[LU-13660] ldiskfs integtrated filesystem snapshot Created: 10/Jun/20  Updated: 26/Mar/21

Status: Open
Project: Lustre
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: New Feature Priority: Minor
Reporter: Andreas Dilger Assignee: WC Triage
Resolution: Unresolved Votes: 1
Labels: None

Issue Links:
Duplicate
Related
Rank (Obsolete): 9223372036854775807

 Description   

The performance of ldiskfs+LVM snapshots quickly declines when there are multiple snapshots in existence, since the block-layer snapshot requires data copies for each snapshot copy.

Rather than having a layered ldiskfs on top of a block-layer snapshot device, it would be better to have an integrated fileystem-level snapshot in ldiskfs, more similar to that of ZFS, so that snapshot blocks/inodes can be COW-referenced within the filesystem without a data copy.

It is desirable to have a whole-filesystem snapshot mechanism, unlike DL-Snap in LU-11512, which requires whole-filesystem traversal in order to make an explicit copy of every directory/file being snapshotted. That can take a considerable time for large filesystems, and multiplies metadata consumption for each additional snapshot created.


Generated at Sat Feb 10 03:03:08 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.