Mounting Alluxio with FUSE

Slack Docker Pulls

Alluxio-FUSE is a feature that allows mounting the distributed Alluxio File System as a standard file system on most flavors of Unix. By using this feature, standard bash tools (for example, ls, cat or mkdir) will have basic access to the distributed Alluxio data store. More importantly, with FUSE your application written in any language like C, C++, Python, Ruby, Perl, Java etc can interact with Alluxio by using standard POSIX APIs like open, write, read, without any Alluxio client integration or set up.

Alluxio-FUSE is based on the project Filesystem in Userspace (FUSE), and most basic file system operations are supported. However, given the intrinsic characteristics of Alluxio, like its write-once/read-many-times file data model, the mounted file system will not have full POSIX semantics and will have specific limitations. Please read the section of limitations for details.

Requirements

  • JDK 1.8 or newer
  • libfuse 2.9.3 or newer (2.8.3 has been reported to also work - with some warnings) for Linux
  • osxfuse 3.7.1 or newer for MacOS

Usage

Mount Alluxio-FUSE

After having properly configured and started the alluxio cluster, and from the node where you wish to mount Alluxio, point a shell to your $ALLUXIO_HOME and run:

$ integration/fuse/bin/alluxio-fuse mount mount_point [alluxio_path]

This will spawn a background user-space java process (alluxio-fuse) that will mount the alluxio path specified at alluxio_path to the local file system on the specified mount_point.

For example, the following command will mount the alluxio path /people to the folder /mnt/people in the local file system.

$ integration/fuse/bin/alluxio-fuse mount /mnt/people /people
Starting alluxio-fuse on local host. Alluxio-fuse mounted at /mnt/people. See /lib/alluxio/logs/fuse.log for logs

When alluxio_path is not given, Alluxio-FUSE defaults it to root (/). Note that the mount_point must be an existing and empty directory in your local file system hierarchy and that the user that runs the alluxio-fuse.sh script must own the mount point and have read and write permissions on it. You can mount Alluxio to multiple mount points. All of these alluxio-fuse processes share the same log output at $ALLUXIO_HOME/logs/fuse.log, which is useful for troubleshooting when errors happen on operations under the mount point.

Unmount Alluxio-FUSE

To umount a previously mounted Alluxio-FUSE file system, go to the node where the file system is mounted and run umount from the $ALLUXIO_HOME directory.

$ integration/fuse/bin/alluxio-fuse umount mount_point

This unmounts the file system at the mount point and stops the corresponding Alluxio-FUSE process. For example,

$ integration/fuse/bin/alluxio-fuse umount /mnt/people
Unmount fuse at /mnt/people (PID:97626).

Check the Alluxio-FUSE mount status

To list the Alluxio-FUSE mount points, go to the node where the file system is mounted and run stat from the $ALLUXIO_HOME directory

$ integration/fuse/bin/alluxio-fuse stat

This outputs the pid, mount_point, alluxio_path of all the running Alluxio-FUSE processes.

For example, the output will be like:

pid	mount_point	alluxio_path
80846	/mnt/people	/people
80847	/mnt/sales	/sales

Optional configuration steps

Alluxio-FUSE is based on the standard Java client API alluxio-core-client-fs to perform its operations. You might want to customize the behaviour of the alluxio client used by Alluxio-FUSE the same way you would for any other client application.

One possibility, for example, is to edit $ALLUXIO_HOME/conf/alluxio-site.properties and set your specific alluxio client options. Note that these changes should be made before Alluxio-FUSE starts.

Assumptions and limitations

Currently, most basic file system operations are supported. However, due to Alluxio implicit characteristics, please be aware that:

  • Files can be written only once, only sequentially, and never be modified. That means overriding a file is not allowed, and an explicit combination of delete and then create is needed. For example, cp command will fail when the destination file exists.
  • Alluxio does not have hard-link and soft-link concepts, so the commands like ln are not supported, neither the hardlinks number is displayed in ll output.
  • The user and group are mapped to the Unix user and group. The FUSE integration does not currently support other user/group mappings such as Kerberos. If alluxio.security.group.mapping.class is overridden to a non-shell-based mapping, chown and chgrp will be no-ops, and ll will return the user and group of the user who started the alluxio-fuse process.

Performance considerations

Due to the conjunct use of FUSE and JNR, the performance of the mounted file system is expected to be worse than what you would see by using the Alluxio Java client directly.

Most of the overhead comes from the fact that there are several memory copies going on for each call on read or write operations, and that FUSE caps the maximum granularity of writes to 128KB. This could be probably improved to a large extent by leveraging the FUSE cache write-backs feature introduced in kernel 3.15 (not supported yet, however, by libfuse 2.x userspace libs).

Configuration Parameters For Alluxio-FUSE

These are the configuration parameters for Alluxio-FUSE.

ParameterDefault ValueDescription

Acknowledgements

This project uses jnr-fuse for FUSE on Java.