[update] Nodes do not marked by error status when update fails

[Tatyanka]

{"build_id": "2014-06-20_00-31-14", "mirantis": "yes", "build_number": "262", "ostf_sha": "2f30e5cab5bec1f1e2fd80e26e4da771a8ffe2d4", "nailgun_sha": "0c5e3b94fdd6bc9a50d5f840bf5151f95a23d908", "production": "docker", "api": "1.0", "fuelmain_sha": "4f547561532baf5f26733bf66db692dc5b61806d", "astute_sha": "694b5a55695e01e1c42185bfac9cc7a641a9bd48", "release": "5.1", "fuellib_sha": "25eb618a33a2ec87bc56f6bad16dc25b1837f0f0"}

Steps to reproduce:
1. Deploy envirnment
2. rsync tarball on master node
3. untar it
4. Run upgrade script (with changes needed for update i/e/ upgrade - > openstack)
5. run update on UI (should fail)
6. See if cluster has error state

Expected Result:
nodes marked by error status

Actual result:

Nodes stays in deploying status(everywhere including UI :))
[root@nailgun update]# fuel nodes
id | status | name | cluster | ip | mac | roles | pending_roles | online
---|-----------|---------------------|---------|--------------|-------------------|------------|---------------|-------
3 | deploying | slave-01_controller | 1 | 10.108.0.4 | 64:f8:7e:45:92:b4 | controller | | True
1 | deploying | slave-02_compute | 1 | 10.108.0.3 | 64:6a:31:9c:a3:76 | compute | | True
4 | deploying | slave-03_cinder | 1 | 10.108.0.5 | 64:71:c0:a2:7b:2b | cinder | | True

[Tatyanka (tatyana-leontovich)]

btw not reproduced on latest master - so move to incomplete to check if this issue stil exist

[Tatyanka (tatyana-leontovich)]

Reproduced on (it was 5.026 upgraded to 5.0.1 and upgraded to ) {"build_id": "2014-07-16_14-11-11", "ostf_sha": "9863db951a6e159f4fa6e6861c8331e1af069cf8", "build_number": "51", "auth_required": false, "api": "1.0", "nailgun_sha": "3bbb3e762d783aa93c977f578feea2de1900d0ed", "production": "docker", "fuelmain_sha": "d6244a8f31ef3ac17ec6df192c5259f575af35b8", "astute_sha": "d90cad0130da014eded5c21fa5f31054ce999dac", "feature_groups": ["mirantis"], "release": "5.1", "fuellib_sha": "7dd412c26db1bae0bb465a2dbe13ecb449eca71b"
Patching fail with Timeout of deployment is exceeded. Nodes not marked with error (see attached image)

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/108327

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/108327
Committed: https://git.openstack.org/cgit/stackforge/fuel-web/commit/?id=9dcc6fa0f0415e2a88945feda1a047171cd7a61b
Submitter: Jenkins
Branch: master

commit 9dcc6fa0f0415e2a88945feda1a047171cd7a61b
Author: Igor Kalnitsky <email address hidden>
Date: Mon Jul 21 12:40:02 2014 +0300

    Mark nodes to error if update was failed

    We have to mark all nodes to error explicitly in case of failed cluster
    updating. Such behaviour was implemented for regular deploy process
    but was mistakenly omitted for the update process.

    Please keep in mind, that all nodes are marked as error only in case
    Astute didn't sent per-node status information.

    Change-Id: I945c96d8fc7a27cb7779894a1f69baaafd096689
    Closes-Bug: #1332494

[Tatyanka (tatyana-leontovich)]

verified {"build_id": "2014-07-21_14-04-23", "ostf_sha": "9863db951a6e159f4fa6e6861c8331e1af069cf8", "build_number": "441", "auth_required": false, "api": "1.0", "nailgun_sha": "438cbb8687bc0c205475e0cfd7a6d637b3e46356", "production": "docker", "fuelmain_sha": "539a5bf7493a5d14690a34bb18c3ad1c75b4f37f", "astute_sha": "fd9b8e3b6f59b2727b1b037054f10e0dd7bd37f1", "feature_groups": ["mirantis"], "release": "5.1", "fuellib_sha": "1ec799bc6c8b08b8c9c6243c426507cb7a46459b"}