Troubleshooting

Most failures fall into a small number of buckets. This page maps symptom to likely cause to fix.

Connect fails immediately

Symptom	Cause	Fix
Permission denied (publickey)	SSH auth wrong	Verify the key matches a line in remote ~/.ssh/authorized_keys; ssh -i <key> user@host from a terminal first
Host key verification failed	Plugin known-host store rejected	Open settings → trust dialog. If a known host changed, see Host-key trust
Connection refused	sshd not listening	Verify the port; nc -zv <host> <port> to confirm reachability
Connection timeout	Network / firewall	Same hop reachable via ping? sshd port open in firewall?
Daemon failed to start (after binary upload)	Daemon crashed at startup	See Daemon won’t start below

Daemon won’t start

Check the daemon log: Settings → Daemon → “View log”, or directly:

ssh user@host 'cat ~/.obsidian-remote/server.log'

Common patterns:

• permission denied opening the socket → check ~/.obsidian-remote/ exists with the right ownership; recreate with mkdir -p ~/.obsidian-remote && chmod 700 ~/.obsidian-remote.
• bind: address already in use on the socket → another daemon is already running. Kill it (pkill -f obsidian-remote-server) or pick a different socket path in the profile.
• vault root does not exist → set the profile’s “Remote vault path” to a path that exists; the daemon won’t create it.
• inotify_add_watch: too many open files → raise fs.inotify.max_user_watches (see fs.watch caveats).

Files don’t sync

Symptom	Cause	Fix
Local edits don’t appear on remote	Plugin write is failing silently — open developer console (Cmd+Opt+I / Ctrl+Shift+I), look for [remote-ssh] errors	Often a permission issue on the remote vault dir
Remote edits don’t appear locally	fs.watch subscription not active, or hit the inotify watch limit	Settings → Daemon → Restart; raise fs.inotify.max_user_watches (see fs.watch caveats)
Specific file always conflicts	Two-way edit collision	See Conflict handling
Big binary files take forever	Plugin downloads on-demand; first open is slow	Expected — local cache survives across reopens

Performance feels slow

Likely culprits, in order of frequency:

1. High SSH latency — ping <host> to estimate RTT. Per-RPC RTT is roughly 2× ping. Anything > 50 ms makes typing feel laggy.
2. inotify limits — daemon can’t subscribe to all dirs. Symptom: remote edits delay until you switch focus. Fix per fs.watch.
3. Vault on slow disk (SD card on Pi) — fs.walk cold-cache is the slowest op. Move to USB SSD if possible.
4. First connect on a new host — binary upload is one-time, ~5 MB. Subsequent connects skip it.

For deep perf debug: enable Debug logging and check <plugin>/console.log (rotated console.log + .1 + .2 + .3) for per-op timings.

The full perf-tuning playbook (network / disk / inotify / daemon-side cache, with measurement commands) lives at Performance tuning.

How to ask for help

When opening an issue, paste:

1. Plugin version: Settings → Community plugins → “Remote SSH” entry shows the version.
2. Daemon version: Settings → Daemon panel → status badge.
3. Plugin log (last 50 lines from <vault>/.obsidian/plugins/remote-ssh/console.log).
4. Daemon log (last 50 lines from ~/.obsidian-remote/server.log).
5. Local OS + remote OS / arch (uname -a on remote).
6. What you expected vs what happened.

Issues at: github.com/sotashimozono/obsidian-remote-ssh/issues.

Next: Logs & telemetry.