Filesystem (CephFS)

CephFS is a shared POSIX filesystem. Unlike a block volume, which one machine mounts at a time, a CephFS share can be mounted read-write by several clients at once. That makes it a good fit for shared scratch space, common datasets, and applications that expect a normal filesystem path rather than an S3 API.

How it is offered

CephFS shares are reachable only from inside the VDC, over an internal storage network. You attach an OpenStack instance to that network and mount the share from inside the guest. The share is exported by the native CephFS driver, so clients mount it with the standard kernel ceph filesystem type.

CephFS is internal to the VDC, unlike S3

Unlike object storage (S3), which is reachable from anywhere on the internet, CephFS is not exposed outside the VDC. We may be able to provide CephFS access to your virtual machines in OpenStack, but it cannot be mounted from a server on the public internet. If you need shared-filesystem access for a project, contact Sunet so we can size and provision the share with you. If you need network access to your data from outside the VDC, use S3 instead.

When to use what

Want	Use
One disk for one VM	Block storage
A filesystem mounted by several VMs at the same time	CephFS
Network/HTTP access to objects from anywhere	Object storage (S3)

Backup

As with the other Ceph storage types, CephFS is durable against hardware failure but is not backed up by Sunet by default. See Durability, quotas and billing.