Dunno. There is a dedicated section for that right, not just those words? It’s just as important as all the other sections. Can you explain what you experienced so I can better understand (I assume you did not resize the control plane)? Were you following some other guide?
Yes, it is well documented, and I was following this guide. I missed that first time around as I was only doing two systems and skimming the doc is not the Right Thing ™ to do.
The issue is that Openstack fails if this step is not done, and there is no easy way for a user to trace it back to this. I mean other than the fact that new VMs cannot be launched after joining a new node into the cluster.
It is also at the same level as fetching the admin creds and launching a sample VM, both of which are optional, so not sure if there is a better way of indicating in this doc what is actually mandatory and what is not.
There is also a Scaling up the cluster page that covers the adding of nodes to an existing cluster, which includes the resize step. I think what you are calling for is a troubleshooting resource that lives outside of the how-tos. I’ve started a draft (unpublished) document here: General troubleshooting.
@pmatulis I’d suggest the “Obtain credentials” section is moved to the “Using the OpenStack CLI” how-to guide as it is not essential to complete the tutorial.
Done. Thank you.
We probably should rethink openstack configure --openrc <file>. The user does not need to interact with that file during the install either. If one is needed for the installation to proceed a default should just be used. We can then implement an init file generator (for both unprivileged and admin user).
Hello, I tried to install microstack multi-noded but since the last week I cant complete an install without errors: when I try to sunbeam cluster boostrap --role control --role compute --role storage I always got a stuck install (timed out maybe). For example im stuck in Deploying OpenStack Control Plane to Kubernetes (this may take a while) ... waiting for services to come online (15/24)
Do you have any ideas what is the problem? Thank you
Hum, maybe.
Should I try to improve my setup especially the hard disk ?
It’s hard to know without more information. What error are you getting?
For the 3rd node, the address should be sunbeam03.example.com:
sunbeam cluster add --name sunbeam03.example.com
1 Like
As recently @jamespage announced, the 2023.2 has just been released, but the deployment page is still on 2023.1 by:
sudo snap install openstack --channel 2023.1
1 Like
Thank you. The documentation will be updated shortly.
1 Like
Following up on this. The project will probably be staying with the 2023.1 as the default documented version due to it being associated with the upstream SLURP release (allows for upgrading to 2024.1 directly). It is the more production-grade release in other words. Sunbeam features and bug fixes will be backported from 2023.2 to 2023.1.
We’re working on how to version the docs better.
1 Like
Hello, I followed the guide, but I preferred to converged infrastructure, when I use only compute role to bootstrap microstack, it doesn’t ask for metallb ips, first this should be fixed. After it’s created, I manually updated ips from kubectl but I think there should be modification commands using juju or sunbeam commands too. Sunbeam configure command doesn’t cover metallb ips.
Hello, I have followed the document exact as it is mentioned for Multi-Node Deployment everything worked as expected but when i run this command on my main node
sunbeam cluster resize
All services get up except 1 service
What can be the issue, dashboard is accessible everything is working fine but i am unable to create any instance
Getting timeout error after a while waiting
What does SLURP stand for?
after some “fight” for setup a 3 node cluster, I succesfully configured changing the order of suggested guide, so:
- sudo snap install openstack --channel 2024.1/edge
- sunbeam prepare-node-script | bash -x && newgrp snap_daemon
- sunbeam cluster bootstrap --role control --role compute --role storage (answer in according whit infrastructure)
- sunbeam configure --openrc demo-openrc (answer in according whit infrastructure)
- sunbeam cluster add & join on other nodes (3 in may case)
- sunbeam cluster resize
if i follow the procedure, joining before configure, the first node don’t associate the nic interface (dont ask anything) and the system was isolated.
1 Like
I got the same problem. just now. In my environment, it timed out at (31/32).
2024.1 is scheduled to be released in this summer, so I wait until then.