Backup & Recovery

You take a storage snapshot of a running HANA system without telling the database. On recovery, why does HANA treat it as a crash, and what would a database-aware snapshot have done differently?

Before a restore you want to confirm the backup files are not just present but actually uncorrupted. Why is hdbbackupdiag not enough, and which tool do you use?

Under sustained high transaction load, single-segment log backups start queuing up. What HANA feature fixes this, and what is the consequence of leaving max_log_backup_size at its default?

From the system database you need to back up only tenant S14. What is the statement, and what is the easily-skipped detail that changes which database gets backed up?

For a production database you must lose as little data as possible. Which HANA recovery mode do you pick, and why is "to a specific data backup" the wrong choice there?

You want your third-party (Backint) HANA backups to land on a specific path. Why does the usual file-backup path parameter not help here?

You need log backups to fire on a fixed time schedule, not just when a segment fills. Which log_backup_interval_mode value do you set, and what does the default mode do instead?

You recover a HANA database onto a different host. Why does the license break, and why does the original license type decide whether you have 90 days of breathing room or an emergency?

A colleague says "the data backup contains everything, so a from-scratch HANA restore just needs that one file." What is missing, and what design property makes this true?

Which direction of HANA backup portability is allowed — single-tenant <-> tenant of an MDC system — and what version constraint always applies?

After a from-scratch HANA reinstall and restore, the database comes up but behaves wrong (missing parameters). What was never in the backup, and where should it have been captured?

Your HANA backup catalog is lost. What kind of recovery is still possible, and crucially what becomes impossible?

Which parameter relocates HANA file-based data backups, and what is the operationally important fact about when the change applies and the per-file size cap?

A stakeholder says "we have a 15-minute failover, so our data is safe." Which objective are they describing, and which one actually governs how much data they can lose?

SAP says reserve space for at least two HANA file backups even when you keep only one. Why does a single retained backup still need double the space at the moment of backup?

Your HANA backup catalog is corrupt and you run hdbbackupdiag --generate to rebuild it. Why should you not treat the rebuilt catalog as complete?

When is HANA's "Initialize Log Area" the correct recovery option, and what is the cost that makes it wrong to use just to speed things up?

Right after a HANA system-replication takeover, are the old primary's backups still valid, and what must happen before delta backups work on the new primary?

In a scale-out (multihost) HANA system, what infrastructure must be in place for a file-system backup, and what happens if it isn't?

You need near-zero-impact HANA backups and someone proposes data snapshots over Backint. What two properties of snapshots make that a risky default?

Before running a HANA file backup you must confirm there's enough target space. Which view gives the estimate, and what verifies the target can actually hold it?

Recovering a HANA database that uses full + differential + incremental + log backups — what is the exact apply order, and why is delta-then-log faster than log-only?

If HANA's savepoints + redo logs already restart the database without losing committed data, why are backups still mandatory?

You perform a point-in-time recovery of the primary in a HANA system-replication pair. Why can't you simply resume replication, and what's the correct sequence?

Someone deleted data by mistake hours ago in production. PiT recovery is the only option that can bring it back — so why is running it on production the wrong first move?

In which three situations does HANA back up (save) a log segment?

Encrypted HANA backups are spiking CPU/memory on an already-loaded system. Which parameter eases the impact, and what do you trade away — and what does it NOT change?

HANA backup encryption uses AES 256-bit with the backup root key. What must you back up separately, and what fails if you don't?

Recovering an entire multitenant HANA system — what is the mandatory order, and what must you NOT attempt to save time?

What is the SQL workflow to create and confirm a HANA data snapshot, and what concurrency rule will silently block your other backups during it?

What are the three phases of a HANA recovery in order, and why must page restore precede log replay rather than the reverse?

HANA's retention policy will auto-delete old backup generations by age. Why does the "Minimum No. of Retained Backup Generations" (default 2) exist, and what must you watch when it deletes?

Differential vs incremental HANA delta backups: one restores from fewer files, the other writes smaller backups. Which is which, and what trade-off does that force?

What does HANA's auto-added backup-file suffix encode, and why must you never rename a backup file after it's created?

A single tenant is misbehaving and you want a fast, low-impact point-in-time copy of just that tenant. Why is a data snapshot the wrong tool, and how does that constrain recovery?

Disk is filling with old HANA backups. Which form of BACKUP CATALOG DELETE actually reclaims space, and what does the other form leave behind?

Among HANA's backup monitoring views, why can M_BACKUP_PROGRESS mislead you if you use it to confirm last night's backups?