EqualLogic driver makes cinder-volume service unavailable if eqlx server closes the ssh connection

[Dunrong Huang]

The eqlx driver would identify one CLI command completion by looking for the system name prompt in the output ("ARRAY_NAME>"). But some cases where there is no "ARRAY_NAME>" in the output. The eqlx server closes the ssh connection immediately after sends such message, which lead to a infinite loop in _get_output() and 100% CPU utilization.

For example, by default, the maximum ssh connections cinder-volume will use is 5, if we have one more cinder-volume service, the server will send eqlx driver below message(no "ARRAY_NAME>") and close the connection.
       "Error: Number of open cli sessions reached maximum allowed value of 7"

Another example is that sometimes eqlx server just sends a message like this:
             Welcome to Group Manager
        Copyright 2001-2014 Dell Inc.
then close the connection.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/147070

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/147070
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=a762665595f0f38bacb50d1a032df1c82b3a4c60
Submitter: Jenkins
Branch: master

commit a762665595f0f38bacb50d1a032df1c82b3a4c60
Author: Dunrong Huang <email address hidden>
Date: Wed Jan 14 13:56:43 2015 +0800

    Fix eqlx endless loop when server closes the connection

    The eqlx driver would identify one CLI command completion by looking
    for the system name prompt in the output ("ARRAY_NAME>"). But some
    cases where there is no "ARRAY_NAME>" in the output. The eqlx server
    closes the ssh connection immediately after sends such message, which
    lead to a infinite loop in _get_output() and 100% CPU utilization.

    What the patch does is to check the return of chan.recv(), if the
    length of returned string is zero, which means server has closed the
    connection, the loop will be exited.

    Closes-Bug: #1410627

    Signed-off-by: Dunrong Huang <email address hidden>
    Change-Id: I600c38d98f0f808c98df010a89ac58ca8e43f1a3

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/juno
Review: https://review.openstack.org/147958

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/147958
Committed: https://git.openstack.org/cgit/openstack/cinder/commit/?id=7ebaf68a900ddf33dab086d331da78a9c8bd5603
Submitter: Jenkins
Branch: stable/juno

commit 7ebaf68a900ddf33dab086d331da78a9c8bd5603
Author: Dunrong Huang <email address hidden>
Date: Wed Jan 14 13:56:43 2015 +0800

    Fix eqlx endless loop when server closes the connection

    The eqlx driver would identify one CLI command completion by looking
    for the system name prompt in the output ("ARRAY_NAME>"). But some
    cases where there is no "ARRAY_NAME>" in the output. The eqlx server
    closes the ssh connection immediately after sends such message, which
    lead to a infinite loop in _get_output() and 100% CPU utilization.

    What the patch does is to check the return of chan.recv(), if the
    length of returned string is zero, which means server has closed the
    connection, the loop will be exited.

    Closes-Bug: #1410627

    Signed-off-by: Dunrong Huang <email address hidden>
    Change-Id: I600c38d98f0f808c98df010a89ac58ca8e43f1a3
    (cherry picked from commit a762665595f0f38bacb50d1a032df1c82b3a4c60)