Add 3 articles; update nav and index (2026-04-29)

New articles: - Python smtplib: Missing Date/Message-ID Headers Break Mail Clients - Fantastical MCP: Permission Denied (macOS Quarantine) - Ubuntu dist-upgrade Repo Quarantine Updated: troubleshooting index, SUMMARY.md nav, WOL article edits. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-29 22:45:17 -04:00 · 2026-04-29 22:45:17 -04:00 · 1524ca66d5
commit 1524ca66d5
parent 91455fac39
6 changed files with 291 additions and 7 deletions
--- a/02-selfhosting/dns-networking/wake-on-lan-router-ssh.md
+++ b/02-selfhosting/dns-networking/wake-on-lan-router-ssh.md
@ -1,14 +1,21 @@
 ---
 title: Wake-on-LAN via Router SSH
 description: Send WOL magic packets through an Asus router over SSH
-tags: [networking, wol, asus, ssh]
+tags:
+  - networking
+  - wol
+  - asus
+  - ssh
 created: 2026-04-19
+updated: 2026-04-27T00:53
 ---

 # Wake-on-LAN via Router SSH

 Most Asus routers running AsusWRT (or Merlin) include `ether-wake` in their BusyBox environment. Combined with SSH access, this lets you wake machines remotely from anywhere — even over a VPN like Tailscale — without needing a dedicated WOL tool on the LAN.

+> **Status: Deployed 2026-04-27.** `wake-majormac` script live on MajorMac and MajorRig. Router SSH confirmed working on RT-AX82U (AsusWRT 388, `ether-wake` at `/usr/sbin/ether-wake`). Credentials stored in Ansible vault as `router_username` / `router_password`.
+
 ## Prerequisites

 - **SSH enabled on the router** — Administration → System → Enable SSH → LAN only
@ -25,9 +32,11 @@ Asus routers use a non-standard SSH port by default. Check your router's SSH set
 Host router
  HostName 192.168.50.1
  Port 1025
-  User <your-username>
+  User majorlinux
 ```

+> RT-AX82U confirmed on port 1025. `sshpass` required for non-interactive use — install via `brew install sshpass` on Mac.
+
 ## Sending a WOL Packet

 ```bash
@ -43,7 +52,7 @@ sshpass -p "$ROUTER_PASS" ssh router 'ether-wake -i br0 AA:BB:CC:DD:EE:FF'

 ## Scripting with Ansible Vault

-If your router password is stored in an Ansible vault, you can pull it at runtime:
+Router credentials are stored in the Ansible vault as `router_username` and `router_password`. The `wake-majormac` script at `~/.local/bin/wake-majormac` (deployed on MajorMac and MajorRig) handles this automatically:

 ```bash
 #!/usr/bin/env bash
@ -51,11 +60,21 @@ set -euo pipefail

 VAULT_FILE="$HOME/MajorAnsible/group_vars/all/vault.yml"
 VAULT_PASS_FILE="$HOME/.ansible/vault_pass"
+MAC="9c:76:0e:3f:10:58"  # MajorMac (Mac Studio) en0

 password=$(ansible-vault view "$VAULT_FILE" --vault-password-file "$VAULT_PASS_FILE" 2>/dev/null \
  | grep '^router_password:' | sed 's/^router_password: *"\{0,1\}\([^"]*\)"\{0,1\}/\1/')

-sshpass -p "$password" ssh router "ether-wake -i br0 AA:BB:CC:DD:EE:FF"
+echo "Sending WOL magic packet to MajorMac ($MAC)..."
+sshpass -p "$password" ssh -o StrictHostKeyChecking=no -o ConnectTimeout=5 \
+  -p 1025 majorlinux@192.168.50.1 "ether-wake -i br0 $MAC"
+echo "Done. MajorMac should wake within ~30 seconds."
+```
+
+To wake MajorMac from any machine on the fleet:
+
+```bash
+wake-majormac
 ```

