cinder volume(IBM SVC Driver) service pending

[Liu Qing]

issue env: IBM SVC(or Vxxx)

cinder.conf
[svc_driver_section]
ssh_max_pool_conn = 1

test steps
make sure cinder volume service is ok

shutdown switch or disable port to the SVC management ip(san_ip in the cinder.conf)

periodic task will be pending
waiting for more than 15min(tcp timeout default value?)
ssh command error will be seen in the log
a new periodic task will start
pending for ever

----
this is because ssh_utils.py does not decrease the current_size correctly.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/468634

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/468634
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=c33f84a8df58d45a31afdeeb453627f6eac2d748
Submitter: Jenkins
Branch: master

commit c33f84a8df58d45a31afdeeb453627f6eac2d748
Author: Liu Qing <email address hidden>
Date: Sat May 27 17:39:19 2017 +0800

    Fix SSHPool current_size not work correctly

    SSHPool will create a new SSHClient if the client it got from the pool
    is not active. If the create failed, then the current_size should minus
    one, as this connection will never be put back to the pool.

    Closes-Bug: 1693984

    Change-Id: Ic5fa53d6426a18138009fdf7de04e3ed6120231b
    Signed-off-by: Liu Qing <email address hidden>

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/cinder 11.0.0.0b3 development milestone.