Patch fault-injector errors

[fabubaker]

What changes were proposed in this pull request?

Fixes all fault-injector/LazyFS bugs that show up on CI.

Quick summary of changes:

• Call sync_all on the .meta file.
• Use atomic rename when modifying the Config file.
• Run a clean shutdown only if the DB was in Running state.
• Update the WAL position correctly when replaying a WAL that has no entries.

Why are the changes needed?

These changes should fix any durability bugs revealed by LazyFS.

Does this PR introduce any user-facing change? If yes is this documented?

No.

How was this patch tested?

Running the fault-injector CI on pometry-storage: https://github.com/Pometry/pometry-storage/pull/304

Are there any further changes required?

No.

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Rust Benchmark'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 2.

Benchmark suite	Current: e31454f	Previous: 9823ef7	Ratio
lotr_graph/num_edges	4 ns/iter (± 0)	0 ns/iter (± 0)	+∞
lotr_graph/num_nodes	4 ns/iter (± 0)	1 ns/iter (± 0)	4
lotr_graph/has_node_nonexisting	5 ns/iter (± 0)	2 ns/iter (± 0)	2.50
lotr_graph/graph_latest	3 ns/iter (± 0)	0 ns/iter (± 0)	+∞
lotr_graph_materialise/materialize	8151899 ns/iter (± 51430)	1564816 ns/iter (± 35303)	5.21
lotr_graph_window_100/num_nodes	14 ns/iter (± 0)	5 ns/iter (± 0)	2.80
lotr_graph_window_100_materialise/materialize	8204897 ns/iter (± 50151)	1669150 ns/iter (± 10700)	4.92
lotr_graph_window_10/has_node_existing	142 ns/iter (± 10)	62 ns/iter (± 11)	2.29
lotr_graph_window_10_materialise/materialize	3286052 ns/iter (± 11788)	971980 ns/iter (± 4278)	3.38
lotr_graph_subgraph_10pc/has_node_nonexisting	5 ns/iter (± 0)	2 ns/iter (± 0)	2.50
lotr_graph_subgraph_10pc_materialise/materialize	1983648 ns/iter (± 18657)	334634 ns/iter (± 1287)	5.93
lotr_graph_subgraph_10pc_windowed/has_node_existing	147 ns/iter (± 8)	62 ns/iter (± 14)	2.37
lotr_graph_subgraph_10pc_windowed_materialise/materialize	1189298 ns/iter (± 12368)	230399 ns/iter (± 2617)	5.16
lotr_graph_window_50_layered/num_edges_temporal	149374 ns/iter (± 3778)	70121 ns/iter (± 7586)	2.13
lotr_graph_window_50_layered/has_node_existing	425 ns/iter (± 24)	129 ns/iter (± 12)	3.29
lotr_graph_window_50_layered/has_node_nonexisting	5 ns/iter (± 0)	2 ns/iter (± 0)	2.50
lotr_graph_window_50_layered/graph_latest	87045 ns/iter (± 1603)	36649 ns/iter (± 916)	2.38
lotr_graph_window_50_layered_materialise/materialize	31779269 ns/iter (± 227986)	3488825 ns/iter (± 24948)	9.11
lotr_graph_persistent_window_50_layered/num_edges_temporal	641946 ns/iter (± 17728)	192686 ns/iter (± 1569)	3.33
lotr_graph_persistent_window_50_layered/has_node_existing	444 ns/iter (± 417)	174 ns/iter (± 83)	2.55
lotr_graph_persistent_window_50_layered/has_node_nonexisting	5 ns/iter (± 0)	2 ns/iter (± 0)	2.50
lotr_graph_persistent_window_50_layered/iterate_exploded_edges	3737106 ns/iter (± 79196)	1659940 ns/iter (± 19402)	2.25
lotr_graph_persistent_window_50_layered/graph_latest	141526 ns/iter (± 6515)	57549 ns/iter (± 4809)	2.46
lotr_graph_persistent_window_50_layered_materialise/materialize	55966716 ns/iter (± 281628)	5298035 ns/iter (± 147912)	10.56

This comment was automatically generated by workflow using github-action-benchmark.

[ljeub-pometry]

Looks good