{"id":471185,"date":"2025-08-19T03:00:16","date_gmt":"2025-08-19T03:00:16","guid":{"rendered":"http:\/\/savepearlharbor.com\/?p=471185"},"modified":"-0001-11-30T00:00:00","modified_gmt":"-0001-11-29T21:00:00","slug":"","status":"publish","type":"post","link":"https:\/\/savepearlharbor.com\/?p=471185","title":{"rendered":"My way of a full system backup without external software: incremental rsync plus btrfs with zstd compression<\/span>"},"content":{"rendered":"
<\/div>\n
\n
\n
\n
\n

The repo of this script is https:\/\/gitlab.com\/vitaly\u2011zdanevich\/full\u2011backup\/\u2011\/blob\/master\/full\u2011backup.sh<\/a> <\/p>\n

Incremental with hard links means that if a file is not changed, on the next backup it will link to the same underlying data, like deduplication. Hard links<\/a>\u00a0\u2014 its usual files. <\/p>\n

Also, this script ignores .gitignore<\/strong> of every folder. <\/p>\n

Run this script from another system. <\/p>\n

# Start this script from the git folder of this script # This script accepts the path to the partition to backup (mounted) # Use this script NOT from the system you want to backup - boot from another system  if [ -z \"$1\" ]; then echo \"Missing path of the source (what to backup)\" exit 1 fi  BEFORE=$(df -h)  STARTED=$(date)  DATE=`date \"+%Y-%m-%d\"`  p=$(pwd)  cd $1 # Because rsync --exclude-from accepts only relative paths, not absolute  rsync --archive --acls --xattrs --delete --progress --verbose --exclude-from=$p\/exclude.txt --filter=':- .gitignore' --link-dest=$p\/last --mkpath $1 $p\/$DATE  ln --symbolic --force --no-dereference $DATE last  echo \"Started at:   \" $STARTED echo \"Current time: \" $(date)  echo \"Before:  $BEFORE  Now: \"  df -h  cd -  # How to restore: # # rsync --archive --acls --xattrs --progress --verbose <from> <to> # \/etc\/fstab: alter UUID # grub install (if new SSD) https:\/\/wiki.gentoo.org\/wiki\/Grub: #     grub-install # Note that if SSD is more than 2 GB - you cannot use MBR, only GPT with EFI partition (ThinkPad T430 supports UEFI boot) # emerge-webrsync # emaint sync guru   # Documentation: # https:\/\/jumpcloud.com\/blog\/how-to-backup-linux-system-rsync # https:\/\/wiki.archlinux.org\/title\/rsync<\/code><\/pre>\n
      exclude.txt<\/strong>:<\/p>\n
dev\/* proc\/* sys\/* run\/* # If run on the system that you want to backup  var\/db\/repos\/gentoo var\/db\/repos\/guru var\/cache\/distfiles\/  tmp\/* var\/tmp  lost+found mnt\/*  home\/vitaly\/.npm\/ home\/vitaly\/.cache\/  home\/vitaly\/.zoom\/  home\/vitaly\/.mozilla\/firefox\/*\/storage home\/vitaly\/.config\/google-chrome\/ home\/vitaly\/.config\/chromium\/ home\/vitaly\/.config\/Microsoft\/Microsoft Teams\/ home\/vitaly\/.config\/teams-for-linux\/ home\/vitaly\/.config\/Slack\/ home\/vitaly\/.thumbnails\/  home\/vitaly\/.cache\/ home\/vitaly\/.local\/share\/TelegramDesktop\/tdata\/user_data\/cache\/ home\/vitaly\/.local\/share\/TelegramDesktop\/tdata\/user_data\/media_cache\/ home\/vitaly\/.local\/share\/Steam\/ home\/vitaly\/.googleearth\/Cache\/ home\/vitaly\/.local\/share\/OMaps\/ home\/vitaly\/.config\/Audius\/Cache\/ home\/vitaly\/.config\/YandexMusic\/Cache\/ home\/vitaly\/.config\/vesktop\/sessionData\/Cache\/ home\/vitaly\/.config\/bilibili\/Cache\/ home\/vitaly\/.local\/share\/Trash\/  home\/vitaly\/go\/pkg\/mod\/cache\/ home\/vitaly\/.cargo\/registry\/  home\/vitaly\/record\/out\/  home\/vitaly\/Desktop\/unreal-5-4-4\/<\/code><\/pre>\n
My HDD for backups is on btrfs<\/a> for ZSTD compression<\/a>, how to prepare the disk: <\/p>\n
sudo apt install btrfs-progs -y # https:\/\/details.nl\/install-and-use-btrfs-on-ubuntu-20-04-lts  sudo mkfs.btrfs \/dev\/sdX # Partitions are not needed  sudo mount -o compress-force=zstd:15 \/dev\/sdX \/mnt # Max compression level is 15 # Default compression you can add to fstab<\/code><\/pre>\n
After backup\u00a0\u2014 you can see the compression ratio: <\/p>\n
sudo apt install btrfs-compsize sudo compsize \/mnt<\/code><\/pre>\n
Output example:<\/p>\n
   Processed 54036 files, 42027 regular extents (42028 refs), 27150 inline.    Type       Perc     Disk Usage   Uncompressed Referenced    TOTAL       73%      211G         289G         289G    zstd        28%       42M         148M         148M    none       100%      174G         174G         174G<\/code><\/pre>\n
The main column here is Perc<\/strong>\u00a0\u2014 lower is better compression. <\/p>\n
Besides this backup solution, I store every config in a separate git repository<\/a>, not one repo for all dot files. <\/p>\n
Sometimes I use Clonezilla<\/a>\u00a0\u2014 with xz compression of the full drive, for Windows and Linux. <\/p>\n
This article is based on my Russian article https:\/\/habr.com\/en\/articles\/929182<\/a>, where people also recommended their backup solutions: <\/p>\n
 https:\/\/github.com\/jimsalterjrs\/sanoid\/wiki\/Syncoid<\/a> <\/p>\n
 https:\/\/github.com\/jimsalterjrs\/sanoid<\/a> <\/p>\n
 https:\/\/github.com\/bit\u2011team\/backintime<\/a> <\/p>\n
 https:\/\/github.com\/ei\u2011grad\/trinkup<\/a> <\/p>\n
 https:\/\/restic.net<\/a> <\/p>\n
 https:\/\/www.urbackup.org<\/a> <\/p>\n
 https:\/\/www.borgbackup.org<\/a> <\/p>\n
 https:\/\/kopia.io<\/a> <\/p>\n
 https:\/\/duplicati.com<\/a> <\/p>\n
 http:\/\/dar.linux.free.fr<\/a> <\/p>\n
 BTRFS snapshots<\/a> <\/p>\n<\/div>\n<\/div>\n<\/div>\n
<\/div>\n

