Feature: Add multinode support #103
Conversation
WalkthroughMakefile targets were reworked to delegate cluster lifecycle to a new Changes
Sequence Diagram(s)sequenceDiagram
autonumber
participant User
participant Makefile
participant ClusterMgr as cluster_manager.sh
participant Podman
participant TopoLVM
User->>Makefile: make run
Makefile->>ClusterMgr: ./src/cluster_manager.sh create
alt TopoLVM requested
ClusterMgr->>TopoLVM: topolvm-create
end
ClusterMgr->>Podman: create network, provision node-1, start microshift
ClusterMgr-->>Makefile: exit/status
rect rgb(240,250,240)
Note right of ClusterMgr: multi-node flow (CI or user)
User->>Makefile: make add-node / CI loop
Makefile->>ClusterMgr: ./src/cluster_manager.sh add-node
ClusterMgr->>Podman: provision node-N, join cluster
ClusterMgr->>ClusterMgr: run-ready / run-healthy checks
ClusterMgr-->>Makefile: exit/status
end
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes
Possibly related PRs
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 (8)
🚧 Files skipped from review as they are similar to previous changes (3)
🧰 Additional context used🧠 Learnings (4)📚 Learning: 2025-10-23T11:55:38.018ZApplied to files:
📚 Learning: 2025-10-29T07:41:49.717ZApplied to files:
📚 Learning: 2025-10-29T07:57:10.176ZApplied to files:
📚 Learning: 2025-10-17T07:44:32.742ZApplied to files:
🪛 markdownlint-cli2 (0.18.1)docs/run.md101-101: Table pipe style (MD055, table-pipe-style) 102-102: Table pipe style (MD055, table-pipe-style) 145-145: Multiple headings with the same content (MD024, no-duplicate-heading) ⏰ 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). (7)
🔇 Additional comments (9)
Comment |
|
@CodeRabbit review the changes |
|
@pacevedom: I'll review the changes in this PR for adding multinode support. I'll analyze the code changes and provide feedback shortly. ✅ Actions performedReview triggered.
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (2)
Makefile (1)
148-150: Consider adding a confirmation message for consistency.The
multinode-deletetarget successfully cleans up the cluster, but unlikemultinode-create, it doesn't echo a confirmation message after completion.Apply this diff to add a confirmation message:
.PHONY: multinode-delete multinode-delete: @USHIFT_IMAGE="${USHIFT_IMAGE}" LVM_DISK=${LVM_DISK} LVM_VOLSIZE=${LVM_VOLSIZE} VG_NAME=${VG_NAME} ./src/multinode/multinode.sh delete + @echo "Cluster has been deleted"src/multinode/multinode.sh (1)
251-273: Consider using consistent grep syntax for maintainability.The function correctly cleans up all cluster resources. However, Line 252 uses BRE syntax (
[0-9]\+) while other commands in the script (lines 191, 217, 225, 238) use ERE syntax with the-Eflag ([0-9]+). While both work, consistency improves readability.Apply this diff for consistency:
- containers=$(sudo podman ps -a --format '{{.Names}}' | grep "^${NODE_BASE_NAME}[0-9]\+") || true + containers=$(sudo podman ps -a --format '{{.Names}}' | grep -E "^${NODE_BASE_NAME}[0-9]+") || true
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (7)
Makefile(2 hunks)packaging/microshift-builder.Containerfile(1 hunks)packaging/microshift-runner.Containerfile(2 hunks)src/image/postinstall.sh(1 hunks)src/multinode/dropins/microshift-multinode.conf(1 hunks)src/multinode/multinode.sh(1 hunks)src/multinode/multinode.spec(1 hunks)
🧰 Additional context used
🧠 Learnings (1)
📚 Learning: 2025-10-17T10:53:08.437Z
Learnt from: ggiguash
PR: microshift-io/microshift#57
File: .github/workflows/builders.yaml:75-84
Timestamp: 2025-10-17T10:53:08.437Z
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:
packaging/microshift-builder.Containerfile
🪛 RuboCop (1.81.1)
src/multinode/multinode.spec
[fatal] 4-4: unterminated string meets end of file
(Using Ruby 2.7 parser; configure using TargetRubyVersion parameter, under AllCops)
(Lint/Syntax)
🔇 Additional comments (20)
src/image/postinstall.sh (1)
56-60: System tuning for multinode workloads looks good.The inotify limit increases (max_user_watches and max_user_instances) are appropriate for a Kubernetes cluster with multiple nodes. These settings are standard to prevent "too many open files" errors from kubelet and container runtimes. Configuration is idempotent and follows standard sysctl naming conventions.
