1237267 - curl --request POST http://$CLOUD_HOST/device/release didn't reply anything

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Bitbar, defect)

Description

[Johan Lorenzo [:jlorenzo]]

https://tools.taskcluster.net/task-inspector/#VQoeqvVyTeKAQqlWRY1nMA/0

> Summary
> =======
> 
> Ran 164 tests
> Expected results: 6
> Unexpected results: 0
> Skipped: 158
> 
> [32mOK(B[m
> ========= Finished Running tests (results: 0, elapsed: 773 secs) (at 2016-01-05 20:18:20.666511) =========
> Releasing device.
> curl: (52) Empty reply from server
> [taskcluster] === Task Finished ===
> [taskcluster] Unsuccessful task run with exit code: 52 completed in 1204.915 seconds

I'm not too sure if this curl call comes from https://github.com/taskcluster/testdroid-proxy or https://github.com/mozilla-b2g/testdroid-client/

Comment 1

[Greg Arndt [:garndt]]

that comes from here: 
https://dxr.mozilla.org/mozilla-central/source/testing/docker/tester-device/bin/entrypoint#43

It's so that the session is released in testdroid.  I'm not immediately sure why this is starting to happen though and wasn't happening before.

Comment 2

[Greg Arndt [:garndt]]

This is the endpoint being called:
https://github.com/taskcluster/testdroid-proxy/blob/7263d8a5ca1dd228ebde79f66054363ee68c03d9/src/bin/server.js#L130

Comment 3

[Johan Lorenzo [:jlorenzo]]

Thanks for the details Greg. Moving the bug to Bitbar as 
> curl --request POST http://$CLOUD_HOST/device/release
didn't reply anything.

From man curl:
> 52     The server didn't reply anything, which here is considered an error.

Comment 4

[Greg Arndt [:garndt]]

I think this is something on our side, not bitbar.

This should have returned something from our proxy.  Does this happen with every testdroid job or just some?

This should be the reply from our proxy:
https://github.com/taskcluster/testdroid-proxy/blob/7263d8a5ca1dd228ebde79f66054363ee68c03d9/src/bin/server.js#L136

Comment 5

[Johan Lorenzo [:jlorenzo]]

(In reply to Greg Arndt [:garndt] from comment #4)
> This should have returned something from our proxy.  Does this happen with
> every testdroid job or just some?
I just noticed this failure once. It seems okay the other times.

Comment 6

[Johan Lorenzo [:jlorenzo]]

Firefox OS labs are being winded down. This issue won't be solved. Closing.