On Thu, Nov 7, 2019 at 12:57 PM Roy Golan <rgolan@redhat.com> wrote:

On Thu, 7 Nov 2019 at 12:28, Nathanaël Blanchet <blanchet@abes.fr> wrote:

Le 07/11/2019 à 11:16, Roy Golan a écrit :

On Thu, 7 Nov 2019 at 11:23, Nathanaël Blanchet <blanchet@abes.fr> wrote:

Le 07/11/2019 à 07:18, Roy Golan a écrit :

On Thu, 7 Nov 2019 at 00:10, Nathanaël Blanchet <blanchet@abes.fr> wrote:

Le 05/11/2019 à 21:50, Roy Golan a écrit :

On Tue, 5 Nov 2019 at 22:46, Roy Golan <rgolan@redhat.com> wrote:

On Tue, 5 Nov 2019 at 20:28, Nathanaël Blanchet <blanchet@abes.fr> wrote:

Le 05/11/2019 à 18:22, Roy Golan a écrit :

On Tue, 5 Nov 2019 at 19:12, Nathanaël Blanchet <blanchet@abes.fr> wrote:

Le 05/11/2019 à 13:54, Roy Golan a écrit :

On Tue, 5 Nov 2019 at 14:52, Nathanaël Blanchet <blanchet@abes.fr> wrote:

I tried openshift-install after compiling but no ovirt provider is available... So waht do you mean when you say "give a try"? Maybe only provisionning ovirt with the terraform module?

[root@vm5 installer]# bin/openshift-install create cluster
? Platform [Use arrows to move, space to select, type to filter, ? for more help]
> aws
azure
gcp
openstack

Its not merged yet. Please pull this image and work with it as a container
quay.io/rgolangh/openshift-installer

A little feedback as you asked:

[root@openshift-installer ~]# docker run -it 56e5b667100f create cluster
? Platform ovirt
? Enter oVirt's api endpoint URL https://air-dev.v100.abes.fr/ovirt-engine/api
? Enter ovirt-engine username admin@internal
? Enter password **********
? Pick the oVirt cluster Default
? Pick a VM template centos7.x
? Enter the internal API Virtual IP 10.34.212.200
? Enter the internal DNS Virtual IP 10.34.212.100
? Enter the ingress IP 10.34.212.50
? Base Domain oc4.localdomain
? Cluster Name test
? Pull Secret [? for help] *************************************
INFO Creating infrastructure resources...
INFO Waiting up to 30m0s for the Kubernetes API at https://api.test.oc4.localdomain:6443...
ERROR Attempted to gather ClusterOperator status after installation failure: listing ClusterOperator objects: Get https://api.test.oc4.localdomain:6443/apis/config.openshift.io/v1/clusteroperators: dial tcp: lookup api.test.oc4.localdomain on 10.34.212.100:53: no such host
INFO Pulling debug logs from the bootstrap machine
ERROR Attempted to gather debug logs after installation failure: failed to create SSH client, ensure the proper ssh key is in your keyring or specify with --key: failed to initialize the SSH agent: failed to read directory "/output/.ssh": open /output/.ssh: no such file or directory
FATAL Bootstrap failed to complete: waiting for Kubernetes API: context deadline exceeded

6 vms are successfully created thin dependent from the template

each vm is provisionned by cloud-init

the step "INFO Waiting up to 30m0s for the Kubernetes API at https://api.test.oc4.localdomain:6443..." fails. It seems that the DNS pod is not up at this time.

Right this moment, there is no more visibility on what is done, what goes wrong... what's happening there? supposing a kind of playbook downloading a kind of images...

