Introduction to Heat
Heat is the main orchestration engine for OpenStack, and is used my OpenStack director to install an OpenStack Overcloud environment.
When we run the “openstack deploy overcloud” command, we are specifically
telling RHEL OSP director that we want it to use the pre-defined Heat templates from
/usr/share/openstack-tripleo-heat-templates/. OSP director will manage the
deployment of a new overcloud heat stack, using files from this directory.
When RHEL OSP director calls the Heat stack, it needs the following data…
- A top-level Heat template to use that describes the overall environment and the
resources required.
- An environment/resource registry to tell Heat where to find resource
definitions for non-standard Heat elements, e.g. TripleO components.
- A set of parameters to declare the deployment-specific options (via -e)
The most important files for us to focus on are in our deployment directory, these are the default files that get called by OSP director.
- The top-level Heat template that OSP director uses for deployment is
/usr/share/openstack-tripleo-heat-templates/overcloud-without-mergepy.yaml
- The resource registry, which tells Heat where to find the templates for
deployment resources is
/usr/share/openstack-tripleo-heat-templates/overcloud-resource-registry-puppet.
yaml
Creating a Heat Stack
To create the stack we run the command below. This command instructs heat to use
the templates in ~/my_templates/, as well as the override templates specified
with the ‘-e’ option.
This is just an example of what I am using in my lab environment, your deploy command will be much different. Also note that I have copied the templates from /usr/share/openstack-tripleo-heat-templates to ~/my_templates/.
#openstack overcloud deploy –debug –templates ~/my_templates/ \
–ntp-server 10.1.0.1 –control-scale 3 –compute-scale 2 \
-e ~/my_templates/advanced-networking.yaml
Troubleshooting a Failed Heat Stack
Unfortunately our deploy failed with the following errors.
Exception: Heat Stack create failed.
DEBUG: openstackclient.shell clean_up DeployOvercloud
DEBUG: openstackclient.shell got an error: Heat Stack create failed.
ERROR: openstackclient.shell Traceback (most recent call last):
File “/usr/lib/python2.7/site-packages/openstackclient/shell.py”, line 176, in
run
return super(OpenStackShell, self).run(argv)
File “/usr/lib/python2.7/site-packages/cliff/app.py”, line 230, in run
result = self.run_subcommand(remainder)
File “/usr/lib/python2.7/site-packages/cliff/app.py”, line 295, in
run_subcommand
result = cmd.run(parsed_args)
File “/usr/lib/python2.7/site-packages/cliff/command.py”, line 53, in run
self.take_action(parsed_args)
File
“/usr/lib/python2.7/site-packages/rdomanager_oscplugin/v1/overcloud_deploy.py”,
line 864, in take_action
self._deploy_tripleo_heat_templates(stack, parsed_args)
File
“/usr/lib/python2.7/site-packages/rdomanager_oscplugin/v1/overcloud_deploy.py”,
line 535, in _deploy_tripleo_heat_templates
parsed_args.timeout)
File
“/usr/lib/python2.7/site-packages/rdomanager_oscplugin/v1/overcloud_deploy.py”,
line 478, in _heat_deploy
raise Exception(“Heat Stack create failed.”)
Exception: Heat Stack create failed.
We can verify that the deploy failed with the command below.
stack@undercloud] # heat stack-list
+————————————–+————+—————+—| id | stack_name | stack_status | creation_time |
+————————————–+————+—————+—
| ce993847-b0ee-4ea2-ac15-dc0ddc81825a | overcloud | CREATE_FAILED |
2016-02-29T20:40:54Z |
Since the stack deploy has failed, let’s take a closer look at the stack resources
and see if we can determine which resources failed.
Here we will make things simple by viewing only failed resources.
[stack@undercloud] # heat resource-list overcloud | grep -i failed
| Compute | c032c668-755f-422f-8ad1-4abf46b022ff
| OS::Heat::ResourceGroup | CREATE_FAILED |
2016-02-29T20:40:55Z |
| Controller | 668d27e0-9ab1-4dbe-8445-1d1ee8839265
| OS::Heat::ResourceGroup | CREATE_FAILED |
2016-02-29T20:40:55Z |
The failed resources are named “Compute” and “Controller“. Lets take a closer
look at those using the “resource-show” argument.
