Hi Tony,
with an EC-backed RBD image you always refer to the replicated pool
when it comes to operations. You'll notice that you can't really list
RBDs in the EC pool:
ceph01:~ # rbd -p volumes-hdd-ec ls
ceph01:~ #
But there are in fact lots of images in that pool, just with their
metadata in the replicated pool:
ceph01:~ # rbd info images/volume-765812f0-230f-4029-9f32-e0e9d4c64e12
| grep -E "data_pool|block_name"
data_pool: volumes-hdd-ec
block_name_prefix: rbd_data.1.6a3351848af8da
ceph01:~ # rados -p volumes-hdd-ec ls | grep -c rbd_data.1.6a3351848af8da
17
So everything you want to do with that image, you'll always have to
use the replicated pool to address those images. We use Ceph for
OpenStack as well, and for our EC cinder backend we have a dedicated
user that has a default rbd pool configured (excerpt from ceph.conf on
openstack nodes):
[client.openstack-hdd-ec]
rbd default data pool = volumes-hdd-ec
keyring = /etc/ceph/ceph.client.openstack-hdd-ec.keyring
Hope this helps!
Eugen
Zitat von Tony Liu <tonyliu0592@xxxxxxxxxxx>:
Hi,
https://docs.ceph.com/en/reef/rados/operations/erasure-code/#erasure-coding-with-overwrites
A replica pool is required to store metadata for EC pool?
My case is RBD with OpenStack. With EC pool, the RBD image has to be
created like this?
rbd create --size 1G --data-pool ec_pool replicated_pool/image_name
What about all other RBD operations, like snapshot, import, export, etc.
Thanks!
Tony
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx