2:40 p.m.
On Thu, Oct 20, 2022 at 1:59 PM Matthew J Black <matthew(a)peregrineit.net> wrote:
Hi Everyone,
Could someone please help me - I've been trying to do an install of oVirt for *weeks*
(including false starts and self-inflicted wounds/errors) and it is still not working.
My setup:
- oVirt v4.5.3
- A brand new fresh vanilla install of RockyLinux 8.6 - all working AOK
- 2*NICs in a bond (802.3ad) with a couple of sub-Interfaces/VLANs - all working AOK
- All relevant IPv4 Address in DNS with Reverse Lookups - all working AOK
- All relevant IPv4 Address in "/etc/hosts" file - all working AOK
- IPv6 (using "method=auto" in the interface config file) enabled on the
relevant sub-Interface/VLAN - I'm not using IPv6 on the network, only IPv4, but
I'm trying to cover all the bases.
- All relevant Ports (as per the oVirt documentation) set up on the firewall
- ie firewall-cmd --add-service={{ libvirt-tls | ovirt-imageio | ovirt-vmconsole |
vdsm }}
- All the relevant Repositories installed (ie RockyLinux BaseOS, AppStream, &
PowerTools, and the EPEL, plus the ones from the oVirt documentation)
I have followed the oVirt documentation (including the special RHEL-instructions and
RockyLinux-instructions) to the letter - no deviations, no special settings, exactly as
they are written.
All the dnf installs, etc, went off without a hitch, including the "dnf install
centos-release-ovirt45", "dnf install ovirt-engine-appliance", and
"dnf install ovirt-hosted-engine-setup" - no errors anywhere.
Here is the results of a "dnf repolist":
- appstream Rocky Linux 8 - AppStream
- baseos Rocky Linux 8 - BaseOS
- centos-ceph-pacific CentOS-8-stream - Ceph Pacific
- centos-gluster10 CentOS-8-stream - Gluster 10
- centos-nfv-openvswitch CentOS-8 - NFV OpenvSwitch
- centos-opstools CentOS-OpsTools - collectd
- centos-ovirt45 CentOS Stream 8 - oVirt 4.5
- cs8-extras CentOS Stream 8 - Extras
- cs8-extras-common CentOS Stream 8 - Extras common
packages
- epel Extra Packages for
Enterprise Linux 8 - x86_64
- epel-modular Extra Packages for Enterprise
Linux Modular 8 - x86_64
- ovirt-45-centos-stream-openstack-yoga CentOS Stream 8 - oVirt 4.5 - OpenStack Yoga
Repository
- ovirt-45-upstream oVirt upstream for CentOS Stream
8 - oVirt 4.5
- powertools Rocky Linux 8 - PowerTools
So I kicked-off the oVirt deployment with: "hosted-engine --deploy --4
--ansible-extra-vars=he_offline_deployment=true".
I used "--ansible-extra-vars=he_offline_deployment=true" because without that
flag I was getting "DNF timout" issues (see my previous post `Local (Deployment)
VM Can't Reach "centos-ceph-pacific" Repo`).
I answer the defaults to all of questions the script asked, or entered the
deployment-relevant answers where appropriate. In doing this I double-checked every answer
before hitting <Enter>. Everything progressed smoothly until the deployment reached
the "Wait for the host to be up" task... which then hung for more than 30
minutes before failing.
From the ovirt-hosted-engine-setup... log file:
- 2022-10-20 17:54:26,285+1100 ERROR otopi.ovirt_hosted_engine_setup.ansible_utils
ansible_utils._process_output:113 fatal: [localhost]: FAILED! => {"changed":
false, "msg": "Host is not up, please check logs, perhaps also on the
engine machine"}
I checked the following log files and found all of the relevant ERROR lines, then checked
several 10s of proceeding and succeeding lines trying to determine what was going wrong,
but I could not determine anything.
- ovirt-hosted-engine-setup...
- ovirt-hosted-engine-setup-ansible-bootstrap_local_vm...
- ovirt-hosted-engine-setup-ansible-final_clean... - not really relevant, I believe
I can include the log files (or the relevant parts of the log files) if people want - but
that are very large: several 100 kilobytes each.
You can upload to some file sharing service (dropbox, google drive,
etc.) and share a link. If it's just not-too-long text you can use
some pastebin (might be more or less convenient for some readers, I
didn't do a poll).
I also googled "oVirt Host is not up" and found several entries, but after
reading them all the most relevant seems to be a thread from these mailing list: `Install
of RHV 4.4 failing - "Host is not up, please check logs, perhaps also on the engine
machine"` - but this seems to be talking about an upgrade and I didn't gleam
anything useful from it - I could, of course, be wrong about that.
So my questions are:
- Where else should I be looking (ie other log files, etc, and possible where to find
them)?
As the error message says - check also the engine machine. Logs from
it are copied to under /var/log/ovirt-hosted-engine-setup/engine-logs*
by the deploy process, but you might find more by ssh to it - search
local_vm_ip in the setup logs to find its local IP address.
See also e.g. thread "[ovirt-users] hosted-engine-setup --deploy fail
on Centos Stream 8".
- Does anyone have any idea why this isn't working?
Not without more information, sorry.
- Does anyone have a work-around (including a completely manual
process to get things working - I don't mind working in the CLI with virsh, etc)?
If all you want is a workaround, I think you should already had one,
no? At that point, the deploy process should not have failed, but
emitted something like:
You can now connect to {{ bootstrap_engine_url }} and
check the status of this host and
eventually remediate it, please continue only when the
host is listed as 'up'
Did you get this? Did you try to login? If not, there might have been
some other issue - please search the log for
he_pause_after_failed_add_host and host_result_up_check.
- What am I doing wrong?
Please, I'm really stumped with this, and I really do need help.
Good luck and best regards,
--
Didi