Questions & Answers

Creating PNDA cluster but stuck at [deploy_package]

0 votes
asked Nov 24, 2017 by Jing Lu (280 points)

Hi all,

I am trying to deploy PNDA.IO@release 3.5 on OpenStack.
I have followed the instructions in the pnda guide
After resolving a lot of problems , the point that I am stuck in is that when I run ../heat_cli.py -e cation -n 1 -o 1 -k 1 -z 1 -f pico -s pnda_key create, the pnda cluster is being provisoning.
But it is so slow and stuck at the [deploy_package] step. After running for two hours it gives me error like this one:

  2017-11-23 03:45:22Z [router_interface]: CREATE_COMPLETE  state changed
  2017-11-23 03:46:18Z [saltmaster_server]: CREATE_COMPLETE  state changed
  2017-11-23 03:46:19Z [deploy_package]: CREATE_IN_PROGRESS  state changed
  2017-11-23 05:45:01Z [deploy_package]: CREATE_FAILED  CREATE aborted (Task create from SoftwareDeployment "deploy_package" Stack "cation" [68e95894-d97b-4cf2-8e5a-340d1d33ff45] Timed out)
  2017-11-23 05:45:01Z [cation]: CREATE_FAILED  Timed out
 
     Stack cation CREATE_FAILED

Then I login the cation-saltmaster VM with the key, and find in /var/log/os-collect-config.log there are errors connecting swift:
 
  Nov 23 06:00:25 cation-saltmaster os-collect-config: HTTPConnectionPool(host='10.1.0.222', port=8080): Max retries exceeded with url: /v1/AUTH_f062ed4f08324b96898ecaf7095bce16/cation-saltmaster_server-iowqnvsh5aut/305d095a-8bea-45c1-ac05-88278b8c0f98?temp_url_sig=a143d5bf4c716dcdc704bdc5c07103cee27fe4e2&temp_url_expires=2147483586 (Caused by ConnectTimeoutError(<urllib3.connection.HTTPConnection object at 0x7f15b4728b50>, 'Connection to 10.1.0.222 timed out. (connect timeout=10.0)'))
  Nov 23 06:00:25 cation-saltmaster os-collect-config: Source [request] Unavailable.
  Nov 23 06:00:25 cation-saltmaster os-collect-config: /var/lib/os-collect-config/local-data not found. Skipping
  Nov 23 06:00:25 cation-saltmaster os-collect-config: No local metadata found (['/var/lib/os-collect-config/local-data'])

and in /etc/os-collect-config.conf the metadata url is set to

  [request]
  metadata_url = http://10.1.0.222:8080/v1/AUTH_3018a8b172f84e749c0e565c4672ad2c/cation-saltmaster_server-zpmw7sc7zqqp/fc5a1523-1e07-4d0d-8229-d1449557d6f3?temp_url_sig=261fbfc1a3a31f085ede4e704b014833c4571992&temp_url_expires=2147483586

The network is unreachable for 10.1.0.222:8080 as this is the internal endpoint url for swift.
Is this the problem cause me stuck at [deploy_package]?

The public endpoint url to swift is 198.168.23.222:8080 and is reachable for the VM.

So my question is should the matadata_url be configured to internal url, can it be set to public endpoint.

Any help or support is highly appreciated!

1 Answer

0 votes
answered Nov 30, 2017 by trsmith2 (3,440 points)
Did you resolve this issue?
commented Nov 30, 2017 by Jing Lu (280 points)
Yes, It has been resolved.
commented Dec 11, 2017 by amfooladgar (200 points)
Could you please explain how you resolved this issue?
I have similar issue, the only difference is that my SaltMaster VM could read the added swift object "cation-deploy_package-qhz3vmib2xlu" but I am not sure why it is empty. So I see the logs in the VM as below:

Dec 11 01:32:21 cation-saltmaster os-collect-config: /var/lib/os-collect-config/local-data not found. Skipping
Dec 11 01:32:21 cation-saltmaster os-collect-config: No local metadata found (['/var/lib/os-collect-config/local-data'])

I am not sure if it is related to the image or if I am missing something.
Thanks in advance for your comments!
commented Dec 12, 2017 by Jing Lu (280 points)
The logs that you see are OK. I see them in my bastion VM's log too. I think my problem is a network issue,  I switched to another openstack platform (osp) then the problem is solved.
...