 ## Troubleshooting
@ -64,8 +83,10 @@ sshpass -p "$password" ssh router "ether-wake -i br0 AA:BB:CC:DD:EE:FF"
 |-------|-----|
 | `Connection refused` | SSH not enabled on router, or wrong port |
 | `Permission denied` | Wrong username/password |
-| Machine doesn't wake | Check WOL is enabled in BIOS; verify MAC address; ensure machine is plugged in (not on battery) |
+| `sshpass: command not found` | `brew install sshpass` on Mac; `apt install sshpass` on Debian |
+| Machine doesn't wake | Check `womp 1` in `pmset -g`; verify MAC address; machine must be plugged in |
 | `ether-wake: not found` | Router firmware may not include it — check with `which ether-wake` |
+| SSH times out | Router SSH is LAN-only; must be on LAN or send via a LAN-connected host |

 ## Why Not Use a Dedicated WOL Tool?

--- a/05-troubleshooting/fantastical-mcp-permission-denied.md
+++ b/05-troubleshooting/fantastical-mcp-permission-denied.md
@ -0,0 +1,98 @@
+---
+title: "Fantastical MCP Server: Permission Denied on Launch (macOS Quarantine)"
+domain: troubleshooting
+category: productivity
+tags: [fantastical, mcp, claude, macos, gatekeeper, quarantine, cowork]
+status: published
+created: 2026-04-26
+updated: 2026-04-26
+---
+
+# Fantastical MCP Server: Permission Denied on Launch (macOS Quarantine)
+
+Fantastical's MCP server fails to connect in Claude/Cowork with a `Server disconnected` error and no dialog or prompt to explain why. The binary is installed but macOS Gatekeeper silently blocks it from executing.
+
+## The Short Answer
+
+```bash
+xattr -d com.apple.quarantine "/Users/majorlinux/Library/Application Support/Claude/Claude Extensions/ant.dir.gh.flexibits.fantastical-mcp/server/FantasticalMCP.app"
+```
+
+Fully quit and reopen Cowork. Fantastical MCP reconnects cleanly.
+
+If the quarantine attribute isn't present, also try setting the executable bit:
+
+```bash
+chmod +x "/Users/majorlinux/Library/Application Support/Claude/Claude Extensions/ant.dir.gh.flexibits.fantastical-mcp/server/FantasticalMCP.app/Contents/MacOS/FantasticalMCP"
+```
+
+## Why This Happens
+
+macOS automatically tags downloaded files with a `com.apple.quarantine` extended attribute. When you launch an app yourself, macOS shows a Gatekeeper dialog — click Open, and the quarantine flag is cleared. But the FantasticalMCP binary is never launched by the user directly; Claude/Cowork spawns it as a subprocess. There's no dialog, and Gatekeeper just returns `Permission denied`. Claude sees the process die immediately and logs `Server disconnected`.
+
+This recurs after any Fantastical update that replaces the MCP binary — the new binary comes in quarantined again.
+
+## Diagnosis
+
+The log tells the whole story. Check:
+
+```bash
+tail -n 50 ~/Library/Logs/Claude/mcp-server-Fantastical.log
+```
+
+If you see this sequence, it's the quarantine issue:
+
+```
+Server started and connected successfully
+Failed to spawn process: Permission denied
+Server transport closed
+Server disconnected.
+```
+
+## Full Fix
+
+**Step 1 — Remove the quarantine flag:**
+
+```bash
+xattr -d com.apple.quarantine \
+  "/Users/majorlinux/Library/Application Support/Claude/Claude Extensions/ant.dir.gh.flexibits.fantastical-mcp/server/FantasticalMCP.app"
+```
+
+**Step 2 — Verify the attribute is gone:**
+
+```bash
+xattr "/Users/majorlinux/Library/Application Support/Claude/Claude Extensions/ant.dir.gh.flexibits.fantastical-mcp/server/FantasticalMCP.app"
+```
+
+Should return empty or only non-quarantine attributes. If `com.apple.quarantine` is still listed, re-run step 1.
+
+**Step 3 — Fully quit and reopen Cowork:**
+
+Cmd+Q (not just close the window). Closing the window leaves the MCP host process running — it won't retry the failed server until the app fully relaunches.
+
+**Step 4 — Verify connection:**
+
+Check the log again:
+
+```bash
+tail -n 10 ~/Library/Logs/Claude/mcp-server-Fantastical.log
+```
+
+You should see `Server started and connected successfully` with no `Permission denied` line following it.
+
+## After a Fantastical Update
+
+If this breaks again after Fantastical auto-updates, re-run the `xattr -d` command from Step 1. The update replaces the binary and macOS re-quarantines the new one.
+
+## MCP Log Locations
+
+| Log | Path |
+|-----|------|
+| Fantastical MCP | `~/Library/Logs/Claude/mcp-server-Fantastical.log` |
+| All MCP servers | `~/Library/Logs/Claude/mcp*.log` |
+| Combined MCP log | `~/Library/Logs/Claude/mcp.log` |
+
+## Related
+
+- [[Fantastical Google Sync Error Flood — Phantom Calendars Fixed via syncselect]]
+- MCP debugging docs: https://modelcontextprotocol.io/docs/tools/debugging
--- a/05-troubleshooting/index.md
+++ b/05-troubleshooting/index.md
@ -1,6 +1,6 @@
 ---
 created: 2026-03-15T06:37
-updated: 2026-04-25T12:57
+updated: 2026-04-29T22:40
 ---
 # 🔧 General Troubleshooting

