glusterFS Upgrade
다음 문서를 참고한다. https://docs.gluster.org/en/latest/Upgrade-Guide/upgrade_to_4.1/
online 업그레이드와 offline 업그레이드 두 종류가 있으나 online upgrade를 사용할것이다.
online upgrade
- Online upgrade is only possible with replicated and distributed replicate volumes. 우리는 distributed replicate volumes 이므로 진행해도 된다.
- 꼭 서버 한대씩 진행해야한다.
- replication을 꼭 다른서버에 하고 있는경우만 online upgrade가 가능하다.
다음 과정을 모든서버에 반복한다.
- Stop all gluster services, either using the command below, or through other means,
killall glusterfs glusterfsd glusterd systemctl stop glustereventsd - Stop all applications that run on this server and access the volumes via gfapi (qemu, NFS-Ganesha, Samba, etc.)
- 새 GlusterFS를 설치한다.
yum update yum clean all yum install centos-release-gluster -y yum update glusterfs-fuse gluster --version #새버전 확인1.새로 설치된 daemon을 실행한다.
systemctl start glusterd systemctl enable glusterd - self-heal을 실행한다.
for i in `gluster volume list`; do gluster volume heal $i; done - no heal backlog가 있는지 확인한다.
gluster volume heal <volname> info - Restart any gfapi based application stopped previously in step (2)
Post upgrade steps
-
client들도 전부 버전 업데이트를 한다.
-
enable the option
gluster volume set <volname> fips-mode-rchecksum on
Upgrade procedure for clients
- unmount 한다.
umount /glusterfs - Stop all applications that access the volumes via gfapi (qemu, etc.)
- Install new Gluster
- Mount all gluster shares
mount -t glusterfs gluster01:/gfs /glusterfs - Start any applications that were stopped previously in step