Project

General

Profile

Actions

Feature #47774

closed

Dashboard - Tasks #46756: mgr/dashboard: orchestration workflows & enhancements

orch,cephadm: host search with filters

Added by Ernesto Puerta over 3 years ago. Updated about 2 years ago.

Status:
Resolved
Priority:
Low
Assignee:
Category:
orchestrator
Target version:
% Done:

0%

Source:
Tags:
dashboard
Backport:
Reviewed:
Affected Versions:
Pull request ID:

Description

For the bulk addition of nodes, a search command (or add --dry-run) would be useful. Additionally, for dashboard integration, a list of potential nodes should be presented in order to cherry-pick the specific nodes.

With masks/lengths and fnmatch for exclusion patterns:

> ceph orch host search 192.0.2.0/255.255.255.0 --exclude 192.0.2.1 --exclude 192.0.2.[5-9]

> ceph orch host search 192.0.2.0/24 --exclude 192.0.2.1 --exclude 192.0.2.[5-9]

# or, reusing a existing command:

> ceph orch host add --dry-run 192.0.2.0/255.255.255.0 --exclude 192.0.2.1 --exclude 192.0.2.[5-9]

> ceph orch host add --dry-run 192.0.2.0/255.255.255.0 --exclude 192.0.2.1 --exclude 192.0.2.[5-9]

Ideally this should also involve that the host add command supports wildcard, but it's ok if the hosts are still added one by one.

Actions #1

Updated by Ernesto Puerta over 3 years ago

  • Parent task set to #46756
Actions #2

Updated by Sebastian Wagner about 3 years ago

  • Priority changed from Normal to Low
Actions #3

Updated by Sebastian Wagner about 3 years ago

actually I'd like to go the other way around and enable remote hosts to join the cluster by themselves. Similar to kubeadm

Actions #4

Updated by Sebastian Wagner almost 3 years ago

That would require some magic tricks to generate the list of hosts, like https://docs.ansible.com/ansible/latest/collections/index_inventory.html or similar.I'm somewhat skeptic adding something like this.

Actions #5

Updated by Paul Cuzner almost 3 years ago

Do we really need this? In most enterprise environments I;ve worked in this type of thing gets jumped on by security teams and network admins.

Actions #6

Updated by Ernesto Puerta almost 3 years ago

Paul Cuzner wrote:

Do we really need this? In most enterprise environments I;ve worked in this type of thing gets jumped on by security teams and network admins.

Without this, the only way of adding multiple hosts in large clusters (N ~ 100s) is launch-and-wait or, even worse, forcing users to write scripts (risks are even bigger... from my old days as a Ceph user I sadly remember myself writing a script to delete & create OSDs with asynchronous calls and causing a lots of issues).

Actions #7

Updated by Sebastian Wagner over 2 years ago

  • Status changed from New to Fix Under Review
  • Assignee set to Adam King
  • Pull request ID set to 44020
Actions #8

Updated by Sebastian Wagner over 2 years ago

  • Subject changed from orch,cephadm: host search/add with masks to orch,cephadm: host search with filters
  • Status changed from Fix Under Review to Pending Backport
Actions #9

Updated by Sebastian Wagner about 2 years ago

  • Status changed from Pending Backport to Resolved
Actions

Also available in: Atom PDF