Add topolvm e2e test #154
Add topolvm e2e test #154
Conversation
WalkthroughThis PR adds end-to-end testing infrastructure for TopoLVM. It introduces a new test script that validates TopoLVM functionality by creating a test workload in Kubernetes, then integrates the test into the CI/CD pipeline alongside a new configuration patch for lvmd settings. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes
Suggested reviewers
Pre-merge checks and finishing touches❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
📜 Recent review detailsConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro 📒 Files selected for processing (5)
🚧 Files skipped from review as they are similar to previous changes (3)
🔇 Additional comments (3)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (3)
src/topolvm/tests/createWorkloads.sh (1)
62-66: Simplify the pod-name check syntax.The parameter expansion
${podName=}is non-idiomatic; use"${podName}"for clarity since the variable is already initialized.-while [[ "${podName=}" == "" && ${iter} -gt 0 ]]; do +while [[ "${podName}" == "" && ${iter} -gt 0 ]]; dosrc/cluster_manager.sh (2)
389-390: Add defensive check for /etc/hosts.Line 389 assumes
/etc/hostsexists. While standard on most systems, verify it exists before copying to avoid failures in minimal/container environments.+if [ ! -f /etc/hosts ]; then + echo "ERROR: /etc/hosts not found" >&2 + exit 1 +fi cp /etc/hosts "${workdir}/hosts"
419-419: Remove redundant cleanup.Line 419 is unreachable; trap at line 378 already removes
${workdir}on exit.fi fi - rm -rf "${workdir}" }
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (5)
Makefile(5 hunks)src/cluster_manager.sh(5 hunks)src/topolvm/assets/kustomization.yaml(1 hunks)src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml(1 hunks)src/topolvm/tests/createWorkloads.sh(1 hunks)
🧰 Additional context used
🧠 Learnings (2)
📚 Learning: 2025-10-29T07:57:10.203Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/cluster_manager.sh:327-327
Timestamp: 2025-10-29T07:57:10.203Z
Learning: In src/cluster_manager.sh, the file `/var/lib/microshift/resources/kubeadmin/kubeconfig` (without a node name) is guaranteed to exist if MicroShift has started and is ready, as it is used internally. This generic kubeconfig path is appropriate for cluster-wide operations like querying cluster status, while node-specific paths like `/var/lib/microshift/resources/kubeadmin/${node_name}/kubeconfig` are used for node-specific operations like joining nodes.
Applied to files:
src/topolvm/tests/createWorkloads.shsrc/cluster_manager.sh
📚 Learning: 2025-10-29T07:41:49.737Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/cluster_manager.sh:112-128
Timestamp: 2025-10-29T07:41:49.737Z
Learning: In src/cluster_manager.sh, the file `/var/lib/microshift/resources/kubeadmin/$node_name/kubeconfig` is guaranteed to exist as part of the MicroShift startup process. If MicroShift is ready, this kubeconfig file will be present, so validation checks before copying are not required in the `_join_node` function.
Applied to files:
src/topolvm/tests/createWorkloads.shMakefilesrc/cluster_manager.sh
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (4)
- GitHub Check: centos10-bootc
- GitHub Check: quick-rpm-and-clean (ubuntu-24.04-arm)
- GitHub Check: quick-start-and-clean (ubuntu-24.04)
- GitHub Check: quick-start-and-clean (ubuntu-24.04-arm)
🔇 Additional comments (3)
Makefile (1)
156-159: LGTM.The env target correctly delegates to cluster_env() with sensible defaults: empty pod_name triggers default to NODE_BASE_NAME}1, and empty CMD triggers interactive shell. Documentation is clear.
src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml (1)
1-8: LGTM.ConfigMap properly aligns with cluster configuration (volume-group: myvg1 matches VG_NAME). YAML formatting and Kustomize integration are correct.
src/topolvm/assets/kustomization.yaml (1)
9-9: LGTM.Patch path is correctly formatted and appended. No ordering conflicts with existing patches.
