Project

General

Profile

Actions
Copy link

Bug #20799

closed

Races when multiple MDS boot at once

Added by Douglas Fuller almost 7 years ago. Updated about 5 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Douglas Fuller
Category:
-
Target version:
-
% Done:

0%

Source:
Development
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(FS):
Labels (FS):
multimds
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

There is a race in MDSRank::starting_done() between MDCache::open_root() and MDLog::start_new_segment()

An MDS in boot state will ignore discover messages, so if another MDS that is booting simultaneously tries to discover the root, it will hang waiting for the reply.

Actions
Copy link

Also available in: Atom PDF