src/multinode/dropins/microshift-multinode.conf (1)
1-4: LGTM! Systemd drop-in is correctly structured.The drop-in follows the standard pattern of clearing the inherited
ExecStartbefore setting a new one, which is necessary to avoid systemd attempting to run both commands.packaging/microshift-builder.Containerfile (1)
64-67: LGTM! Multinode RPM build follows established patterns.The multinode RPM build sequence correctly mirrors the TopoLVM and Kindnet patterns, ensuring consistency in the build process.
packaging/microshift-runner.Containerfile (2)
21-21: LGTM! Environment variable declaration follows the established pattern.The
WITH_MULTINODEflag is properly aligned with other feature flags (WITH_KINDNET,WITH_TOPOLVM,WITH_OLM) and correctly defaults to0(disabled) for backward compatibility.
38-40: LGTM! Conditional installation is correctly implemented.The multinode package installation follows the same pattern as other optional components. Note that unlike other components,
microshift-multinodedoesn't install a separate release-info package, which appears intentional.Makefile (2)
84-84: LGTM! Multinode flag is properly propagated to the image build.The
WITH_MULTINODEenvironment variable is correctly passed to the container build, enabling conditional multinode package installation in the runner image.
143-146: LGTM! The multinode-create target is well-structured.The target properly passes all necessary environment variables to the multinode script and provides helpful user feedback.
src/multinode/multinode.spec (1)
1-41: LGTM! RPM spec file is correctly structured.The spec file properly defines the multinode subpackage with appropriate dependencies and file installations. The path references align with the builder Containerfile's copy operations.
Note: The RuboCop "unterminated string" error is a false positive—this is an RPM spec file, not Ruby code.
src/multinode/multinode.sh (12)
1-14: LGTM! Script initialization is properly configured.The script uses appropriate error handling with
set -euo pipefailand provides sensible defaults for all configuration variables while allowing environment overrides.
16-26: LGTM! TopoLVM backend creation is correctly implemented.The function properly handles the reuse case and creates the LVM backend with appropriate commands. Error handling is provided by the script's
set -econfiguration.
28-47: LGTM! Kubeconfig handling functions are correctly implemented.The wait timeout of 3000 seconds (50 minutes) is generous for initial cluster bootstrap, and the copy function cleanly extracts the kubeconfig.
72-88: LGTM! Node joining logic is correctly implemented.The function properly orchestrates the node join process by copying the primary kubeconfig and executing the
microshift add-nodecommand with appropriate service stops.
90-101: LGTM! Node readiness check is properly implemented.The function waits for the greenboot healthcheck with a reasonable timeout and handles errors gracefully.
103-120: LGTM! Network and cluster existence checks are correctly implemented.Both functions properly use Podman commands to check and create resources as needed.
122-137: Verify IP address allocation doesn't conflict with Podman's gateway.The functions correctly extract the subnet and calculate node IP addresses by setting the fourth octet to
node_id + 10(e.g., .11, .12, .13). This allocation strategy should avoid conflicts with Podman's typical gateway at .1, but please verify that Podman's default IPAM doesn't use IPs in the .10+ range for other purposes.
139-180: LGTM! Cluster creation logic is well-structured.The function properly validates inputs, prevents duplicate clusters, and correctly orchestrates node creation with appropriate differentiation between the first node (which must be ready before others join) and subsequent nodes.
