Project

General

Profile

Bug #40649

set_mon_vals failed to set cluster_network = 10.1.2.0/24: Configuration option 'cluster_network' may not be modified at runtime

Added by Vandeir Eduardo over 4 years ago. Updated about 4 years ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
% Done:

0%

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

Description

When using any rbd command on a client machine, for example, "rbd ls poolname", those messages are always displayed:

2019-07-03 10:12:06.595 7f5165661700 -1 set_mon_vals failed to set cluster_network = 10.1.2.0/24: Configuration option 'cluster_network' may not be modified at runtime
2019-07-03 10:12:06.595 7f5165661700 -1 set_mon_vals failed to set public_network = 10.1.1.0/24: Configuration option 'public_network' may not be modified at runtime
2019-07-03 10:12:06.599 7f5164e60700 -1 set_mon_vals failed to set cluster_network = 10.1.2.0/24: Configuration option 'cluster_network' may not be modified at runtime
2019-07-03 10:12:06.599 7f5164e60700 -1 set_mon_vals failed to set public_network = 10.1.1.0/24: Configuration option 'public_network' may not be modified at runtime

After this, the normal output of the command (rbd names), is displayed.

If I run the same command on a ceph node, those "warning/informational??" messages are not displayed.

I only found something similar to this behavior on this thread: https://www.spinics.net/lists/ceph-devel/msg42656.html

Already tried commands: "ceph config rm global cluster_network" and "ceph config rm global public_network" without success.

I remember, at some point in cluster configuration, that I used the command "ceph config assimilate-conf ..."

Client machine ceph.conf:
[global]
fsid = xxxxxxxxxxxxxxxxxxxxxxxxxxx
mon_host = [v2:10.1.1.41:3300/0,v1:10.1.1.41:6789/0] [v2:10.1.1.42:3300/0,v1:10.1.1.42:6789/0] [v2:10.1.1.43:3300/0,v1:10.1.1.43:6789/0]
auth_client_required = cephx

Ceph nodes ceph.conf:
fsid = xxxxxxxxxxxxxxxxxxxxxxxxxxx
mon_initial_members = ceph2a,ceph2b,ceph2c
mon_host = 10.1.1.41,10.1.1.42,10.1.1.43
auth_cluster_required = cephx
auth_service_required = cephx
auth_client_required = cephx
public_network = 10.1.1.0/24
cluster_network = 10.1.2.0/24

rbd_default_features = layering,striping,exclusive-lock,object-map,fast-diff
osd_pool_default_size = 2
osd_pool_default_min_size = 1

osd_objectstore = bluestore
osd_memory_target = 2147483648

[osd]
osd_scrub_chunk_min = 1
osd_scrub_chunk_max = 1
osd_scrub_sleep = 0.1
osd_scrub_begin_hour = 23
osd_scrub_end_hour = 7

[mon]
mon_osd_down_out_subtree_limit = host

Output of "ceph config dump":
WHO MASK LEVEL OPTION VALUE RO
global advanced auth_client_required cephx *
global advanced auth_cluster_required cephx *
global advanced auth_service_required cephx *
global advanced cluster_network 10.1.2.0/24 *
global basic osd_memory_target 2147483648
global advanced osd_objectstore bluestore *
global advanced osd_pool_default_min_size 1
global advanced osd_pool_default_size 2
global advanced public_network 10.1.1.0/24 *
global advanced rbd_default_features 31
mon advanced mon_osd_down_out_subtree_limit host
osd advanced osd_scrub_begin_hour 23
osd advanced osd_scrub_chunk_max 1
osd advanced osd_scrub_chunk_min 1
osd advanced osd_scrub_end_hour 7
osd advanced osd_scrub_sleep 0.100000

Ceph version:
ceph version 14.2.1 (d555a9489eb35f84f2e1ef49b77e19da9d113972) nautilus (stable)


Related issues

Related to RADOS - Bug #40282: set_mon_vals failed to set New 06/11/2019

History

#1 Updated by Vandeir Eduardo over 4 years ago

As described by Manuel Rios in https://tracker.ceph.com/issues/40282 , the workaround is include configs:

public_network = 10.1.1.0/24
cluster_network = 10.1.2.0/24

also into ceph.conf on client machines.

#2 Updated by Vandeir Eduardo over 4 years ago

Anyway, I generated the ceph.conf file used on client machines using the command "ceph config generate-minimal-conf".

The configs public_network e cluster_network are not included by the command.

If the correct configuration on client machines should include those configs, I think the command to generate de minimal config should have included them.

#3 Updated by Greg Farnum over 4 years ago

  • Project changed from Ceph to RADOS

#4 Updated by Frank Ritchie about 4 years ago

This also happened to me during an upgrade from Luminous to Nautilus.

The cluster/public networks were not defined when I ran

ceph config generate-minimal-conf
and
ceph config assimilate-conf

Now I have global read only settings in the central config database producing messages with every ceph command:


-1 set_mon_vals failed to set cluster_network = 10.192.80.0/24: Configuration option 'cluster_network' may not be modified at runtime
-1 set_mon_vals failed to set public_network = 10.192.80.0/24: Configuration option 'public_network' may not be modified at runtime


  1. ceph config dump | grep network
    global advanced cluster_network 10.192.80.0/24 *
    global advanced public_network 10.192.80.0/24 *

There doesn't seem to be a way to remove these settings.

#5 Updated by Frank Ritchie about 4 years ago

FYI, I was able to remove the config settings with:

$ ceph config rm <who> <what>

followed by

$ ceph config key rm <key>

#6 Updated by Nathan Cutler about 4 years ago

  • Related to Bug #40282: set_mon_vals failed to set added

Also available in: Atom PDF