Bug 1489439

PR[1] #450 fixes the issue.

One can use the following config file:
#####################
[hosts]
node 1

[pv]
action=create
devices=vdb,vdc

[vg1]
action=create
vgname=RHS_vg1
pvname=vdb,vdc
#####################

Creates a vg as following:

[root@dhcp42-210 ~]# vgs
  VG              #PV #LV #SN Attr   VSize   VFree 
  RHS_vg1           2   0   0 wz--n-  99.99g 99.99g
  rhgs_dhcp42-210   1   2   0 wz--n- <19.00g     0 

[1] https://github.com/gluster/gdeploy/pull/450

Thanks,
Devyani

Recently I had a discussion with Sahina, regarding multiple PVs for a volume group. Sahina would you want this feature in 3.3.1 or 3.4.0?

RHHI setup allows for multiple bricks in a thinpool as we don't require the gluster snapshot feature. This has been tested and qualified in RHHI 1.0

The requirement for multiple disks in a VG is when users want to attach lvmcache (may have only 1 NVMe SSD) to thinpool. If the server has multiple disks these can be formed into 1 VG -> 1 thinpool and cachepool attached to this.

Devyani,

Tested the same with Gdeploy 2.0.2.22 build using the sample conf file provided by you in comment #3. This seems failing for me.


# rpm -qa | grep gdeploy
gdeploy-2.0.2-22.el7rhgs.noarch


# rpm -qa | grep ansible
ansible-2.3.2.0-1.el7.noarch


Gdeploy.conf -

[hosts]
dhcp37-121.lab.eng.blr.redhat.com

[pv]
action=create
devices=sdb,sdc,sdd,vda,vdb

[vg1]
action=create
vgname=RHS_vgs1
pvname=sdb,sdc,sdd,vda,vdb


Output=

************************************

# gdeploy  -c backend_Gdeploysetup2.conf 

PLAY [gluster_servers] ****************************************************************************************************************************************

TASK [Clean up filesystem signature] **************************************************************************************************************************
skipping: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/sdb) 
skipping: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/sdc) 
skipping: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/sdd) 
skipping: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/vda) 
skipping: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/vdb) 

TASK [Create Physical Volume] *********************************************************************************************************************************
changed: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/sdb)
changed: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/sdc)
changed: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/sdd)
changed: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/vda)
changed: [dhcp37-121.lab.eng.blr.redhat.com] => (item=/dev/vdb)

PLAY RECAP ****************************************************************************************************************************************************
dhcp37-121.lab.eng.blr.redhat.com : ok=1    changed=1    unreachable=0    failed=0   


PLAY [gluster_servers] ****************************************************************************************************************************************

TASK [Create volume group on the disks] ***********************************************************************************************************************
changed: [dhcp37-121.lab.eng.blr.redhat.com] => (item={u'brick': u'/dev/sdb', u'vg': u'RHS_vgs1'})
failed: [dhcp37-121.lab.eng.blr.redhat.com] (item={u'brick': u'/dev/sdc', u'vg': u'RHS_vgs1'}) => {"changed": false, "failed": true, "item": {"brick": "/dev/sdc", "vg": "RHS_vgs1"}, "msg": "A volume group called RHS_vgs1 already exists", "rc": 1}
failed: [dhcp37-121.lab.eng.blr.redhat.com] (item={u'brick': u'/dev/sdd', u'vg': u'RHS_vgs1'}) => {"changed": false, "failed": true, "item": {"brick": "/dev/sdd", "vg": "RHS_vgs1"}, "msg": "A volume group called RHS_vgs1 already exists", "rc": 1}
failed: [dhcp37-121.lab.eng.blr.redhat.com] (item={u'brick': u'/dev/vda', u'vg': u'RHS_vgs1'}) => {"changed": false, "failed": true, "item": {"brick": "/dev/vda", "vg": "RHS_vgs1"}, "msg": "A volume group called RHS_vgs1 already exists", "rc": 1}
failed: [dhcp37-121.lab.eng.blr.redhat.com] (item={u'brick': u'/dev/vdb', u'vg': u'RHS_vgs1'}) => {"changed": false, "failed": true, "item": {"brick": "/dev/vdb", "vg": "RHS_vgs1"}, "msg": "A volume group called RHS_vgs1 already exists", "rc": 1}
	to retry, use: --limit @/tmp/tmpt98PU1/vgcreate.retry

PLAY RECAP ****************************************************************************************************************************************************
dhcp37-121.lab.eng.blr.redhat.com : ok=0    changed=0    unreachable=0    failed=1   

Ignoring errors...
*******************************


Cli output-

# pvs
  PV         VG              Fmt  Attr PSize   PFree  
  /dev/sda2  rhel_dhcp37-121 lvm2 a--  <19.00g      0 
  /dev/sdb   RHS_vgs1        lvm2 a--  <20.00g <20.00g
  /dev/sdc                   lvm2 ---   20.00g  20.00g
  /dev/sdd                   lvm2 ---   20.00g  20.00g
  /dev/vda                   lvm2 ---   20.00g  20.00g
  /dev/vdb                   lvm2 ---   20.00g  20.00g
# vgs
  VG              #PV #LV #SN Attr   VSize   VFree  
  RHS_vgs1          1   0   0 wz--n- <20.00g <20.00g
  rhel_dhcp37-121   1   2   0 wz--n- <19.00g      0

Hi manisha,
The config file pasted in comment #3 works for me, with the following output[1].
The PR #450[2] as mentioned in the same comment, fixes the issue.

After comment #8, sachi mentioned that it was fixed in gdeploy-2.0.2-20.
I checked the source code for gdeploy-2.0.2-22, the changes in the patch do not exist.(Will check 2.0.2-20's source code)
If it wasn't working for you, am presuming the patch didn't make it to the build.
So, what we can do for now is we make sure we add this patch in the next build.
We are working on a new build including more fixes, we'll make sure to cherry-pick this patch as well(by next week).

[1] https://paste.opensuse.org/94844657

Cli Output:
[root@dhcp43-218 ~]# vgs
  VG           #PV #LV #SN Attr   VSize   VFree 
  RHS_vg1        2   0   0 wz--n-  99.99g 99.99g
  rhgs_network   1   2   0 wz--n- <19.00g     0 

[2] https://github.com/gluster/gdeploy/pull/450

Thanks Devyani for the update.I haven't tested the same with gdeploy-2.0.2-20.
But accordingly it should work with gdeploy-2.0.2-22.el7rhgs.noarch build as well.


Since its failing for me on gdeploy-2.0.2-22.el7rhgs.noarch ,Based on comment #9 I am moving this BZ to assigned state in order to track the same for the next Gdeploy build ( having this patch included)

ack!

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHEA-2018:1958