OpenEBS Local PV RawFile

Overview

There are a few reasons to consider using node-based (rather than network-based) storage architecture:

• Performance: Almost no network-based storage solution can keep up with baremetal disk performance in terms of IOPS/latency/throughput combined. And you’d like to get the best out of the SSD you’ve got!
• On-premise Environment: You might not be able to afford the cost of upgrading all your networking infrastructure, to get the best out of your network-based storage solution.
• Complexity: Network-based solutions are distributed systems. And distributed systems are not easy! You might want to have a system that is easier to understand and to reason about. Also, with less complexity, you can fix unpredicted issues more easily.

Warning
We’re still pre-v1, meaning our helm api or storage class parameters might have breaking changes

RawFile vs HostPath

Another OpenEBS provisioner, the LocalPV-HostPath makes it pretty easy to automatically provision HostPath PVs and use them in your workloads. That being said, it has the following known limitations:

• You can’t monitor volume usage: There are hacky workarounds to run “du” regularly, but that could prove to be a performance killer, since it could put a lot of burden on your CPU and cause your filesystem cache to fill up. Not really good for a production workload.
• You can’t enforce hard limits on your volume’s size: Again, you can hack your way around it, with the same caveats.
• You are stuck with whatever filesystem your kubelet node is offering.
• You can’t customize your filesystem.

Important
All the above issues stem from the same root cause:
HostPath/LocalPVs are simple bind-mounts from the host filesystem into the pod.

The idea behind RawFile-LocalPV

To use a Filesystem based extent file as the emulated block device (i.e. a soft-LUN block device), and leverage LINUX loop devices to associate that soft-LUN file as a complete flexible block device (i.e. an emulated soft disk device).

At this point you can create a PV with a filesystem on it, which adds the following benefits:

• You can monitor volume usage by running df -hT in O(1) since each soft-LUN block device is mounted separately on the local node (displaying utilization status/metrics or each mount point).
• The size limit is enforced by the operating system, based on the backing file system capacity and soft-lun device file size.
• Since volumes are backed by different files, each soft-lun device file can be formatted using different filesystems, and/or customized with different filesystem options.

How to Use

Please follow the instructions from the Install Guide.

Features

• Direct I/O: Near-zero disk performance overhead
• Dynamic provisioning
• Enforced volume size limit
• Access Modes
  • ReadWriteOnce
  • ReadWriteOncePod
  • ReadOnlyMany
  • ReadWriteMany
• Volume modes
  • Filesystem mode
  • Block mode
    • The readOnly attribute in the PVC template is not currently handled properly
• Volume metrics
• Supports fsTypes: ext4, btrfs, xfs
• Online expansion: If fs supports it (e.g. ext4, btrfs, xfs)
• Online shrinking: If fs supports it (e.g. btrfs)
• Offline expansion/shrinking
• Volume Snapshots
  • Filesystem-level snapshots: btrfs supported
  • Restores and Clones are not currently implemented