SoftwareDeploymentGroupTest fails with NotFound error

[Thomas Herve]

Seen here: http://logs.openstack.org/50/373850/1/check/gate-heat-dsvm-functional-orig-mysql-lbaasv2/603074a/console.html

In the middle of the test, the stack is delete (we can see it in the api logs), and then we fail trying to get it. No idea what's going here.

[Zane Bitter (zaneb)]

Same issue as https://bugs.launchpad.net/heat/+bug/1625921/comments/1

Fixed by Rabi's patch for that bug which I just approved.

[Zane Bitter (zaneb)]

Actually, I don't think that patch was sufficient to fix it. I unduplicated the bug because it will be tidier to fix it as a separate issue, rather than mixing it with the original report in bug 1625921, which seems unrelated other than affecting the same test.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/474004
Committed: https://git.openstack.org/cgit/openstack/heat/commit/?id=41b8e44d1e89440dca994bb927ecb35784d94e34
Submitter: Jenkins
Branch: master

commit 41b8e44d1e89440dca994bb927ecb35784d94e34
Author: Zane Bitter <email address hidden>
Date: Tue Jun 13 19:38:39 2017 -0400

    Fix races in SoftwareDeploymentGroupTest

    Don't assume that we can get the physical IDs of all of the
    SoftwareDeployment resources as soon as the stack becomes
    CREATE_IN_PROGRESS. 4dd67bb1aa2df4f5270f79600ac1f888b0bd9a5f reads them
    again once the stack is COMPLETE; this patch also uses the same physical
    resource IDs to verify the update.

    Also, make sure all of the resources are IN_PROGRESS before trying to
    signal them, because the signal_resources() utility method only signals
    resources that are IN_PROGRESS.

    Change-Id: I9787a5de5e4272a3ab370f653182aa9283ae01c0
    Closes-Bug: #1697794
    Closes-Bug: #1626073
    Closes-Bug: #1625921

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/heat 9.0.0.0b3 development milestone.