eslutsky
force-pushed
the
add_topolvm_test
branch
from
cf92b88 to
562bf88
Compare
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
.github/actions/build/action.yaml (1)
120-121: Remove unnecessary blank lines for consistency.Lines 120–121 introduce extra whitespace that may be unintentional. Consider removing excess blank lines to maintain consistent spacing with the rest of the action file.
make env "CMD=bash ./src/topolvm/tests/createWorkloads.sh" - - for i in $(seq 2 ${{ inputs.node-count }}); do
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (4)
.github/actions/build/action.yaml(2 hunks)src/topolvm/assets/kustomization.yaml(1 hunks)src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml(1 hunks)src/topolvm/tests/createWorkloads.sh(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml
- src/topolvm/tests/createWorkloads.sh
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (11)
- GitHub Check: quick-rpm-and-clean (ubuntu-24.04)
- GitHub Check: quick-rpm-and-clean (ubuntu-24.04-arm)
- GitHub Check: quick-start-and-clean (ubuntu-24.04)
- GitHub Check: quick-start-and-clean (ubuntu-24.04-arm)
- GitHub Check: isolated-network (ovnk, ubuntu-24.04, 1)
- GitHub Check: fedora-bootc (ubuntu-24.04-arm)
- GitHub Check: fedora-bootc (ubuntu-24.04)
- GitHub Check: ubuntu-rpm2deb
- GitHub Check: isolated-network (kindnet, ubuntu-24.04, 0)
- GitHub Check: centos10-bootc
- GitHub Check: centos9-bootc
🔇 Additional comments (2)
src/topolvm/assets/kustomization.yaml (1)
9-9: Verify the patch file exists atsrc/topolvm/assets/topolvm_configmap-lvmd_patch.yamland contains valid YAML.The patch reference syntax is correct and follows the existing pattern in the kustomization file, but confirm the referenced patch file exists with valid Kustomize patch content.
.github/actions/build/action.yaml (1)
107-108: Verify that the test script andmake envtarget exist.The TopoLVM test is properly positioned after basic MicroShift validation but before node scaling. However, confirm that:
- The script
src/topolvm/tests/createWorkloads.shexists and is executable- The
make envtarget supports theCMDparameter as expected
eslutsky
force-pushed
the
add_topolvm_test
branch
from
1e37e13 to
20ccb40
Compare
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
.github/actions/build/action.yaml (1)
120-121: Remove extra blank lines.Lines 120–121 contain only whitespace and should be removed to maintain clean formatting.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
.github/actions/build/action.yaml(3 hunks).github/actions/prebuild/action.yaml(1 hunks)
🔇 Additional comments (3)
.github/actions/prebuild/action.yaml (1)
36-36: Verifyawkprerequisite is actually used.The addition of
awkaligns with the PR's new test and env functionality, but confirm it's actively used in the scripts or Makefile targets added in this PR (especially in the new cluster_env implementation or test scripts)..github/actions/build/action.yaml (2)
90-90: VerifyEXPOSE_KUBEAPI_PORTis properly handled in Makefile andcluster_manager.sh.The variable is set but ensure the Makefile
runtarget actually consumes it and correctly propagates it downstream, as indicated in the PR objectives.
107-108: Verify themake envtarget and test script exist and work as expected.This step depends on:
- A
make envtarget in the Makefile that accepts aCMDparameter- The script at
./src/topolvm/tests/createWorkloads.shexists and is executable- The script will appropriately fail if TopoLVM issues occur
Confirm these prerequisites are in place and functional.