#heat resource-show overcloud Compute
| resource_status_reason | ResourceUnknownStatus: Resource failed – Unknown
status FAILED due to “Resource CREATE failed: ResourceUnknownStatus:
Resource failed – Unknown status FAILED due to “Resource CREATE failed:
StackValidationFailed: Property error : OsNetConfigImpl: config The Parameter
(BondInterfaceOvsOptions) was not provided.”” |
Let’s now do the same for Controller.
#heat resource-show overcloud Controller
| resource_status_reason | ResourceUnknownStatus: Resource failed – Unknown
status FAILED due to “Resource CREATE failed: ResourceUnknownStatus:
Resource failed – Unknown status FAILED due to “Resource CREATE failed:
StackValidationFailed: Property error : OsNetConfigImpl: config The Parameter
(BondInterfaceOvsOptions) was not provided.”” |
Apparently I have some issues with my OVS bonding options, so I need to get those straight before I can continue.
Deleting a Failed Heat Stack
Since our last deploy failed, we need to delete the failed stack before we can kick off another stack deploy. Below is an example of that command – note we are using the UUID of the stack.
[stack@vz-undercloud] # heat stack-delete 2b0da4f6-e6f8-41cd-89e8-bf070d0e0d15
+--------------------------------------+------------+-------------------| id | stack_name | stack_status | creation_time |
+--------------------------------------+------------+-------------------
| 2b0da4f6-e6f8-41cd-89e8-bf070d0e0d15 | overcloud | DELETE_IN_PROGRESS |
2016-03-01T17:21:58Z |
+--------------------------------------+------------+-------------------
then…
[stack@vz-undercloud] # heat stack-list
+--------------------------------------+------------+---------------+---
| id | stack_name | stack_status | creation_time |
+--------------------------------------+------------+---------------+---
| 2b0da4f6-e6f8-41cd-89e8-bf070d0e0d15 | overcloud | DELETE_FAILED |
2016-03-01T17:21:58Z |
+--------------------------------------+------------+---------------+---
Now lets kick off another deploy
#openstack overcloud deploy –debug –templates ~/my_templates/ \
–ntp-server 10.1.0.1 –control-scale 3 –compute-scale 2 \
-e ~/my_templates/advanced-networking.yaml
Unfortunately, this deploy failed as well.
Ok, let’s take a look at /var/log/heat/heat/heat-engine.log for more details. I also suggest opening another ssh session and tailing the log while the delete is attempting to do its thing.
If the output is too verbose to follow, I suggest attempting to thin out the output using the command below
#tail -f /var/log/heat/heat-engine.log | egrep ‘error|fatal’
This lead me to the following error.
2016-03-01 13:46:12.366 18554 ERROR heat.engine.resource [-] Error marking
resource as failed
2016-03-01 13:46:12.366 18554 TRACE heat.engine.resource DBConnectionError:
(_mysql_exceptions.OperationalError) (2003, “Can’t connect to MySQL server on
‘172.16.0.10’ (111)”)
Mysql is down? So now we need to look at the mariadb logs – where we see the following.
160301 12:16:24 [Warning] Failed to setup SSL
160301 12:16:24 [Warning] SSL error: SSL_CTX_set_default_verify_paths failed
Apparently SELinux is blocking the reads for the certificates.
There are two ways to work around this issue. You can run “restorecon -v
/path/to/certs/“, or you can work around by disabling selinux by running
“setenforce 0” or by editing the /etc/selinux/config file and setting ‘SELINUX=DISABLED’.
You may need to rerun the delete, in my case it was stuck in
“DELETE_IN_PROGRESS”. I restarted all heat releated services to force the delete to error.
#systemctl restart openstack-heat-engine.service openstack-heat-api.service
openstack-heat-api-cloudwatch.service openstack-heat-api-cfn.service
This will cause the delete to error. You can then retry the delete.
