Skip to content

Creating an EC Datastore

Warning

The following Erasure Coding profiles are the only officially supported HyperCloud EC profiles:

* 4 + 2 (k = 4, m = 2) (66.67% Storage Efficiency)
* 8 + 4 (k = 8, m = 4) (66.67% Storage Efficiency)
* 8 + 3 (k = 3, m = 3) (72.73% Storage Efficiency)

For best resiliency, SoftIron recommends k + m + 1, which allows recovery in the case where a whole storage node goes offline. Setting m to a value of 2 will create a system with minimum robustness and increase the risk of inaccessibility.

Danger

The CRUSH failure domain of host is the only supported configuration.

Warning

Deviation from these configurations incurs risk of cluster failure.

This process is manually accomplished via the CLI. To create a Ceph backend pool, you will need to ssh into the dashboard VM and run the following commands:

root@hypercloud-dashboard:~# ceph osd erasure-code-profile set ec83 k=8 m=3  crush-failure-domain=host

root@hypercloud-dashboard:~# ceph osd pool create ec83 2048 2048 erasure ec83
pool 'ec83' created

root@hypercloud-dashboard:~# ceph osd pool set ec83 allow_ec_overwrites true
set pool 12 allow_ec_overwrites to true

root@hypercloud-dashboard:~# rbd pool init --pool ec83

root@hypercloud-dashboard:~# ceph osd pool create ecdata
pool 'ecdata' created

root@hypercloud-dashboard:~# ceph osd pool set ecdata size 3
set pool 13 size to 3

root@hypercloud-dashboard:~# ceph osd pool application enable ecdata rbd
enabled application 'rbd' on pool 'ecdata'

Warning

If rbd is not enabled on the ecdata pool, the following warning from ceph -s will occur:

# ceph -s
cluster:
    id:     bad29d1f-c916-4352-816e-499df8b276ae
    health: HEALTH_WARN
            1 pool(s) do not have an application enabled

After this initial setup, you will then create a template:

  1. Navigate to /dashboard/staging/tmp

  2. Create a file for the template (e.g. tmp.tmpl)

    NAME         = ecdata
DS_MAD       = ceph
TM_MAD       = ceph
DISK_TYPE    = rbd
BRIDGE_LIST  = hypercloud-dashboard
POOL_NAME    = ecdata
EC_POOL_NAME = ec83
STAGING_DIR  = /dashboard/staging/tmp

  3. Create a second template file (e.g. tmp.tmpl2)

    BRIDGE_LIST         = "hypercloud-dashboard"
DISK_TYPE           = "RBD"
CLONE_TARGET        = "SELF"
CLONE_TARGET_SHARED = "SELF"
CLONE_TARGET_SSH    = "SYSTEM"
DISK_TYPE_SHARED    = "RBD"
DISK_TYPE_SSH       = "FILE"
LN_TARGET           = "NONE"
LN_TARGET_SHARED    = "NONE"
LN_TARGET_SSH       = "SYSTEM"
TM_MAD_SYSTEM       = "ssh,shared"

  4. Run the following commands to create the templates for the datastore:

    root@hypercloud-dashboard:~# hypercloud datastore create tmp.tmpl 
root@hypercloud-dashboard:~# hypercloud datastore update ecdata tmp.tmpl2 --append

  5. The newly created datastore is now available for use.