There are really only two ways to do snapshots that I know of and they have trade-offs:
COW into the snapshot (like VMware, Ceph, etc):
When a write is committed, the changes are committed to a diff file and the base file is left untouched. This only has a single write penalty, if you want to discard the child, it is fast as you just delete the diff file. The negative side effects is that reads may have to query each diff file before being satisfied, and if you want to delete the snapshot, but keep the changes (merge the snapshot into the base), then you have to copy all the diff blocks into the base image.
COW into the base image (like most Enterprise disk systems with snapshots for backups):
When a write is committed, the system reads the blocks to be changed out of the base disk and places those original blocks into a diff file, then writes the new blocks directly into the base image. The pros to this approach is that snapshots can be deleted quickly and the data is "merged" already. Read access for the current data is always fast as it only has to search one location. The cons are that each write is really a read and two writes, recovering data from a snapshot can be slow as the reads have to search one or more snapshots.
My experience is that you can't have your cake and eat it too. If you have the choice, you choose the option that fits your use case best. Ceph doesn't have the ability to select which snapshot method it uses (most systems don't).
I hope that helps explain why the request is not easily fulfilled.
On Tue, Dec 16, 2014 at 9:04 AM, Alexandre DERUMIER <aderumier@xxxxxxxxx> wrote:
Hi,
>>That is normal behavior. Snapshotting itself is a fast process, but
>>restoring means merging and rolling back.
Any future plan to add something similar to zfs or netapp,
where you can instant rollback a snapshot ?
(Not sure it's technically possible to implement such snapshot with distributed storage)
----- Mail original -----
De: "aderumier" <aderumier@xxxxxxxxx>
À: "Wido den Hollander" <wido@xxxxxxxx>
Cc: "ceph-users" <ceph-users@xxxxxxxxxxxxxx>
Envoyé: Mardi 16 Décembre 2014 17:02:12
Objet: Re: rbd snapshot slow restore
Alexandre Derumier
Ingénieur système et stockage
Fixe : 03 20 68 90 88
Fax : 03 20 68 90 81
45 Bvd du Général Leclerc 59100 Roubaix
12 rue Marivaux 75002 Paris
MonSiteEstLent.com - Blog dédié à la webperformance et la gestion de pics de trafic
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
