TypeDB 3.7.0-rc0 panic: [STO10] Failed to recover from checkpoint on startup (Podman volume)

[CaliLuke]

Description

TypeDB server in typedb/typedb:3.7.0-rc0 crashes on startup with a checkpoint recovery failure for an existing database. No manual shutdowns or host sleeps occurred; only local test runs between restarts.

Environment

1. TypeDB distribution: Core (container image typedb/typedb:3.7.0-rc0)
2. TypeDB version: 3.7.0-rc0
3. Environment: macOS host running Podman (podman-machine applehv, 2 GiB RAM, 100 GiB disk)
4. Client and version: typedb console bundled with the 3.7.0-rc0 image
5. Other details: data volume auto-k-server_typedb_data mounted at /opt/typedb/server/data

Reproducible Steps

1. Set up
   • Start TypeDB with existing volume: podman compose -f docker-compose.yml up -d typedb
   • Volume mount path (inside VM): /var/home/core/.local/share/containers/storage/volumes/auto-k-server_typedb_data/_data
2. Execute
   • Container immediately restarts; status shows Up <few seconds> (starting)
3. Test/Query
   • Probe: ~/.typedb/typedb console --tls-disabled --address=localhost:1729 --username=admin --password=password --command="database list"
4. Unexpected result
   • Console fails with transport error: connection error: stream closed because of a broken pipe
   • Container logs show checkpoint recovery panic (STO10/CLO5) for project_4b8fc2079ccc471bb682f41fab5dd09e

Expected result

TypeDB server should start and list databases; existing project database should open without checkpoint errors.

Additional information

Relevant logs from TypeDB or Driver:

2025-11-30T07:29:54.044013Z ERROR A panic occurred panic.payload="called `Result::unwrap()` on an `Err` value: [SRO6] Could not open database.
Cause:
      [DBO4] Error opening storage layer.
Cause:
      [STO10] Failed to recover from checkpoint for database 'project_4b8fc2079ccc471bb682f41fab5dd09e'.
Cause:
      [CLO5] Error restoring checkpoint in directory '"/opt/typedb-server-linux-arm64/server/data/project_4b8fc2079ccc471bb682f41fab5dd09e/checkpoint/1764487485088949"'.)
Cause:
      Os { code: 2, kind: NotFound, message: "No such file or directory" }" panic.location="main.rs:55:101"

The data volume currently appears empty when inspected inside the Podman VM; the database folder noted in the panic is missing. A full volume wipe/reseed recovers the service, but the issue recurs after routine test runs, suggesting checkpoint corruption or volume handling instability.

[CaliLuke]

note: the error happens using docker too, moving to podman was my attempt at troubleshooting if the source of instability was the notoriously flakey docker desktop for mac.

[flyingsilverfin]

@dmitrii-ubskii