@ -39,7 +39,12 @@ Practical fixes for common Linux, networking, and application problems.
 ## 💾 Storage
 - [mdadm RAID Recovery After USB Hub Disconnect](storage/mdadm-usb-hub-disconnect-recovery.md)

+## 📧 Email & SMTP
+- [Python smtplib: Missing Date/Message-ID Headers Break Mail Clients](python-smtplib-missing-rfc-headers.md)
+
 ## 📝 Application Specific
+- [Fantastical MCP: Permission Denied (macOS Quarantine)](fantastical-mcp-permission-denied.md)
+- [Fantastical Google Sync: Phantom Calendars via syncselect](fantastical-google-phantom-calendar-syncselect.md)
 - [Obsidian Vault Recovery — Loading Cache Hang](obsidian-cache-hang-recovery.md)
 - [Gemini CLI Manual Update](gemini-cli-manual-update.md)

--- a/05-troubleshooting/python-smtplib-missing-rfc-headers.md
+++ b/05-troubleshooting/python-smtplib-missing-rfc-headers.md
@ -0,0 +1,60 @@
+---
+title: "Python smtplib: Missing Date/Message-ID Headers Break Mail Clients"
+domain: troubleshooting
+category: general
+tags: [email, python, smtplib, spam, rfc, spark]
+status: published
+created: 2026-04-29
+updated: 2026-04-29
+---
+# Python smtplib: Missing Date/Message-ID Headers Break Mail Clients
+
+## Problem
+
+Emails sent via Python's `smtplib` and `EmailMessage` appear on some mail clients but not others. The emails are delivered to the server and visible in Maildir, but specific clients silently suppress them.
+
+## Root Cause
+
+Python's `EmailMessage` does **not** automatically add `Date:` or `Message-ID:` headers. These are required by RFC 5322. Without them:
+
+- **SpamAssassin** flags `MISSING_DATE` and `MISSING_MID`, and may set `X-Spam-Flag: YES` even if the overall score is below the spam threshold
+- **Mail clients** (e.g., Spark) may filter on the spam flag header and silently hide the message — no Junk folder, just invisible
+- **Other clients** (e.g., iPhone Mail, some Spark builds) may be more lenient and display the message anyway
+
+This creates a confusing situation where the same email appears on one device but not another, despite both using the same IMAP account.
+
+## Fix
+
+Always include `Date` and `Message-ID` headers when constructing emails with `EmailMessage`:
+
+```python
+import smtplib
+from email.message import EmailMessage
+from email.utils import formatdate, make_msgid
+
+msg = EmailMessage()
+msg['Subject'] = 'Your subject here'
+msg['From'] = 'sender@example.com'
+msg['To'] = 'recipient@example.com'
+msg['Date'] = formatdate(localtime=True)
+msg['Message-ID'] = make_msgid(domain='example.com')
+msg.set_content('Email body here')
+
+with smtplib.SMTP('mail.example.com', 25) as s:
+    s.send_message(msg)
+```
+
+## Verification
+
+After applying the fix, check that SpamAssassin no longer flags the headers:
+
+```bash
+# Check email headers on the mail server
+grep -E 'MISSING_DATE|MISSING_MID|X-Spam' /var/vmail/domain/user/cur/<message-file>
+```
+
+A clean message should show `X-Spam-Status: No` with no `MISSING_DATE` or `MISSING_MID` in the test list.
