'Noop run' feature behaviour changed (9.1 snapshot 286 -> 287)

[Alexandra]

Hello,

There are several differences for noop run of fuel tasks between 9.1 snapshot 286 and newer snapshots. All automated tests are impacted

1) Previously it was not possible to launch 'fuel2 env nodes deploy' task with --noop in case of non operational environment

9.1 snapshot 286:
[root@nailgun ~]# fuel env
id | status | name | release_id
---+--------+----------------------+-----------
1 | new | noop-nova-cinder-lvm | 2
[root@nailgun ~]# fuel node
id | status | name | cluster | ip | mac | roles | pending_roles | online | group_id
---+----------+-------------------------+---------+------------+-------------------+-------+-----------------+--------+---------
 6 | discover | slave-06_mongo | 1 | 10.109.7.8 | 64:21:2b:1e:4f:64 | | mongo | 1 | 1
 4 | discover | slave-01_controller | 1 | 10.109.7.3 | 64:69:fe:ef:e5:c0 | | controller | 1 | 1
 3 | discover | slave-03_compute_cinder | 1 | 10.109.7.5 | 64:24:c6:fd:3f:db | | cinder, compute | 1 | 1
 5 | discover | slave-05_ironic | 1 | 10.109.7.7 | 64:de:73:49:0d:8a | | ironic | 1 | 1
 1 | discover | slave-02_controller | 1 | 10.109.7.4 | 64:1a:06:7d:49:fa | | controller | 1 | 1
 2 | discover | slave-04_compute_cinder | 1 | 10.109.7.6 | 64:16:f5:af:93:29 | | cinder, compute | 1 | 1
[root@nailgun ~]# fuel2 env nodes deploy -e 1 -n 1 -f --noop
400 Client Error: Bad Request for url: http://10.109.7.2:8000/api/v1/clusters/1/deploy/?nodes=1&force=1&noop_run=1 (Deployment operation cannot be started. Nodes with uids [1] are not provisioned yet. )

9.1 snapshot 292:
[root@nailgun ~]# fuel env
id | status | name | release_id
---+--------+----------------------+-----------
1 | new | noop-nova-cinder-lvm | 2
[root@nailgun ~]# fuel node
id | status | name | cluster | ip | mac | roles | pending_roles | online | group_id
---+----------+-------------------------+---------+-------------+-------------------+-------+-----------------+--------+---------
 6 | discover | slave-05_ironic | 1 | 10.109.19.7 | 64:c0:8c:1c:7a:e0 | | ironic | 1 | 1
 2 | discover | slave-03_compute_cinder | 1 | 10.109.19.5 | 64:74:03:55:5b:b8 | | cinder, compute | 1 | 1
 4 | discover | slave-04_compute_cinder | 1 | 10.109.19.6 | 64:e7:05:a4:88:86 | | cinder, compute | 1 | 1
 1 | discover | slave-01_controller | 1 | 10.109.19.3 | 64:f3:86:e7:2f:87 | | controller | 1 | 1
 3 | discover | slave-02_controller | 1 | 10.109.19.4 | 64:79:a7:99:a3:c2 | | controller | 1 | 1
 5 | discover | slave-06_mongo | 1 | 10.109.19.8 | 64:03:5e:8c:9a:c6 | | mongo | 1 | 1
[root@nailgun ~]# fuel2 env nodes deploy -e 1 -n 1 -f --noop
Deployment task with id 18 for the nodes 1 within the environment 1 has been started.

2) Previously 'fuel2 env nodes deploy' task with --noop called 'dry_run_deployment' task. Now it calls 'deploy' (and 'deployment') tasks => node status became running (it's wrong according to feature requirements)

9.1 snapshot 286:
[root@nailgun ~]# fuel node
id | status | name | cluster | ip | mac | roles | pending_roles | online | group_id
---+--------+-------------------------+---------+------------+-------------------+-----------------+---------------+--------+---------
 4 | ready | slave-01_controller | 1 | 10.109.7.3 | 64:69:fe:ef:e5:c0 | controller | | 1 | 1
 5 | ready | slave-05_ironic | 1 | 10.109.7.7 | 64:de:73:49:0d:8a | ironic | | 1 | 1
 3 | ready | slave-03_compute_cinder | 1 | 10.109.7.5 | 64:24:c6:fd:3f:db | cinder, compute | | 1 | 1
 1 | ready | slave-02_controller | 1 | 10.109.7.4 | 64:1a:06:7d:49:fa | controller | | 1 | 1
 2 | ready | slave-04_compute_cinder | 1 | 10.109.7.6 | 64:16:f5:af:93:29 | cinder, compute | | 1 | 1
 6 | ready | slave-06_mongo | 1 | 10.109.7.8 | 64:21:2b:1e:4f:64 | mongo | | 1 | 1
[root@nailgun ~]# fuel2 env nodes deploy -e 1 -n 1 -f --noop
Deployment task with id 13 for the nodes 1 within the environment 1 has been started.

