09 Nov 2022, Vitaliy Filippov
This is the fork of original "Grive" (https://github.com/Grive/grive) Google Drive client with the support for the new Drive REST API and partial sync.
Grive simply downloads all the files in your Google Drive into the current directory. After you make some changes to the local files, run grive again and it will upload your changes back to your Google Drive. New files created locally or in Google Drive will be uploaded or downloaded respectively. Deleted files will also be "removed". Currently Grive will NOT destroy any of your files: it will only move the files to a directory named .trash or put them in the Google Drive trash. You can always recover them.
There are a few things that Grive does not do at the moment:
- continously wait for changes in file system or in Google Drive to occur and upload. A sync is only performed when you run Grive (there are workarounds for almost continuous sync. See below).
- symbolic links support.
- support for Google documents.
These may be added in the future.
When Grive is run for the first time, you should use the "-a" argument to grant permission to Grive to access to your Google Drive:
cd $HOME mkdir google-drive cd google-drive grive -a
A URL should be printed. Go to the link. You will need to login to your Google account if you haven't done so. After granting the permission to Grive, the authorization code will be forwarded to the Grive application and you will be redirected to a localhost web page confirming the authorization.
If everything works fine, Grive will create .grive and .grive_state files in your current directory. It will also start downloading files from your Google Drive to your current directory.
To resync the direcory, run
grive in the folder.
cd $HOME/google-drive grive
Exclude specific files and folders from sync: .griveignore
Rules are similar to Git's .gitignore, but may differ slightly due to the different implementation.
- lines that start with # are comments
- leading and trailing spaces ignored unless escaped with \
- non-empty lines without ! in front are treated as "exclude" patterns
- non-empty lines with ! in front are treated as "include" patterns and have a priority over all "exclude" ones
- patterns are matched against the filenames relative to the grive root
- a/**/b matches any number of subpaths between a and b, including 0
- **/a matches
ainside any directory
- b/** matches everything inside
b, but not b itself
- * matches any number of any characters except /
- ? matches any character except /
- .griveignore itself isn't ignored by default, but you can include it in itself to ignore
Scheduled syncs and syncs on file change events
There are tools which you can use to enable both scheduled syncs and syncs when a file changes. Together these gives you an experience almost like the Google Drive clients on other platforms (it misses the almost instantious download of changed files in the google drive).
Grive installs such a basic solution which uses inotify-tools together with
systemd timer and services. You can enable it for a folder in your
directory (in this case the
First install the
inotify-tools (seems to be named like that in all major distros):
test that it works by calling
Prepare a Google Drive folder in your $HOME directory with
# 'google-drive' is the name of your Google Drive folder in your $HOME directory systemctl --user enable grive@$(systemd-escape google-drive).service systemctl --user start grive@$(systemd-escape google-drive).service
You can enable and start this unit for multiple folders in your
directory if you need to sync with multiple google accounts.
You can also only enable the time based syncing or the changes based syncing
by only directly enabling and starting the corresponding unit:
grive-changes@$(systemd-escape google-drive).service or
Files and folders which are shared with you don't automatically show up in your folder. They need to be added explicitly to your Google Drive: go to the Google Drive website, right click on the file or folder and chose 'Add to My Drive'.
Different OAuth2 client to workaround over quota and google approval issues
Google recently started to restrict access for unapproved applications: https://developers.google.com/drive/api/v3/about-auth?hl=ru
Grive2 is currently awaiting approval but it seems it will take forever. Also even if they approve it the default Client ID supplied with grive may exceed quota and grive will then fail to sync.
You can supply your own OAuth2 client credentials to work around these problems by following these steps:
- Go to https://console.developers.google.com/apis/api/drive.googleapis.com
- Choose a project (you might need to create one first)
- Go to https://console.developers.google.com/apis/library/drive.googleapis.com and "Enable" the Google Drive APIs
- Go to https://console.cloud.google.com/apis/credentials and click "Create credentials > Help me choose"
- In the "Find out what credentials you need" dialog, choose:
- Which API are you using: "Google Drive API"
- Where will you be calling the API from: "Other UI (...CLI...)"
