Project

General

Profile

Feature #61903

pybind/mgr/volumes: add config to turn off subvolume deletion

Added by Patrick Donnelly 8 months ago. Updated about 2 months ago.

Status:
Fix Under Review
Priority:
High
Assignee:
Category:
Administration/Usability
Target version:
% Done:

0%

Source:
Development
Tags:
Backport:
reef,quincy,pacific
Reviewed:
Affected Versions:
Component(FS):
mgr/volumes
Labels (FS):
task(easy), task(intern)
Pull request ID:

Description

Sometimes we want to be able to turn off asynchronous subvolume deletion during cluster recovery scenarios. Add a config to turn the threads off.

History

#1 Updated by Patrick Donnelly 8 months ago

  • Tracker changed from Bug to Feature

#2 Updated by Venky Shankar 8 months ago

Patrick Donnelly wrote:

Sometimes we want to be able to turn off asynchronous subvolume deletion during cluster recovery scenarios. Add a config to turn the threads off.

Maybe add functionality to scale down the purge threads and then setting it to 0 would disable async purge.

Also, how about doing the same for async threads? Those are heavy on the file system too.

#3 Updated by Patrick Donnelly 8 months ago

Venky Shankar wrote:

Patrick Donnelly wrote:

Sometimes we want to be able to turn off asynchronous subvolume deletion during cluster recovery scenarios. Add a config to turn the threads off.

Maybe add functionality to scale down the purge threads and then setting it to 0 would disable async purge.

I'm not a fan of doing that because "restoring" the setting to a sane default is left to the admin who doesn't know better.

Also, how about doing the same for async threads? Those are heavy on the file system too.

You mean the cloner threads? i thought we already have a config for that.

#4 Updated by Venky Shankar 7 months ago

Patrick Donnelly wrote:

Venky Shankar wrote:

Patrick Donnelly wrote:

Sometimes we want to be able to turn off asynchronous subvolume deletion during cluster recovery scenarios. Add a config to turn the threads off.

Maybe add functionality to scale down the purge threads and then setting it to 0 would disable async purge.

I'm not a fan of doing that because "restoring" the setting to a sane default is left to the admin who doesn't know better.

Also, how about doing the same for async threads? Those are heavy on the file system too.

You mean the cloner threads? i thought we already have a config for that.

I meant supporting turning it off (by setting it to zero).

#5 Updated by Patrick Donnelly 7 months ago

  • Assignee deleted (Patrick Donnelly)

#6 Updated by Venky Shankar 7 months ago

  • Assignee set to Rishabh Dave

Rishabh, please take this one.

#7 Updated by Venky Shankar 7 months ago

From GChat

Rishabh Dave, 1:40 PM
https://tracker.ceph.com/issues/61903
Should I add an option to turn off threading or add an option to allow specifying number of thread user wants?

Right - an option that disables async subvolume deletion. As Patrick mentioned in note-3, there isn't a need to allow scaling the number of purge threads.

#8 Updated by Venky Shankar 7 months ago

Venky Shankar wrote:

From GChat

Rishabh Dave, 1:40 PM
https://tracker.ceph.com/issues/61903
Should I add an option to turn off threading or add an option to allow specifying number of thread user wants?

Right - an option that disables async subvolume deletion. As Patrick mentioned in note-3, there isn't a need to allow scaling the number of purge threads.

Rishabh, another point to add to this - One the operator disables async subvolume deletions, there needs to be a way for the operator to check if the purge threads have backed out performing the deletes and nothing related to async deletions are in-progress. This is required IMO as disabling async deletions is essentially asking the purge threads to voluntarily back-out doing purges.

#9 Updated by Rishabh Dave about 2 months ago

  • Status changed from New to Fix Under Review

Also available in: Atom PDF