PostgreSQL TOAST corruption
PostgreSQL stores oversized column values outside the main table row in a related TOAST table. If one chunk is missing, duplicated, or points to the wrong value, a query may fail only when it reads that specific text, JSON, bytea, geometry, or other large field.
Understand the symptom
The main row can hold an external pointer while pg_toast stores numbered pieces. Recovery identifies the base relation, toast relation, value ID, expected sequence, and transaction context, then evaluates backups, replicas, WAL, indexes, and readable columns without assuming that VACUUM or REINDEX can recreate missing payload bytes.
What you may see
A SELECT fails with a pg_toast relation, value ID, and expected chunk.
Queries succeed until they detoast the damaged large value.
Logical export reaches the corrupt value and stops even though many tables remain readable.
Access paths may encounter affected tuples in a different order or avoid the large column.
Protect the source
Other columns and related records may still contain essential business information.
Tuple cleanup and freezing alter the cluster while the corruption scope is still unknown.
TOAST identities, chunks, indexes, and MVCC state must remain internally consistent.
Do this now
Pause VACUUM, schema changes, bulk loads, and applications that update the affected database.
Save relation OID, value ID, chunk number, SQL, server log, version, and first occurrence.
Copy PGDATA, tablespaces, WAL archives, configuration, extensions, and backups together.
List the affected table, key ranges, large columns, dependencies, and acceptable recovery point.
NEPFOSYS process
NEPFOSYS maps the base row, TOAST table, indexes, and physical files within the preserved cluster.
Readable tuples, replicas, backups, WAL, and alternate row versions are evaluated on copies.
Recoverable rows and columns are extracted with damaged values clearly separated for application review.
A clear next step
Start a case with the exact symptoms, incident timeline, prior attempts, encryption details, and the files that matter most. NEPFOSYS will route the case to the appropriate DFW intake or nationwide mail-in path.
Start with a case reviewQuestions, answered
It stores pieces of oversized values that do not fit efficiently inside the main heap row.
No. An index can be rebuilt from existing rows, but it cannot recreate payload bytes that are absent from the TOAST relation.
Often, readable columns can still be extracted, but the exact result depends on tuple, page, transaction, and storage condition.
DFW local · Nationwide mail-in