Chapter 20. Managing Resource Usage

Procedure 20.1. Limiting glusterd resources on RHEL7 based Red Hat Gluster Storage

1. Stop all gluster processes

   # systemctl stop glusterd

   Important

   If glusterd crashes, there is no functionality impact to this crash as it occurs during the shutdown. For more information, see Section 24.3, “Resolving glusterd Crash”

2. Create a service configuration directory for glusterd

   # mkdir /etc/systemd/system/glusterd.service.d

3. Create a service configuration file

   # echo "[Service]
   CPUAccounting=yes
   Slice=glusterfs.slice" >> /etc/systemd/system/glusterd.service.d/99-cpu.conf

4. Create a slice file

   The following defines a slice that sets CPUQuota to the recommended value of 400% (four cores).

   # echo "[Slice]
   CPUQuota=400%" >> /etc/systemd/system/glusterfs.slice

   You can alter the percentage to suit your environment by editing the value in the slice file:

   # systemctl set-property glusterfs.slice CPUQuota=value

5. Restart the system daemon

   # systemctl daemon-reload

6. Start gluster processes

   # systemctl start glusterd