182-214: LGTM! Node addition logic correctly extends existing clusters.The function properly determines the last node ID and adds new nodes sequentially with appropriate joining logic.
216-230: LGTM! Start and stop commands are correctly implemented.Both commands properly enumerate containers and handle them appropriately. The
--time 0flag ensures immediate container shutdown.
232-249: LGTM! Status command provides useful cluster information.The function properly checks cluster state and uses
oc get nodesto display detailed node information when the cluster is running.
275-319: LGTM! Usage documentation and main entrypoint are well-structured.The usage help is comprehensive, documenting all commands and environment variables. The main function provides useful visibility into the runtime configuration before executing commands.
pacevedom
force-pushed
the
80
branch
2 times, most recently
from
d5d3eef to
9b4cff9
Compare
|
@coderabbitai review this change |
|
@pacevedom: I'll review the changes in this PR. Let me analyze the multinode support implementation across all the modified files. ✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (2)
src/multinode/multinode.sh (2)
16-26: Consider verifying VG exists when LVM_DISK file exists.If
LVM_DISKexists but the loop device or volume group don't (e.g., after a system reboot), subsequent operations may fail. Consider checking VG existence or recreating the loop device when the file is found.
268-270: Fix indentation to use spaces consistently.Lines 268-270 use tabs while the rest of the file uses spaces. Maintain consistent indentation.
- sudo vgremove -y "${VG_NAME}" || true - local -r device_name="$(sudo losetup -j "${LVM_DISK}" | cut -d: -f1)" - [ -n "${device_name}" ] && sudo losetup -d "${device_name}" || true + sudo vgremove -y "${VG_NAME}" || true + local -r device_name="$(sudo losetup -j "${LVM_DISK}" | cut -d: -f1)" + [ -n "${device_name}" ] && sudo losetup -d "${device_name}" || true
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (7)
Makefile(2 hunks)packaging/microshift-builder.Containerfile(1 hunks)packaging/microshift-runner.Containerfile(2 hunks)src/multinode/dropins/microshift-multinode.conf(1 hunks)src/multinode/multinode.sh(1 hunks)src/multinode/multinode.spec(1 hunks)src/rpm/postinstall.sh(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (3)
- Makefile
- packaging/microshift-runner.Containerfile
- packaging/microshift-builder.Containerfile
🧰 Additional context used
🧠 Learnings (1)
📚 Learning: 2025-10-23T11:55:38.018Z
Learnt from: pacevedom
PR: microshift-io/microshift#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/rpm/postinstall.sh
🪛 RuboCop (1.81.1)
src/multinode/multinode.spec
[fatal] 4-4: unterminated string meets end of file
(Using Ruby 2.7 parser; configure using TargetRubyVersion parameter, under AllCops)
(Lint/Syntax)
⏰ 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). (3)
- GitHub Check: isolated-network-ovnk
- GitHub Check: isolated-network-kindnet
- GitHub Check: ubuntu-rpm2deb
🔇 Additional comments (8)
src/rpm/postinstall.sh (2)
43-46: Firewall rules already reviewed and justified.These firewall rules were discussed in previous reviews. The public zone exposure is intentional for multinode deployments across different networks.
Based on learnings.
47-51: LGTM!The inotify limits are appropriate for Kubernetes workloads and prevent resource exhaustion issues with cAdvisor and kubelet.
src/multinode/dropins/microshift-multinode.conf (1)
1-4: LGTM!The systemd drop-in correctly clears and replaces ExecStart to enable multinode mode. The comment helpfully explains the systemd requirement.
src/multinode/multinode.spec (1)
25-31: LGTM!The multinode subpackage definition is correct with appropriate version dependency and noarch specification.
src/multinode/multinode.sh (4)
1-15: LGTM!Proper error handling with
set -euo pipefailand well-defined environment variable defaults with override capability.
49-70: LGTM!The issues flagged in the previous review have been fixed:
vol_optsnow has proper local declaration, and${device}uses correct Bash syntax.
