Just installed a single host HCI with gluster, with only the engine vm running.

Is it expected this situation below?

# virsh -r list
 Id    Name                           State
----------------------------------------------------
 2     HostedEngine                   running

and

# virsh -r dumpxml 2

. . .

    <disk type='file' device='disk' snapshot='no'>
      <driver name='qemu' type='raw' cache='none' error_policy='stop' io='native' iothread='1'/>
      <source file='/var/run/vdsm/storage/e4eb6832-e0f6-40ee-902f-f301e5a3a643/fc34d770-9318-4539-9233-bfb1c5d68d14/b151557e-f1a2-45cb-b5c9-12c1f470467e'>
        <seclabel model='dac' relabel='no'/>
      </source>
      <backingStore/>
      <target dev='vda' bus='virtio'/>
      <serial>fc34d770-9318-4539-9233-bfb1c5d68d14</serial>
      <alias name='ua-fc34d770-9318-4539-9233-bfb1c5d68d14'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x07' function='0x0'/>
    </disk>
. . .

where

# ll /var/run/vdsm/storage/e4eb6832-e0f6-40ee-902f-f301e5a3a643/
total 24
lrwxrwxrwx. 1 vdsm kvm 133 Jan 16 14:44 39df7b45-4932-4bfe-b69e-4fb2f8872f4f -> /rhev/data-center/mnt/glusterSD/10.10.10.216:_engine/e4eb6832-e0f6-40ee-902f-f301e5a3a643/images/39df7b45-4932-4bfe-b69e-4fb2f8872f4f
lrwxrwxrwx. 1 vdsm kvm 133 Jan 16 14:44 5ba6cd9e-b78d-4de4-9b7f-9688365128bf -> /rhev/data-center/mnt/glusterSD/10.10.10.216:_engine/e4eb6832-e0f6-40ee-902f-f301e5a3a643/images/5ba6cd9e-b78d-4de4-9b7f-9688365128bf
lrwxrwxrwx. 1 vdsm kvm 133 Jan 16 15:56 8b8e41e0-a875-4204-8ab1-c10214a49f5c -> /rhev/data-center/mnt/glusterSD/10.10.10.216:_engine/e4eb6832-e0f6-40ee-902f-f301e5a3a643/images/8b8e41e0-a875-4204-8ab1-c10214a49f5c
lrwxrwxrwx. 1 vdsm kvm 133 Jan 16 15:56 c21a62ba-73d2-4914-940f-cee6a67a1b08 -> /rhev/data-center/mnt/glusterSD/10.10.10.216:_engine/e4eb6832-e0f6-40ee-902f-f301e5a3a643/images/c21a62ba-73d2-4914-940f-cee6a67a1b08
lrwxrwxrwx. 1 vdsm kvm 133 Jan 16 14:44 fc34d770-9318-4539-9233-bfb1c5d68d14 -> /rhev/data-center/mnt/glusterSD/10.10.10.216:_engine/e4eb6832-e0f6-40ee-902f-f301e5a3a643/images/fc34d770-9318-4539-9233-bfb1c5d68d14
lrwxrwxrwx. 1 vdsm kvm 133 Jan 16 14:44 fd73354d-699b-478e-893c-e2a0bd1e6cbb -> /rhev/data-center/mnt/glusterSD/10.10.10.216:_engine/e4eb6832-e0f6-40ee-902f-f301e5a3a643/images/fd73354d-699b-478e-893c-e2a0bd1e6cbb

so hosted engine not using libgfapi?

Also, on hosted engine

[root@hciengine ~]# engine-config -g LibgfApiSupported
LibgfApiSupported: false version: 4.1
LibgfApiSupported: false version: 4.2
LibgfApiSupported: false version: 4.3
[root@hciengine ~]#

So that if I import a CentOS7 Atomic host image from glance repo as template and create a new vm from it,when running this VM I get

# virsh -r dumpxml 3

. . .

    <disk type='file' device='disk' snapshot='no'>
      <driver name='qemu' type='qcow2' cache='none' error_policy='stop' io='native' iothread='1'/>
      <source file='/rhev/data-center/mnt/glusterSD/10.10.10.216:_data/601d725a-1622-4dc8-a24d-2dba72ddf6ae/images/e4f92226-0f56-4822-a622-d1ebff41df9f/c6b2e076-1519-433e-9b37-2005c9ce6d2e'>
        <seclabel model='dac' relabel='no'/>
      </source>
      <backingStore/>
      <target dev='vda' bus='virtio'/>
      <serial>e4f92226-0f56-4822-a622-d1ebff41df9f</serial>
      <boot order='1'/>
      <alias name='ua-e4f92226-0f56-4822-a622-d1ebff41df9f'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x06' function='0x0'/>
    </disk>
. . .

I remember there was an "old" bug opened causing this default of not enabling libgfapi

Does this mean it was not solved yet?

If I remember correctly the bugzilla was this one related to HA:

that is still in new status.... since almost 2 years

Is this the only one open?

Thanks,

Gianluca