AsyncLocalStorage benchmark broken

[mcollina]

• Version: master
• Platform: Linux
• Subsystem: async_hooks / AsyncLocalStorage

What steps will reproduce the bug?

$ ../node run.js async_hooks | tee out

async_hooks/async-resource-vs-destroy.js
async_hooks/async-resource-vs-destroy.js n=1000000 duration=5 connections=500 path="/" asyncMethod="callbacks" type="async-resource" benchmarker="wrk": 22,246.15
async_hooks/async-resource-vs-destroy.js n=1000000 duration=5 connections=500 path="/" asyncMethod="async" type="async-resource" benchmarker="wrk": 18,063.78
async_hooks/async-resource-vs-destroy.js n=1000000 duration=5 connections=500 path="/" asyncMethod="callbacks" type="destroy" benchmarker="wrk": 20,765.57
async_hooks/async-resource-vs-destroy.js n=1000000 duration=5 connections=500 path="/" asyncMethod="async" type="destroy" benchmarker="wrk": 15,926.41
async_hooks/async-resource-vs-destroy.js n=1000000 duration=5 connections=500 path="/" asyncMethod="callbacks" type="async-local-storage" benchmarker="wrk": 22,208.39
/home/matteo/repositories/node/benchmark/async_hooks/async-resource-vs-destroy.js:119
    return store.state;
                 ^

TypeError: Cannot read property 'state' of undefined
    at getCLS (/home/matteo/repositories/node/benchmark/async_hooks/async-resource-vs-destroy.js:119:18)
    at /home/matteo/repositories/node/benchmark/async_hooks/async-resource-vs-destroy.js:149:39
    at FSReqCallback.readFileAfterClose [as oncomplete] (internal/fs/read_file_context.js:63:3)

What is the expected behavior?

It seems it's losing track of the promise chain.

[mcollina]

cc @vdeturckheim

[vdeturckheim]

Wow, I will take a look today.
cc @puzpuzpuz

[puzpuzpuz]

Thanks for CCing me @vdeturckheim

I'll also try to reproduce the problem on my machine today.

[himself65]

with macOS 10.15.3

[mcollina]

My bad, I was running this on the Linux box, I have updated the bug report.

[puzpuzpuz]

@mcollina thanks for reporting this issue. For some reasons, I was able to reproduce it only with wrk, but not with autocannon. Nevertheless, it should be fixed by #32503.

Update. Forgot to mention that it was an issue in the benchmark itself, not a context loss.

[vdeturckheim]

@puzpuzpuz you just outspeeded me :D good finding!

[puzpuzpuz]

As #32503 is merged, this issue should be closed.