The" pull secret step" is not clear: we must have a redhat account to https://cloud.redhat.com/openshift/install/ to get a key like

{"auths":{"cloud.openshift.com":{"auth":"b3BlbnNoaWZ0LXJlbGVhc2UtZGV2K2V4cGxvaXRhYmVzZnIxdGN0ZnR0dmFnMHpuazMxd2IwMnIwenV1MDg6TE9XVzFQODM1NzNJWlI4MlZDSUEyTFdEVlJJS0U5VTVWM0NTSUdOWjJH********************==","email":"exploit@abes.fr"},"quay.io":{"auth":"b3BlbnNoaWZ0LXJlbGVhc2UtZGV2K2V4cGxvaXRhYmVzZnIxdGN0ZnR0dmFnMHpuazMxd2IwMnIwenV1MDg6TE9XVzFQODM1NzNJWlI4MlZDSUEyTFdEVlJJS0U5VTVWM0NTSUdOWjJH********************==","email":"exploit@abes.fr"},"registry.connect.redhat.com":{"auth":"NTI0MjkwMnx1aGMtMVRDVEZUVFZBRzBaTkszMXdCMDJSMFp1VTA4OmV5SmhiR2NpT2lKU1V6VXhNaUo5LmV5SnpkV0lpT2lJMk4ySTJNREV3WXpObE1HSTBNbVE0T1RGbVpUZGxa**********************","email":"exploit@abes.fr"},"registry.redhat.io":{"auth":"NTI0MjkwMnx1aGMtMVRDVEZUVFZBRzBaTkszMXdCMDJSMFp1VTA4OmV5SmhiR2NpT2lKU1V6VXhNaUo5LmV5SnpkV0lpT2lJMk4ySTJNREV3WXpObE1HSTBNbVE0T1RGbVpUZGxa**********************","email":"exploit@abes.fr"}}}

Can you tell me if I'm doing wrong?

What is the template you are using? I don't think its RHCOS(Red Hat CoreOs) template, it looks like Centos?

Use this gist to import the template https://gist.github.com/rgolangh/adccf6d6b5eaecaebe0b0aeba9d3331b

Unfortunately, the result is the same with the RHCOS template...

Make sure that:

- the IPs supplied are taken, and belong to the VM network of those master VMs

- localdomain or local domain suffix shouldn't be used

- your ovirt-engine is version 4.3.7 or master

I didn't mention that you can provide any domain name, even non-existing.

When the bootstrap phase will be done, the instllation will teardown the bootsrap mahchine.

At this stage if you are using a non-existing domain you would need to add the DNS Virtual IP

you provided to your resolv.conf so the installation could resolve api.$CLUSTER_NAME.$CLUSTER_DOMAIN.

Also, you have a log under your $INSTALL_DIR/.openshift_install.log

I tried several things with your advices, but I'm still stuck at the https://api.test.oc4.localdomain:6443/version?timeout=32s test

with logs:

time="2019-11-06T20:21:15Z" level=debug msg="Still waiting for the Kubernetes API: the server could not find the requested resource"

So it means DNS resolution and network are now good and ignition provisionning is is OK but something goes wrong with the bootstrap vm.

Now if I log into the bootstrap vm, I can see a selinux message, but it may be not relevant...

SELinux: mount invalid. Same Superblock, different security settings for (dev nqueue, type nqueue).

Some other cluewWith journalctl:

journalctl -b -f -u bootkube

Nov 06 21:55:40 localhost bootkube.sh[2101]: {"level":"warn","ts":"2019-11-06T21:55:40.661Z","caller":"clientv3/retry_interceptor.go:61","msg":"retrying of unary invoker failed","target":"endpoint://client-7beef51d-daad-4b46-9497-8e135e528f7c/etcd-1.test.oc4.localdomain:2379","attempt":0,"error":"rpc error: code = DeadlineExceeded desc = latest connection error: connection error: desc = \"transport: Error while dialing dial tcp: lookup etcd-1.test.oc4.localdomain on 10.34.212.101:53: no such host\""}
Nov 06 21:55:40 localhost bootkube.sh[2101]: {"level":"warn","ts":"2019-11-06T21:55:40.662Z","caller":"clientv3/retry_interceptor.go:61","msg":"retrying of unary invoker failed","target":"endpoint://client-03992fc6-5a87-4160-9b87-44ec6e82f7cd/etcd-2.test.oc4.localdomain:2379","attempt":0,"error":"rpc error: code = DeadlineExceeded desc = latest connection error: connection error: desc = \"transport: Error while dialing dial tcp: lookup etcd-2.test.oc4.localdomain on 10.34.212.101:53: no such host\""}
Nov 06 21:55:40 localhost bootkube.sh[2101]: {"level":"warn","ts":"2019-11-06T21:55:40.662Z","caller":"clientv3/retry_interceptor.go:61","msg":"retrying of unary invoker failed","target":"endpoint://client-00db28a7-5188-4666-896b-e37c88ad3ae9/etcd-0.test.oc4.localdomain:2379","attempt":0,"error":"rpc error: code = DeadlineExceeded desc = latest connection error: connection error: desc = \"transport: Error while dialing dial tcp: lookup etcd-0.test.oc4.localdomain on 10.34.212.101:53: no such host\""}
Nov 06 21:55:40 localhost bootkube.sh[2101]: https://etcd-1.test.oc4.localdomain:2379 is unhealthy: failed to commit proposal: context deadline exceeded
Nov 06 21:55:40 localhost bootkube.sh[2101]: https://etcd-2.test.oc4.localdomain:2379 is unhealthy: failed to commit proposal: context deadline exceeded
Nov 06 21:55:40 localhost bootkube.sh[2101]: https://etcd-0.test.oc4.localdomain:2379 is unhealthy: failed to commit proposal: context deadline exceeded
Nov 06 21:55:40 localhost bootkube.sh[2101]: Error: unhealthy cluster
Nov 06 21:55:40 localhost podman[61210]: 2019-11-06 21:55:40.720514151 +0000 UTC m=+5.813853296 container died 7db3014e3f19c61775bac2a7a155eeb8521a6b78fea0d512384dd965cb0b8b01 (image=registry.svc.ci.openshift.org/origin/4.3-2019-10-29-180250@sha256:696a0ba7d344e625ec71d915d74d387fc2a951b879d4d54bdc69d460724c01ae, name=etcdctl)
Nov 06 21:55:40 localhost podman[61210]: 2019-11-06 21:55:40.817475095 +0000 UTC m=+5.910814273 container remove 7db3014e3f19c61775bac2a7a155eeb8521a6b78fea0d512384dd965cb0b8b01 (image=registry.svc.ci.openshift.org/origin/4.3-2019-10-29-180250@sha256:696a0ba7d344e625ec71d915d74d387fc2a951b879d4d54bdc69d460724c01ae, name=etcdctl)
Nov 06 21:55:40 localhost bootkube.sh[2101]: etcdctl failed. Retrying in 5 seconds...

It seems to be again a dns resolution issue.

[user1@localhost ~]$ dig api.test.oc4.localdomain +short
10.34.212.201

[user1@localhost ~]$ dig etcd-2.test.oc4.localdomain +short
nothing

So what do you think about that?

Key here is the masters - they need to boot, get ignition from the bootstrap machine and start publishing their IPs and hostnames.

Connect to a master, check its hostname, check its running or failing containers `crictl ps -a` by root user.

You were right:
# crictl ps -a
CONTAINER ID        IMAGE                                                                                                                                CREATED             STATE               NAME                 ATTEMPT             POD ID
744cb8e654705       e77034cf36baff5e625acbba15331db68e1d84571f977d254fd833341158daa8                                                                     4 minutes ago       Running             discovery            75                  9462e9a8ca478
912ba9db736c3       e77034cf36baff5e625acbba15331db68e1d84571f977d254fd833341158daa8                                                                     14 minutes ago      Exited              discovery            74                  9462e9a8ca478

# crictl logs 744cb8e654705
E1107 08:10:04.262330       1 run.go:67] error looking up self for candidate IP 10.34.212.227: lookup _etcd-server-ssl._tcp.test.oc4.localdomain on 10.34.212.51:53: no such host

# hostname
localhost

Conclusion: discovery didn't publish IPs and hostname to coreDNS because the master didn't get its name master-0.test.oc4.localdomain during provisionning phase.