eslutsky
force-pushed
the
add_topolvm_test
branch
2 times, most recently
from
a6e6303 to
aeb5770
Compare
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (2)
src/cluster_manager.sh (2)
381-392: Simplify pod name / command disambiguation.Lines 381–387 implicitly treat the first argument as a pod name if a container with that name exists; otherwise, it's treated as part of the command. This works but is non-obvious. Consider an explicit check or clearer logic:
- if [ -z "${pod_name}" ]; then - pod_name="${NODE_BASE_NAME}1" - elif ! _is_container_created "${pod_name}" 2>/dev/null; then - # First argument is not a container, treat it as part of the command - command="${pod_name} ${command}" - pod_name="${NODE_BASE_NAME}1" - fi + # If no pod_name provided or it's not a container, use default and treat first arg as command + if [ -z "${pod_name}" ] || ! _is_container_created "${pod_name}" 2>/dev/null; then + if [ -z "${pod_name}" ]; then + pod_name="${NODE_BASE_NAME}1" + else + command="${pod_name} ${command}" + pod_name="${NODE_BASE_NAME}1" + fi + fiOr document this behavior with an inline comment to aid future readers.
409-419: Dead code path afterexec bash.Line 418 (
rm -rf "${workdir}") is unreachable when the interactive shell is invoked (line 416:exec bash -i). The trap at line 397 already handles cleanup, so this redundancy can be removed. For the command-execution path (line 412), the trap still fires on exit.if [ -n "${command}" ]; then # Execute the command and exit echo "Executing command in environment with kubeconfig..." sh -c "${command}" else # Start interactive shell echo "Starting shell environment with kubeconfig..." exec bash -i fi - rm -rf "${workdir}"
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
.github/actions/build/action.yaml(3 hunks).github/actions/prebuild/action.yaml(1 hunks)src/cluster_manager.sh(5 hunks)src/topolvm/assets/kustomization.yaml(1 hunks)src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml(1 hunks)src/topolvm/tests/createWorkloads.sh(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (3)
- .github/actions/build/action.yaml
- src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml
- src/topolvm/tests/createWorkloads.sh
🧰 Additional context used
🧠 Learnings (3)
📚 Learning: 2025-10-23T11:55:38.018Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/image/postinstall.sh:52-54
Timestamp: 2025-10-23T11:55:38.018Z
Learning: In MicroShift multinode deployments at src/image/postinstall.sh, etcd ports 2379 and 2380 are exposed in the public firewall zone because nodes may be deployed across different networks and need to communicate with each other for cluster formation and operation.
Applied to files:
src/cluster_manager.sh
📚 Learning: 2025-10-29T07:57:10.203Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/cluster_manager.sh:327-327
Timestamp: 2025-10-29T07:57:10.203Z
Learning: In src/cluster_manager.sh, the file `/var/lib/microshift/resources/kubeadmin/kubeconfig` (without a node name) is guaranteed to exist if MicroShift has started and is ready, as it is used internally. This generic kubeconfig path is appropriate for cluster-wide operations like querying cluster status, while node-specific paths like `/var/lib/microshift/resources/kubeadmin/${node_name}/kubeconfig` are used for node-specific operations like joining nodes.
Applied to files:
src/cluster_manager.sh
📚 Learning: 2025-10-29T07:41:49.737Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/cluster_manager.sh:112-128
Timestamp: 2025-10-29T07:41:49.737Z
Learning: In src/cluster_manager.sh, the file `/var/lib/microshift/resources/kubeadmin/$node_name/kubeconfig` is guaranteed to exist as part of the MicroShift startup process. If MicroShift is ready, this kubeconfig file will be present, so validation checks before copying are not required in the `_join_node` function.
Applied to files:
src/cluster_manager.sh
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (9)
- GitHub Check: isolated-network (kindnet, ubuntu-24.04, 0)
- GitHub Check: fedora-bootc (ubuntu-24.04-arm)
- GitHub Check: fedora-bootc (ubuntu-24.04)
- GitHub Check: centos9-bootc
- GitHub Check: quick-start-and-clean (ubuntu-24.04)
- GitHub Check: quick-rpm-and-clean (ubuntu-24.04)
- GitHub Check: quick-start-and-clean (ubuntu-24.04-arm)
- GitHub Check: quick-rpm-and-clean (ubuntu-24.04-arm)
- GitHub Check: centos10-bootc
🔇 Additional comments (7)
src/topolvm/assets/kustomization.yaml (1)
9-9: Verify the patch file exists and is properly formatted.The patch entry addition looks syntactically correct and maintains consistency with existing entries. However, ensure that
topolvm_configmap-lvmd_patch.yamlis present in the same directory and contains a valid Kubernetes manifest..github/actions/prebuild/action.yaml (1)
36-36: Dependency addition looks good.The
gawkpackage aligns with the_get_ip_address()function incluster_manager.sh(line 81) that uses awk for subnet calculations.src/cluster_manager.sh (5)
135-142:_get_hostname()implementation looks solid.Error handling and early exit on failure are appropriate. The function follows the established pattern in the script.
