segfault after unexpected server reboot #680

New issue

Closed

opened 2023-12-22 09:16:07 +00:00 by lpapier · 2 comments

lpapier commented

2023-12-22 09:16:07 +00:00

Hi,
After an unexpected server reboot, garage failed to start with a segfault error.

garage version is

garage v0.8.4 [features: k2v, sled, lmdb, sqlite, consul-discovery, kubernetes-discovery, metrics, telemetry-otlp, bundled-libs]

See included log file generated with RUST_LOG=garage=trace garage -c ./disaster.toml server.

My config is a single garage node.

metadata_dir = "/data/disaster/meta"
data_dir = "/data/disaster/data"
db_engine = "lmdb"
block_size = 10485760
replication_mode = "1"
compression_level = 2

Running mdb_stat -a -e -f db.lmdb gives no error.

Could it be a garage bug or a lmdb corruption?
If you need more info or trace, just ask for.

Thanks.
Laurent

Hi, After an unexpected server reboot, `garage` failed to start with a segfault error. `garage` version is ``` garage v0.8.4 [features: k2v, sled, lmdb, sqlite, consul-discovery, kubernetes-discovery, metrics, telemetry-otlp, bundled-libs] ``` See included log file generated with `RUST_LOG=garage=trace garage -c ./disaster.toml server`. My config is a single garage node. ```toml metadata_dir = "/data/disaster/meta" data_dir = "/data/disaster/data" db_engine = "lmdb" block_size = 10485760 replication_mode = "1" compression_level = 2 ``` Running `mdb_stat -a -e -f db.lmdb` gives no error. Could it be a garage bug or a lmdb corruption? If you need more info or trace, just ask for. Thanks. Laurent

garage_crash_trace.log

13 KiB