imain / ocp-doit Goto Github PK

0curl: (6) Could not resolve host: releases-redhat-coreos-dev.cloud.paas.upshift.redhat.com; Name or service not known

All of the environment variables have been removed from the installer, so we're not forced to answer everything interactively.

I think we need to move to the installconfig yaml format instead, but not looked into the details yet.

For me 05_build_ocp_dependencies.sh is broken, the podman run starting the registry fails like:

ERRO[0005] Error while applying layer: ApplyLayer exit status 1 stdout:  stderr: lchown /etc/shadow: invalid argument 
Failed
unable to pull docker.io/library/registry: unable to pull image: Error committing the finished image: error adding layer with blob "sha256:d6a5679aa3cfc52503fac8e2345790783537564df5af43355ec662d7703aacbf": ApplyLayer exit status 1 stdout:  stderr: lchown /etc/shadow: invalid argument

This appears to be a permissions issue, running the podman run with sudo works, but then it fails like:

54ec7d2ab09c  docker.io/library/registry:latest  /entrypoint.sh /e...  About a minute ago  Exited (1) About a minute ago  0.0.0.0:8787->5000/tcp  registry
[shardy@dell-r630-007 ocp-doit]$ sudo podman logs 54ec7d2ab09c
time="2018-12-10T12:08:41Z" level=warning msg="No HTTP secret provided - generated random secret. This may cause problems with uploads if multiple registries are behind a load-balancer. To provide a shared secret, fill in http.secret in the configuration file or set the REGISTRY_HTTP_SECRET environment variable." go.version=go1.7.6 instance.id=ac290285-dcdb-4229-8c8a-ccefeb94f142 version=v2.6.2 
time="2018-12-10T12:08:41Z" level=info msg="redis not configured" go.version=go1.7.6 instance.id=ac290285-dcdb-4229-8c8a-ccefeb94f142 version=v2.6.2 
time="2018-12-10T12:08:41Z" level=info msg="Starting upload purge in 23m0s" go.version=go1.7.6 instance.id=ac290285-dcdb-4229-8c8a-ccefeb94f142 version=v2.6.2 
time="2018-12-10T12:08:41Z" level=info msg="using inmemory blob descriptor cache" go.version=go1.7.6 instance.id=ac290285-dcdb-4229-8c8a-ccefeb94f142 version=v2.6.2 
time="2018-12-10T12:08:41Z" level=fatal msg="open /certs/registry.crt: no such file or directory" 

Does this work for anybody else, or should we disable the registry for now?

Reported by apuimedo that deploying breaks with an error about missing DesignatePools like:

2018-11-23 14:56:07Z [standalone.StandaloneServiceChain.ServiceChain.15]: CREATE_FAILED StackValidationFailed: resources[15]: Property error: resources.DesignateCentralBase.properties: Property DesignatePools not assigned

I think it's due to this commit:
openstack-archive/tripleo-heat-templates@3abb38a

So we either need to disable designate or pass a valid config like this sample:
https://github.com/openstack/tripleo-heat-templates/blob/master/environments/designate-config.yaml

Hi
When I reinstall standalone OSP, after I clean existed one(work well for this), and then reinstall with ./02_run_all_in_one.sh, I got the mysql error.

I am thinking if tripleo_cleanup.sh clean all things has to be cleaned.

No idea if others met similar issue.

After I've yum updated and restarted my server and reinstalled the undercloud (02_run_all_in_one.sh), new VMs ended in the ERROR state with No valid host was found. as the reason.

I found this message in /var/log/containers/nova/nova-scheduler.log:

EndpointNotFound: ['internal'] endpoint for placement service in RegionOne region not found

The region our undercloud is configured with is called regionOne and the names are case-insensitive. Looks like the generated placement config has the wrong value:

$ grep RegionOne -r /var/lib/config-data/puppet-generated/

The workaround (thanks @flaper87!) is to change it back to regionOne and restart the services:

$ sed -i 's/RegionOne/regionOne/' /var/lib/config-data/puppet-generated/nova*/etc/nova/nova.conf
$ docker ps -q | xargs docker restart

I'm not sure what's causing this. We'll want to get the proper fix in, but in the meantime, we can at least add the workaround to the allinone script.

Every command including the OpenShift deployment is run as the OpenStack admin user / project. This is neither necessary, nor desirable (the installer should work for regular cloud users and therefore we should test it as such.

The script below creates a normal user with one extra privilege (we do have to add the swiftoperator role to be able to create containers).

It expects two clouds.yaml entries: standalone-admin (the admin user, currently called standalone in ocp-doit) and standalone -- the new user we're going to be creating.

This creates the necessary networks, security group rules and so on to be able to run openshift-installer. I have tested it and it all seems to be working.

I suggest we change ocp-doit to always do this and run the installer as a regular user, but that would require changes to clouds.yaml and a bunch of the scripts so I'm opening this issue first. Maybe we just want to ship this script instead and not disrupt the existing ones.

#!/bin/bash

set -eux
set -o pipefail

# Create & configure the user, project and quota
export OS_CLOUD=standalone-admin

openstack project create openshift
openstack user create --password 'password' openshift
openstack role add --user openshift --project openshift _member_
openstack role add --user openshift --project openshift swiftoperator
# NOTE(shadower): minimum secgroup: 6, secgroup-rules: 60
openstack quota set --secgroups 100 --secgroup-rules 1000 openshift

# Set the user's key, router, subnet, tempurl access, etc.
export OS_CLOUD=standalone

openstack keypair create --public-key ~/.ssh/id_rsa.pub default
openstack network create --internal private
openstack subnet create private --subnet-range 192.168.24.0/24 --network private
openstack router create internets
openstack router set internets --external-gateway public
openstack router add subnet internets private
openstack object store account set --property Temp-URL-Key=superkey

openstack security group rule create default --protocol tcp --dst-port 22:22 --remote-ip 0.0.0.0/0
openstack security group rule create --protocol icmp default
openstack security group rule create --protocol udp --dst-port 53:53 default