Database & Tenant Administration

On SPS 01+, you drop a tenant database and assume its backups are gone with it. Why might that assumption now cost you disk space or a compliance miss?

After copying a tenant by SAP HANA database replication, the encrypted target won't start / loses customer config. What did replication NOT bring across that you must restore manually?

You must move ONE tenant (not the whole system) between SAP HANA systems with near-zero downtime. Which copy/move method fits, and why not the others?

You try to set up SAP HANA database replication and it refuses to start. Which two tenant-side preconditions are most often the cause?

You need to revert a tenant to its fallback snapshot. What is the step practitioners forget that makes the single START statement fail?

You're restoring an encrypted SAP HANA database to a DIFFERENT system. Why does ordering matter, and what fails if you get it wrong?

A client needs to reach a HANA tenant but the tenant's individual SQL port may change if it's moved or recreated. Which connection style survives that, and what's the trade-off?

During troubleshooting you need ONE tenant to stay down while the rest of the HANA system restarts. Why is manually stopping it the wrong move?

What deadline does a SAP HANA system rename silently start, and what happens if you miss it?

You upgrade an old single-container HANA system to 2.0 SPS 01+. What silently changes about its architecture, and where does your existing data end up?

SAP recommends separate network zones for HANA. Which two zones are most likely to be wrongly collapsed into one, and which pairing is actually acceptable?

You see I/O checksum errors versus duplicate keys / NULLs in non-NULL columns. Which is physical and which is logical corruption — and why does the distinction change your response?

Before a risky data load you want a quick way to roll back a tenant. Why must you NOT treat a fallback snapshot as your backup, and what one-per-tenant limit bites you?

A data-volume reclaim aborts with "Shrink canceled, probably because of snapshot pages" on a system-replication system. What's the real cause and what must you check before retrying?

You want to scan the HANA persistence layer for physical corruption with hdbpersdiag. Why can't you just run it on production, and what's the standard workaround?

You're consolidating many tenants onto one HANA instance and hit a wall around 20. What's the real ceiling and which parameter raises it?

You plan to copy a tenant via backup-and-restore using backups taken partly with the file system and partly with a third-party tool. Why will this fail, and what does SAP prefer instead?

A database-replication job reaches ACTIVE. The finalize SQL differs for a copy versus a move — what's the one-clause difference and why does it matter operationally?

SAP HANA runs consistency checks at startup. Why is that not enough on its own, and what should you still schedule?

You enable systemdb_separated_sql_port=true to isolate system-DB SQL. What breaks for existing clients, and what must NOT be exposed externally?

Database-replication REPLICATION_STATUS shows SYNCING. Is that the normal initial-load state, and how do you tell it apart from a genuine problem?

A colleague says HANA data/log volumes 'always' live under /hana/data/<SID>. When is that true, when is it not, and which parameters actually decide?

Corruption is reported. Do you reach for CHECK_CATALOG or CHECK_TABLE_CONSISTENCY — and which one must you NOT run on a schedule?

In CHECK_TABLE_CONSISTENCY, what does NULL vs a name in the schema/table arguments mean, and how do you skip one specific sub-check?

Data-volume striping defaults to a 2,000 GB cap per striped file. Where does that number come from, and why shouldn't you casually raise it?

To parallelize HANA volume I/O across disks you plan to stripe both data and log volumes. Why is half of that plan impossible?

You stop a tenant with ALTER SYSTEM STOP DATABASE. What does the IMMEDIATE option actually do, and when would that be the wrong choice?

Renaming a tenant fails. Which precondition is most often overlooked, and what side effect does a successful rename have on the OS and cockpit?

You want to grant an operator the ability to stop/start a tenant but NOT create, rename, or delete tenants. Which privilege is enough, and which is overkill?

You created a new tenant from the system database and it came up instantly. Why can't you administer it yet, and what was set up during creation?

Why would you disable features like import/export or backups on a tenant, and where do you find the full list of what can be disabled?

You archived a large amount of data but disk usage didn't drop. Why doesn't HANA reclaim the space automatically, and what must precede the reclaim?

Why does a more frequent savepoint interval make HANA restart faster, and what is actually replayed on restart?

A tenant admin gets "Change not allowed for tenant database." What is the blacklist actually protecting against — and what is it NOT about?

Of a savepoint's three phases, why is Phase 2 the one that must stay under a second — and what's the real difference from Phases 1 and 3?

You want to move a tenant between systems with zero client-side changes. What must you have set up beforehand for that to work, and what's the trade-off?

A tenant is missing a service. Can you add it from within that tenant, or does the operation require the system database?