Reinstallation of a failed compute
Bug #1487440 reported by
Veronica Krayneva
This bug affects 2 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Invalid
|
High
|
Ivan Ponomarev |
Bug Description
Scenario:
1. Create a cluster
2. Add 3 nodes with controller roles
3. Add a node with compute and cinder roles
4. Add a node with mongo role
5. Provision the compute node
(fuel node --node-id <> --provision)
6. For the compute node put inappropriate task to be executed to cause a failure on deployment(fuel node --node <> --tasks hiera). Start deployment of the node (fuel node --node-id <> --deploy)
7. Wait until deployment fails
Expected result:
All steps passed
Actual result:
On step 7 deployment doesn't fails
build_number: "187"
build_id: "2015-08-
Changed in fuel: | |
importance: | Undecided → High |
assignee: | nobody → Ivan Ponomarev (ivanzipfer) |
milestone: | none → 7.0 |
Changed in fuel: | |
status: | New → Confirmed |
description: | updated |
To post a comment you must log in.
Checked on 7.0 iso #197 On step 7 after executing hiera task compute node is in error state:
13 | error | Untitled (1b:60) | 3 | 10.109.0.14 | 64:1c:19:7c:1b:60 | cinder, compute | | True | 3
Which is expected because you skip all the required tasks so compute node doesn't even have puppet manifests.
During full deployment it also failed on cinder role, here's the list of tasks for the problem env:
22 | ready | verify_networks | 3 | 100 | 3d6fb21d- b8e4-4b36- 839a-d593d2499e bb dcfe-46cb- b2cc-9f09a2bf61 64 availability | 3 | 100 | 3477078c- 1972-442b- 9918-e670e6cd3a f6 availability_ with_setup | 3 | 100 | e9e38b69- 80dc-4f4d- 82a6-8a774d0011 e9 a084-4fbe- 92c2-afa3e470b5 f7 e37d-48e7- 8d5e-f9e0714a81 90 045b-49bf- 80c1-ba100d2f92 5a
23 | ready | check_dhcp | 3 | 100 | 4cd59030-
24 | ready | check_repo_
25 | ready | check_repo_
28 | error | deployment | 3 | None | 5f4686a2-
26 | ready | provision | 3 | 100 | 2af659c4-
27 | error | deployment | 3 | 100 | 67d7f495-
It has deployed compute role without errors. Which is also expected behavior. Compute role does not perform any interaction with keystone/ corosync/ DB/etc, it just installs packages, configures services to connect to endpoints/ messanger/ etc, so there is nothing to fail.
So I'm marking this bug as invalid. And the test case itself does not look valid to me.