nova-scheduler insufficient log details

[Scott Devoid]

File this under non-debug log usability. If your system is full (i.e. a schedule attempt runs out of hosts to use), the following log message is printed if debug is disabled:

2014-07-02 17:13:55.584 7567 WARNING nova.scheduler.driver [req-e8749ca6-6ab9-4598-a75e-2cdc0ba82ad6 5ef4ddd7c64545eda6a37cbd0da77c35 030d99a0fd014dd1859a1f633822ce29] [instance: 4e7c44d6-6c8a-43b5-851b-b7f0676850c8] Setting instance to ERROR state.

And ONLY that message. It would be nice if the log indicated why it was putting the instance in an error state. Even "failed to find suitable host" would be nice.

[Michael Turek (mjturek)]

Hello Scott, I ran into this while working on another bug report. I'm going to investigate a bit and see if there's any way for the log to better indicate what went wrong.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/110314

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/110314
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=2aed78705763edb723a4fd4598eae855b5847650
Submitter: Jenkins
Branch: master

commit 2aed78705763edb723a4fd4598eae855b5847650
Author: Michael Turek <email address hidden>
Date: Mon Aug 4 19:04:04 2014 +0000

    More informative nova-scheduler log after NoValidHost is caught.

    This patch adds a warning log entry for when a NoValidHost
    exception causes a scheduling error. The NoValidHost
    exception's message will be logged before reporting that
    the machine went to error. Previously, all other exceptions
    would show a stack trace. However, since the NoValidHost case
    is skipped, a user may not understand why his/her machine
    went to an ERROR state. This patch seeks to remove any
    confusion.

    Change-Id: I0f01b1095251f5c659fe010c7376f3f6bc56babe
    Closes-Bug: 1336977