122-130: Config injection flow is clear.The conditional logic around
EXPOSE_KUBEAPI_PORT, file creation, copy, restart, and cleanup is well-structured. The use of_get_hostnameensures consistency when subjectAltNames are injected.
106-109: Port mapping logic is correct.The conditional injection of
-p 6443:6443based onEXPOSE_KUBEAPI_PORTfollows the established pattern for optional flags (e.g.,network_opts,vol_opts).
455-468: CLI routing and usage update are consistent.The
envcommand is properly added to the main dispatch and reflected in the usage message. The pass-through of arguments via"$@"is correct.
16-16:EXPOSE_KUBEAPI_PORTconfiguration is well-placed.New variable follows the existing pattern (default value, environment override). No concerns.
eslutsky
force-pushed
the
add_topolvm_test
branch
from
aeb5770 to
e0bb2db
Compare
There was a problem hiding this comment.
Actionable comments posted: 2
♻️ Duplicate comments (1)
src/cluster_manager.sh (1)
373-406: Fix kubeconfig path: use node name instead of hostname. Line 393 attempts to copy kubeconfig using the host's FQDN from_get_hostname(), but kubeconfigs inside the container are organized by node name (e.g.,microshift-okd-1), not the host's FQDN. This will cause thepodman cpcommand to fail. Additionally, add validation that the copy succeeds before attempting to use the kubeconfig.Apply this diff to use the correct kubeconfig path:
cluster_env() { local command="${1:-}" # Set first_container from the first value of containers array local -r first_container=$(_get_running_containers | head -n1) if [ -z "${first_container}" ]; then echo "ERROR: No running cluster containers found." >&2 exit 1 fi # Verify that ${API_SERVER_PORT} is open for connections on the host if ! sudo ss -ltn "( sport = :${API_SERVER_PORT} )" | grep -q ":${API_SERVER_PORT}"; then echo "ERROR: API server port ${API_SERVER_PORT} is closed, make sure EXPOSE_KUBEAPI_PORT is set to 1." >&2 exit 1 fi local -r workdir=$(mktemp -d /tmp/kubeconfig-XXXXXX) # shellcheck disable=SC2064 trap "rm -rf '${workdir}'" RETURN echo "Copying kubeconfig from ${first_container}..." - sudo podman cp "${first_container}:/var/lib/microshift/resources/kubeadmin/$(_get_hostname)/kubeconfig" "${workdir}/kubeconfig" + sudo podman cp "${first_container}:/var/lib/microshift/resources/kubeadmin/kubeconfig" "${workdir}/kubeconfig" sudo chown "$(whoami):$(whoami)" "${workdir}/kubeconfig" export KUBECONFIG="${workdir}/kubeconfig" if [ -n "${command}" ]; then # Execute the command and exit echo "Executing command in environment with kubeconfig..." sh -c "${command}" else # Start interactive shell echo "Starting shell environment with kubeconfig..." bash -li fi }This uses the generic kubeconfig path that is guaranteed to exist when MicroShift is running, as documented in learnings about
cluster_manager.sh.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (7)
.github/actions/build/action.yaml(2 hunks)Makefile(5 hunks)src/cluster_manager.sh(6 hunks)src/topolvm/assets/kustomization.yaml(1 hunks)src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml(1 hunks)src/topolvm/generate_manifests.sh(1 hunks)src/topolvm/tests/createWorkloads.sh(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- src/topolvm/assets/topolvm_configmap-lvmd_patch.yaml
- src/topolvm/tests/createWorkloads.sh
🧰 Additional context used
🧠 Learnings (6)
📓 Common learnings
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/image/postinstall.sh:52-54
Timestamp: 2025-10-23T11:55:38.018Z
Learning: In MicroShift multinode deployments at src/image/postinstall.sh, etcd ports 2379 and 2380 are exposed in the public firewall zone because nodes may be deployed across different networks and need to communicate with each other for cluster formation and operation.
