Bug #42779: mgr/dashboard: cluster configuration 'boolean' and other improvements - Dashboard - Ceph

Actions

Copy link

Bug #42779

closed

mgr/dashboard: cluster configuration 'boolean' and other improvements

Added by Ernesto Puerta over 4 years ago. Updated about 3 years ago.

Status:

Duplicate

Priority:

Normal

Assignee:

Category:

Target version:

% Done:

Source:

Tags:

Backport:

Regression:

Severity:

3 - minor

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

On QE we have found 2 UX/UI issues with this.

As configuration options already indicate which services are consuming each configuration option, we shouldn't allow users modify options that are going to have no effect (like enabling mon_allow_pool_delete on mgr level). However, Ceph core does only provide a 'service' metadata, which sometimes seems to be enough (for mon_allow_pool_delete, services = [mon]), but not for others (rbd_move_to_trash_on_remove has services = [rbd], but rbd is not a valid who in ceph config set <who> rbd_move_to_trash_on_remove ...)

Additionally, for boolean config options, the check boxes seem a bit misleading (QE engineers reported that they 'felt' that booleans couldn't be changed). For non boolean settings the layour is a bit different and more intuitive:

In the following screenshot, both issues are depicted:

A proposal for fixing both things could be as follows:

Files

Download all files

service_mon.png (21.2 KB) service_mon.png		Ernesto Puerta, 11/12/2019 05:55 PM
boolean_proposal.png (24.1 KB) boolean_proposal.png		Ernesto Puerta, 11/12/2019 06:09 PM
client_cache_size.png (37.8 KB) client_cache_size.png		Ernesto Puerta, 11/12/2019 06:15 PM

Related issues 2 (1 open — 1 closed)

Actions

Copy link

Updated by Tatjana Dehler over 4 years ago

Ernesto Puerta wrote:

On QE we have found 2 UX/UI issues with this.

As configuration options already indicate which services are consuming each configuration option, we shouldn't allow users modify options that are going to have no effect (like enabling mon_allow_pool_delete on mgr level). However, Ceph core does only provide a 'service' metadata, which sometimes seems to be enough (for mon_allow_pool_delete, services = [mon]), but not for others (rbd_move_to_trash_on_remove has services = [rbd], but rbd is not a valid who in ceph config set <who> rbd_move_to_trash_on_remove ...)

Is already filed here: https://tracker.ceph.com/issues/37669

Additionally, for boolean config options, the check boxes seem a bit misleading (QE engineers reported that they 'felt' that booleans couldn't be changed).

We also need to consider https://tracker.ceph.com/issues/41776 when improving the config editor for boolean values. Maybe we can use the switch supporting three states (true, false, default)

Actions

Copy link