Project

General

Profile

Feature #37085

add command to bring cluster down rapidly

Added by Patrick Donnelly about 1 month ago.

Status:
New
Priority:
High
Category:
-
Target version:
Start date:
11/13/2018
Due date:
% Done:

0%

Source:
Development
Tags:
Backport:
Reviewed:
Affected Versions:
Component(FS):
MDSMonitor
Labels (FS):
Pull request ID:

Description

We now have a command to nicely bring the cluster down via `ceph fs set <name> down true`. This does sequential deactivations until all ranks are stopped.

What we need next is a command that does the same process but with "mds fail" in preparation for deleting the file system or just to bring it down immediately for whatever reason. The current process for doing this depends on the number of ranks in the file system which is annoying and makes automation harder:

ceph fs set <name> joinable false
ceph mds fail <name>:<rank> # for each rank

Also available in: Atom PDF