- What data will you be accessing: "User Data"
- In the next steps create a client id (name doesn't matter) and setup the consent screen (defaults are ok, no need for any URLs)
- The needed "Client ID" and "Client Secret" are either in the shown download or can later found by clicking on the created credential on https://console.developers.google.com/apis/credentials/
- When you change client ID/secret in an existing Grive folder you must first delete
grive -a --id <client_id> --secret <client_secret>and follow the steps to authenticate the OAuth2 client to allow it to access your drive folder.
For the detailed instructions, see http://yourcmc.ru/wiki/Grive2#Installation
You need the following libraries:
- yajl 2.x
- Boost (Boost filesystem, program_options, regex, unit_test_framework and system are required)
There are also some optional dependencies:
- CppUnit (for unit tests)
- libbfd (for backtrace)
- binutils (for libiberty, required for compilation in OpenSUSE, Ubuntu, Arch and etc)
On a Debian/Ubuntu/Linux Mint machine just run the following command to install all these packages:
sudo apt-get install git cmake build-essential libgcrypt20-dev libyajl-dev \ libboost-all-dev libcurl4-openssl-dev libexpat1-dev libcppunit-dev binutils-dev \ debhelper zlib1g-dev dpkg-dev pkg-config
sudo dnf install git cmake libgcrypt-devel gcc-c++ libstdc++ yajl-devel boost-devel libcurl-devel expat-devel binutils zlib
pkg install git cmake boost-libs yajl libgcrypt pkgconf cppunit libbfd
Build Debian packages
On a Debian/Ubuntu/Linux Mint you can use
dpkg-buildpackage utility from
to build grive. Just clone the repository,
cd into it and run
dpkg-buildpackage -j4 --no-sign
Grive uses cmake to build. Basic install sequence is
mkdir build cd build cmake .. make -j4 sudo make install
Alternativly you can define your own client_id and client_secret during build
mkdir build cd build cmake .. "-DAPP_ID:STRING=<client_id>" "-DAPP_SECRET:STRING=<client_secret>" make -j4 sudo make install
- Implement Google OAuth loopback IP redirect flow
- Various small fixes
- Support for .griveignore
- Automatic sync solution based on inotify-tools and systemd
- no-remote-new and upload-only modes
- Ignore regexp does not persist anymore (note that Grive will still track it to not accidentally delete remote files when changing ignore regexp)
- Added options to limit upload and download speed
- Faster upload of new and changed files. Now Grive uploads files without first calculating md5 checksum when file is created locally or when its size changes.
- Added -P/--progress-bar option to print ASCII progress bar for each processed file (pull request by @svartkanin)
- Added command-line options to specify your own client_id and client_secret
- Now grive2 skips links, sockets, fifos and other unusual files
- Various small build fixes
- Much faster and more correct synchronisation using local modification time and checksum cache (similar to git index)
- Automatic move/rename detection, -m option removed
- force option works again
- Instead of crashing on sync exceptions Grive will give a warning and attempt to sync failed files again during the next run.
- Revision support works again. Grive 0.4.x always created new revisions for all files during sync, regardless of the absence of the --new-rev option.
- Shared files now sync correctly
- Option to exclude files by perl regexp
- Reimplemented HTTP response logging for debug purposes
- Use multipart uploads (update metadata and contents at the same time) for improved perfomance & stability
- Bug fixes
- Simple option to move/rename files and directories, via
grive -m oldpath newpath(by Dylan Wulf, email@example.com)
- force option does not work as documented #51
- Bug fixes
First fork release, by Vitaliy Filippov / vitalif at mail*ru
- Support for the new Google Drive REST API (old "Document List" API is shut down by Google 20 April 2015)
- REAL support for partial sync: syncs only one subdirectory with
grive -s subdir
- Major refactoring - a lot of dead code removed, JSON-C is not used anymore, API-specific code is split from non-API-specific
- Some stability fixes from Visa Putkinen https://github.com/visap/grive/commits/visa
- Slightly reduce number of syscalls when reading local files.
Bug fix & minor feature release. Fixed bugs:
- #93: missing reference count increment in one of the Json constructors
- #82: retry for HTTP error 500 & 503
- #77: Fixed a bug where grive crashed on the first run.