Skip to content

[Test] Test stabilization fixes for 3.16.1: patching, build-image no-internet, multi_az_fsx and AZ allowlisting - #7609

Merged
gmarciani merged 11 commits into
aws:integ-tests-3.16.1from
gmarciani:wip/mgiacomo/3161/integ-tests-stabilize-0904-1
Sep 8, 2026
Merged

[Test] Test stabilization fixes for 3.16.1: patching, build-image no-internet, multi_az_fsx and AZ allowlisting#7609
gmarciani merged 11 commits into
aws:integ-tests-3.16.1from
gmarciani:wip/mgiacomo/3161/integ-tests-stabilize-0904-1

Conversation

@gmarciani

@gmarciani gmarciani commented Sep 4, 2026

Copy link
Copy Markdown
Contributor

Description of changes

Chery-picked

plus:

  • fix to AZ allowlisting to make dynamic capacity reservation honor the allowlisting. This fix is required otherwise the change in [Test] Networking setup: honor the AZ allowlisting everywhere in the networking setup and exclude eu-west-2d. #7603 could cause failures due to capacvity reservation targeting AZs that the VPC stack would not allow.
  • additional improvements to proxy infra to prevent failures in verifications made by proxy client (the cherry-picked commits above stabilized the proxy instance, but not the proxyclient instance).
  • pinning parent image used by build-image tests for RHEL9 to version 9.8. This pinning is needed because previous versions require the paid Extended Update Support (EUS) repos to be configured to install required packages, such as kernel-devel.
  • build_image_no_internet: make the build-image config used by the test honor the flag to enable/disable Lustre installation. As a result, we disable Lustre installation on ubuntu, the same way we do for test_build_image.

Tests

SUCCESS

test-suites:
  createami:
    test_createami.py::test_build_image_no_internet:
      dimensions:
        # RHEL family (rhel8, rhel9) -> Red Hat Enterprise Linux reservation platform
        - regions: [{{ g4dn_2xlarge_CAPACITY_RESERVATION_2_INSTANCES_3_HOURS_NOPG_rhel9 }}]
          instances: ["g4dn.2xlarge"]
          oss: ["rhel9"]
        # Non-RHEL family (alinux2023, ubuntu2204, ubuntu2404, rocky8, rocky9) -> Linux/UNIX reservation platform
        - regions: [{{ g4dn_2xlarge_CAPACITY_RESERVATION_5_INSTANCES_3_HOURS_NOPG_ubuntu2404 }}]
          instances: ["g4dn.2xlarge"]
          oss: ["alinux2023", "ubuntu2204", "ubuntu2404"]
  storage:
    test_fsx_lustre.py::test_multi_az_fsx:
      dimensions:
        - regions: ["eu-west-2"]
          instances: {{ common.INSTANCES_DEFAULT_X86 }}
          oss: [{{ LUSTRE_OS_X86_1 }}]
          schedulers: ["slurm"]
  patching:
    test_patching.py::test_patching_cluster:
      dimensions:
        - regions: [{{ g4dn_8xlarge_CAPACITY_RESERVATION_4_INSTANCES_2_HOURS_NOPG_ubuntu2404 }}]
          instances: ["g4dn.8xlarge"]
          oss: ["ubuntu2204","ubuntu2404"]
          schedulers: ["slurm"]
          flags: ["patching:full-capped"]
        - regions: [{{ g4dn_8xlarge_CAPACITY_RESERVATION_4_INSTANCES_2_HOURS_NOPG_rhel9 }}]
          instances: ["g4dn.8xlarge"]
          oss: ["rhel8","rhel9"]
          schedulers: ["slurm"]
          flags: ["patching:full-capped"]
        - regions: [{{ g4dn_8xlarge_CAPACITY_RESERVATION_2_INSTANCES_2_HOURS_NOPG_alinux2023 }}]
          instances: ["g4dn.8xlarge"]
          oss: ["alinux2023"]
          schedulers: ["slurm"]
          flags: ["patching:full"]

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

…ules loaded before/after patching to prevent false alarms.

Now the test fails the kernel module verification if:
1. a mandatory kernel module is not loaded, OR
2. the module is not loaded (may be lazily loaded), but it exists and fails to load.

