Manila-CI: glusterfs (NFS) driver CI fails because backend Glusterfs volume runs out of space

[Ramana Raja]

glusterfs (NFS) driver CI uses directory mapped layout. Shares are created within GlusterFS volume. During the CI run, the GlusterFS volume runs out of available space, so shares are not created, and tempest tests fail.

Pasting the log of manila scheduler service,

 [req-b46ed283-e753-4d4d-b98a-4d968cb2b9a4 ] Insufficient free space for share creation on host devstack-trusty-rax-dfw-5045249@glusternfs1#GLUSTERFS (requested / avail): 1/0.0
[req-b46ed283-e753-4d4d-b98a-4d968cb2b9a4 ] Filter CapacityFilter returned 0 host(s)
 [req-b46ed283-e753-4d4d-b98a-4d968cb2b9a4 ] Failed to schedule create_share: No valid host was found. .

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/226639
Committed: https://git.openstack.org/cgit/stackforge/devstack-plugin-glusterfs/commit/?id=830a57308b05159b21553b9c3351b9cbd94de3c5
Submitter: Jenkins
Branch: master

commit 830a57308b05159b21553b9c3351b9cbd94de3c5
Author: Ramana Raja <email address hidden>
Date: Wed Sep 23 12:52:38 2015 +0530

    Manila-CI: increase size of backend GlusterFS vol

    The glusterfs (NFS) driver CI that currently uses directly mapped
    layout sometimes fails since the backend GlusterFS volume within
    which the driver creates shares does not have sufficient space to
    allow Tempest to create shares for testing.

    Closes-Bug: #1498792
    Change-Id: If9ac508834637908bb91df55e856e08bca01e88f