Two application pods can update the same database row at the same time because each pod has its own process memory. A mutex inside one pod cannot coordinate with another. To protect a read, decision, and write in PostgreSQL, put the sequence in a transaction and lock the row with FOR UPDATE before using its current values.
Why two pods can race
Pods are independent processes: they do not share in-memory locks or state. If both handle requests that read and then change the same database row, the shared database—not either pod’s memory—is the coordination point. PostgreSQL transactions and row locks govern how those concurrent operations interact. PostgreSQL’s Explicit Locking documentation notes: “Row-level locks do not affect data querying; they block only writers and lockers to the same row.”
As an Amazon Associate I earn from qualifying purchases.
This is a data-correctness problem, distinct from keeping enough pods running. A Kubernetes PodDisruptionBudget constrains voluntary pod disruptions for availability; it does not lock a database row or make concurrent updates atomic. Kubernetes documents disruption budgets as part of voluntary disruption management.
Free tools Windows power users keep installed
One-click scans. No signup required.
Lock the row before making the decision
When a business rule depends on a row’s current value, acquire the row lock within the same transaction that validates and changes it. For example, a balance adjustment could follow this shape:
#1 Best Overall
BEGIN;
SELECT balance FROM accounts WHERE id = :id FOR UPDATE;
-- validate the business rule using the locked row
UPDATE accounts SET balance = :new_balance WHERE id = :id;
COMMIT;The transaction must remain open from the locking read through the update and commit. PostgreSQL holds the lock until the transaction ends, so another transaction that tries to lock or modify that row may have to wait. Keep this interval short: contention can delay other work, and locking can add disk writes. PostgreSQL’s locking guidance describes these effects.
The SQL illustrates the coordination pattern, not a complete application implementation. The exact transaction API, validation, rollback, and error handling depend on the application’s database driver and business rule.
Rank #2
Isolation level changes what happens after a wait
A lock does not remove the need to understand the transaction isolation level. In PostgreSQL Read Committed, a locking operation that waits for a concurrent updater can proceed against the updated row version after the other transaction commits. Under Repeatable Read, trying to lock a row changed since the transaction began can instead produce a serialization error. Check the behavior against the PostgreSQL major version you deploy; the cited isolation guidance is for PostgreSQL 18, while the cited explicit-locking page is PostgreSQL 14 documentation. PostgreSQL’s transaction isolation documentation explains the isolation behavior.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →If the selected isolation level can reject a transaction with a serialization failure, the application needs a retry path for that failure. Retrying should rerun the relevant transaction logic rather than merely repeat the final write, because the business decision may need to be recalculated from the new database state. PostgreSQL specifically says applications using Serializable isolation should be prepared to retry serialization failures. See PostgreSQL’s isolation guidance.
Choose between explicit locking and Serializable isolation
Both approaches can help preserve correctness, but conflicts surface differently. Explicit locking gives the application control over which rows to lock and can make competing operations wait. Serializable isolation asks PostgreSQL to ensure transactions behave as if they ran in a serial order; it can abort a transaction that cannot safely be serialized, so the application must handle retries.
| Approach | How conflicts are handled | Application work | Scope and trade-off |
|---|---|---|---|
Explicit row lock (FOR UPDATE) |
Competing writers or lockers on the locked row can block until the lock is released. | Lock the relevant rows before relying on their values; handle transaction errors and keep the transaction short. | Coordinates only the rows and transaction scope actually locked; blocking can increase contention. |
| Serializable transaction isolation | PostgreSQL may reject a transaction with a serialization failure rather than allow an unsafe outcome. | Be prepared to retry the transaction after a serialization failure. | Enforces serializable behavior across the transaction, but retries and transaction aborts affect application behavior. |
Neither method is universally faster or safer for every workload. The right choice depends on the invariant, the data participating in it, and how the application handles blocking or retries. PostgreSQL’s documented locking and isolation semantics are specific to PostgreSQL and can differ across database products and versions.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Know what the lock does not protect
A row lock protects only the rows actually locked and the work coordinated within that transaction. If a rule depends on another row, table, or predicate that is read without suitable coordination, locking one row does not automatically make the wider invariant safe.
Recommended Free Tools
PostgreSQL documents an example in which a locked row and an unlocked subquery read can observe inconsistent privilege information. Its suggested mitigations have permission and performance trade-offs. For multi-row operations such as transfers, consider locking all relevant rows in a stable order, using an appropriate isolation level, or choosing another database-supported atomic strategy. Handle rollback and any required retry behavior. PostgreSQL’s application-level consistency guidance discusses the limitation and mitigation trade-offs.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
