ovirt-hosted-engine deploy hangs
This is on a brand new Oracle Linux ver8.10 install on a physical server. I have completed all the perquisite steps in the instructions at https://docs.oracle.com/en/virtualization/oracle-linux-virtualization-manage... I run the deployment script: hosted-engine --deploy --4 Answer all the questions and let it run. It runs for about an hour and then hangs at the step: [ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Create temporary lock file] [ INFO ] changed: [localhost -> localhost] [ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Pause execution until /tmp/ansible.yo6aij1p_he_setup_lock is removed, delete it once ready to proceed] The first time I let it run all night and it was still hung in the morning. I have tried this using the cli command above. I have also tried it using the cockput gui method, both procedures hang at the same place. I have to run /usr/sbin/ovirt-hosted-engine-cleanup after each failure before trying again I can't find any clues in the logs but perhaps someone can tell me what to look for. I am happy to post the logs here as well. I assume this is at the point where the temporary Self Hosted engine has been created and is ready to be migrated to the shared storage. The lock placed there until it is successfully moved. But the instructions say that I should be prompted for the "storage you would like to use" I don't get that promtp. It is hung somewhere before it gets to that stage. What I do know is that the temporary Self hosted engines is created because I can ssh into and login as root Next I will try again but this time use the --ansible-extra-vars option to fill in all the variables at the beginning. I don't know what else to try
during the time waiting for my first time newbie post to be approved, I did get this working. I changed 2 things on the final attempt, I am not sure which of these 2 was the key to making this work (since it takes over an hour for the script to fail, I didn't limit my changes to one per attempt): -1) disabled the interface that was configured for ipv4 only. I had read in the documentation that the oVirt engine requires ipv6 so it must not be disabled even if you don't use it. So I had made sure ipv6 was enabled on the interface assigned to the oVirt engine. But there is another interface on the host that was setup without ipv6 for use in connecting guest VMs to another network. I didn't think it would matter if a non ipv6 interface existed as long as it is not the one selected during the hosted engine deploy script. But in desperation I disabled this extra interface to see if that made a difference. -2) fixed the rngd service (Randon Number Generator) service not starting. I don't know if the oVirt engine setup is dependant on the nrgd service and installs it during setup but I did notice nrgd not starting after the setup failures. It turned out it could not start as the host has neither a TPM nor a CPU supporting rdrand. So I did the not recommended hack of pointing the nrgd service to /drv/urandom rather than letting it find a supported hardware source of random.
participants (1)
-
bleuze＠enkon.com