 \u0441\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u043e\u0440\u0438\u0433\u0438\u043d\u0430\u043b \u0441\u0442\u0430\u0442\u044c\u0438  https:\/\/habr.com\/ru\/articles\/938448\/<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"
<\/div>\n
\n
\n
\n
\n
The repo of this script is https:\/\/gitlab.com\/vitaly\u2011zdanevich\/full\u2011backup\/\u2011\/blob\/master\/full\u2011backup.sh<\/a> <\/p>\n
Incremental with hard links means that if a file is not changed, on the next backup it will link to the same underlying data, like deduplication. Hard links<\/a>\u00a0\u2014 its usual files. <\/p>\n
Also, this script ignores .gitignore<\/strong> of every folder. <\/p>\n
Run this script from another system. <\/p>\n
# Start this script from the git folder of this script # This script accepts the path to the partition to backup (mounted) # Use this script NOT from the system you want to backup - boot from another system  if [ -z \"$1\" ]; then echo \"Missing path of the source (what to backup)\" exit 1 fi  BEFORE=$(df -h)  STARTED=$(date)  DATE=`date \"+%Y-%m-%d\"`  p=$(pwd)  cd $1 # Because rsync --exclude-from accepts only relative paths, not absolute  rsync --archive --acls --xattrs --delete --progress --verbose --exclude-from=$p\/exclude.txt --filter=':- .gitignore' --link-dest=$p\/last --mkpath $1 $p\/$DATE  ln --symbolic --force --no-dereference $DATE last  echo \"Started at:   \" $STARTED echo \"Current time: \" $(date)  echo \"Before:  $BEFORE  Now: \"  df -h  cd -  # How to restore: # # rsync --archive --acls --xattrs --progress --verbose <from> <to> # \/etc\/fstab: alter UUID # grub install (if new SSD) https:\/\/wiki.gentoo.org\/wiki\/Grub: #     grub-install # Note that if SSD is more than 2 GB - you cannot use MBR, only GPT with EFI partition (ThinkPad T430 supports UEFI boot) # emerge-webrsync # emaint sync guru   # Documentation: # https:\/\/jumpcloud.com\/blog\/how-to-backup-linux-system-rsync # https:\/\/wiki.archlinux.org\/title\/rsync<\/code><\/pre>\n
      exclude.txt<\/strong>:<\/p>\n
dev\/* proc\/* sys\/* run\/* # If run on the system that you want to backup  var\/db\/repos\/gentoo var\/db\/repos\/guru var\/cache\/distfiles\/  tmp\/* var\/tmp  lost+found mnt\/*  home\/vitaly\/.npm\/ home\/vitaly\/.cache\/  home\/vitaly\/.zoom\/  home\/vitaly\/.mozilla\/firefox\/*\/storage home\/vitaly\/.config\/google-chrome\/ home\/vitaly\/.config\/chromium\/ home\/vitaly\/.config\/Microsoft\/Microsoft Teams\/ home\/vitaly\/.config\/teams-for-linux\/ home\/vitaly\/.config\/Slack\/ home\/vitaly\/.thumbnails\/  home\/vitaly\/.cache\/ home\/vitaly\/.local\/share\/TelegramDesktop\/tdata\/user_data\/cache\/ home\/vitaly\/.local\/share\/TelegramDesktop\/tdata\/user_data\/media_cache\/ home\/vitaly\/.local\/share\/Steam\/ home\/vitaly\/.googleearth\/Cache\/ home\/vitaly\/.local\/share\/OMaps\/ home\/vitaly\/.config\/Audius\/Cache\/ home\/vitaly\/.config\/YandexMusic\/Cache\/ home\/vitaly\/.config\/vesktop\/sessionData\/Cache\/ home\/vitaly\/.config\/bilibili\/Cache\/ home\/vitaly\/.local\/share\/Trash\/  home\/vitaly\/go\/pkg\/mod\/cache\/ home\/vitaly\/.cargo\/registry\/  home\/vitaly\/record\/out\/  home\/vitaly\/Desktop\/unreal-5-4-4\/<\/code><\/pre>\n
