Glusterfs_Upgrade_Procedure.md

#GlusterFS prep procedure

• clone disk from 01
• on 02 and 01,add gluster:server_install recipe
• on 02, apt-get install glusterfs-server
• on 01, service glusterfs-server stop
• on 01, killall glusterfsd updates the brick-port from 24009(3.1.x) to 49152(3.4.x)
• test+verify connection to 02(if fail, start 01 service)
• on 01, apt-get install glusterfs-server
• test+verify connection by writing a temp file
• on 02, apt-get install glusterfs-server
• on 02, apt-get install glusterfs-server
• on 02, killall glusterfsd updates the brick-port from 24009(3.1.x) to 49152(3.4.x)

#GlusterFS Expanding procedure

• Create 2 new servers, attach separate disk(500gb) to be used for glusterFS
• bootstrap in chef
  knife bootstrap <ip_address> -N <fqdn> -E production -r 'recipe[apt::default],recipe[gluster::server_install]' -x ubuntu --sudo
• configure disk on new host(s)

(echo n; echo p; echo 1; echo; echo; echo w) | fdisk /dev/vdb
mkfs.xfs -i size=512 /dev/vdb1
mkdir -p /GlusterFS/brick1
mount -t xfs /dev/vdb1 /GlusterFS/brick1/

• verify connections

ping -c 3 <host03>
ping -c 3 <host04>

• verify current connected peer count is 1
  gluster peer status

• add new host(s) to trusted storage pool

gluster peer probe <host03>
gluster peer probe <host04>

• verify peer connected count is 3
  gluster peer status

• IF Peer Rejected error
  gluster peer detach <node> on 01
service glusterfs-server stop on 03 and 04
  rm -rf /var/lib/gluster/* on 03 and 04
  service glusterfs-server start on 03 and 04
  gluster volume reset <volume>
  retry peer probe

• verify the current volume info

gluster volume info

• add new bricks to volume gluster volume add-brick vol01 <host1>:/GlusterFS/brick1 <host2>:/GlusterFS/brick1

• verify new bricks on the volume(if going from 2 => 4, type should also change from replicated to distributed-replicated) gluster volume info

• rebalance the volume gluster volume rebalance vol01 start