UPI/Bare Metal - The master node cannot finish the installation ("config/master", "timeout awaiting response headers"/"EOF") #696
Replies: 14 comments 12 replies
-
The state of the bootstrap node for more than 10 hours...
New error on master node...
Any, help!? =D |
Beta Was this translation helpful? Give feedback.
-
See if this helps... This output occurs in okd_master_3 (10.3.0.7)....
Connecting okd_master_2 (10.3.0.6) from okd_services (10.3.0.14)... NOTE: The okd_master_2 (10.3.0.6) was able to boot (reached login screen).
Pinging the okd_bootstrap (10.3.0.4) from okd_master_2 (10.3.0.6)...
Calling the problematic URL from okd_master_2 (10.3.0.6)...
ABOUT THE INFRASTRUCTURE: Virtual machines...
Network layout...
Thanks! =D |
Beta Was this translation helpful? Give feedback.
-
Configuration files... BIND 9 (DNS): . db.10.3.0
. db.okd.local
. named.conf.local
. named.conf
HAProxy (load balancer): . haproxy.cfg
OpenShift (OKD) "*.yaml" files: . htpasswd_provider.yaml
. install-config.yaml
. registry_pv.yaml
Thanks! =D |
Beta Was this translation helpful? Give feedback.
-
Done! =D
|
Beta Was this translation helpful? Give feedback.
-
Yes! I can ping the entire infrastructure. Pings performed from okd_boostrap (10.3.0.4)...
|
Beta Was this translation helpful? Give feedback.
-
"did you change config to roundrobin" -> Yes. =D |
Beta Was this translation helpful? Give feedback.
-
Further question... Are these names important (below)? In other words, do they need to follow some kind of convention that is expected by OKD (OpenShift) within the infrastructure? . BIND 9 Configurations:
. HAProxy Configurations:
|
Beta Was this translation helpful? Give feedback.
-
The guide I'm following is this https://itnext.io/guide-installing-an-okd-4-5-cluster-508a2631cbee "the names a(re) important" -> So I can't change something like "okd4-control-plane-1" to "okd-master-1" and something like "okd4_https_ingress_traffic_be" to "okd_https_ingress_traffic_be", right? (This question may sound redundant, but I really need to confirm this....) |
Beta Was this translation helpful? Give feedback.
-
I adjusted the names according to the "defaults" above and now all machines (nodes) reach the login splash screen. One of the nodes I had to redo the process (reinstall) to work. However, now we have two problems... 1# The nodes are all the time printing information on the screen that is mixed with the traditional components of the initial login screen. I don't know if this is normal...; 2# According to the command below the bootstrap process is not finishing...
... so it doesn't go over the line "INFO Waiting up to 40m0s for bootstrapping to complete...". Further question... The fact that I use this configuration...
... could it be cause of a problem? |
Beta Was this translation helpful? Give feedback.
-
"1# The nodes are all the time printing information on the screen that is mixed with the traditional components of the initial login screen. I don't know if this is normal...;" -> An image may help... |
Beta Was this translation helpful? Give feedback.
-
Yes this means the server have finally connect to the bootstrap and the configuration have started |
Beta Was this translation helpful? Give feedback.
-
I also encountered the same problem, but I used the form of static configuration IP, I added kernel parameters through Fedora-coreos->tag |
Beta Was this translation helpful? Give feedback.
-
Don't waste time with the guide you provided. If you want to install
OpenShift (OKD) 4.X cluster over UPI/"bare-metal". This guide is the way to
go!
Install the OpenShift (OKD) 4.X cluster (UPI/"bare-metal")
<https://github.com/eduardolucioac/okd_bare_metal>
😉
Em seg., 19 de dez. de 2022 às 16:02, p0th3ad ***@***.***>
escreveu:
… Hi pals,
I'm following this guide
<https://itnext.io/guide-installing-an-okd-4-5-cluster-508a2631cbee> by
keeping the very same subnets and names, but slightly different versions:
***@***.*** ~]$ oc version
Client Version: 4.11.0-0.okd-2022-12-02-145640
Kustomize Version: v4.5.4
***@***.*** ~]$ openshift-install version
openshift-install 4.11.0-0.okd-2022-12-02-145640
built from commit c258fd3a0027ef3fa4ed685e093e044b1d4014cd
release image
***@***.***:924958ed5ba93ea9d72e4b5a319d0d4df220edbb891816c106215cbf318a0a40
release architecture amd64
I also have the same problem with okd4-control-plane-1
[image: image]
<https://user-images.githubusercontent.com/121056246/208499909-2d36937e-3064-4d88-a9d8-5f5284d8ada4.png>
DHCP and DNS looks good to me. IP addresses and names are correct.
Did you fix your issue gonhidalgo?
—
Reply to this email directly, view it on GitHub
<#696 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABRX4JNV3KQWJT36Z5LEPULWOCWMVANCNFSM463WUYGA>
.
You are receiving this because you were mentioned.
|
Beta Was this translation helpful? Give feedback.
-
Hi, i am trying to instll okd 4.13 snc on my Esxi 8. and tryed same on my high end laptop also. trying to run "virsh console okd4-snc-master", it is unable to get api server. please help me. |
Beta Was this translation helpful? Give feedback.
-
Simple question, but so far very difficult to answer... =-[
I am trying to deploy OpenShift (OKD) 4.5 or 4.7 as directed here Guide: Installing an OKD 4.5 Cluster. Look at the "Starting the control plane nodes" section.
I'm trying to create the cluster using an UPI (User Provisioned Infrastructure)/Bare Metal (KVM).
PROBLEM:
The "master" node cannot finish installation after reboot. It keeps showing the following error...
For version 4.5 we use "Fedora CoreOS 32.20200715.3.0".
The "master" node cannot finish installation after reboot. It keeps showing the following error...
For version 4.7 we use "Fedora CoreOS 34.20210518.3.0".
I've waited hours and hours and the master nodes are still in the same situation. What can I do to resolve this?
Thanks! =D
Beta Was this translation helpful? Give feedback.
All reactions