Make garbage collection test less failure prone (#3599)

## Issue Addressed

NA

## Proposed Changes

This PR attempts to fix the following spurious CI failure:

```
---- store_tests::garbage_collect_temp_states_from_failed_block stdout ----
thread 'store_tests::garbage_collect_temp_states_from_failed_block' panicked at 'disk store should initialize: DBError { message: "Error { message: \"IO error: lock /tmp/.tmp6DcBQ9/cold_db/LOCK: already held by process\" }" }', beacon_node/beacon_chain/tests/store_tests.rs:59:10
```

I believe that some async task is taking a clone of the store and holding it in some other thread for a short time. This creates a race-condition when we try to open a new instance of the store.

## Additional Info

NA
This commit is contained in:
Paul Hauner 2022-09-23 03:52:44 +00:00
parent 3a3dddc5fb
commit 9246a92d76

View File

@ -26,6 +26,7 @@ use store::{
HotColdDB, LevelDB, StoreConfig,
};
use tempfile::{tempdir, TempDir};
use tokio::time::sleep;
use tree_hash::TreeHash;
use types::test_utils::{SeedableRng, XorShiftRng};
use types::*;
@ -1985,16 +1986,16 @@ async fn pruning_test(
check_no_blocks_exist(&harness, stray_blocks.values());
}
#[test]
fn garbage_collect_temp_states_from_failed_block() {
#[tokio::test]
async fn garbage_collect_temp_states_from_failed_block() {
let db_path = tempdir().unwrap();
// Wrap these functions to ensure the variables are dropped before we try to open another
// instance of the store.
let mut store = {
let store = get_store(&db_path);
let harness = get_harness(store.clone(), LOW_VALIDATOR_COUNT);
// Use a `block_on_dangerous` rather than an async test to stop spawned processes from holding
// a reference to the store.
harness.chain.task_executor.clone().block_on_dangerous(
async move {
let slots_per_epoch = E::slots_per_epoch();
let genesis_state = harness.get_current_state();
@ -2021,9 +2022,19 @@ fn garbage_collect_temp_states_from_failed_block() {
store.iter_temporary_state_roots().count(),
block_slot.as_usize() - 1
);
},
"test",
);
store
};
// Wait until all the references to the store have been dropped, this helps ensure we can
// re-open the store later.
loop {
store = if let Err(store_arc) = Arc::try_unwrap(store) {
sleep(Duration::from_millis(500)).await;
store_arc
} else {
break;
}
}
// On startup, the store should garbage collect all the temporary states.
let store = get_store(&db_path);