Scale-Out Systems & Host Auto-Failover

What are the SAP S/4HANA scale-out node minimums and the maximum cluster size?

You have a small config table that joins with many table groups, and a large fact table that joins with it. Which gets a replicated copy on every node and which gets placed once — and what's the deciding property?

A standby took over after MASTER1's name server failed; the cluster ran fine on the substitute for days. After a planned full restart, who is master and what happens to index-server roles?

When distributing tenant databases across a scale-out HANA cluster, what does SAP tell you NOT to do, and what runs on all nodes regardless?

Why must a single standby host have access to ALL data volumes in a scale-out HANA cluster, not just one worker's?

All else equal, when does SAP say to choose scale-up over scale-out for HANA, and what is the single biggest performance reason?

An admin assumes HANA scale-out works like a traditional shared-disk DB cluster (same data on every node). What's wrong, and what's the operational implication?

A 9 TB transaction table won't fit one host AND it constantly joins a small master-data table. Which of the three table-arrangement techniques applies to each part?

Scale-out is transparent to client apps — except for one thing they must be told. What, and what breaks if you skip it?

Beyond serving hot vs warm data, what failover problem does host grouping prevent in a scale-out HANA system?

Combining HANA system replication with a scale-out system: which host count MUST match between primary and secondary, and which may differ?

Up to three HANA hosts can be name-server master candidates. What is the master name server uniquely responsible for that the slaves are not?

In a SAN scale-out setup, what does the Fiber Channel Storage Connector do at takeover to prevent two nodes writing the same volumes?

Besides the index server, which HANA components carry worker/standby role pairs in a scale-out system?

HANA host auto-failover uses both network- and storage-based heartbeats. Why have both instead of relying on the network ping alone?

Must every node in a distributed HANA system have the same CPUs and memory, and how do you safely run mixed hardware?

You're converting a single-host HANA system to multi-host. What hdblcm action should you run first, and what does it change?

While a standby node in a scale-out HANA cluster is idle, what is it actually doing — and what is it NOT doing?

In M_LANDSCAPE_HOST_CONFIGURATION, which column tells you a host's CONFIGURED name-server priority versus its CURRENT role — and why does the distinction matter after a failover?

In host auto-failover, why is fencing required in addition to the heartbeat before a standby takes over?

Cost-wise, when can a scale-out HANA deployment beat one large scale-up host — and what hidden cost of the scale-up option are people forgetting?

A single service (e.g. the index server) crashes on a worker host. Does host auto-failover take over the host?

Standard SAP sizing assumes data fills what fraction of memory, what's the rest for, and how does the rule relax for warm data?

What must a HANA JDBC connection string contain so the client survives a host takeover, and what's the failure mode if you list only one host?

How does SAP NetWeaver AS hold the list of HANA hosts for takeover support, and when does the app server actually report a connection error?

Across NFSv3, NFSv4/GPFS, and SAN, when do you NEED a storage connector for HANA failover and what kind?

How do you start/stop ALL nodes of a scale-out HANA system together, and which command lists each host's role?

Does HANA host auto-failover need an external cluster manager, and what infrastructure prerequisite does it absolutely require instead?

For a distributed HANA system, what are the two valid listeninterface values, which does SAP recommend, and what's the catch with the other?

What distinguishes a STANDBY index server from a WORKER, and what does that imply about takeover speed?

For a HANA scale-out cluster, how do NAS and SAN differ in what the standby can reach, and what extra mechanism does SAN need at takeover?

After deploying a HANA patch once, all hosts see the new version, yet each host still keeps its own logs/traces. Which directory is shared and which is local — and which one explains each behavior?