD log disks. Do not combine it with Ultra.
- Never use ReadWrite host caching for HANA data or log.
- The temp disk is not for HANA files.
- Azure Files SMB is not supported for HANA data.
- HCMT validates storage; Quick Sizer sizes the system.
Practice questions
Choose an answer, then review the explanation to check your understanding.
1. A HANA VM uses Premium SSD disks for /hana/log on an M-series VM. Which cache setting do you use?
Show answer
Answer: B.Write Accelerator (M-series) gives low-latency log writes on Premium SSD. Host caching is not used on log disks.
2. You want to improve throughput for /hana/data on several Premium SSD disks. What do you do?
Show answer
Answer: C.Use LVM or mdadm striping across several disks so the volume gets the combined IOPS and throughput.
3. A HANA system needs shared NFS volumes with sub-millisecond latency for data and log, with little disk management. What do you choose?
Show answer
Answer: B.Azure NetApp Files (NFS 4.1) is supported for HANA data, log and shared and gives low latency without striping.
4. A 1 TB HANA database on an M64ms VM grows and performance drops because of storage. What helps most?
Show answer
Answer: C.When storage is the limit, more disks in the stripe raise IOPS and throughput. Network and vCPU changes do not fix disk limits.