Unable to move cluster resource using PCS

Question

I am using the corosync, pacemaker and PCS tool to manage the cluster service for SMTP outgoing servers.

I am finding a weird issue, when I try to move the cluster using resource group, it throws an error as below. not able to identify what I am missing.

[~]# pcs resource move prmcnzzeml02 Error: Unable to move resource 'ausprmcnzzeml02' using a location constraint. Current location of the resource may be affected by some other constraint. 

This is the pcs status:

[root@prnnxeml02 ~]# pcs status Cluster name: PRNZZEML02 Cluster Summary: * Stack: corosync (Pacemaker is running) * Current DC: prnnxeml02.host.com (version 2.1.7-5.0.1.el9_4-0f7f88312) - partition with quorum * Last updated: Mon Jun 24 18:05:09 2024 on prnnxeml02.host.com * Last change: Mon Jun 24 15:04:19 2024 by root via root on prnnxeml02.host.com * 2 nodes configured * 3 resource instances configured Node List: * Online: [ prnnxeml02.host.com prnnyeml02.host.com ] Full List of Resources: * Resource Group: prnzzeml02: * Address (ocf:heartbeat:IPaddr2): Started prnnxeml02.host.com * Sendmail (systemd:sendmail): Started prnnxeml02.host.com * OpenDkim (systemd:opendkim): Started prnnxeml02.host.com Daemon Status: corosync: active/enabled pacemaker: active/enabled pcsd: active/enabled 

[root@prmcnnxeml02 ~]# pcs constraint config --full Colocation Constraints: resource 'Sendmail' with resource 'Address' (id: colocation-Sendmail-Address-INFINITY) score=INFINITY resource 'OpenDkim' with resource 'Sendmail' (id: colocation-OpenDkim-Sendmail-INFINITY) score=INFINITY Order Constraints: start resource 'Address' then start resource 'Sendmail' (id: order-Address-Sendmail-Mandatory) kind=Mandatory start resource 'Sendmail' then start resource 'OpenDkim' (id: order-Sendmail-OpenDkim-Mandatory) kind=Mandatory 

Both server have similar versions installed. prnnxeml02 [root@prnnxeml02 ~]# yum list installed | grep pacemaker pacemaker.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-cli.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-cluster-libs.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-libs.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-schemas.noarch 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons [root@prnnxeml02 ~]# [root@prnnxeml02 ~]# [root@prnnxeml02 ~]# yum list installed | grep corosync corosync.x86_64 3.1.8-1.el9 @oraclelinux9-x86_64-addons corosynclib.x86_64 3.1.8-1.el9 @oraclelinux9-x86_64-appstream [root@prnnxeml02 ~]# [root@prnnxeml02 ~]# [root@prnnxeml02 ~]# yum list installed | grep pcs pcs.x86_64 0.11.7-2.el9_4 @oraclelinux9-x86_64-addons prnnyeml02 [root@prnnyeml02 ~]# yum list installed | grep pacemaker pacemaker.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-cli.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-cluster-libs.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-libs.x86_64 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons pacemaker-schemas.noarch 2.1.7-5.0.1.el9_4 @oraclelinux9-x86_64-addons [root@prnnyeml02 ~]# [root@prnnyeml02 ~]# [root@prnnyeml02 ~]# yum list installed | grep corosync corosync.x86_64 3.1.8-1.el9 @oraclelinux9-x86_64-addons corosynclib.x86_64 3.1.8-1.el9 @oraclelinux9-x86_64-appstream [root@prnnyeml02 ~]# [root@prnnyeml02 ~]# [root@prnnyeml02 ~]# yum list installed | grep pcs pcs.x86_64 0.11.7-2.el9_4 @oraclelinux9-x86_64-addons 

Let me know if you need any more details from pacemaker, corosync logs, I can paste it.

Answer 1

You don't want any additional colocation and ordering constraints. In your configuration, the resource group itself will do the heavy lifting. It will ensure that all of its resources are running on the same node and that they are started in order (and stopped in the reverse order).