If the delete is taking a long time, you can dig a bit deeper into the delete
using the command below.
#heat resource-list overcloud
Now drill down more with the command below.
#heat event-list overcloud
Make note of the resource_name and its id and use them in the next command.
Note that stack name is still overcloud.
heat event-show overcloud Compute d9e13b02-07b0-4beb-8442-f25de0e7ef8b
I have found that rebooting the undercloud will clear out any in-progress
tasks, you can then run the delete again.
You can also try to manually delete each node from Ironic by mimic’ing what the
nova driver in Ironic does. This is shown below for reference.
$ ironic node-set-provision-state <node uuid> deleted
And to remove the instance_uuid
$ ironic node-update <node uuid> remove instance_uuid
Troubleshooting Failed and In Progress Deployments.
Updated 6/28/2018
While your deploy is running, you can watch for stuck or failed deployments.
$ heat deployment-list | grep -vi complete
WARNING (shell) "heat deployment-list" is deprecated, please use "openstack software deployment list" instead
+--------------------------------------+--------------------------------------+--------------------------------------+--------+-------------+----------------------+------------------+
| id | config_id | server_id | action | status | creation_time | status_reason |
+--------------------------------------+--------------------------------------+--------------------------------------+--------+-------------+----------------------+------------------+
| 143b25ce-9c1e-48b0-bd92-2569640b5208 | 5b9843a5-2092-4c75-8eae-2e042d7b34d2 | f08232e7-7a14-4885-86ab-443b4581afe6 | CREATE | IN_PROGRESS | 2017-05-17T19:46:38Z | None |
| 27d2a61e-e5cd-4313-967d-cc7c385320cb | c113ead1-bb9a-46ca-a992-4527c2e9fda1 | 67dc5457-5ae7-4290-ad77-36bf8bc81d56 | CREATE | IN_PROGRESS | 2017-05-17T19:46:38Z | None |
| fabab6ae-8848-4a35-8066-d83f82335e95 | 763c4760-3e8c-43c1-8afd-d1b8d6bbb55b | e51061ed-f96f-43ec-b18a-4dc079d961d8 | CREATE | IN_PROGRESS | 2017-05-17T19:46:38Z | None |
| 0b6993c1-a320-4dc7-980c-e789d87a0c84 | 50f1e303-123a-451b-9d86-f29c32e6ce0b | 8d87818a-943f-4f69-8112-7450dfa26bb7 | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| 0e537823-f17d-4009-b2e4-6d11adcdfeb7 | 1d7938ad-a6e5-4d8f-9604-489408295a59 | 98847475-ed1c-44b2-aa4e-bbddfe81a1d1 | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| 1d89b9cf-87cf-4050-a00f-496c82fd9432 | 4f378bc3-10a7-4a47-b269-2c7de5d75dd7 | 69257ad8-a3ee-453b-a532-db3c293a34b9 | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| 3d5915a6-1470-4ffe-a0f1-a356df33e1f9 | dbc8c62c-f5cc-4cfa-93af-ef7b9f02fd19 | 75b49872-c1fb-4c83-87d1-579dcd027bc8 | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| a2e57b50-dc03-42a8-b851-b33ade36d591 | 0c7133d8-20bc-481d-b9c5-2d0867a656e7 | 000794e9-4ce2-4c8a-aa69-02aadfd19ce6 | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| ab463b0f-fd6e-4169-9283-3e8bbcc1bb4b | 627f1288-3823-4e21-b61d-4bafcbdc446a | b6cf41e8-e60c-43ca-a3ed-d24c078031c8 | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| ad05853f-2e70-4f11-b8c0-26d2d4c324fb | 8cd656e7-129f-40d4-a1bc-ed0181b9e768 | 24017696-102d-43ab-a101-f3a33a091faa | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| b6418aec-385a-4b6f-845b-84c87fa10be7 | 04747e35-f0cc-45a4-b251-78ea61a08c83 | ffe36a72-2293-48eb-93ed-bfe1077a0245 | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| e741d37d-6119-4d37-8b91-519ecfad8397 | 271139d7-c29c-47df-ad11-26bbc3373c4c | 167dc5ee-0831-4fa8-a7ca-8db6d2b3d4ea | CREATE | IN_PROGRESS | 2017-05-17T19:46:39Z | None |
| 0910d714-027b-40b0-acf3-140f5ab90837 | 9317c6ca-b767-49f2-a56c-45b61bbee2d5 | 34f6b4fd-fa98-46b6-8ace-b6f76af50162 | CREATE | IN_PROGRESS | 2017-05-17T19:46:40Z | None |
| 508e585d-19ec-41e8-bc1e-049d8206f614 | 0917bf21-54f8-46ce-89c8-6461001840a5 | c8aa01f8-5627-46a4-885c-dc61e0886c09 | CREATE | IN_PROGRESS | 2017-05-17T19:46:40Z | None |
+--------------------------------------+--------------------------------------+--------------------------------------+--------+-------------+----------------------+------------------+
Next I can inspect a stuck/hung deployment using the command below. Note that server id above directly corresponds to a nova uuid.
