Remove M1-specific docs - Airbyte works on M1 Macs now

[evantahler]

In a fresh environment without any special environment variables, Airbyte works just fine locally now, using the most recent published containers.

To confirm this:

# Erase any docker images & volumes you may have 
docker system prune -a # hit yes
docker volume prune # hit yes

# unset any special environment variables you may have
unset DOCKER_BUILD_PLATFORM
unset DOCKER_BUILD_ARCH
unset ALPINE_IMAGE
unset POSTGRES_IMAGE
unset JDK_VERSION

# check out the master branch
git checkout master # in the airbyte repo

# see that it just works (tm)
docker compose up

┆Issue is synchronized with this Monday item by Unito

[evantahler]

As a note when testing, it looks like POSTGRES_IMAGE isn't in use, so that can be removed as well (#11860)

[sherifnada]

Seminal work

[jrhizor]

Have you tried running acceptance tests locally on the published amd64 images on an M1 machine? I think most of the M1-related issues people were seeing when not building the images themselves were transient exceptions or random stall outs that aren't present when building for ARM.

If you can reliably run the acceptance tests this is 100% good to go.

[evantahler]

Ah! On my local machine (M1 Macbook Pro), I've got one consistent failure.

$ uname -a
# Yep, really using an M1 chip (ARM64 something something)
Darwin evan-airbyte-mac.local 21.4.0 Darwin Kernel Version 21.4.0: Fri Mar 18 00:46:32 PDT 2022; root:xnu-8020.101.4~15/RELEASE_ARM64_T6000 arm64


$ SUB_BUILD=PLATFORM ./gradlew clean build
# ...
BUILD SUCCESSFUL in 4m 38s
355 actionable tasks: 318 executed, 24 from cache, 13 up-to-date


$ SUB_BUILD=PLATFORM ./gradlew :airbyte-tests:acceptanceTests
Building Airbyte Sub Build: PLATFORM

> Task :airbyte-tests:acceptanceTests
AcceptanceTests > testGetDestinationSpec() PASSED
AcceptanceTests > testFailedGet404() PASSED
AcceptanceTests > testGetSourceSpec() PASSED
AcceptanceTests > testCreateDestination() PASSED
AcceptanceTests > testDestinationCheckConnection() PASSED
AcceptanceTests > testCreateSource() PASSED
AcceptanceTests > testSourceCheckConnection() PASSED
AcceptanceTests > testDiscoverSourceSchema() PASSED
AcceptanceTests > testCreateConnection() PASSED
AcceptanceTests > testManualSync() FAILED
    org.opentest4j.AssertionFailedError: expected: <succeeded> but was: <pending>
        at app//org.junit.jupiter.api.AssertionUtils.fail(AssertionUtils.java:55)
        at app//org.junit.jupiter.api.AssertionUtils.failNotEqual(AssertionUtils.java:62)
        at app//org.junit.jupiter.api.AssertEquals.assertEquals(AssertEquals.java:182)
        at app//org.junit.jupiter.api.AssertEquals.assertEquals(AssertEquals.java:177)
        at app//org.junit.jupiter.api.Assertions.assertEquals(Assertions.java:1141)
        at app//io.airbyte.test.acceptance.AcceptanceTests.waitForSuccessfulJob(AcceptanceTests.java:1652)
        at app//io.airbyte.test.acceptance.AcceptanceTests.testManualSync(AcceptanceTests.java:498)

11 tests completed, 1 failed, 1 skipped

> Task :airbyte-tests:acceptanceTests FAILED

Initially, I couldn't even boot the app, so I thought the problemspace was more binary. I'll try and see if I can get that test passing.

[evantahler]

Any advice on how to get the Github tests passing @jrhizor ?

[davinchia]

can you try ./gradlew format?

[davinchia]

@jrhizor are we actually publishing M1 images with each release now or are these changes only limited to the developing locally portion?

[evantahler]

I'm still doing discovery on this topic (notes @ https://www.notion.so/MacOS-M1-Problems-6c3c441d951f4410bdf47a8125fbb195), but so far the only custom image I've found for M1/ARM is airbyte/temporal-auto-setup, which may itself no longer be needed, if we upgrade.

[davinchia]

there is an open issue on the infrastructure side to publish Airbyte M1 images that is in the last stages: https://github.com/airbytehq/airbyte-cloud/issues/801

this is separate from actually running Airbyte on M1 macs, which sounds like is part of your goals too?

[evantahler]

./gradlew format is OK (edit: my fault)

[jrhizor]

These problems running syncs with Airbyte go away when using M1-built images for the platform. We probably just need to wait for https://github.com/airbytehq/airbyte-cloud/issues/801 and then we can merge this.

[sherifnada]

it looks like ARM64 should be supportable on self-hosted runners: https://docs.github.com/en/actions/hosting-your-own-runners/about-self-hosted-runners#architectures

so that's probably the first step actually, like Jared said

[davinchia]

Yeap. We can switch the AMIs out as we need to. We should build both architectures so we don't need to immediately switch those out, though we should eventually do so since AWS ARM instances are 25% cheaper than x86.

[CLAassistant]

All committers have signed the CLA.