AI & Agent Dev Bug Sandbox Back to Radar

Production PostgreSQL Unavailable After Disk Exhaustion And Expansion; Crash Recovery Not Bringing Database Online

A disk-full event during a migration caused PostgreSQL to crash with SQLSTATE 53100. After expanding disk to 8GB, crash recovery checkpoint completed and read-only mode ended, but all connection methods (Supavisor, direct, REST, SQL Editor) remain timed out, indicating the PostgreSQL backend is not accepting connections or is stuck in a state that prevents normal operation.

criticalConfidence 65%Supabase

Origin Analysis

Disk exhaustion during a WAL write forced the startup process to exit and shut down PostgreSQL. After disk expansion, PostgreSQL entered crash recovery but likely encountered a missing or corrupt WAL segment required for redo, causing the startup process to hang or fail before opening the listen socket. Consequently, Supavisor and authentication services cannot reach the backend, resulting in timeouts across all connection methods.

1. Fill the PostgreSQL data volume during a write-heavy migration until SQLSTATE 53100 is raised. 2. Observe postmaster shutdown. 3. Expand disk to 8GB and restart database/project. 4. See 'crash-recovery checkpoint complete' and 'read-only off' in dashboard logs. 5. Attempt connections via Supavisor session/transaction, direct PostgreSQL, REST, and SQL Editor; all time out.

Fixing Code Block

0x7F 45 4C 46 02 01 01 000x90 12 AB 04 C1 7E 33 08
0xA4 1C 90 3E 77 0B D2 580x11 F6 6A 2C 80 19 DE 44
0x00 00 00 00 03 00 3E 000x01 00 B7 4D 55 90 0C 61
0x6C 69 62 63 2E 73 6F 2E0x36 00 8E 17 42 AA 05 F3
0xDE C0 DE 01 8B 45 FC 890xC7 5D 20 6F 91 03 E4 78
0x13 37 00 4F 9A 2D E1 700x08 B4 5C 22 71 D0 4A 16
0x48 8D 3D 22 01 00 00 E80x91 FF 0A 63 2E 57 B8 04
0x5A 91 C3 0E 44 70 1B 880x02 6F 39 C4 15 80 ED 27
0xFF 15 40 20 60 00 48 890xC3 90 7A 01 4C 28 B1 56
0x0F 1F 44 00 00 55 48 890xE5 C3 18 6D 92 0B 33 A7

Edge Case Audit

Do not run pg_resetwal, delete WAL files, or restore an older backup without explicit data-loss acceptance; these actions can corrupt the cluster or discard current WAL history. A forced restart may interrupt in-progress recovery and should be preceded by a filesystem snapshot. Rollback: if the hotfix worsens the situation, restore from the latest base backup and replay available WAL, but do not overwrite the current data directory unless absolutely necessary.