Landscape Planning, Sizing & Deployment Models

A team assumes SAP BTP gives them raw infrastructure to install and administer their own stack. Why is that wrong, and what does BTP actually take off their plate?

You need to recover a single tenant in an SAP HANA MDC system. Which database do you initiate the recovery from, and why does it matter operationally?

PaaS removes system admin, performance, load-balancing and HA work from developers. What is the trade-off you accept in exchange?

Vertical vs horizontal standardization in an SAP landscape: which one targets business functions and which targets technologies?

Moving SAP HANA to IaaS — a stakeholder expects the cloud provider to handle OS patching and HANA administration. Correct them: who actually owns those, and what does that imply for headcount?

When would you still choose SCOS (single component on one system) for SAP HANA despite it costing the most, and what do you get for that cost?

An issue hits your HANA storage/network. Who is your support contact — and how does that differ between an appliance and TDI?

Which ABAP report estimates memory/disk requirements when migrating SAP ERP or SAP S/4HANA to SAP HANA, and what is its sizing ceiling?

After running Quick Sizer you have CPU, memory, disk, I/O and a SAPS number. Which single output does the hardware vendor actually need, and why?

Who must fill in the Quick Sizer, and why does handing it to the BASIS team produce a bad sizing?

In the HANA sizing report, the 'Size data that is currently unloaded' option appears. When must you NOT uncheck it, and why?

Why does HANA hardware certification enforce a core-to-memory ratio, and why isn't there a single ratio for all systems?

vMotion/DRS/HA/FT are supported with SAP HANA from vSphere 5.5 — but version alone doesn't make a deployment compliant. What additional rules must hold?

In HANA easy/lazy sizing, what are the data- and log-volume formulas, and which one has a hard ceiling you can hit?

On a large system the HANA sizing report keeps dying mid-run. What two things must you set, and which is the one people forget?

Which operating systems are supported to run the SAP HANA database server itself (not the client)?

From which HANA release did MDC become the default, and what happens to an existing single-tenant system when you upgrade across that line?

Which IBM POWER generations are supported for production SAP HANA, and which is restricted to nonproduction only?

Someone proposes VMware Fault Tolerance to protect a HANA VM for 'zero-downtime' failover. Give two independent reasons it's a non-starter.

You're sizing SAP BW on HANA / BW/4HANA. Why not just use the generic HANA sizing report, and which report do you use instead?

A site relies on plain HANA system replication (no cluster software) and assumes failover is automatic. What actually happens when the primary fails?

In a distributed HANA system, which name server actually persists the topology to disk — and what do the others hold?

What exactly triggers a split-brain in an HA cluster, and why is the result data corruption rather than just downtime?

What does the enqueue replication server (ERS) protect, and what specifically would be lost on an ASCS takeover without it?

Why must you manually adjust the HANA global allocation limit under MCOS, and what's the failure if you don't?

Under HANA TDI, how much disk do you size for the installation on a single-node system, and where does the formula stop scaling?

The two on-premise HANA deployment options are Appliance and TDI. Which one ships preconfigured on partner hardware, and which runs on your existing certified kit?

What single limitation makes MCOD unsuitable for production HANA, even though apps have separate schemas?

In a distributed HANA system, distinguish statement routing, connection selection, and command splitting — what does each do differently?

What does MCOS mean for HANA, and when should you specifically avoid it?

Why does choosing the HANA appliance model rule out virtualization and storage replication, and what DR option are you left with?

MDC vs MCOD for HANA: which gives you independent backup/recovery per application, and why is that the deciding factor for production?

The SAP message server fails. What's the user impact — and what's the counterintuitive part about who is and isn't affected?

In HANA easy/lazy sizing, what are the shared- and backup-volume formulas, and what does a distributed (multi-node) setup change about the shared volume?

What's the operational difference between converged and hyperconverged infrastructure that affects how you scale?

Sizing the data volume for a HANA scale-out cluster — how are standby nodes treated, and what temporary spike must you plan for?

What's the key advantage of HANA TDI over the appliance, and what capability does it unlock that the appliance can't offer?

Which four network zones does SAP recommend for a HANA landscape, and what is gained by isolating the fourth?

In an SAP S/4HANA system, which components are true SPOFs needing HA — and which one do people wrongly add to that list?

VMware HA is protecting your HANA VM. What is its blind spot, and what failure would go completely unnoticed?

After a HANA takeover you must point clients at the secondary. IP redirection vs DNS binding — which works across networks, and what's the catch with each?

Users report a freshly restarted HANA is 'available' but some queries are slow. What's happening during startup, and why does that argue against restarting large instances casually?

Under HANA TDI, a customer assumes SAP or the vendor will confirm their storage/network is HANA-certified. Who is actually responsible, and with what tool?

Under HANA TDI, how do you size the log volume, and what's the breakpoint where the formula switches?

On a HANA appliance the hardware vendor installs everything. So who is your support contact when something breaks — and why is that not the vendor?

In a HANA fiber-channel storage zone, which component remaps the LUNs during failover, and what extra storage piece does a distributed system require?

A single HANA appliance is approaching its scale-up memory limit. What's the documented ceiling, and what are your only two moves when you hit it?

In a distributed HANA system, why must you list ALL potential master hosts in the client connection string, not just the current one?

Why must an HA-clustered application/database use a logical (virtual) hostname instead of a node's physical hostname?

What are the minimum LPAR sizes for production vs nonproduction HANA on IBM POWER, and does the core-to-memory ratio apply to both?

SAP supports co-deploying SAP HANA with the NetWeaver AS on the same hardware (since HANA 1.0 SP 07). Why is it still advised against?

How is HANA installation disk space sized for a scale-out cluster, and why isn't it 1 × RAM per node?

On a HANA appliance, what are the data- and log-volume sizing formulas relative to RAM — and which multiplier goes with which?

Which LPAR types are supported for HANA on IBM POWER, and which common LPAR mode is NOT allowed?

Under HANA TDI, who is allowed to perform the installation, and what's the one exemption?

In HANA 'easy and lazy' memory sizing, RAM = ((source footprint + dynamic objects) ÷ 7) × 2. What do the 7 and the 2 actually represent?

HANA CPU sizing uses a default user mix. What is it, what average resource factor does it yield, and what's the risk in the default?

You are sizing SAP HANA memory for data copied from a non-Unicode source system. What padding does SAP recommend, and what specifically drives it?

How does a cluster file system let a standby HANA node reach the data/log volumes after a failover, without a shared SAN?