PRAGMA integrity_check; reports “database disk image is malformed (11)” after Litestream restore, but passes after re-restoring later

[ashish01]

I’m observing an inconsistent behavior when restoring a SQLite database from Litestream.

Steps to reproduce:

1. Restore the database using:

   litestream restore -config litestream.yml -o data/db.db /app/data/db.db

2. Open the restored database in sqlite3 and run:

   PRAGMA integrity_check;

3. The result is:

   Parse error: database disk image is malformed (11)
   

4. However, when I run the restore command again about 30 minutes later and then check integrity on the newly restored database,
   the integrity check passes successfully.

Litestream config:

access-key-id: ${LITESTREAM_ACCESS_KEY_ID}
secret-access-key: ${LITESTREAM_SECRET_ACCESS_KEY}

dbs:
  - path: /app/data/db.db
    replicas:
      - type: s3
        bucket: ${B2_BUCKET}
        path: db
        endpoint: https://${B2_ENDPOINT}
        region: us-west-004
        sync-interval: 10s
        retention: 168h  # 7 days
        retention-check-interval: 24h

Question:
Under what conditions could a Litestream restore produce a SQLite database that fails integrity check with
Parse error: database disk image is malformed (11), only for a later restore from the same replica to succeed?