SAP HANA on AWS

An architect provisions slow st1 HDD for /hana/data on a production HANA node, reasoning that HANA runs in memory so disk speed is irrelevant. Why is this wrong?

What are the four main SAP HANA deployment offerings on AWS?

You must size an AWS EC2 target for an existing on-prem SAP HANA system versus a brand-new greenfield implementation. Which sizing input do you reach for in each case, and what's the trap?

In a typical HANA EC2 storage layout, why is /backup placed on st1 (throughput-optimized HDD) while /hana/data and /hana/log sit on gp3 or io2 Block Express?

You've outgrown a single HANA host. What decides whether you go scale-up (bigger box) versus scale-out (more nodes) on AWS, and what extra dependency does scale-out drag in?

For production HANA you can choose gp3 or io2 Block Express, and you can stripe volumes with RAID 0. Which volume do you pick for production, and what caps the RAID 0 gain?

When configuring Amazon FSx for NetApp ONTAP for SAP HANA, which two FSx features must you turn OFF, and why are the default FSx backups not enough?

HANA system replication can be multi-tier or multi-target. Which one chains replication sequentially, and why does that distinction change your failure exposure on AWS?

Which AWS storage services are supported for SAP HANA scale-up vs scale-out (data/log vs /hana/shared vs backups)?

Why is EC2 auto-recovery insufficient as the sole HA mechanism for production SAP HANA, and where is it actually appropriate?

HANA System Replication on AWS runs either performance-optimized (Preload ON) or cost-optimized (Preload OFF). What single decision does this fork hinge on, and what must the secondary's size be in each case?

What OS options are certified for SAP HANA on AWS, and which flavor do you pick for an HA setup?

To avoid paying for an idle cost-optimized HSR secondary, you run a non-prod HANA DB on it via MCOS. What is the step people skip that turns this into an outage?

HANA scale-out host auto-failover to a standby node on AWS is only supported on one storage service. Which one, and what would break if you tried it on the others?

What does the AWS Backint Agent change about how you back up SAP HANA to Amazon S3 compared with the conventional approach?

You plan to use gp3 for a production HANA volume. On which instances is that NOT allowed, and what's the giveaway?

Walking the HANA HA/DR options from cheapest to most resilient on AWS, which rung first gives you a LOW RPO, and which adds true out-of-region DR?

Why do you keep synchronous HANA HSR within one region (across AZs) and never stretch it across regions on AWS?

Your infra team tunes down the EBS IOPS/throughput on a production HANA node to save money, and HANA still runs fine in testing. Why can this quietly cost you SAP support?

You need to pick an EC2 instance for a production HANA system and be sure it stays SAP-supported. Which references do you consult, and what step do people skip after picking from the list?

When would you reach for Amazon FSx for NetApp ONTAP for HANA instead of plain EBS, and what makes it different at the protocol level?

An x2iedn.32xlarge is listed at 2.7 TiB under one HANA spec and 4 TiB under another. You run S/4HANA and want the full 4 TiB — what's the catch above 2.7 TB?

You're replatforming an existing database to SAP HANA and need a first-pass memory estimate before detailed sizing. What quick rule of thumb does the book give, and what else should you factor in?

SAP HANA TDI moved from v4 to v5. What rule did v5 relax, why did OLTP customers push for it, and what do you give up in exchange?

You want automated, application-consistent EBS snapshots of a HANA node so a restore isn't caught mid-write. Which AWS tool provides that, and what's the scope limitation to note?

How large can a single SAP HANA (BYOL) system go on AWS, and why can't you quote one number — what changes between scale-up, S/4HANA scale-out, and OLAP scale-out?

A production HANA node on gp3 is hitting its IOPS ceiling during month-end. Why is gp3 attractive for handling that spike, and what's the operational advantage over provisioning up front?

Your cost-optimized HSR plan is to resize the smaller secondary up to production size only at takeover time. What assumption in that plan can fail exactly when you need it?

You size a cost-optimized HSR secondary just by turning off preload — but you're running logreplay. What extra sizing factor does logreplay force, and when might you switch to delta_datashipping instead?

You must encrypt HANA data and log backups going to S3 via AWS Backint with your own KMS key. Which parameter and config file wire that up?

You want a serverless function to run SQL against HANA, and separately to bulk-extract HANA data into an AWS pipeline. What's the connectivity trick for Lambda, and which services extract at the database layer?

What CPU architecture and EC2 instance families does SAP HANA require on AWS, and what marks an instance as 'high-memory'?

In the cheapest HANA HA/DR pattern on AWS — backup/restore via Backint to S3 — walk through what actually happens on an AZ failure, and which factors decide how long recovery takes.

For a HANA scale-out deployment on AWS you want the nodes to talk to each other with the lowest possible network latency. Which placement group do you use, and what happens to it after an auto-recovery event?

You can mount Amazon S3 to an EC2 instance, so an engineer proposes putting HANA data files on S3 to cut storage cost. Which design principle says no, and what does AWS give you instead?

Before you ever configure Pacemaker or HSR, HANA itself has two built-in fault-tolerance features on AWS. What are they, and what layer of failure does each cover?

You're choosing which AZs to place a synchronous HANA HSR primary and secondary in. What latency target does SAP set, and how do you actually measure it before committing?

You want to lift-and-shift an existing HANA system to AWS without changing the OS/DB combination. What minimum HANA and SLES versions must you be on for that to be a supported config?

For a performance-optimized HANA scale-up HA setup on AWS, which three HSR settings define the configuration, and how is the secondary sized?