Bug #11504: CephFS restriction on removing cache tiers is overly strict - CephFS - Ceph

Actions

Copy link

Bug #11504

closed

CephFS restriction on removing cache tiers is overly strict

Added by John Spray almost 9 years ago. Updated over 8 years ago.

Status:

Resolved

Priority:

Low

Assignee:

Category:

Target version:

% Done:

Source:

Community (user)

Tags:

Backport:

hammer

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Component(FS):

Labels (FS):

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

We prevent removing any cache tier that sits on a cephfs pool, even if the tier has been safely set to 'forward' and flushed.

We should permit removal of the tier if it is set to forward and PGMonitor statistics indicate that it is empty.

Related issues 1 (0 open — 1 closed)

Actions

Copy link

Updated by John Spray almost 9 years ago

Source changed from other to Community (user)

Ref http://comments.gmane.org/gmane.comp.file-systems.ceph.user/19795

Actions

Copy link

Updated by Greg Farnum almost 9 years ago

Mmm, the thing to be careful of here is that we can't remove cache tiers if they're over an EC pool. That whole saga is presumably why the code is overly-restrictive to begin with.

Actions

Copy link

Updated by John Spray almost 9 years ago

Oh yeah, now I remember why we added it :-)

So the condition for permitting removal is that it's set to forward or readonly, it's flushed, and the underlying pool is usable by cephfs (i.e. replicated)

Actions

Copy link