Memory Management & Table Buffering

A row in a buffered table is updated. Is the new value written into the buffer, and what determines how much of the buffer is discarded?

Buffer/memory size parameters are static, but RSMEMORY changes memory allocation at runtime. Exactly what can it change, and what can it NOT?

You open ST06 to investigate a server load problem from yesterday, but the History view is empty while Snapshot and Previous Hours work. What is the cause, and what do you do?

You increase a buffer size parameter in RZ10 to fix swaps. When does the change take effect, and what unit pitfall can silently mis-size it?

Where is a table's buffering configured, and what are the three buffering settings?

In a multi-instance system a buffered table is changed on one instance. Why can another instance still serve the old value — and what does that rule out for buffering candidates?

On a busy HANA system the SQL plan cache is constantly full and evicting plans. Is that a correctness problem, and what is the real cost of the eviction?

What read-to-write threshold should a table clear before you buffer it, and why is a low write rate alone not enough?

Among the heap parameters, what is the distinct role of abap/heaplimit versus the abap/heap_area_* limits?

You buffer a large table with single-record buffering, but performance doesn't improve and the database is still hit constantly. What did you overlook?

Across dialog steps a new work process serves the next step. What does the roll mechanism actually move between work processes, and why does that design matter?

A dialog work process exhausts extended memory or hits its per-process quota. What does it do next, and why can't it keep handing its context to other work processes?

Both SAP buffers and SAP memory areas live in instance memory. What single property distinguishes them, and why does it matter for tuning?

Right after an instance restart, ST02 shows poor buffer hit ratios. Should you resize the buffers based on this reading?

What are the main SAP instance buffers, and which one is the largest?

Despite the word 'initial' in em/initial_size_MB, what does this parameter actually control — and what's the trap?

Buffering can speed data access by up to a factor of 100. What kind of data earns that factor, and why doesn't it generalize to all tables?

HANA's memory usage repeatedly reaches the allocation limit. Beyond slowness, what failures does this threshold trigger?

What are the three table buffering methods, and what does each load?

Why are PRIV-mode dialog work processes a system-wide danger and not just a per-transaction nuisance, and how does the system cap the damage?

ST02 shows swaps in red. For which buffers is a non-zero swap count actually acceptable, and what's the general target?

Slow queries are clearly running, yet HANA logs no expensive SQL statements. What's the cause, and what does an empty list NOT mean?

You suspect a table should be buffered. Which transaction finds the candidate, which tells you the right buffering type, and why can't one do both?

On a single-host S/4HANA where HANA and the ABAP instance share the same server, what stops the two from starving each other of memory?

A report keeps aborting from memory shortage and you need a fast, temporary bump to a dynamically-changeable memory parameter without editing profiles. Which transaction helps?