📚 Learning: 2025-10-29T07:57:10.203Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/cluster_manager.sh:327-327
Timestamp: 2025-10-29T07:57:10.203Z
Learning: In src/cluster_manager.sh, the file `/var/lib/microshift/resources/kubeadmin/kubeconfig` (without a node name) is guaranteed to exist if MicroShift has started and is ready, as it is used internally. This generic kubeconfig path is appropriate for cluster-wide operations like querying cluster status, while node-specific paths like `/var/lib/microshift/resources/kubeadmin/${node_name}/kubeconfig` are used for node-specific operations like joining nodes.
Applied to files:
src/cluster_manager.sh
📚 Learning: 2025-10-29T07:41:49.737Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/cluster_manager.sh:112-128
Timestamp: 2025-10-29T07:41:49.737Z
Learning: In src/cluster_manager.sh, the file `/var/lib/microshift/resources/kubeadmin/$node_name/kubeconfig` is guaranteed to exist as part of the MicroShift startup process. If MicroShift is ready, this kubeconfig file will be present, so validation checks before copying are not required in the `_join_node` function.
Applied to files:
src/cluster_manager.shMakefile
📚 Learning: 2025-10-29T07:44:21.893Z
Learnt from: pacevedom
Repo: microshift-io/microshift PR: 103
File: src/cluster_manager.sh:77-81
Timestamp: 2025-10-29T07:44:21.893Z
Learning: MicroShift is not designed to support clusters with 245+ nodes running as containers on the same system. IP address allocation in cluster_manager.sh does not need overflow validation for this reason.
Applied to files:
src/cluster_manager.sh
📚 Learning: 2025-10-17T10:31:57.408Z
Learnt from: ggiguash
Repo: microshift-io/microshift PR: 57
File: src/debian/install.sh:12-0
Timestamp: 2025-10-17T10:31:57.408Z
Learning: In the MicroShift project, the Ubuntu version in src/debian/install.sh is intentionally hardcoded to "xUbuntu_20.04" because CRI-O builds are not available for all the latest Ubuntu versions. This should not be changed to dynamic detection.
Applied to files:
src/cluster_manager.sh
📚 Learning: 2025-10-17T10:53:08.461Z
Learnt from: ggiguash
Repo: microshift-io/microshift PR: 57
File: .github/workflows/builders.yaml:75-84
Timestamp: 2025-10-17T10:53:08.461Z
Learning: In the microshift repository, the RPM-to-Debian conversion workflow follows this pattern: `convert.sh` receives an RPM directory as input and outputs .deb files to a `deb/` subdirectory within that directory. The `install.sh` script expects to receive the parent RPM directory (not the deb subdirectory) and uses recursive `find` commands to locate the .deb files in subdirectories.
