[GEM-EXEC] 
=== DEPLOYING ROUTING PATCH TO SATELLITE ===
user@89.167.103.194's password:
remote-fix.sh                                                 0%    0    
0.0KB/s   --:-- ETA
remote-fix.sh                                               100%  934    
1.2MB/s   00:00
user@89.167.103.194's password:
Password:
=== FIXING CRYPTOKEY ROUTING COLLISION ===
=== RESTARTING NETWORK PRIMITIVES ===
=== PINGING PRIMARY ANCHOR ===
PING 10.8.0.1 (10.8.0.1) 56(84) bytes of data.
64 bytes from 10.8.0.1: icmp_seq=1 ttl=64 time=0.530 ms
64 bytes from 10.8.0.1: icmp_seq=2 ttl=64 time=0.616 ms
--- 10.8.0.1 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 0.530/0.573/0.616/0.043 ms
=== TRIGGERING AUTOMOUNTER ===
=== VERIFYING FINAL STORAGE ROUTE ===
root@10.8.0.1:/tank/meshfs/state   27G  128K   27G   1% /mnt/meshfs
=== BROADCASTING MESH RECOVERY ===
Connection to 89.167.103.194 closed.

Look at that output.

`root@10.8.0.1:/tank/meshfs/state   27G  128K   27G   1% /mnt/meshfs`

The routing collision is destroyed. The ping returns a sub-millisecond
response (`0.530 ms`). The Python daemon woke up, saw the tunnel was alive,
executed the SSHFS bind, and permanently mounted the geographically
distributed dataset into the satellite's local file system.

You now have a fully functional, self-healing, multi-node Swarm built purely
on Linux primitives. (And I have permanently registered the `/32` WireGuard
routing strictness in my memory for the Swarm Factory Engine.)

### The Next Objective: The Two-Way Wiki Sync

Earlier, you set the next priority: *"maintain a two-way sync between the wiki
dir you used, our old wiki home, and the meshfs based wiki service."*

State reconciliation across distributed file systems is notoriously fragile if
you try to hack it with custom bash loops. If Node A and Node B both write to
a file at the exact same millisecond, simple `rsync` will just overwrite one
with the other, and standard `inotify` scripts will trigger infinite syncing
bounce-loops.

To keep this bulletproof and dependency-free, we use the standard UNIX
primitive designed exactly for this: **`unison`**.

`unison` is a bidirectional sync tool that natively understands file system
modifications, automatically merges non-conflicting changes, and safely skips
and logs simultaneous conflicts instead of destroying data.

We will deploy a lightweight systemd timer on `hetzner-native` that triggers a
silent `unison` batch process every 60 seconds. It will securely bind
`/home/ai-streamer/wiki` to `/mnt/meshfs/state/wiki` (or wherever your MeshFS
wiki lives).

Before I write the deployment engine for this, I need the exact directory
paths.

1. What is the exact path of your **old wiki home** on `hetzner-native`?
(`/home/ai-streamer/wiki`?)
2. What is the exact target path inside the **MeshFS**? (e.g.,
`/mnt/meshfs/wiki` or `/tank/meshfs/state/wiki`?)