Don't require AWS access keys for S3 pytests

[arpad-m]

Don't require AWS access keys (AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY) for S3 usage in the pytests, and also allow AWS_PROFILE to be passed.

One of the two methods is required however.

This allows local development like:

aws sso login --profile dev
export ENABLE_REAL_S3_REMOTE_STORAGE=nonempty REMOTE_STORAGE_S3_REGION=eu-central-1 REMOTE_STORAGE_S3_BUCKET=neon-github-ci-tests AWS_PROFILE=dev
cargo build_testing && RUST_BACKTRACE=1 ./scripts/pytest -k debug-pg16 test_runner/regress/test_tenant_delete.py::test_tenant_delete_smoke

related earlier PR for the cargo unit tests of the remote_storage crate: #6202

[arpad-m]

sadly it doesn't work however, the test fails and the pageserver log says it's an authentication issue:

2024-01-31T16:51:13.730780Z  INFO Using s3 bucket 'neon-github-ci-tests' in region 'eu-central-1' as a remote storage, prefix in bucket: 'Some("929215e1-150a-4671-a5f0-fa8511a54bc0/test-tenant-delete-smoke-debug-pg16-/pageserver")', bucket endpoint: 'None', profile: Some("dev"), access_key_id: None
[...]
2024-01-31T16:51:16.558695Z  INFO attach_preload{tenant_id=354d6017ac6faba459d228e056d30cd1 shard_id=0000}: list timelines for 354d6017ac6faba459d228e056d30cd1 failed, will retry (attempt 1): Failed to download a remote file: Failed to list S3 prefixes

Caused by:
    0: dispatch failure
    1: other
    2: the credential provider was not enabled
    3: no providers in chain provided credentials

Apparently the key does get picked up by the pageserver but for some reason it still doesn't pick up the auth information...

[arpad-m]

found the issue, had to also import the HOME env variable. See the new commit.

It works now locally, although i had to increase the iteration number, probably due to higher latencies from my box to S3:

diff --git a/test_runner/regress/test_tenant_delete.py b/test_runner/regress/test_tenant_delete.py
index b4e5a550f..392b8add6 100644
--- a/test_runner/regress/test_tenant_delete.py
+++ b/test_runner/regress/test_tenant_delete.py
@@ -89,7 +89,7 @@ def test_tenant_delete_smoke(
     iterations = poll_for_remote_storage_iterations(remote_storage_kind)
 
     assert ps_http.get_metric_value("pageserver_tenant_manager_slots") == 2
-    tenant_delete_wait_completed(ps_http, tenant_id, iterations)
+    tenant_delete_wait_completed(ps_http, tenant_id, iterations * 10)
     assert ps_http.get_metric_value("pageserver_tenant_manager_slots") == 1
 
     tenant_path = env.pageserver.tenant_dir(tenant_id)

[github-actions]

2376 tests run: 2263 passed, 0 failed, 113 skipped (full report)

---

Flaky tests (8)

Postgres 16

• test_gc_cutoff: debug
• test_emergency_mode: release

Postgres 15

• test_gc_cutoff: debug
• test_emergency_mode: debug
• test_location_conf_churn[3]: debug

Postgres 14

• test_gc_cutoff: debug
• test_timeline_deletion_with_files_stuck_in_upload_queue: debug
• test_timeline_size_quota_on_startup: release

Code coverage (full report)

• functions: 54.5% (11228 of 20610 functions)
• lines: 81.5% (63274 of 77650 lines)

---

_{The comment gets automatically updated with the latest test results
d2e1d44 at 2024-02-01T18:35:15.038Z :recycle:}

[bayandin]

Nice! It's going to be much more convenient to use it locally