Applied to files:
Makefile
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (11)
- GitHub Check: quick-rpm-and-clean (ubuntu-24.04-arm)
- GitHub Check: quick-rpm-and-clean (ubuntu-24.04)
- GitHub Check: quick-start-and-clean (ubuntu-24.04)
- GitHub Check: quick-start-and-clean (ubuntu-24.04-arm)
- GitHub Check: fedora-bootc (ubuntu-24.04-arm)
- GitHub Check: ubuntu-rpm2deb
- GitHub Check: fedora-bootc (ubuntu-24.04)
- GitHub Check: isolated-network (ovnk, ubuntu-24.04, 1)
- GitHub Check: centos10-bootc
- GitHub Check: centos9-bootc
- GitHub Check: isolated-network (kindnet, ubuntu-24.04, 0)
🔇 Additional comments (4)
src/topolvm/generate_manifests.sh (1)
95-104: Verify the lvmd.yaml ConfigMap data format.The ConfigMap patch uses escaped newlines (
\n) in a heredoc with single quotes (line 96), which means they'll be literal\nstrings rather than actual newlines. Confirm this is the intended format for lvmd configuration consumption.Check how the lvmd service consumes this ConfigMap data and whether it expects literal
\nor actual newlines in the YAML structure.Makefile (1)
180-182: LGTM. Theenvtarget correctly depends onrun-readyto ensure the cluster is ready before exposing the environment.src/cluster_manager.sh (1)
86-93: Good error handling for hostname resolution. The_get_hostname()function validates the FQDN and exits gracefully if unavailable..github/actions/build/action.yaml (1)
107-108: The TopoLVM test script atsrc/topolvm/tests/createWorkloads.shexists and is properly referenced in the workflow.
| local port_opts="" | ||
| local mount_opts="" | ||
| if [ "${EXPOSE_KUBEAPI_PORT}" = "1" ]; then | ||
| port_opts="-p ${API_SERVER_PORT}:${API_SERVER_PORT}" | ||
| echo -e "apiServer:\n subjectAltNames:\n - $(_get_hostname)" | sudo tee "${EXTRA_CONFIG}" >/dev/null | ||
| mount_opts="--volume ${EXTRA_CONFIG}:/etc/microshift/config.d/api_server.yaml:ro" | ||
| fi |
There was a problem hiding this comment.
Add validation for API config file creation. When EXPOSE_KUBEAPI_PORT=1, the code writes to ${EXTRA_CONFIG} without validating the directory exists or the write succeeds.
Apply this diff to ensure the directory exists and handle failures:
local port_opts=""
local mount_opts=""
if [ "${EXPOSE_KUBEAPI_PORT}" = "1" ]; then
port_opts="-p ${API_SERVER_PORT}:${API_SERVER_PORT}"
+ sudo mkdir -p "$(dirname "${EXTRA_CONFIG}")"
+ if ! echo -e "apiServer:\n subjectAltNames:\n - $(_get_hostname)" | sudo tee "${EXTRA_CONFIG}" >/dev/null; then
+ echo "ERROR: Failed to write API server configuration to ${EXTRA_CONFIG}" >&2
+ exit 1
+ fi
- echo -e "apiServer:\n subjectAltNames:\n - $(_get_hostname)" | sudo tee "${EXTRA_CONFIG}" >/dev/null
mount_opts="--volume ${EXTRA_CONFIG}:/etc/microshift/config.d/api_server.yaml:ro"
fi🤖 Prompt for AI Agents
In src/cluster_manager.sh around lines 117 to 123, when EXPOSE_KUBEAPI_PORT=1
the script writes an apiServer YAML to ${EXTRA_CONFIG} without ensuring the
parent directory exists or that the write succeeded; update the code to create
the parent directory (mkdir -p) for ${EXTRA_CONFIG} before writing, check the
tee/write command exit status and on failure emit an error message to stderr (or
the script logger) and exit non‑zero, and only set mount_opts after verifying
the write was successful.
070827b to
3e924b0
Compare
Signed-off-by: Evgeny Slutsky <eslutsky@redhat.com>
Signed-off-by: Evgeny Slutsky <eslutsky@redhat.com>
eslutsky
force-pushed
the
add_topolvm_test
branch
from
3e924b0 to
3ae6fab
Compare
2 participants
Resolves #150
Depends on #153
This tests can be executed with:
make env "CMD=bash ./src/topolvm/tests/createWorkloads.sh"Summary by CodeRabbit
Release Notes
New Features
Tests
✏️ Tip: You can customize this high-level summary in your review settings.