I have a 4 node setup (centos 7) with hosted engine on glusterfs (replica 3 arbiter1).
Gluster fs is like this
ohost01 104G brick (real data)
ohost02 104g brick (real data)
ohost04 104g brick (arbiter)
ohost05 104g partition used as nfs-storage.
hosted engine is on gluster. I also have an fc domain of 3,6 TB
mount of gluster is like this
storage=172.16.224.10:/engine
mnt_options=backup-volfile-servers=172.16.224.11:172.16.224.13
172.16.224.10 is ohost01 storage network
172.16.224.12 is ohost02 storage network
172.16.224.13 is ohost04 storage network
Today I upgraded all nodes. I did it like this:
hosted-engine was running on ohost05 at the time
Put ohost04 (arbiter) on maintenance and did upgrade (ok)
Same with ohost02
Ohost01 was spm so I did ohost05 spm then put ohost01 on maintenance and then upgraded it.
I notticed that engine VM paused during the process (which usualy does not happen) as I
have backup-volfile-servers mount option. But today I notticed that this option is
ignored. On the hosts I also noted that mount is like this
172.16.224.10:/engine on /rhev/data-center/mnt/glusterSD/172.16.224.10:_engine type
fuse.glusterfs
(rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
so reduduncy is gone from gluster and I cannot figure out why.
If I restart ohost01 (after maintenance) hosted engine get paused until ohost01 comes back
up.
How can I solve this issue