Fedora Account System
Red Hat Associate
Red Hat Customer
Description of problem: Cockpit installation. When I try to use different devices for different hosts the ansible playbook for gluster deployment is generated wrong. On every host pvname is the same. For example: I want: host1 - should use /dev/sdc host2 - should use /dev/sdd host3 - should use /dev/sde but generated playbook has: gluster_infra_volume_groups: - vgname: gluster_vg_sdc pvname: /dev/sdc for every node. The same situation is with volumes size, where I want to have different size for arbiter node. The settings from the first host dominate. Version-Release number of selected component (if applicable): RHHI: 1.8 RHV: 4.4.1
(In reply to apoczeka from comment #0) > Description of problem: > > Cockpit installation. > > When I try to use different devices for different hosts the ansible playbook > for gluster deployment is generated wrong. > On every host pvname is the same. For example: > > I want: > host1 - should use /dev/sdc > host2 - should use /dev/sdd > host3 - should use /dev/sde > > but generated playbook has: > gluster_infra_volume_groups: > - vgname: gluster_vg_sdc > pvname: /dev/sdc > for every node. > > The same situation is with volumes size, where I want to have different size > for arbiter node. > > The settings from the first host dominate. > > > Version-Release number of selected component (if applicable): > RHHI: 1.8 > RHV: 4.4.1 Hey! The flow is designed such a way that whatever changes you make to the first hosts bricks, are replicated to the second and third hosts bricks automatically. Are you by any chance, changing the 1st host bricks configurations after making the changes to the 2nd and 3rd host? Just in case you are, can you try changing the configuration as follows - 1. Make changes to the first host bricks 2. From the drop down, select the second host, you should be seeing your changes to the first host here. Now make the changes for the second host 3. After changing the second host brick details, select the third host from the drop down. You should see the changes made to the second host over here, you can edit the configurations for the third host, and all hosts should have the required details. Please try this once and let me know in case the error still persists
Tested with cockpit-ovirt-dashboard-0.14.12 When creating gluster volume with non-uniform disks across the 3 nodes ( different disks from each host -for example sdb from host1, sdc from host2, sdd from host3 ), the inventory files are created in the right way and deployment is successful <snip> gluster_infra_volume_groups: - vgname: gluster_vg_sdc pvname: /dev/sdc - vgname: gluster_vg_sdf pvname: /dev/sdf - vgname: gluster_vg_sdi pvname: /dev/sdi </snip>