28-248: LGTM!Cluster management functions are well-structured with appropriate error handling and timeouts for multinode operations.
276-320: LGTM!Clear usage documentation and well-organized command dispatch with helpful environment summary.
|
@coderabbitai do a final round of review. Also update the PR description with a concise summary of this change. |
🧩 Analysis chain@pacevedom: I'll perform a final review of the changes and have updated the PR description with a concise summary of the multinode support implementation. 🏁 Script executed: Length of output: 502 🏁 Script executed: Length of output: 52 @pacevedom: I've successfully updated the PR description with a comprehensive summary of the multinode support changes. The description now covers:
I'm now triggering a final review of all the changes. ✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (3)
src/multinode/multinode.sh (2)
32-39: Add informative error message on kubeconfig wait timeout.When the wait times out, the function returns 1 with no indication of what failed, making debugging difficult.
wait_for_kubeconfig() { local -r name="${1}" local -r kubeconfig="/var/lib/microshift/resources/kubeadmin/${name}/kubeconfig" for _ in $(seq 600); do if sudo podman exec -i "${name}" /bin/test -f "${kubeconfig}"; then return 0 fi sleep 5 done + echo "ERROR: Timeout waiting for kubeconfig in container ${name}" >&2 return 1 }
248-248: Use variable for kubeconfig path in cmd_status.Line 248 hardcodes the kubeconfig path instead of using a configuration variable, making the code less maintainable and inconsistent with the rest of the script.
cmd_status() { if ! sudo podman container exists "${NODE_BASE_NAME}1"; then echo "Cluster is not initialized." exit 1 fi local -r running_containers=$(sudo podman ps --format '{{.Names}}' | grep -E "^${NODE_BASE_NAME}[0-9]+$" || true) if [ -z "${running_containers}" ]; then echo "Cluster is down. No cluster containers are currently running." return 0 fi local -r first_container=$(echo "${running_containers}" | head -n1) echo "Cluster is running." - sudo podman exec -i "${first_container}" oc --kubeconfig=/var/lib/microshift/resources/kubeadmin/kubeconfig get nodes -o wide + local -r kubeconfig="/var/lib/microshift/resources/kubeadmin/kubeconfig" + sudo podman exec -i "${first_container}" oc --kubeconfig="${kubeconfig}" get nodes -o wide exit $? }src/multinode/README.md (1)
209-211: Specify language identifier for fenced code block.Line 209 starts a code block without specifying the language, which impacts readability and syntax highlighting in rendered documentation.
-``` +```conf /etc/systemd/system/microshift.service.d/multinode.conf</blockquote></details> </blockquote></details> <details> <summary>📜 Review details</summary> **Configuration used**: Path: .coderabbit.yaml **Review profile**: CHILL **Plan**: Pro <details> <summary>📥 Commits</summary> Reviewing files that changed from the base of the PR and between 9b4cff95497f52e6245a4976fb709dd6e6bcd020 and 367cb99436862b955377de298434e76a55c4557b. </details> <details> <summary>📒 Files selected for processing (4)</summary> * `.github/actions/build/action.yaml` (3 hunks) * `.github/workflows/builders.yaml` (1 hunks) * `src/multinode/README.md` (1 hunks) * `src/multinode/multinode.sh` (1 hunks) </details> <details> <summary>🧰 Additional context used</summary> <details> <summary>🪛 LanguageTool</summary> <details> <summary>src/multinode/README.md</summary> [grammar] ~207-~207: Ensure spelling is correct Context: ... The multinode RPM installs the service dropin file that enables multinode mode: ``` ... (QB_NEW_EN_ORTHOGRAPHY_ERROR_IDS_1) --- [grammar] ~213-~213: Ensure spelling is correct Context: ...hift.service.d/multinode.conf ``` This dropin configures MicroShift to run with the `... (QB_NEW_EN_ORTHOGRAPHY_ERROR_IDS_1) </details> </details> <details> <summary>🪛 markdownlint-cli2 (0.18.1)</summary> <details> <summary>src/multinode/README.md</summary> 209-209: Fenced code blocks should have a language specified (MD040, fenced-code-language) </details> </details> </details> <details> <summary>⏰ 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). (7)</summary> * GitHub Check: isolated-network-ovnk * GitHub Check: isolated-network-kindnet * GitHub Check: multinode-bootc * GitHub Check: ubuntu-rpm2deb * GitHub Check: fedora-bootc * GitHub Check: centos9-bootc * GitHub Check: centos10-bootc </details> <details> <summary>🔇 Additional comments (3)</summary><blockquote> <details> <summary>src/multinode/multinode.sh (1)</summary><blockquote> `49-71`: **Critical issues from previous review have been fixed.