Don’t Be Like Me — Use rsync

Recently, my home Jellyfin server ran out of storage. To fix that, I added a new, larger drive to store the Jellyfin media data. After mounting the new drive, I used the mv command to transfer the old data into the new storage.

Unfortunately, my Windows laptop, which I was using to SSH into the server decided to restart for an update mid-transfer, interrupting the process halfway.

---

🕳 The Problem

Because the transfer was done using mv, the operation did not survive the disconnection. I also wasn’t using screen or tmux, so I couldn’t reconnect to the session and resume.

I had to manually inspect the files to see what had been moved and what hadn’t.

To make things worse, I thought mv wouldn’t work properly if the destination folder already contained files, so I assumed I had to delete everything and start over.

That’s when I learned (thanks to ChatGPT) that:

💡 You can use rsync to transfer files even on the same machine, not just over a network.

I used to think rsync was only for remote transfers like scp. I was wrong.

---

Why rsync is a better choice

When a mv is interrupted:

• It doesn’t keep track of what was already moved.
• It can’t resume where it left off.
• It may leave the destination in a half-copied state.

Meanwhile, rsync:

• Can resume partial transfers.
• Skips files that are already copied.
• Works whether you’re transferring locally or remotely.
• Can optionally delete the source files only after each successful copy

-remove-source-files

---

What I did to recover

1. Checked the target folder:

ls -lah /mnt/1tb-drive/hlp-media/jellyfin/data/

Output (simplified):

Movies/   → 37G
Music/    → 1.6G
TV/       → 105G

2. Compared with the source:

ls -lah /srv/sftp/jellyuser/data/
du -sh /srv/sftp/jellyuser/data/*

Output:

Movies/   → 74G
Learning/ → 63G (not transferred yet)
Music/    → 1.6G
TV/       → 105G

---

My first instinct (which wasn’t great)

Originally, I thought I had to delete the entire target folder and start over:

rm -rf /mnt/1tb-drive/hlp-media/jellyfin/data/*
mv -v /srv/sftp/jellyuser/data /mnt/1tb-drive/hlp-media/jellyfin

But that would be slow, risky, and wasteful especially if the large TV and Music folders were already correctly copied.

---

🛠 The Better Way: tmux + rsync

Here’s the safer and smarter method:

Step 1: Start a persistent shell using tmux

tmux new -s transferjob

If your SSH session disconnects, you can resume it later with:

tmux attach -t transferjob

Step 2: Run rsync to resume the transfer without deleting anything

sudo rsync -av /srv/sftp/jellyuser/data/ /mnt/1tb-drive/hlp-media/jellyfin/data/

This:

• Copies only missing or updated files
• Keeps already-copied files
• Doesn’t require wiping the destination
• Lets you resume if interrupted again

---

Lessons Learned

• rsync is not just for remote transfers — it’s great for local transfers too
• Always use tmux or screen for long-running SSH tasks
• mv is dumb; rsync is smart
• You don’t need to delete target directories just because they’re not empty

I am just an idiot that learned something the hard way.

Thanks to ChatGPT, I am glad that I stumble upon this mistake at home where the stakes were low and the risk was minimal.

Now I’ve learned a valuable habit for the future, and maybe saved myself from a much bigger mess down the road.