Open Stack

Hi,

Cinder backup service discards excess bytes when restoring based on configuration.
https://github.com/openstack/cinder/blob/master/cinder/backup/drivers/ceph.py#L317
https://github.com/openstack/cinder/blob/master/cinder/backup/drivers/ceph.py#L284


Lisa

From: 刘庆 [mailto:winglq at gmail.com]
Sent: Friday, June 24, 2016 11:05 AM
To: openstack-dev at lists.openstack.org
Subject: [openstack-dev] [cinder]restore method while ceph backup backend is used

Hi all,
This is Liu Qing, new to this community. I have a question about the ceph backup backend.
From the cinder/backup/manager.py, the restore process is done by attach the original device(restore dest) to the cinder service host. And in the ceph driver the full_restore will discard the additional space if restore dest(by volume extend) is larger than the backup. There are two ways to discard the unused space, by ceph discard or write zeroes to the unused space. As the dest is attached to the host, the dest is not recognized as an ceph volume. Ceph discard will never be used in volume restore, right? But ceph discard is much more efficient than writing zeroes to the unused space. Is there any plan to use the ceph discard way? Thanks.

Liu Qing
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20160726/0dcc6305/attachment.html>