This document in intended to capture common problems and their mitigations. Newest additions will be at the top of the document. At some point, xtoph_deploy problems and mitigations will be moved to the xtoph_deploy repo and simply linked here, but for now it is one document
libgcrypt in 8.2 (GA) has a bug that creates a symptom of not being able to deploy a vm into libvirt with virt-install. Error reports as "Host does not support any virtualization options".
|
ℹ️
|
resolved by updating libgcrypt with yum update libgcrypt AND don’t use RHEL 8.2
|
Libvirtd introduced it’s own zone into the firwalld ruleset. The latest xtoph_deploy ansible role has a shell script to detect the existence of the ruleset and will trigger the duplication of the firewall rules for the libvirt zone.
|
ℹ️
|
resolved by updating to the latest RHEL 8.3 releases with yum update AND using the lastest workshop playbooks with git pull.
|
The bastion runs the openshift installer which creates a certificate that is valid for 24 hours. Simply trying to rerun the deployer after the 24 hour window can have many odd symptoms.
|
ℹ️
|
To delete all VMs and simply start again, run xtoph_deploy.sh redeploy.
|
|
ℹ️
|
To preserve the existing bastion VM and it’s current downloads: manually delete all OCP cluster nodes, run opc4_workshop.sh unlock followed by xtoph_deploy.sh deploy. This will permit the openshift installer to generate a new certificate without having to download everything anew.
|
This usually tracks back to faulty DNS configuration, or even another DHCP server on the network providing faulty DNS info. Check the workshop configs and the xtoph_deploy configs, and more importantly run some DNS tests to see if everything works as expected.
|
ℹ️
|
These playbooks assume a fresh "@Base" installation on the target hosts. There is no excessive validation checking before installation/deployment is attempted. |