How to start additional self-heal daemon to speed up the healing progress in RHGS ?

Environment

• Red Hat Enterprise Linux 7.x
• Red Hat Gluster Storage 3.x

Issue

• Healing progress is slow and taking long time to heal remaining entries.
• Is there way to speed up healing as volume option settings not helping much ?

Resolution

1. Get the UUID of the gluster node

   # cat /var/lib/glusterd/glusterd.info|grep UUID|awk -F'=' '{print $2}'
   

   Copy output of above command as a UUID of the node and to be used in <gluster-node-uuid>

2. Generate new socketid

   # uuidgen|sed 's/-//g'
   

   Copy output of above command as a socketid to be used in <glustershd-socketid>

3. Run below command to start additional self-heal daemon:

   # /usr/sbin/glusterfs -s localhost --volfile-id gluster/glustershd -p /var/run/gluster/glustershd/<glustershd-pidfile> -l /var/log/glusterfs/<glustershd-logfile> -S /var/run/gluster/<glustershd-socketid>.socket --xlator-option *replicate*.node-uuid=<gluster-node-uuid>
   

   Replace below values with new and earlier captured data.
   <glustershd-pidfile> : New pidfile to be used by self-heal daemon
   <glustershd-logfile> : New logfile to be used by self-heal daemon
   <glustershd-socketid> : Captured in uuidgen command
   <gluster-node-uuid> : Captured from glusterd.info file

Root Cause

Sometimes number of files to be healed will be too many and single self-heal daemon is not sufficient to process these unhealed entries and may take significant time to heal it. We can fork out new self-heal daemon to parallely start healing pending entries.

Diagnostic Steps

• Verify that additional self-heal daemon process is running on node

  # ps -aef|grep glustershd
  

This solution is part of Red Hat’s fast-track publication program, providing a huge library of solutions that Red Hat engineers have created while supporting our customers. To give you the knowledge you need the instant it becomes available, these articles may be presented in a raw and unedited form.