In this way we prevent the false alarms on kernel modules that legitimately do not exist anymore on the new kernel.
Before this fix, some kernel packages could escape the pinning. With this fix, none of them can.
@gmarciani gmarciani added skip-changelog-update Disables the check that enforces changelog updates in PRs 3.x Test labels Sep 4, 2026
@gmarciani gmarciani changed the title Wip/mgiacomo/3161/integ tests stabilize 0904 1 [Test] Porting test stabilization fixes for 3.16.1 Sep 4, 2026
…risk of ICEs and instance types not being supported.
…ilure message to facilitate troubleshooting.
…and disable unattended upgrades.

 This is meant to prevent failures in Proxy and ProxyClient bootstrap caused by boot-time upgrades holding the dpkg lock.
@gmarciani
gmarciani force-pushed the wip/mgiacomo/3161/integ-tests-stabilize-0904-1 branch from accf787 to 86bde88 Compare September 4, 2026 16:09
…-connect-timeout 5 --max-time 15) to explicitly fail on unresponsive endpoints, consistently for all the tested endpoints.
…nst network glitches.

The client now verifies only what it was supposed to verify: that every expected endpoint is reachable through the proxy. IF the endpoint is temporarily unresponsive, that does not matter.
This pinning is needed because previous versions require the paid Extended Update Support (EUS) repos to be configured to install required packages, such as kernel-devel.
@gmarciani gmarciani changed the title [Test] Porting test stabilization fixes for 3.16.1 [Test] Test stabilization fixes for 3.16.1: patchibng, build-image no-internet, multi_az_fsx Sep 4, 2026
…networking setup.

Before this change, only the default AZ was honoring the allowlisting, so other components such as the ODCR stack had chances to select unexpected AZs.

With this change, by default all the region's AZs are usable, but if the region has an explicit allowlisting, only those allowlisted AZs will be considered.

As part of this change we set the allowlisting for eu-west-2, where we exclude eu-west-2d which is a new AZ that does not support our default instance type c5.xlarge.
@gmarciani
gmarciani force-pushed the wip/mgiacomo/3161/integ-tests-stabilize-0904-1 branch 2 times, most recently from 05b6d5f to 25590e8 Compare September 4, 2026 22:10
This fix is needed the shared test VPC is only built in the allowlisted AZs, so a
reservation placed outside them would pin the test to an AZ with no matching subnet.
…he test honor the flag to enable/disable Lustre installation.

As a result, we disable Lustre installation on ubuntu, the same way we do for test_build_image.
@gmarciani
gmarciani force-pushed the wip/mgiacomo/3161/integ-tests-stabilize-0904-1 branch from 25590e8 to afc67c5 Compare September 8, 2026 15:13
@gmarciani
gmarciani marked this pull request as ready for review September 8, 2026 15:13
@gmarciani
gmarciani requested review from a team as code owners September 8, 2026 15:13
@codecov

codecov Bot commented Sep 8, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (integ-tests-3.16.1@6cce701). Learn more about missing BASE report.

Additional details and impacted files
@@                  Coverage Diff                  @@
##             integ-tests-3.16.1    #7609   +/-   ##
=====================================================
  Coverage                      ?   89.94%           
=====================================================
  Files                         ?      180           
  Lines                         ?    16255           
  Branches                      ?        0           
=====================================================
  Hits                          ?    14621           
  Misses                        ?     1634           
  Partials                      ?        0           
Flag Coverage Δ
unittests 89.94% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@gmarciani gmarciani changed the title [Test] Test stabilization fixes for 3.16.1: patchibng, build-image no-internet, multi_az_fsx [Test] Test stabilization fixes for 3.16.1: patching, build-image no-internet, multi_az_fsx and AZ allowlisting Sep 8, 2026
@gmarciani
gmarciani merged commit 65fbae0 into aws:integ-tests-3.16.1 Sep 8, 2026
24 checks passed
@gmarciani
gmarciani deleted the wip/mgiacomo/3161/integ-tests-stabilize-0904-1 branch September 8, 2026 18:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

3.x skip-changelog-update Disables the check that enforces changelog updates in PRs Test

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants