Continuous benchmark stalled

[joanhey]

https://tfb-status.techempower.com/

[michaelhixson]

I don't think it is stalled. Right now it is going through the four variants of vibed, which each timeout after one hour of building. Hopefully we'll see new results in a half hour or so after it gets through the last variant of vibed.

From the current run's benchmark.log file:

--------------------------------------------------------------------------------
Benchmarking vertx-web-susom-postgres
--------------------------------------------------------------------------------
BENCHMARKING DB ...
Benchmark results:
Complete
BENCHMARKING QUERY ...
Benchmark results:
Complete
BENCHMARKING UPDATE ...
Benchmark results:
Complete
BENCHMARKING FORTUNE ...
Benchmark results:
Complete
vertx-web-susom-postgres: Benchmarking time: 8m 18s
vertx-web-susom-postgres: Build time: 1s
vertx-web-susom-postgres: Time starting database: 1s
vertx-web-susom-postgres: Time until accepting requests: 1s
vertx-web-susom-postgres: Verify time: 8m 56s
vertx-web-susom-postgres: Total test time: 9m 8s
tfb: Total time building so far: 4h53m 26s
tfb: Total time verifying so far: 81h14m 19s
tfb: Total time benchmarking so far: 76h26m 7s
tfb: Total execution time so far: 89h3m 25s
vibed: ERROR: Problem starting vibed
vibed: Total test time: 1h1m 0s
tfb: Total time building so far: 5h53m 26s
tfb: Total time verifying so far: 81h14m 19s
tfb: Total time benchmarking so far: 76h26m 7s
tfb: Total execution time so far: 90h4m 26s
vibed-dmd-pgsql: ERROR: Problem starting vibed-dmd-pgsql
vibed-dmd-pgsql: Total test time: 1h1s
tfb: Total time building so far: 6h53m 26s
tfb: Total time verifying so far: 81h14m 19s
tfb: Total time benchmarking so far: 76h26m 7s
tfb: Total execution time so far: 91h4m 30s
vibed-ldc: ERROR: Problem starting vibed-ldc
vibed-ldc: Total test time: 1h1m 0s
tfb: Total time building so far: 7h53m 26s
tfb: Total time verifying so far: 81h14m 19s
tfb: Total time benchmarking so far: 76h26m 7s
tfb: Total execution time so far: 92h5m 32s

[joanhey]

Perhaps will be useful to try #4543, to lower that 4 hours.

[NateBrady23]

I'm not sure vibed is a good example of re-using the same dockerfile. Most of the differences happen during the build step; compiling with different configurations.

[michaelhixson]

The run got past vibed and tfb-status shows new data now, so I'm going to close this issue. I agree that we should fix vibed at some point, but I see that as part of the always-present meta-issue of "we should fix or remove all failing frameworks".

[zloster]

@nbrady-techempower @michaelhixson
These two runs seem to need cleaning:

• https://tfb-status.techempower.com/results/7005d40e-50d4-4ea4-8506-6010b6494844
• https://tfb-status.techempower.com/results/b9da3281-5254-4581-89eb-fe5933ed1f44