Adding an erasure-coded pool to a Ceph File System

Edit online

By default, Ceph uses replicated pools for data pools. However, another erasure-coded data pool to the Ceph File System can be added, as needed.

Before you begin

Before you begin, make sure that you have the following prerequisites in place:

A running IBM Storage Ceph cluster.
An existing Ceph File System.
Pools that use BlueStore OSDs.
Root-level access to a Ceph Monitor node.
The attr package installed.

About this task

Ceph File Systems (CephFS) backed by erasure-coded pools use less overall storage compared to Ceph File Systems that are backed by replicated pools. While erasure-coded pools use less overall storage, they also use more memory and processor resources than replicated pools.

Important: For production environments, use the default replicated data pool for CephFS. The creation of inodes in CephFS creates at least one object in the default data pool. It is better to use a replicated pool for the default data to improve small-object write performance, and to improve read performance for updating backtraces.

For more information, see:

Procedure

Create an erasure-coded data pool for CephFS.

ceph osd pool create DATA_POOL_NAME erasure

For example,

[root@mon ~]# ceph osd pool create cephfs-data-ec01 erasure
pool 'cephfs-data-ec01' created

Verify that the pool was added.
```
[root@mon ~]# ceph osd lspools
```

Enable overwrites on the erasure-coded pool.

ceph osd pool set DATA_POOL_NAME allow_ec_overwrites true

For example,

[root@mon ~]# ceph osd pool set cephfs-data-ec01 allow_ec_overwrites true
set pool 15 allow_ec_overwrites to true

Verify the status of the Ceph File System.

ceph fs status FILE_SYSTEM_NAME

For example,

[root@mon ~]# ceph fs status cephfs-ec
cephfs-ec - 14 clients
=========
RANK  STATE              MDS                ACTIVITY     DNS    INOS   DIRS   CAPS
 0    active  cephfs-ec.example.ooymyq  Reqs:    0 /s  8231   8233    891    921
       POOL           TYPE     USED  AVAIL
cephfs-metadata-ec  metadata   787M  8274G
  cephfs-data-ec      data    2360G  12.1T

       STANDBY MDS
cephfs-ec.example.irsrql
cephfs-ec.example.cauuaj

Add the erasure-coded data pool to the existing CephFS.
```
ceph fs add_data_pool FILE_SYSTEM_NAME DATA_POOL_NAME
```
In the following example, the new data pool, cephfs-data-ec01, is added to the existing erasure-coded file system, cephfs-ec.
```
[root@mon ~]# ceph fs add_data_pool cephfs-ec cephfs-data-ec01
```

Verify that the erasure-coded pool was added to the Ceph File System.

ceph fs status FILE_SYSTEM_NAME

For example,

[root@mon ~]# ceph fs status cephfs-ec
cephfs-ec - 14 clients
=========
RANK  STATE              MDS                ACTIVITY     DNS    INOS   DIRS   CAPS
 0    active  cephfs-ec.example.ooymyq  Reqs:    0 /s  8231   8233    891    921
       POOL           TYPE     USED  AVAIL
cephfs-metadata-ec  metadata   787M  8274G
  cephfs-data-ec      data    2360G  12.1T
 cephfs-data-ec01     data       0   12.1T

       STANDBY MDS
cephfs-ec.example.irsrql
cephfs-ec.example.cauuaj

Set the file layout on a new directory.

mkdir PATH_TO_DIRECTORY
setfattr -n ceph.dir.layout.pool -v DATA_POOL_NAME PATH_TO_DIRECTORY

In the following example, all new files that are created in the /mnt/cephfs/newdir directory inherit the directory layout and places the data in the newly added erasure-coded pool.

[root@mon ~]# mkdir /mnt/cephfs/newdir
[root@mon ~]# setfattr -n ceph.dir.layout.pool -v cephfs-data-ec01 /mnt/cephfs/newdir