Openstack update 12 to 13 ("openstack overcloud update stack" missed)

Latest response

Hello folks,

we are facing issue during the update preparation from pike(12) to queens(13)
according to the official update document
"https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/13/html/upgrading_red_hat_openstack_platform/assembly-preparing_for_openstack-platform_upgrade#performing-a-minor-update-of-a-containerized-overcloud"
I should execute this command openstack overcloud update stack
but the command does not exist.

has someone faced same experience with it and can someone provide please a solution?

Thanks and lot in advance
appreciate your help

Br
Niko

NR
Log in to join the conversation

Responses

Daniel Macpherson's picture Red Hat Expert 1065 points
26 July 2018 8:45 AM

Hi Niko,

I'll investigate this issue and get back to you.

Dan

NR Active Contributor 245 points
26 July 2018 9:05 AM

Hi Dan,

thanks for your quick response. I have an additional question:

Should we work through the whole pdf document or just parts of it?

I did notice, that the steps in 2.3, for example, will get repeated in 3.1. What is the reason for this? And when I had a look at 6.1 there I should upgrade the overcloud (like in 2.4) but with the commands that are actually present in our CLI. There are more steps that are dupliacted in some way.

Thanks Br Niko

Daniel Macpherson's picture Red Hat Expert 1065 points
26 July 2018 9:37 AM

Good questions. So the commands in 2.3 and 3.1 are similar, but they do different things at different stages. 2.3 updates to the latest version of OSP12, while 3.1 uses the new OSP13 repository and upgrades the packages accordingly.

It sounds like your undercloud has already been upgraded to OSP13, which might explain why the OSP12 "openstack overcloud update stack" command isn't there. I'll double check on my test environment and get back to you. In the meantime, what version appears when you run the following on the undercloud:

$ sudo rpm -q python-tripleoclient
BK Community Member 34 points
26 July 2018 9:45 AM

Hi Dan,

thank you for your answer.

The version of the package is: python-tripleoclient-9.2.1-13.el7ost.noarch

As you said, that in 2.3 it will update to the latest OSP12, i guess i was a little bit to hasty. Because i included the new openstack 13 rpm-repo to early and therefor he grabbed the newer osp13-tripleo-client and installed this one. Maybe this was kind of a fault.

br, Benjamin

BK Community Member 34 points
26 July 2018 10:13 AM

i reverted the tripleo-client now to the previous version (from OSP12) and now the command is present.

Again - thank you for your help.

Br, Benjamin

NR Active Contributor 245 points
26 July 2018 10:37 AM

Hey Dan,

Benjamin and me are now able continue with setup. We would come back to you in case we will face further issues. We would appreciate getting some recommendations for further setup steps in case we have to be attention something.

Thanks a lot Niko

Daniel Macpherson's picture Red Hat Expert 1065 points
26 July 2018 11:36 AM

Glad it's all worked out, guys! And definitely post any additional questions you have. I'll keep a watch of this thread!

NR Active Contributor 245 points
30 July 2018 7:52 AM

Daniel, we are getting stuck at some someproblem with the iLo-connection between the controller and the undercloud.

we are getting this error right now (even after i tried to introspect them again

Node 7e6cd661-6f40-4114-a909-044629ab8007 did not reach state "manageable", the state is "enroll", error: Failed to get power state for node 7e6cd661-6f40-4114-a909-044629ab8007. Error: [iLO 10.26.6.11] The Redfish controller at "https://10.26.6.11" has thrown error. Error Unable to connect to https://10.26.6.11/redfish/v1/. Error: ("bad handshake: Error([(\'SSL routines\', \'SSL23_GET_SERVER_HELLO\', \'sslv3 alert handshake failure\')],)",)'

any suggestion Daniel? thanks in advance,

Br Niko

BK Community Member 34 points
30 July 2018 11:42 AM

some additional information...

We are running our OSP on HP Gen7 (Controller-nodes) and Gen9 (Compute-nodes) servers. The Gen7 servers only support TLSv1.1 and the Gen9 are supporting TLSv1.2. I can introspect the Compute nodes but not the Controller nodes. Could this be the fault?

Br Benjamin

Daniel Macpherson's picture Red Hat Expert 1065 points
30 July 2018 12:45 PM

There's a chance this might be the case, but here's a couple of things to check/test:

  • Do you have the CAs for both the Gen7s and Gen9s installed on the undercloud's CA store?
  • Can you perform standard cURL requests from the undercloud to the redfish API on the Gen7s?
  • Do the Gen7s introspect when you change redfish_verify_ca to false in the driver_info field for each Gen7 node?
BK Community Member 34 points
30 July 2018 1:14 PM

so when I do a cURL request to the Gen7-servers, it tells me (only working with --insecure):

Object Not Found

Object Not Found

The requested URL '/redfish/v1/' was not found on the HP iLO.

<

p>

They do not introspect with redfish_verify_ca to false.

We tried it with self-signed certs for a quick test, but I will include them now to the CA store on the director. On the other hand it is working without them for the Gen9-server.

Did the introspection-process changed in some way from osp12 to 13?

Thank you again for your quick response. Really appreciate it.

Daniel Macpherson's picture Red Hat Expert 1065 points
30 July 2018 1:31 PM

So it seems like it's not just the introspection process. It that ironic on the undercloud doesn't seem to want to connect to your Gen7s at all. So it can't get the power states nor turn the nodes on or off.

It looks like you're using the redfish driver to manage the Gen7s, but it doesn't look like there's any redfish api on them. Have you tried using the iLo driver or standard IPMI?

BK Community Member 34 points
30 July 2018 1:35 PM

We are using the

"pm_type":"pxe_ilo"

for all of our nodes.

Could it be, that the ironic.conf did change in some way with the undercloud-update to 13?

Daniel Macpherson's picture Red Hat Expert 1065 points
30 July 2018 4:18 PM

Ah sorry. I thought you were using the redfish driver.

Not sure about the ironic.conf, though I'm assuming you had no issues with controlling the Gen7 nodes with OSP12?

BK Community Member 34 points
31 July 2018 4:56 AM

No problem =)

Nope. We didn't have issues with the Gen7 with OSP12. The driver issue started during the upgrade process of the Platform (I think when I started to upgrade the overcloud). This was the first time I saw this error and that all controllers aren't reachable anymore (with the "openstack baremetal node list"-command)

Can I kind of force-disable the redfish-driver?

Daniel Macpherson's picture Red Hat Expert 1065 points
2 August 2018 3:33 AM

So there's probably two things I can suggest at this point:

  1. Try the new next-gen drivers. To do this, add ilo to the enabled_hardware_types list in the undercloud.conf file. Then rerun openstack undercloud install. Register the Gen7s using ilo instead of pxe_ilo.

  2. If that doesn't work, you might have to switch to the ipmi driver.

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.