$ heat deployment-show 05d3c2b7-104c-47fd-beeb-3819d55aafb7
WARNING (shell) "heat deployment-show" is deprecated, please use "openstack software deployment show" instead
{
"status": "COMPLETE",
"server_id": "e1389b2e-ae2c-4095-8654-fd1cac1a8563",
"config_id": "1bfd85e8-217c-440f-b095-f4a2fcce4bfc",
"output_values": {
"deploy_stdout": "ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDVWznMhZnwFKR9KH+7ppJr5f7ZI2LsxwZV6+xPkNOhNE5o2DDNGWGVoGE+2SIiENuhk938LRUixg4pj1coXl9johjv8h2PcB+6rUxVT7jSKGUdpwgTlLsNOEugcR5oJjf0mM55BMzz1nr1yZMTy9THGwU/A43RZcOVvuJXFVA3MIlYGXXUidEOQM6NMEloUpVd/hj2BXUzdaUJV83E0aSGPvED02LixV5uorY0n4OIk2EXC8LlG0863YwJEdQi7hd8/rczLwX18CFFAQJ2mOwkcuJB37V/BEWUqBczHO46e0uc1ZmHhtdPFKgFObgP8gSrT4g0HjOds/BVPgBTtQf5 tripleo-validations\n",
"deploy_stderr": "",
"deploy_status_code": 0
},
"creation_time": "2017-05-17T19:46:38Z",
"updated_time": "2017-05-17T19:49:26Z",
"input_values": {},
"action": "CREATE",
"status_reason": "Outputs received",
"id": "05d3c2b7-104c-47fd-beeb-3819d55aafb7"
Deleting a Zombie Overcloud
Sometimes it is impossible to delete a failed deploy. After attempting a deploy the stack was stuck in the state below.
+--------------------------------------+------------+--------------------+----------------------+----------------------+
| id | stack_name | stack_status | creation_time | updated_time |
+--------------------------------------+------------+--------------------+----------------------+----------------------+
| 34fa8bd8-b80b-4379-925f-26465e366da7 | overcloud | DELETE_IN_PROGRESS | 2018-07-06T15:58:09Z | 2018-07-06T16:15:57Z |
I was given this command to kill it
# openstack stack list --nested --column ID --format value | xargs openstack stack delete --yes
This finally allowed the stack to delete.
[stack@tpavcpp4r1v4uc ~]$ heat stack-list
WARNING (shell) "heat stack-list" is deprecated, please use "openstack stack list" instead
+----+------------+--------------+---------------+--------------+
| id | stack_name | stack_status | creation_time | updated_time |
+----+------------+--------------+---------------+--------------+
+----+------------+--------------+---------------+--------------+
If you run into the error below when attempting to redeploy
Exception creating plan: Unable to create plan. The Mistral environment already exists
You will need to run the command below to delete the mistral plan
#openstack overcloud plan delete overcloud
Additional Resources
https://wiki.openstack.org/wiki/Heat
http://hardysteven.blogspot.com/2015/04/debugging-tripleo-heat-templates.html
Like this:
Like Loading...
