[ovirt-users] Ovirt Node 4.1.1.1 installation, engine on gluster, existing disk sdb not found or filtered, deployment fails

[Oliver Dietzel]

Thx a lot, i already got rid of the multipaths.

Now 5 tries later i try to understand who disk space calc works.

I already understand that the combined GByte limit for my drive sdb is around 530.
 
> sdb                                                        8:16   0 558,9G  0 disk

Now the thin pool creation kicks me! :)

(i do a  vgremove gluster_vg_sdb on all hosts and reboot all three hosts between retries)

TASK [Create LVs with specified size for the VGs] ******************************
failed: [hv1.iw.rto.de] (item={u'lv': u'gluster_thinpool_sdb', u'size': u'530GB', u'extent': u'100%FREE', u'vg': u'gluster_vg_sdb'}) => {"failed": true, "item": {"extent": "100%FREE", "lv": "gluster_thinpool_sdb", "size": "530GB", "vg": "gluster_vg_sdb"}, "msg": "  Insufficient suitable allocatable extents for logical volume gluster_thinpool_sdb: 135680 more required\n", "rc": 5}


-----Ursprüngliche Nachricht-----
Von: knarra [mailto:knarra at redhat.com] 
Gesendet: Mittwoch, 3. Mai 2017 11:17
An: Oliver Dietzel <O.Dietzel at rto.de>; 'users at ovirt.org' <users at ovirt.org>
Betreff: Re: [ovirt-users] Ovirt Node 4.1.1.1 installation, engine on gluster, existing disk sdb not found or filtered, deployment fails

On 05/03/2017 02:06 PM, Oliver Dietzel wrote:
> Hi,
>
> i try to set up a 3 node gluster based ovirt cluster, following this guide:
> https://www.ovirt.org/blog/2017/04/up-and-running-with-ovirt-4-1-and-g
> luster-storage/
>
> oVirt nodes were installed with all disks available in the system, 
> installer limited to use only /dev/sda (both sda and sdb are HPE 
> logical volumes on a p410 raid controller)
>
>
> Glusterfs deployment fails in the last step before engine setup:
>
> PLAY RECAP *********************************************************************
> hv1.iw              : ok=1    changed=1    unreachable=0    failed=0
> hv2.iw              : ok=1    changed=1    unreachable=0    failed=0
> hv3.iw              : ok=1    changed=1    unreachable=0    failed=0
>
>
> PLAY [gluster_servers] 
> *********************************************************
>
> TASK [Clean up filesystem signature] 
> *******************************************
> skipping: [hv1.iw] => (item=/dev/sdb)
> skipping: [hv2.iw] => (item=/dev/sdb)
> skipping: [hv3.iw] => (item=/dev/sdb)
>
> TASK [Create Physical Volume] 
> **************************************************
> failed: [hv3.iw] (item=/dev/sdb) => {"failed": true, 
> "failed_when_result": true, "item": "/dev/sdb", "msg": "  Device 
> /dev/sdb not found (or ignored by filtering).\n", "rc": 5}
> failed: [hv1.iw] (item=/dev/sdb) => {"failed": true, 
> "failed_when_result": true, "item": "/dev/sdb", "msg": "  Device 
> /dev/sdb not found (or ignored by filtering).\n", "rc": 5}
> failed: [hv2.iw] (item=/dev/sdb) => {"failed": true, 
> "failed_when_result": true, "item": "/dev/sdb", "msg": "  Device 
> /dev/sdb not found (or ignored by filtering).\n", "rc": 5}
>
>
> But: /dev/sdb exists on all hosts
>
> [root at hv1 ~]# lsblk
> NAME                                                     MAJ:MIN RM   SIZE RO TYPE  MOUNTPOINT
> sda                                                        8:0    0 136,7G  0 disk
> ...
> sdb                                                        8:16   0 558,9G  0 disk
> └─3600508b1001c350a2c1748b0a0ff3860                      253:5    0 558,9G  0 mpath
>
>
>
> What can i do to make this work?
>
> ___________________________________________________________
> Oliver Dietzel
>
>
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users

Hi Oliver,

     I see that multipath is enabled on your system and for the device sdb it creates mpath and once this is created system will identify sdb as "3600508b1001c350a2c1748b0a0ff3860". To make this work perform the steps below.

1) multipath -l (to list all multipath devices)

2) black list devices in /etc/multipath.conf by adding the lines below, if you do not see this file run the command 'vdsm-tool configure --force' which will create the file for you.

blacklist {
         devnode "*"
}

3) mutipath -F which flushes all the mpath devices.

4) Restart mutipathd by running the command 'systemctl restart multipathd'

This should solve the issue.

Thanks
kasturi.