SUSE High Availability Cluster Services – How to stop, start or view the status
This blog post aims to summarize the starting and stopping options available for SUSE Linux Enterprise Server (SLES) High Availability (HA) cluster stack operations. I will offer brief answers to a few questions that have been commonly asked by our customers and partners.
Before We Start:
First, some background information, SUSE Linux Enterprise (SLE) High Availability Extension (HAE) team sets a specific systemd unit parameter StopWhenUnneeded=yes
inside the corosync.service
systemd unit configurations file /usr/lib/systemd/system/corosync.service
.
What this means is only systemctl stop pacemaker
command execution will stop the corosync
service when corosync
service thinks it is no longer needed. A common misconception is that SLES HA cluster operators think they could restart the whole cluster stack on one node using systemctl restart pacemaker
command, which is not correct.
Tip
For more information about
StopWhenUnneeded=yes
parameter explanation, please refer tosystemd.unit
man page.
Upcoming Change:
The SLE HAE team plans to drop the parameter StopWhenUnneeded=yes
from SLE 15 SP2 and the newer releases. SLES HA cluster operators are encouraged to use the Cluster Resource Manager Shell (CRMSH) commands to avoid any confusion.
The following table lists the main SLES HA cluster stack starting and stopping operations using CRMSH commands:
SLES HA Cluster Operation | CRMSH Command |
---|---|
Starting the cluster stack on one node | crm cluster start |
Stopping the cluster stack on one node | crm cluster stop |
Restarting the cluster stack on one node | crm cluster restart |
Getting the cluster stack status on one node | crm cluster status |
Note
All of the above-mentioned commands in the above table should be executed by an operation user who has the required privileges or by the
root
user.
Tip
For more info about the above-mentioned
crm
commands, please refer tocrm cluster help
help messages.
Cluster Stack Control using systemctl
command:
Although we at SUSE recommend using CRMSH commands for the SLES HA cluster stack operations, you are still able to use the systemd
commands to start, restart and stop SLES HA cluster stack as follows:
SLES HA Cluster Operation | systemd Commands |
---|---|
Starting the cluster stack on one node | systemctl start pacemaker |
Stopping the cluster stack on one node | systemctl stop corosync |
Restarting the cluster stack on one node | systemctl restart corosync |
Note
- Starting
pacemaker
service will first start thecorosync
one.- Stopping
corosync
service will first stop thepacemaker
one.
Summary:
We at SUSE recommend using CRMSH commands that provide a clear option to control your cluster stack.
No comments yet