Storage / Stage 10 of 11

Rsync — mirrors, comparison and deletion boundaries

Not running Starting usually ~20 s

Step 7 of 8 · lab · 25 min

Bring a directory mirror up to date (rsync)

A directory mirror has fallen behind its source. Make it an exact copy and leave both the source and the neighbouring data intact.

Situation

~/sync-lab/mirror is a mirror of the ~/sync-lab/source directory that nobody has refreshed in a while. Tomorrow the data moves from the mirror to a standby server, so the mirror has to match the source byte for byte. Next to them, ~/sync-lab/keep holds data that has nothing to do with the mirror.

Task

Make mirror an exact copy of the contents of source:

mirror has the same files and directories as source, hidden ones included, and nothing extra;
every file's contents, mode, owner and modification time match source;
a dry-run sync from source into mirror that compares contents finds no differences;
source and keep are left as they were.

Constraints

You may delete things only inside mirror.
Work as your own user; this task needs no sudo.

Lab · on your own

No commands are given in this lab: type them in the terminal yourself. Goals are checked on the machine after every command.

Goals

0 of 4

mirror has the same files and directories as source, hidden ones included, and nothing extra — not done yet
File contents, modes, owners and modification times match source — not done yet
A dry-run sync from source into mirror that compares contents finds no differences — not done yet
source and keep are left as they were — not done yet

Ctrl+Enter

Starts the machine node1 · Ubuntu 24.04

The machine is starting

Hints

0 of 3

Report a problem

Lab goals

0 of 4

mirror has the same files and directories as source, hidden ones included, and nothing extra — not done yet
File contents, modes, owners and modification times match source — not done yet
A dry-run sync from source into mirror that compares contents finds no differences — not done yet
source and keep are left as they were — not done yet

node1 Ubuntu 24.04

Lab goals

0 of 4

mirror has the same files and directories as source, hidden ones included, and nothing extra — not done yet
File contents, modes, owners and modification times match source — not done yet
A dry-run sync from source into mirror that compares contents finds no differences — not done yet
source and keep are left as they were — not done yet

Files and directories · picked for this lab

rsync — file synchronization and copying

Cheat sheet

$_ — insert into the terminal

rsync -avzP src/ dst/ sync (changes only, with progress)
rsync -avzP src/ user@host:/dst/ over SSH
rsync -avzn --delete src/ dst/ dry-run a mirror FIRST
rsync -avz --exclude='cache/' src/ dst/
rsync -av --link-dest=/bak/prev src/ /bak/today/ snapshot backup
remember the trailing / on the source

-n (dry-run) shows what would change without writing a single byte to disk — always check a command with --delete this way before removing the flag. -a does not include deletion in the destination: without --delete, rsync only adds and updates files, it never removes anything extra.

The trailing slash matters!

bash

rsync -av src/ dst/    # CONTENTS of src → into dst (dst/file)
rsync -av src  dst/    # the src directory ITSELF → into dst (dst/src/file)

A trailing slash on the source means "contents". No slash means "the directory itself". A frequent source of mistakes.

Commands

bash

rsync -avzP dir/ user@host:/backup/        # to a remote host over SSH, with progress
rsync -avz user@host:/data/ ./local/       # from a remote host
rsync -avzn --delete src/ dst/             # dry-run a mirror FIRST
rsync -avz --delete src/ dst/              # mirroring (deletes extra files in dst!)
rsync -avz --exclude='*.tmp' --exclude='cache/' src/ dst/
rsync -avz -e 'ssh -p 2222' src/ host:/    # non-default SSH port
rsync -avz --bwlimit=5000 big/ host:/      # don't saturate the link (5 MB/s)
# incremental snapshots (unchanged files are hard-linked to yesterday's):
rsync -av --delete --link-dest=/bak/2026-06-29 src/ /bak/2026-06-30/

--delete removed something needed

--delete makes dst an exact copy of src. Always run -n (dry-run) first and check the list. A safety net: --backup --backup-dir=.

Copied to the wrong place / a duplicated directory

The trailing slash on the source (see the slash section above).

It copies everything again every time

Comparison is by mtime and size; if time "drifts" (different filesystems or time zones), add -c (checksum comparison) or --modify-window=1.

Permissions/owners weren't preserved

You need -a (or -o/-g/-p individually); for foreign owners, root rights on the receiving side. On FAT/NTFS, owners are never preserved.

Slow with lots of small files

Normal (rsync checks every file). -z helps over a network; for a one-off archive in a single file, tar is faster.

--delete removes everything in the destination that isn't in the source, with no confirmation prompt and no trash bin: directories swapped by mistake will wipe production data in a single run. Always check such a command with -n (dry-run) and read the list of files to be deleted before removing the flag. A safety net for important data — --backup --backup-dir= instead of a direct delete.

Rsync — mirrors, comparison and deletion boundaries

Situation

Task

Constraints

Signing in is needed to start the machine

The trailing slash matters!

Commands

--delete removed something needed

Copied to the wrong place / a duplicated directory

It copies everything again every time

Permissions/owners weren't preserved

Slow with lots of small files