I changed the master-0 hostname and reinitiates ignition to verify:

# hostnamectl set-hostname master-0.test.oc4.localdomain

# touch /boot/ignition.firstboot && rm -rf /etc/machine-id && reboot

After reboot is completed, no more exited discovery container:

CONTAINER ID        IMAGE                                                                                                                                CREATED             STATE               NAME                 ATTEMPT             POD ID
e701efa8bc583       77ec5e26cc676ef2bf5c42dd40e55394a11fb45a3e2d7e95cbaf233a1eef472f                                                                     20 seconds ago      Running             coredns              1                   cbabc53322ac8
2c7bc6abb5b65       d73eca122bd567a3a1f70fa5021683bc17dd87003d05d88b1cdd0215c55049f6                                                                     20 seconds ago      Running             mdns-publisher       1                   6f8914ff9db35
b3f619d5afa2c       7aa184de043265814f9a775968234ac3280a285056da773f1aba0917e9615370                                                                     21 seconds ago      Running             haproxy-monitor      1                   0e5c209496787
07769ce79b032       7aa184de043265814f9a775968234ac3280a285056da773f1aba0917e9615370                                                                     21 seconds ago      Running             keepalived-monitor   1                   02cf141d01a29
fb20d66b81254       e77034cf36baff5e625acbba15331db68e1d84571f977d254fd833341158daa8                                                                     21 seconds ago      Running             discovery            77                  562f32067e0a7
476b07599260e       86a34bc5edd3e70073313f97bfd51ed8937658b341dc52334fb98ea6896ebdc2                                                                     22 seconds ago      Running             haproxy              1                   0e5c209496787
26b53050a412b       9f94e500f85a735ec212ffb7305e0b63f7151a5346e41c2d5d293c8456f6fa42                                                                     22 seconds ago      Running             keepalived           1                   02cf141d01a29
30ce48453854b       7aa184de043265814f9a775968234ac3280a285056da773f1aba0917e9615370                                                                     22 seconds ago      Exited              render-config        1                   cbabc53322ac8
ad3ab0ae52077       7aa184de043265814f9a775968234ac3280a285056da773f1aba0917e9615370                                                                     22 seconds ago      Exited              render-config        1                   6f8914ff9db35
650d62765e9e1       registry.svc.ci.openshift.org/origin/4.3-2019-10-29-180250@sha256:9a7e8297f7b70ee3a11fcbe4a78c59a5861e1afda5657a7437de6934bdc2458e   13 hours ago        Exited              coredns              0                   2ae0512b3b6ac
481969ce49bb9       registry.svc.ci.openshift.org/origin/4.3-2019-10-29-180250@sha256:768194132b4dbde077e32de8801c952265643da00ae161f1ee560fabf6ed1f8e   13 hours ago        Exited              mdns-publisher       0                   d49754042b792
3594d9d261ca7       registry.svc.ci.openshift.org/origin/4.3-2019-10-29-180250@sha256:8c3b0223699801a9cb93246276da4746fa4d6fa66649929b2d9b702c17dac75d   13 hours ago        Exited              haproxy-monitor      0                   3476219058ba8
88b13ec02a5c1       7aa184de043265814f9a775968234ac3280a285056da773f1aba0917e9615370                                                                     13 hours ago        Exited              keepalived-monitor   0                   a3e13cf07c04f
1ab721b5599ed       registry.svc.ci.openshift.org/origin/4.3-2019-10-29-180250@sha256:629d73fdd28beafda788d2248608f90c7ed048357e250f3e855b9462b92cfe60   13 hours ago

because DNS registration is OK:

[user1@master-0 ~]$ dig etcd-0.test.oc4.localdomain +short
10.34.212.227

CONCLUSION:

none of rhcos vm is correctly provisionned to their targeted hostname, so they all stay with localhost.

What is your engine version? the hostname support for ignition is merged into 4.3.7 and master

4.3.7.1-1.el7

https://gerrit.ovirt.org/c/100397/ merged 2 days ago, so it will apear in 4.3.7.2.

Sandro when is 4.7.3.2 is due?

You can also use the nightly 4.3 snapshot - it's not really nightly anymore - it's updated per every run of CI Change-Queue, IIUC:

https://www.ovirt.org/develop/dev-process/install-nightly-snapshot.html