My HDD for backups is on btrfs<\/a> for ZSTD compression<\/a>, how to prepare the disk: <\/p>\n
sudo apt install btrfs-progs -y # https:\/\/details.nl\/install-and-use-btrfs-on-ubuntu-20-04-lts  sudo mkfs.btrfs \/dev\/sdX # Partitions are not needed  sudo mount -o compress-force=zstd:15 \/dev\/sdX \/mnt # Max compression level is 15 # Default compression you can add to fstab<\/code><\/pre>\n
After backup\u00a0\u2014 you can see the compression ratio: <\/p>\n
sudo apt install btrfs-compsize sudo compsize \/mnt<\/code><\/pre>\n
Output example:<\/p>\n
   Processed 54036 files, 42027 regular extents (42028 refs), 27150 inline.    Type       Perc     Disk Usage   Uncompressed Referenced    TOTAL       73%      211G         289G         289G    zstd        28%       42M         148M         148M    none       100%      174G         174G         174G<\/code><\/pre>\n
The main column here is Perc<\/strong>\u00a0\u2014 lower is better compression. <\/p>\n
Besides this backup solution, I store every config in a separate git repository<\/a>, not one repo for all dot files. <\/p>\n
Sometimes I use Clonezilla<\/a>\u00a0\u2014 with xz compression of the full drive, for Windows and Linux. <\/p>\n
This article is based on my Russian article https:\/\/habr.com\/en\/articles\/929182<\/a>, where people also recommended their backup solutions: <\/p>\n
 https:\/\/github.com\/jimsalterjrs\/sanoid\/wiki\/Syncoid<\/a> <\/p>\n
 https:\/\/github.com\/jimsalterjrs\/sanoid<\/a> <\/p>\n
 https:\/\/github.com\/bit\u2011team\/backintime<\/a> <\/p>\n
 https:\/\/github.com\/ei\u2011grad\/trinkup<\/a> <\/p>\n
 https:\/\/restic.net<\/a> <\/p>\n
 https:\/\/www.urbackup.org<\/a> <\/p>\n
 https:\/\/www.borgbackup.org<\/a> <\/p>\n
 https:\/\/kopia.io<\/a> <\/p>\n
 https:\/\/duplicati.com<\/a> <\/p>\n
 http:\/\/dar.linux.free.fr<\/a> <\/p>\n
 BTRFS snapshots<\/a> <\/p>\n<\/div>\n<\/div>\n<\/div>\n
<\/div>\n

 \u0441\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u043e\u0440\u0438\u0433\u0438\u043d\u0430\u043b \u0441\u0442\u0430\u0442\u044c\u0438  https:\/\/habr.com\/ru\/articles\/938448\/<\/a>
<\/br><\/br><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[],"tags":[],"class_list":["post-471185","post","type-post","status-publish","format-standard","hentry"],"_links":{"self":[{"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=\/wp\/v2\/posts\/471185","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=471185"}],"version-history":[{"count":0,"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=\/wp\/v2\/posts\/471185\/revisions"}],"wp:attachment":[{"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=471185"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=471185"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/savepearlharbor.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=471185"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}