on failure of granular task, bad summary message

[Andrew Woodward]

when a granular task deploys now we get a short message of

---
Deployment has failed. Method granular_deploy. Failed to execute hook .
---

when you show additional information you get

---
Deployment has failed. Method granular_deploy. Failed to execute hook .
---
priority: 500
fail_on_error: true
type: shell
uids:
- '4'
parameters:
  cmd: ruby /etc/puppet/modules/osnailyfacter/modular/astute/upload_cirros.rb
  timeout: 180
.
Inspect Astute logs for the details
---

Expected results: "granular_deploy" should describe the actual task that failed, in this case "upload_cirros"
Expected results: "uids: '4'" should more clearly describe which node failed "node-4"
Expected results "Inspect Astute logs for the details" should contain an actual link to the asute log, and in best case point to the exact message that raised the error.

[OpenStack Infra (hudson-openstack)]

Related fix proposed to branch: master
Review: https://review.openstack.org/185008

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/185008
Committed: https://git.openstack.org/cgit/stackforge/fuel-astute/commit/?id=7e07b873b607e8b91845c465c2884af19cca56a1
Submitter: Jenkins
Branch: master

commit 7e07b873b607e8b91845c465c2884af19cca56a1
Author: Ryan Moe <email address hidden>
Date: Thu May 21 15:22:03 2015 -0700

    Improve shell task error message

    The shell command which has failed is included in the
    error message along with the task timeout.

    Change-Id: I0ea474a899ba86d32b6b4f1a96de4baee17ecdc6
    Related-bug: #1435603

[Vladimir Sharshov (vsharshov)]

Current level of message is better that we could show in current arch. Moving to 8.0, because it is UX problem. Known issue.

[Alexander Kislitsky (akislitsky)]

We passed SCF in 8.0. Moving the bug to 9.0.

[Alex Schultz (alex-schultz)]

I'd also like to ask that we improve the messaging around task timeouts as well. It's not obvious from the error message if a task has failed due to an error in the task or if it has exceeded the task timeout. We should make those two cases have messaging that is clearly different from each other.

[Dmitry Pyzhov (dpyzhov)]

https://review.openstack.org/#/c/279436/

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/279436
Committed: https://git.openstack.org/cgit/openstack/fuel-astute/commit/?id=d09d2bfba97ca9359df1227dcdc8836e38bddfb1
Submitter: Jenkins
Branch: master

commit d09d2bfba97ca9359df1227dcdc8836e38bddfb1
Author: Dmitry Guryanov <email address hidden>
Date: Thu Feb 11 20:38:39 2016 +0300

    fix error reporting from shell and sync hooks

    1. Report exit codes from nodes, where command execution
       failed. I've changed return value of function
       run_shell_command. Now it returns nil in case of
       success and string descrition in case of error.
    2. Show different message for each exception handled
       in shell_hook.
    3. Fix warning about number of retires.

    I also think that it's better to increase timeout in
    nailgun, so timeout will be greater than retries * interval,
    in this case if command failed all times, the error will be
    'Failed to run command', not about timeout.

    Change-Id: I5051a84c9c146999ae7b3ef430ec3013989f1271
    Closes-Bug: #1435603

[Alexander Kurenyshev (akurenyshev)]

For now a high level error shows an info about failed task, for ex.:
"Deployment has failed. All nodes are finished. Failed tasks: Task[connectivity_tests/3] Stopping the deployment process!"
Also there are the info about nodes which were failed with that tasks, if applicable.