Feature request - OSD with VDO

[paradox55]

Can you add VDO integration as an option when creating an OSD?

Thanks!

 

[Alwin]

A vdo device, is a dm mapped device that does compression and dedup. This will kill the performance of Ceph (if it works at all), if OSDs reside on top of a vdo device. An OSD can use compression and de-duplication is handled in Ceph on a different layer (pool). Besides that it is not upstream in the Linux kernel.

In the article, is a description about a VDO device, but in no combination with Ceph (strangely).
https://ceph.io/geen-categorie/shrinking-your-storage-requirements-with-vdo/

 

[Philippe.Grassia]

+1 on the feature request.
If anybody has data about performance drop please share

 

[jw6677]

Just reviving this old thread, as it appears that ceph has supported VDO devices since as far back as at least Mimic.
See the bottom of the LVM flags documentation:

VDO​

A VDO-enabled device is detected when device is getting prepared, and then stored for later checks when activating. This affects mount options by appending the discard mount flag, regardless of mount flags being used.

Example for an enabled VDO device:
ceph.vdo=1

Source: https://docs.ceph.com/en/mimic/dev/ceph-volume/lvm/

+1 on the feature request of VDO enabled OSD creation.