OS Recommendations

Ceph Dependencies

As a general rule, we recommend deploying Ceph on newer releases of Linux.We also recommend deploying on releases with long-term support.

Linux Kernel

• Ceph Kernel Client

If you are using the kernel client to map RBD block devices or mountCephFS, the general advice is to use a “stable” or “longtermmaintenance” kernel series provided by either http://kernel.org oryour Linux distribution on any client hosts.

For RBD, if you choose to track long-term kernels, we currently recommend4.x-based “longterm maintenance” kernel series:

• 4.19.z

• 4.14.z

For CephFS, see CephFS best practices for kernel version guidance.

Older kernel client versions may not support your CRUSH tunables profileor other newer features of the Ceph cluster, requiring the storage clusterto be configured with those features disabled.

Platforms

The charts below show how Ceph’s requirements map onto various Linuxplatforms. Generally speaking, there is very little dependence onspecific distributions aside from the kernel and system initializationpackage (i.e., sysvinit, upstart, systemd).

Nautilus (14.2.z)

Luminous (12.2.z)

Notes

• 1: The default kernel has an older version of btrfs that we do notrecommend for ceph-osd storage nodes. We recommend using XFS.

• 2: The default kernel has an old Ceph client that we do not recommendfor kernel client (kernel RBD or the Ceph file system). Upgrade to arecommended kernel.

• 3: The default kernel regularly fails in QA when the btrfsfile system is used. We do not recommend using btrfs forbacking Ceph OSDs.

Testing

• B: We build release packages for this platform. For some of theseplatforms, we may also continuously build all ceph branches and exercisebasic unit tests.

• I: We do basic installation and functionality tests of releases on thisplatform.

• C: We run a comprehensive functional, regression, and stress test suiteon this platform on a continuous basis. This includes development branches,pre-release, and released code.