+
+## Key Takeaway
+
+Python's `EmailMessage` is a low-level builder — it trusts you to set all required headers. Unlike higher-level mail libraries or webmail interfaces, it will happily send a message with no date or message ID. Always add both explicitly in any script that sends email via `smtplib`.
--- a/05-troubleshooting/ubuntu-dist-upgrade-repo-quarantine.md
+++ b/05-troubleshooting/ubuntu-dist-upgrade-repo-quarantine.md
@ -0,0 +1,98 @@
+---
+title: "Ubuntu dist-upgrade Quarantines Third-Party Repos"
+domain: troubleshooting
+category: ubuntu
+tags: [ubuntu, apt, dist-upgrade, repositories, tailscale, digitalocean]
+status: published
+created: 2026-04-28
+updated: 2026-04-28
+---
+
+# Ubuntu dist-upgrade Quarantines Third-Party Repos
+
+## Problem
+
+When running `do-release-upgrade` (e.g., Jammy 22.04 to Noble 24.04), Ubuntu renames all third-party `.list` files in `/etc/apt/sources.list.d/` to `.list.distUpgrade`. This silently disables every third-party repo — packages from those repos stop receiving updates with no warning.
+
+The upgrade process does this intentionally because it can't guarantee third-party repos will have packages for the new release. Some repos get re-added as `.sources` files during the upgrade, but many don't.
+
+## Symptoms
+
+- `apt list --upgradable` shows nothing for packages you know have updates (e.g., Tailscale stuck on an old version)
+- `apt list --installed` shows packages as `[installed,local]` instead of `[installed]` — the "local" tag means apt has no repo to check for updates
+- `.distUpgrade` files accumulate in `/etc/apt/sources.list.d/` indefinitely
+
+## Diagnosis
+
+Check for quarantined repos:
+
+```bash
+ls /etc/apt/sources.list.d/*.distUpgrade
+```
+
+For each file, check whether a replacement `.list` or `.sources` file already exists:
+
+```bash
+ls /etc/apt/sources.list.d/*.list /etc/apt/sources.list.d/*.sources
+```
+
+## Fix
+
+### Distro-agnostic repos (e.g., DigitalOcean agents)
+
+If the repo URL doesn't reference a distro codename (jammy/noble), just rename:
+
+```bash
+mv /etc/apt/sources.list.d/digitalocean-agent.list.distUpgrade \
+   /etc/apt/sources.list.d/digitalocean-agent.list
+```
+
+### Distro-specific repos (e.g., Tailscale, ondrej-php)
+
+The quarantined file references the old distro (jammy). Re-run the upstream install script to get a correct entry for the new release:
+
+```bash
+# Tailscale
+curl -fsSL https://tailscale.com/install.sh | sh
+
+# Or manually: update the codename
+sed 's/jammy/noble/' /etc/apt/sources.list.d/tailscale.list.distUpgrade \
+  > /etc/apt/sources.list.d/tailscale.list
+apt update && apt upgrade tailscale
+```
+
+### Already replaced by .sources
+
+If the upgrade process already created a `.sources` replacement (common for ubuntu-esm-apps, ondrej-php), the `.distUpgrade` file is just clutter — delete it:
+
+```bash
+rm /etc/apt/sources.list.d/ondrej-ubuntu-php-jammy.list.distUpgrade
