Bug 1552560

+++ This bug was initially created as a clone of Bug #1552559 +++

Description of problem:
-----------------------
Single node RHHI deployment requires the gdeploy configuration file to be executed standalone without cockpit interface.

The gdeploy configuration file as found in https://www.ovirt.org/documentation/gluster-hyperconverged/chap-Single_node_hyperconverged/ is recommended for single node RHHI installation.

There are certain corrections required in this script. This bug is for tracking all the required changes for that configuration file

Note: There is already BZ 1552395 to make this configuration file shipped along with gdeploy-templates of cockpit-ovirt-dashboard

Version-Release number of selected component (if applicable):
-------------------------------------------------------------
cockpit-ovirt-dashboard-0.11.12

How reproducible:
----------------
Not applicable

Steps to Reproduce:
-------------------
Not applicable

Actual results:
---------------
Not applicable

Expected results:
-----------------
gdeploy configuration file should deploy the single node RHHI successfully.

Here are the required corrections.
1. Include the script that validates the non-availability of other partitions on the disk

<expected>
[script1]
action=execute
ignore_script_errors=no
file=/usr/share/gdeploy/scripts/grafton-sanity-check.sh -d @disk@
</expected>

2. VDO section is missing
<expected>
[vdo]
action=create
names=@VDO_Volume_Name@
devices=@disk@
logicalsize=@logical_size@T # Note:logicalsize is 10x physical space on disk
#slabsize=32G               # Note: used only when the physical size is few TBs
blockmapcachesize=128M
readcache=enabled
readcachesize=20M
emulate512=enabled
writepolicy=auto
ignore_vdo_errors=no
</expected>

3. Blacklisting the local devices missing
<expected>
[shell2]
action=execute
command=vdsm-tool configure --force

[script3]
action=execute
file=/usr/share/gdeploy/scripts/blacklist_all_disks.sh
ignore_script_errors=no
</expected>

4. gdeploy should terminate if pv, vg, lv creation fails. So all the corresponding sections in the gdeploy config file should have 'ignore_xx_errors=no'

<expected>
[pv]
...
...
ignore_pv_errors=no

[vg]
...
...
ignore_vg_errors=no

[lv]
...
...
ignore_lv_errors=no

</expected>

5. Standard naming convention should be followed - 'gluster_$component_$disk' across different components in config file
For example:
lv -> gluster_lv_sdb
vg -> gluster_vg_sdb
thinpool -> gluster_thinpool_sdb
mount point of bricks -> /gluster_bricks/engine, /gluster_bricks/data, /gluster_bricks/vmstore

6. chunksize is not required as its auto-computed

7. poolmetadatasize is set to 15GB, it should be 16GB ( as per recommendation in RHGS guide )

8. Size with units
Some of the places where size is to be filled in by user has a tag @SIZE@.
It could be helpful, if we can have @SIZE@GB, or @SIZE@G that could be useful

9. lvmcache
There is a section to extend the SSD in to the existing VG.
This is not required. 'setup-cache' implicitly does that.
Please remove this.

10. Add gluster user to QEMU group
This is not relevant for now, but for futuristic approach, when user migrates to gfapi based storage domain access, this is required
<expected>
[shell3]
action=execute
command=usermod -a -G gluster qemu
</expected>