[root@nailgun ~]# fuel task
id | status | name | cluster | progress | uuid
---+---------+-------------------------+---------+----------+-------------------------------------
5 | ready | deploy | 1 | 100 | 01cac02e-b837-4bbd-bbf5-77967fe2e094
8 | ready | provision | 1 | 100 | e520c9c5-665c-4953-8857-403df7dede21
10 | ready | verify_networks | 1 | 100 | f6675a8f-43e2-4ce0-a5e9-92fa6f8ea30f
9 | ready | deployment | 1 | 100 | 0aa1e75a-6222-4038-9707-bbc80a9b8f2b
11 | ready | check_dhcp | 1 | 100 | 699a04ed-4bf4-405c-a070-dd4909537d5c
12 | ready | check_repo_availability | 1 | 100 | 2a79040f-9c4c-4b90-9722-eaf15d6ea820
13 | running | dry_run_deployment | 1 | 100 | 3f9110a5-36c2-40b3-aefe-70ee27bb69af

9.1 snapshot 292:
root@nailgun ~]# fuel node
id | status | name | cluster | ip | mac | roles | pending_roles | online | group_id
---+--------+-------------------------+---------+-------------+-------------------+-----------------+---------------+--------+---------
 6 | ready | slave-05_ironic | 1 | 10.109.19.7 | 64:c0:8c:1c:7a:e0 | ironic | | 1 | 1
 2 | ready | slave-03_compute_cinder | 1 | 10.109.19.5 | 64:74:03:55:5b:b8 | cinder, compute | | 1 | 1
 4 | ready | slave-04_compute_cinder | 1 | 10.109.19.6 | 64:e7:05:a4:88:86 | cinder, compute | | 1 | 1
 1 | ready | slave-01_controller | 1 | 10.109.19.3 | 64:f3:86:e7:2f:87 | controller | | 1 | 1
 3 | ready | slave-02_controller | 1 | 10.109.19.4 | 64:79:a7:99:a3:c2 | controller | | 1 | 1
 5 | ready | slave-06_mongo | 1 | 10.109.19.8 | 64:03:5e:8c:9a:c6 | mongo | | 1 | 1
[root@nailgun ~]# fuel2 env nodes deploy -e 1 -n 1 -f --noop
Deployment task with id 13 for the nodes 1 within the environment 1 has been started.
[root@nailgun ~]# fuel task
id | status | name | cluster | progress | uuid
---+---------+------------------------------------+---------+----------+-------------------------------------
1 | error | deploy | 1 | 100 | 586f5ee0-9877-4de0-948c-4506d4eefb57
11 | ready | deployment | 1 | 100 | c2808562-bb63-404d-b808-d87a17a2cf65
2 | error | deployment | 1 | 100 | d64e988f-cde2-405d-956e-c00d56990d37
3 | ready | verify_networks | 1 | 100 | ad1cb6ed-75a8-4eb0-b0c9-4547c804ae45
4 | ready | check_dhcp | 1 | 100 | dbc2234a-01de-4060-825c-31aa8e117aef
5 | ready | check_repo_availability | 1 | 100 | bedc08ea-2dad-4f47-8e1d-8b626629ce3c
6 | ready | check_repo_availability_with_setup | 1 | 100 | 05cc605f-a603-49bf-b16c-e0c18472b850
13 | running | deploy | 1 | 2 | 72e1bc74-48f7-45b7-9a76-c3b34e859b56
12 | ready | dump | | 100 | 73ed807d-8e89-4099-bdc2-0920003d481f
7 | ready | deploy | 1 | 100 | f4f66662-e975-40aa-896b-411cd709288c
10 | ready | provision | 1 | 100 | 1efec0c7-892e-44cb-b087-bff91f67af96
14 | running | deployment | 1 | 2 | da56e7d7-f94d-4874-92ac-e04758637862
[root@nailgun ~]# fuel node
id | status | name | cluster | ip | mac | roles | pending_roles | online | group_id
---+-----------+-------------------------+---------+-------------+-------------------+-----------------+---------------+--------+---------
 6 | ready | slave-05_ironic | 1 | 10.109.19.7 | 64:c0:8c:1c:7a:e0 | ironic | | 1 | 1
 2 | ready | slave-03_compute_cinder | 1 | 10.109.19.5 | 64:74:03:55:5b:b8 | cinder, compute | | 1 | 1
 4 | ready | slave-04_compute_cinder | 1 | 10.109.19.6 | 64:e7:05:a4:88:86 | cinder, compute | | 1 | 1
 1 | deploying | slave-01_controller | 1 | 10.109.19.3 | 64:f3:86:e7:2f:87 | controller | | 1 | 1
 3 | ready | slave-02_controller | 1 | 10.109.19.4 | 64:79:a7:99:a3:c2 | controller | | 1 | 1
 5 | ready | slave-06_mongo | 1 | 10.109.19.8 | 64:03:5e:8c:9a:c6 | mongo | | 1 | 1

I'll update list of changes if new differs found.
All automated tests are failed currently

[Alexandra (aallakhverdieva)]

+ minor update:
now it's not possible to use -f for grap execute:
[root@nailgun ~]# fuel2 graph execute -t custom -e 1 -f --noop -n 3
usage: fuel2 graph execute [-h] -e ENV [--force] [--trace]
                           [--dry-run | --noop] -t GRAPH_TYPES
                           [GRAPH_TYPES ...] [-n NODES [NODES ...]]
                           [-T TASK_NAMES [TASK_NAMES ...]]
fuel2 graph execute: error: unrecognized arguments: -f

It's minor because --force flag is OK, but previously both -f and --force were workable

[Vladimir Sharshov (vsharshov)]

It's expected behavior. We use same action as for deploy, but add special modification in boolean params 'noop_run' which setup required behavior.

Also noop run is same task as another. It has his own progress. Nodes also marked as progressing, but after task finished all nodes must get it previous states. We have discussed this behavior in this bug: https://bugs.launchpad.net/fuel/+bug/1621808

Also it is ok to run noop start on all nodes in any state including bootstrap, for example provision in experimental mode using deploy action as deployment, but works on bootstrap. What is why now this action available and make sense for bootstrap node too.