+```
+
+### After all fixes
+
+```bash
+apt update
+apt list --upgradable   # should now show pending updates
+apt upgrade
+```
+
+## Real-World Example: MajorsHouse Fleet (2026-04-28)
+
+Five Ubuntu 24.04 servers were dist-upgraded from Jammy in October 2024. The `.distUpgrade` quarantine was discovered 6 months later when Tailscale's website wouldn't load (Pi-hole was blocking subdomains, but the investigation revealed teelia was stuck on Tailscale 1.76.0 — 20 versions behind — because the repo was disabled).
+
+| Host | Quarantined files | Impact |
+|------|------------------|--------|
+| dcaprod | 8 | Tailscale, DO agents, MySQL, ondrej-php, ESM, vector |
+| teelia | 4 | Tailscale (stuck on 1.76.0), DO agents, certbot bionic PPA |
+| majorlinux | 8 | Tailscale, DO agents, MySQL, ondrej-php, ESM, apt-fast |
+| majortoot | 11 | Tailscale, DO agents, nodesource, PostgreSQL, vector, zabbix, ESM |
+| tttpod | 0 | Clean — was likely rebuilt rather than upgraded |
+
+All files were audited, stale ones deleted, distro-agnostic repos renamed, and distro-specific repos re-added via upstream install scripts. DO agents upgraded from 3.16.11 to 3.18.12, teelia's Tailscale jumped from 1.76.0 to 1.96.4.
+
+## Prevention
+
+- **Post-upgrade audit:** After any `do-release-upgrade`, immediately run `ls /etc/apt/sources.list.d/*.distUpgrade` and resolve each file.
+- **Prefer `.sources` format:** When adding new third-party repos, use the DEB822 `.sources` format — it's what Ubuntu itself uses on Noble and is handled more gracefully during upgrades.
+- **Ansible playbook:** Consider a post-upgrade play that checks for `.distUpgrade` files and alerts or auto-fixes distro-agnostic repos.
--- a/SUMMARY.md
+++ b/SUMMARY.md
@ -1,6 +1,6 @@
 ---
 created: 2026-04-02T16:03
-updated: 2026-04-25T12:57
+updated: 2026-04-26T10:11
 ---
 * [Home](index.md)
 * [Linux & Sysadmin](01-linux/index.md)
@ -90,6 +90,8 @@ updated: 2026-04-25T12:57
    * [Ollama Drops Off Tailscale When Mac Sleeps](05-troubleshooting/ollama-macos-sleep-tailscale-disconnect.md)
    * [Ollama: `ollama run` with Piped Stdin Bypasses Chat Template + SYSTEM Prompt](05-troubleshooting/ollama-chat-template-pipe-stdin-bypass.md)
    * [rsync over Tailscale: Hung in TCP Teardown After Transfer Completes](05-troubleshooting/networking/rsync-tailscale-teardown-stall.md)
+    * [Fantastical MCP: Permission Denied (macOS Quarantine)](05-troubleshooting/fantastical-mcp-permission-denied.md)
+    * [Fantastical Google Sync: Phantom Calendars Fixed via syncselect](05-troubleshooting/fantastical-google-phantom-calendar-syncselect.md)
    * [macOS: Repeating Alert Tone from Mirrored iPhone Notification](05-troubleshooting/macos-mirrored-notification-alert-loop.md)
    * [ClamAV CPU Spike: Safe Scheduling with nice/ionice](05-troubleshooting/security/clamscan-cpu-spike-nice-ionice.md)
    * [Ansible: Vault Password File Not Found](05-troubleshooting/ansible-vault-password-file-missing.md)