TestSnapDeals is flaky

[magik6k]

https://app.circleci.com/pipelines/github/filecoin-project/lotus/27614/workflows/e4e62d74-d527-4791-8c07-1178427a19e6/jobs/905778/steps

Failed
=== RUN   TestSnapDeals
2023-03-27T08:11:55.363Z	WARN	advmgr	sealer/manager_test.go:231	Unspecified proof type, make sure to set LOTUS_TEST_SNAP_DEALS_PROOF_TYPE to '2k', '8M', '512M', '32G' or '64G'
2023-03-27T08:11:55.363Z	WARN	advmgr	sealer/manager_test.go:232	Continuing test with 2k sectors
2023-03-27T08:11:55.380Z	WARN	ffiwrapper	ffiwrapper/sealer_cgo.go:196	AddPiece: cannot close pieceData reader *sealer.NullReader because it is not an io.Closer
PC1
PC2
2023-03-27T08:11:55.507Z	WARN	stores	paths/index.go:369	sector {1000 1} redeclared in bbd884a9-e2fe-4d26-9361-47faec02e879
2023-03-27T08:11:55.507Z	WARN	ffiwrapper	ffiwrapper/sealer_cgo.go:196	AddPiece: cannot close pieceData reader *strings.Reader because it is not an io.Closer
2023-03-27T08:11:55.520Z	WARN	ffiwrapper	ffiwrapper/sealer_cgo.go:196	AddPiece: cannot close pieceData reader *strings.Reader because it is not an io.Closer
RU
RU duration (2KiB): 58.222697ms
PR1
PR1 duration (2KiB): 5.354157ms
PR2
thread '<unnamed>' panicked at 'range end index 527801108 out of range for slice of length 368444928', library/core/src/slice/index.rs:73:5
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
    manager_test.go:290: 
        	Error Trace:	/home/circleci/lotus/storage/sealer/manager_test.go:290
        	Error:      	Received unexpected error:
        	            	storage call error 0: Rust panic: no unwind information [name: ee6baf28488b]: Rust panic: no unwind information
        	Test:       	TestSnapDeals
--- FAIL: TestSnapDeals (0.34s)

somehow ProveReplicaUpdate2 occasionally fails:

lotus/storage/sealer/manager_test.go

Lines 289 to 290 in 2120fae

	proof, err := m.ProveReplicaUpdate2(ctx, sid, sectorKey, out.NewSealed, out.NewUnsealed, vanillaProofs)
	require.NoError(t, err)

[rjan90]

Closing this ticket as outdated. Lotus has moved from CircleCI to GHA, and we are currently monitoring Flaky jobs over time in this Grafana dashboard.

We have also consolidated flaky test reports into a single tracking issue here: #12001