Project

General

Profile

Actions
Copy link

Bug #44003

closed

cephadm: multiple mgrs scheduled on same host

Added by Sage Weil over 4 years ago. Updated about 4 years ago.

Status:
Resolved
Priority:
Urgent
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

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

Description

By default the mgr binds to a bunch of random ports. We shouldn't put more than one on a single host (unless forced to).

This is mostly handled if you are scheduling new mgr instances. But if some already exist, they tend to be laid down on top of the same hosts.

This results in errors like

[ERR] MGR_MODULE_ERROR: 2 mgr modules have failed
    Module 'dashboard' has failed: Timeout('Port 8443 not free on ::.',)
    Module 'prometheus' has failed: Timeout('Port 9283 not free on ::.',)

Related issues 1 (0 open — 1 closed)

Blocked by Orchestrator - Bug #44019: cephadm: rgw update doesn't workResolvedSebastian Wagner

Actions
Actions
Copy link
 #1

Updated by Sebastian Wagner over 4 years ago

this is also true for other daemons, like RGWs, etc. also afaik MONs share a default port.

Actions
Copy link
 #2

Updated by Sebastian Wagner over 4 years ago

  • Blocked by Bug #44019: cephadm: rgw update doesn't work added
Actions
Copy link
 #3

Updated by Sebastian Wagner over 4 years ago

A fix for this should make use of https://github.com/ceph/ceph/pull/33205

Actions
Copy link
 #4

Updated by Sage Weil about 4 years ago

  • Status changed from New to Resolved
Actions
Copy link

Also available in: Atom PDF