** The function now correctly declares `vol_opts` as local (line 54) and uses proper Bash variable expansion `${device}` (line 56) instead of the previous Makefile-style `$${device}`. The shellcheck directive on line 59 is appropriate for the controlled vol_opts expansion. </blockquote></details> <details> <summary>.github/actions/build/action.yaml (1)</summary><blockquote> `120-128`: **Verify multinode testing strategy covers sufficient validation.** When `multinode=1`, the action skips the standard readiness checks (`run-ready`, `run-healthy`) and only runs `multinode-create` and `multinode-delete`. This means multinode builds don't validate that individual MicroShift instances are healthy during cluster lifecycle, only that the orchestration script succeeds. Consider whether additional validation is needed within the multinode test (e.g., checking node status after create, verifying etcd quorum). </blockquote></details> <details> <summary>.github/workflows/builders.yaml (1)</summary><blockquote> `119-135`: **Multinode workflow job is properly integrated.** The new `multinode-bootc` job follows the established pattern for variant builds, correctly passes the `multinode=1` flag to the build action, and runs on appropriate infrastructure. </blockquote></details> </blockquote></details> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
pacevedom
force-pushed
the
80
branch
2 times, most recently
from
8ecb276 to
13679bf
Compare
|
@coderabbitai have a quick review and update the PR description |
🧩 Analysis chain🏁 Script executed: Length of output: 2202 🏁 Script executed: Length of output: 52 @pacevedom: I've updated the PR description with a more concise summary while keeping all the essential information. The description now provides a cleaner overview of the multinode support implementation, including cluster management, build/runtime integration, infrastructure changes, packaging, and documentation. I'm now triggering a quick review of the changes. ✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 6
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (6)
.github/actions/build/action.yaml(2 hunks).github/actions/quick-start-clean/action.yaml(1 hunks).github/workflows/builders.yaml(1 hunks)Makefile(1 hunks)src/cluster_manager.sh(1 hunks)src/rpm/postinstall.sh(1 hunks)
🧰 Additional context used
🧠 Learnings (1)
📚 Learning: 2025-10-23T11:55:38.018Z
Learnt from: pacevedom
PR: microshift-io/microshift#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/rpm/postinstall.sh
🪛 actionlint (1.7.8)
.github/workflows/builders.yaml
134-134: input "multinode" is not defined in action "build-rpms-and-images" defined at "./.github/actions/build". available inputs are "bootc-image-tag", "bootc-image-url", "build", "isolated-network", "multi-node", "okd-version-tag", "ovnk-networking", "ushift-branch"
(action)
⏰ 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). (7)
- GitHub Check: fedora-bootc
- GitHub Check: isolated-network-ovnk
- GitHub Check: isolated-network-kindnet
- GitHub Check: multi-node-bootc
- GitHub Check: ubuntu-rpm2deb
- GitHub Check: centos9-bootc
- GitHub Check: centos10-bootc
🔇 Additional comments (8)
src/rpm/postinstall.sh (1)
43-52: ✓ Multinode connectivity and resource limits configured appropriately.The firewall rules and sysctl adjustments are well-placed. The comment clearly explains why etcd ports are exposed in the public zone for multinode clusters. Inotify limits align with kubelet/cAdvisor requirements in containerized environments.
src/cluster_manager.sh (5)
7-15: ✓ Configuration variables well-designed for environment overrides.Clean use of bash parameter expansion for configuration. Defaults align with typical single-node and multinode deployments.
