Xen snapshot uploads can fail without retry under retryable circumstances

[Christopher Lefelhocz]

If a glance server is completely down, the xen server taking a snapshot will fail and report back as a "non-retryable" exception. This is not correct and the compute node should really go to the next server in the list and retry.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/126973

[Jesse J. Cook (jesse-j-cook)]

Related to: https://review.openstack.org/#/c/128090/

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/126973
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=e03525bf1adc0a9966018ab33288bdda8c610267
Submitter: Jenkins
Branch: master

commit e03525bf1adc0a9966018ab33288bdda8c610267
Author: Christopher Lefelhocz <email address hidden>
Date: Wed Oct 8 13:18:19 2014 -0500

    Fix xen plugin to retry on upload failure

    The xen plugin does not catch all cases where upload can throw
    an exception due to connection failures and timeouts. This
    change extends the try/execpt window to catch these cases.

    Also added original connect which in part keeps us from lazy-loading
    the connection and logging of exception information so we can
    figure out later what happened.

    Change-Id: If643a7da9f602ab7bb5d255031a841c674fa6c6f
    Fixes-bug: 1378903