Installing Blobfuse in Ubuntu to mount Azure Blob as file share

Azure Blob is around a factor x3 cheaper than Azure Files, it doesn't have the current 5 TB limitation that Azure Files has (100 TB is in tech preview), and it has tiering (hot/cold/archive) to reduce storage prices even further. The problem is, though, that Blob is object storage and not block storage so if you need to mount Blob as a file system from e.g. a Linux server you have to put something in front of it.

Avere and Blobfuse lets you do this. With Blofuse it's important to note that it's not 100% POSIX compliant, see link here for limitations. This means that linux permissions won't apply (chmod, chown) and symbolic links will not work either. We will be given full read/right permissions to the container (as the storage account key is used for authentication).

As a side note, you can do copying of data from CLI with Azcopy v10 and Rclone but they work with their own set of commands and do not remote mount Blob as a folder or share.

Prerequisites:

Note that to set this up you need an SSD disk or a dedicated portion of memory (as a cache) in the VM where you set this up. This example will show how to use memory as a ram cache.

Also, you need to have a storage account created in MS Azure (that we will mount).

This MS guide has been used to set up Blobfuse. Ubuntu 18.04 has been used in this setup but the guide shows how to use other OS'es as well.

Install/configure:

Ssh or log in to your server

Su to root:

$ sudo su -

Configure the Microsoft package repository:

# wget https://packages.microsoft.com/config/ubuntu/18.04/packages-microsoft-prod.deb

(or if you're on Ubuntu 16.04: https://packages.microsoft.com/config/ubuntu/16.04/packages-microsoft-prod.deb)

# dpkg -i packages-microsoft-prod.deb
# apt-get update

To install Blobfuse:

# apt-get install blobfuse

Mount an 8 GB temp ram drive for cache (the VM used for this has 16 GB so I'm using half for the cache):

# mkdir /mnt/ramdisk

# mount -t tmpfs -o size=8g tmpfs /mnt/ramdisk

# mkdir /mnt/ramdisk/blobfusetmp

# chown jakob /mnt/ramdisk/blobfusetmp

Go to your home folder:

# cd /home/jakob

Create below file (it doesn't have to be in your home folder but it can be):

# touch fuse_connection.cfg

Edit the file with nano:

The below

# nano fuse_connection.cfg

Add the following content to the file (this info can be found in Azure Portal -> Storage Accounts -> click your storage account -> Access keys. The container name is the name of the virtual folder that you're mounting. This container should already have been created either on in the Azure Portal or the Azure Storage Explorer):

accountName jakobsstorageaccount

accountKey wbMTSWXXXXXXXXXXXXXXXXXXusRtAA==

containerName jakobscontainer1

Save the file and exit.

Change permissions so only root can access it:

# chmod 700 fuse_connection.cfg

Make a folder where the content of container will be mounted:  

# mkdir /mnt/blobfuse/jakobscontainer1

Mount the container:

# blobfuse mnt/blobfuse/jakobscontainer1 --tmp-path=/mnt/ramdisk/blobfusetmp --config-file=/home/jakob/fuse_connection.cfg -o attr_timeout=240 -o entry_timeout=240 -o negative_timeout=120

That's it. Now test that you can list the content of the Blob container:

# ls -alh /mnt/blobfuse/jakobscontainer1

From here you can use cp or Rsync to copy files back and forth.