Move IPVS-to-iptables prerequisite to all platforms in eBPF docs#2650
Move IPVS-to-iptables prerequisite to all platforms in eBPF docs#2650ctauchen merged 3 commits intotigera:mainfrom
Conversation
The warning about switching from kube-proxy IPVS mode to iptables mode before enabling eBPF was nested under the MKE section, making it appear MKE-specific. This is a prerequisite for all platforms. Move it to a caution admonition at the top of the "Configure kube-proxy" section and add guidance about restarting nodes after the switch. Fixes projectcalico/calico#12476 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
✅ Deploy Preview for calico-docs-preview-next ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
✅ Deploy Preview succeeded!Built without sensitive environment variables
To edit notification comments on pull requests, go to your Netlify project configuration. |
There was a problem hiding this comment.
Pull request overview
Moves the IPVS-to-iptables prerequisite for eBPF/kube-proxy migration into a general warning so it’s visible for all platforms, and removes the now-redundant MKE-specific note.
Changes:
- Added a top-level
:::cautionunder “Configure kube-proxy” calling out the required switch from IPVS to iptables mode. - Added guidance to restart/roll after switching modes.
- Removed the duplicated IPVS note from the MKE subsection.
|
|
||
| :::caution | ||
|
|
||
| If you are running `kube-proxy` in IPVS mode, you must switch it to iptables mode before disabling `kube-proxy` or enabling eBPF mode. This applies to all platforms and is required for a successful migration. After switching to iptables mode, restart or roll your nodes before proceeding. |
There was a problem hiding this comment.
“Restart or roll your nodes” is ambiguous and could be interpreted as a full node reboot, which is significantly more disruptive than restarting specific components. Can you clarify exactly what needs to be restarted after switching from IPVS to iptables (for example: kube-proxy DaemonSet pods, calico-node pods, or the nodes themselves) and what “roll” means in this context?
| If you are running `kube-proxy` in IPVS mode, you must switch it to iptables mode before disabling `kube-proxy` or enabling eBPF mode. This applies to all platforms and is required for a successful migration. After switching to iptables mode, restart or roll your nodes before proceeding. | |
| If you are running `kube-proxy` in IPVS mode, you must switch it to iptables mode before disabling `kube-proxy` or enabling eBPF mode. This applies to all platforms and is required for a successful migration. After switching to iptables mode, restart the `kube-proxy` instances on each node before proceeding. For clusters where `kube-proxy` runs as a `DaemonSet`, this means doing a rolling restart so the `kube-proxy` pods are recreated with the updated configuration. For managed environments, use the equivalent provider workflow to roll out the `kube-proxy` configuration change. A full node reboot is not required. |
|
|
||
| :::caution | ||
|
|
||
| If you are running `kube-proxy` in IPVS mode, you must switch it to iptables mode before disabling `kube-proxy` or enabling eBPF mode. This applies to all platforms and is required for a successful migration. After switching to iptables mode, restart or roll your nodes before proceeding. |
There was a problem hiding this comment.
In this caution admonition, iptables is referenced as plain text (“iptables mode”) but later in this same doc (e.g., “kube_proxy_mode to iptables” around line 452) the mode name is formatted as code. Consider using consistent formatting here (for example, iptables mode) to match the rest of the page.
| If you are running `kube-proxy` in IPVS mode, you must switch it to iptables mode before disabling `kube-proxy` or enabling eBPF mode. This applies to all platforms and is required for a successful migration. After switching to iptables mode, restart or roll your nodes before proceeding. | |
| If you are running `kube-proxy` in IPVS mode, you must switch it to `iptables` mode before disabling `kube-proxy` or enabling eBPF mode. This applies to all platforms and is required for a successful migration. After switching to `iptables` mode, restart or roll your nodes before proceeding. |
Apply the same fix from the previous commit to all remaining affected files across Calico Enterprise and versioned docs for both enabling-ebpf.mdx and install.mdx. Affected products/versions: - Calico: current, 3.29, 3.30, 3.31 - Calico Enterprise: current, 3.20-2, 3.21-2, 3.22-2, 3.23-1 Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Summary
enabling-ebpf.mdx+install.mdx)enabling-ebpf.mdx+install.mdx)Fixes projectcalico/calico#12476
Test plan
🤖 Generated with Claude Code