17-151: ✓ Helper functions are logically sound and handle edge cases.LVM backend, network, and node provisioning helpers use defensive patterns (existence checks, error fallbacks, cleanup on failure). Node joining flow correctly sequences kubeconfig copy and microshift add-node execution.
191-230: ✓ Node addition orchestration correctly sequences setup, health checks, and join.Iteration to find the next node ID is correct. Pre-join health check ensures the primary node is stable before adding new replicas.
287-315: ✓ Readiness and health check logic is clear and guards against missing nodes.The distinction between microshift.service "running" (ready) and greenboot-healthcheck "exited" (healthy) is appropriate. Early exit if no containers found avoids spurious timeouts.
344-391: ✓ Main dispatch and directory guard are well-implemented.Case statement cleanly routes actions. The Makefile presence check prevents accidental runs from wrong directory, reducing debugging friction.
.github/actions/build/action.yaml (1)
33-37: ✓ Input definition is correct.The
multi-nodeinput with integer type and default value is properly declared.Makefile (1)
107-150: ✓ Cluster orchestration cleanly centralized through cluster_manager.sh delegation.Environment overrides are consistently passed through all targets. The polling loops for run-ready and run-healthy follow a sensible retry pattern. The transition from inline container logic to centralized script management is well-structured.
pacevedom
force-pushed
the
80
branch
5 times, most recently
from
b928205 to
69d4fb3
Compare
|
@coderabbitai have another round and update again the PR description to match the changes in the commits. |
🧩 Analysis chain🏁 Script executed: Length of output: 2237 🏁 Script executed: Length of output: 1330 🏁 Script executed: Length of output: 854 🏁 Script executed: Length of output: 88 @pacevedom: I've updated the PR description to accurately match the actual changes in the commits. The key corrections include: Corrected:
Actual changes documented:
The description now accurately reflects the 6 files changed in this PR. ✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 2
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
docs/run.md(3 hunks)
🧰 Additional context used
🪛 markdownlint-cli2 (0.18.1)
docs/run.md
101-101: Table pipe style
Expected: leading_and_trailing; Actual: leading_only; Missing trailing pipe
(MD055, table-pipe-style)
102-102: Table pipe style
Expected: leading_and_trailing; Actual: leading_only; Missing trailing pipe
(MD055, table-pipe-style)
145-145: Multiple headings with the same content
(MD024, no-duplicate-heading)
⏰ 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). (7)
- GitHub Check: ubuntu-rpm2deb
- GitHub Check: quick-start-and-clean
- GitHub Check: centos10-bootc
- GitHub Check: isolated-network-kindnet
- GitHub Check: fedora-bootc
- GitHub Check: centos9-bootc
- GitHub Check: isolated-network-ovnk
🔇 Additional comments (1)
docs/run.md (1)
104-111: Documentation properly documents multinode workflow.The updates clearly explain the multinode cluster creation process (initial
make run, thenmake add-nodefor additional nodes), container login via node names, and status checks. ISOLATED_NETWORK constraint is properly documented. References to new make targets align with the cluster manager implementation.Also applies to: 128-143, 161-169, 171-184
3 participants
Resolves #80
This PR adds multinode cluster management infrastructure to MicroShift, enabling creation and management of multi-node Kubernetes clusters using Podman.
Key Changes
Cluster Management Script:
src/cluster_manager.shscript providing cluster lifecycle operationsMakefile Integration:
add-node,start,stop,delete,run-status,run-ready,run-healthyruntarget refactored to use cluster_manager.shcleantarget now depends ondeletefor cluster cleanupInfrastructure Configuration:
src/rpm/postinstall.shCI/CD Enhancements:
multi-nodeinput in build action (integer, default: 0)Changed Files
Summary by CodeRabbit
New Features
Improvements
CI
Documentation
Removed