The last thing I need is the complication of having the engine hosted on
the cluster. It's way too easy to shoot yourself in the foot that way.
On Mon, Jun 13, 2022, 11:32 PM Sketch wrote:
> On Tue, 14 Jun 2022, Guillaume Pavese wrote:
>
> > Not sure about recovering your cluster on a 4.5
On Tue, 14 Jun 2022, Guillaume Pavese wrote:
Not sure about recovering your cluster on a 4.5 install with a 4.4 backup. I
would also like to know if that is possible.
It's definitely possible. I had an issue with my 4.4->4.5 upgrade (always
make a backup first) and wanted to switch my
Not sure about recovering your cluster on a 4.5 install with a 4.4 backup.
I would also like to know if that is possible.
As a separate note, if you really need to install 4.4 instead of 4.5; I
managed to install a 4.4.10 hosted engine recently by
- extracting the ova from the ovirt-appliance rpm
It looks like I've run into this now:
https://bugzilla.redhat.com/show_bug.cgi?id=2077794
https://lists.ovirt.org/archives/list/users@ovirt.org/message/SBCWNXLFLJBKTA3TFJARE7QCYZQ6QMMH/
It looks like a permanent fix was put in place in 4.5.
Should I consider just bumping up to 4.5 instead of
This came in just as the engine came up.
The final pieces of the puzzle were:
The database@server notation expressed in the logs is cool, but is not how
it is expressed to actually connect. That threw up a red herring.
I had to create all of the postgres users
Postgres ident authentication
On Sun, 12 Jun 2022, David Johnson wrote:
There's a version mismatch between the libpq and the postgresql verson:
[root@ovirt1 ~]# dnf info --installed libpq* postgresql*
Installed Packages
Name : libpq5
Version : 14.3
Release : 42PGDG.rhel8
Architecture : x86_64
Size
On reviewing the log from engine-setup:
2022-06-12 23:27:29,507-0400 DEBUG otopi.context context._executeMethod:127
Stage setup METHOD
otopi.plugins.ovirt_engine_common.ovirt_engine.db.connection.Plugin._setup
2022-06-12 23:27:29,510-0400 DEBUG
otopi.ovirt_engine_setup.engine_common.database
Update:
I scanned the centos libraries, and the lowest version of postgres that had
consistent package versions was 13. There was no libpq at the version 12
level in the centos RPM library.
I performed the manual installation process for the postgresql database at
the *Initializing the
Progress ... there's something weird going on with the centos postgres
package(s). After a bunch of removes and installs of postgres the engine
install succeeded and the engine configuration succeeded.
Now it looks like not all of the postgresql libraries were loaded.
[root@ovirt1 ~]#
Other issues when running dnf install ovirt-engine
Failed to get unit file state for openvswitch.service: No such file or
directory
Failed to get unit file state for ovn-northd.service: No such file or
directory
Failed to get unit file state for ovirt-provider-ovn.service: No such file
or
Following
https://lists.ovirt.org/archives/list/users@ovirt.org/thread/CKIRPIOU4BEQGOZ3QZBHK5I7O3NRZVIM/
The /etc/init.d/postgresql directory should be created by dnf install
ovirt-engine. I ran dnf reinstall ovirt-engine and the directory is still
missing.
>
>>
Installing on a pristine Centos 8, I had to manually install the correct
version of ansible and postgres 12, then change the service name from
postgresql-12 to postgresql.
There are scripts looking for init.d/postgresql. Since Centos 8 implements
SystemCtl , there should be no init.d's (and there
12 matches
Mail list logo
