1764901 - Install blackhole-2ch and blackhole-16ch on macOS test machines

Status: RESOLVED FIXED

(Infrastructure & Operations :: RelOps: General, defect, P5)

Description

[Paul Adenot (:padenot)]

It's on brew under those names if it helps, and it's free software (https://github.com/ExistentialAudio/BlackHole, GPL 3.0).

If we need more, we can compile it easily https://github.com/ExistentialAudio/BlackHole/wiki/Running-Multiple-BlackHole-Drivers. 2 more + the builtin should suffice though.

Comment 1

[Paul Adenot (:padenot)]

Dave, I'm told you might be the right person to look into this? If not, do you know who owns this? We need this for proper testing of bug 1238038 on macOS. We can test on Linux right now because PulseAudio has a facility to do this built-in, but on macOS it requires a 3rd party software.

Comment 2

[:dhouse]

(In reply to Paul Adenot (:padenot) from comment #0)

It's on brew under those names if it helps, and it's free software (https://github.com/ExistentialAudio/BlackHole, GPL 3.0).

If we need more, we can compile it easily https://github.com/ExistentialAudio/BlackHole/wiki/Running-Multiple-BlackHole-Drivers. 2 more + the builtin should suffice though.

I downloaded the 2ch, 16ch, and 64ch pkg's. We don't use homebrew currently (removed for security concerns in '20). So the pkg files are perfect (we re-package dmgs or archives as pkg installs for the workers).

I'll set up a test instance with the 2ch and 16ch drivers installed for you to test. If they don't have unique uid's, I'll rebuild the 16ch one with "16ch" in the uid.

Comment 3

[:dhouse]

Attachment: BlackHole2ch.v0.2.10.pkg

Comment 4

[:dhouse]

Attachment: BlackHole16ch.v0.2.10.pkg

Comment 5

[Paul Adenot (:padenot)]

(In reply to :dhouse from comment #2)

I'll set up a test instance with the 2ch and 16ch drivers installed for you to test. If they don't have unique uid's, I'll rebuild the 16ch one with "16ch" in the uid.

They can both be installed at the same time, not going to be a problem. Thanks!

Comment 6

[:dhouse]

Hi :padenot, here is a test pool "releng-hardware/gecko-t-osx-1015-r8-bug1764901". Can you try using the blackhole and see if this works for your tests? I set up the 2ch and 16ch blackhole in puppet and confirmed they are available, and so if this looks good I can roll it out to the production pool.

https://firefox-ci-tc.services.mozilla.com/provisioners/releng-hardware/worker-types/gecko-t-osx-1015-r8-bug1764901

Comment 7

[Paul Adenot (:padenot)]

Thanks a lot!

Chun-Min, can you try your tests there? A command like:

mach try fuzzy --worker-suffix t-osx-1015-r8=-bug1764901

should allow scheduling something on the right worker, but I'm not sure we need to modify something in mochitest.ini or others, I don't remember if we had disabled on all OSes but Linux Desktop.

Comment 8

[C.M.Chang[:chunmin]]

Do you mean the test D143102 added in bug 1238038? That test runs on every platform except Android, as long as the system has at least two audio input devices. I've pushed some mochitest jobs via ./mach try fuzzy --worker-suffix t-osx-1015-r8=-bug1238038. Is this what you expect?

Comment 9

[Paul Adenot (:padenot)]

Yeah, we'll see what's going to happen on those boxes, thanks.

Comment 10

[Paul Adenot (:padenot)]

Dave, it seems like something went wrong with https://treeherder.mozilla.org/jobs?repo=try&revision=93f77969a2484ec6c11a53b0bae2b46a68110803, do you know how we would go about verifying this work, do we need anything from you, or should we try to simply re-push? Thanks!

Comment 11

[:dhouse]

:padenot i didn't get back to you on this! I've recreated the test pool as it had gone away after my long delay. It looks like the push from comment 8 was pointed at a different worker suffix/pool (the comment 8 mach appears to point at a pool named after the test's bug instead of this bug; my bad for naming it with a bug number!)
So the try push needs to be pushed like in comment 7: mach try fuzzy --worker-suffix t-osx-1015-r8=-bug1764901

Comment 12

[Michelle Goossens [:masterwayz]]

Is this still needed?

Comment 13

[Paul Adenot (:padenot)]

Yes please, sorry for not answering earlier, this likely got buried in our bugmail without the NEEDINFO.

Comment 14

[Michelle Goossens [:masterwayz]]

Ryan and I are slowly deploying this on machines

Comment 15

[Paul Adenot (:padenot)]

Chun-Min, maybe you could try when you have a minute?