Error: Transaction test error:
file /usr/lib64/libmariadb.so.3 conflicts between attempted installs of MariaDB-shared-10.5.2-1.el8.x86_64 and mariadb-connector-c-3.0.7-1.el8.x86_64
I can’t get Virtualmin to install on CentOS8. It keeps failing at Phase 2 of the installation with a transaction test error.
No, I don’t. We’re about to fix all current issues first and only then move to checking other issues.
Error: Transaction test error: file /usr/lib64/libmariadb.so.3 conflicts between attempted installs of MariaDB-shared-10.5.2-1.el8.x86_64 and mariadb-connector-c-3.0.7-1.el8.x86_64
Please make sure that you are installing it on fresh, minimum CentOS 8 image.
Yes, it’s being installed on a fresh installation of Centos 8.
I installed in this order: CentOS 8 > NGINX > MariaDB > PHP 7 > Webmin > VirtualMin
That’s how I got the mariadb error.
I tried it also in a different order: CentOS 8 > NGINX > PHP 7 > Webmin > VirtualMin > MariaDB
When I tried installing Virtualmin BEFORE installing MariaDB, but I ran into another issue where I got a PREIN error.
for example…
Importing GPG key 0x60D62A6B:
Userid : “Virtualmin, Inc. (Package signing key for Virtualmin 6) <security@virtualmin”
Fingerprint: E36F 0664 7D8E BD2B E364 2BCE D9F9 0107 60D6 2A6B
From : /etc/pki/rpm-gpg/RPM-GPG-KEY-virtualmin-6
Error in PREIN scriptlet in rpm package wbm-virtualmin-nginx-ssl
Error in PREIN scriptlet in rpm package wbm-virtualmin-awstats
Error: Transaction failed
…but if I install it --force, it completes, sometimes.
I found a workaround for the PREIN scriptlet error by manually installing Virtualmin from within Webmin.
I think @sungdze that @Ilia means Virtualmin should be installed on a fresh CentOS 8 installation, without anything else installed between CentOS 8 and Virtualmin.
Hence CentOS 8 > Virtualmin instead of CentOS 8 > NGINX > MariaDB > PHP 7 > Webmin > VirtualMin or even CentOS 8 > NGINX > PHP 7 > Webmin > VirtualMin > MariaDB. Could you try that and let us know how it goes?
Yes, I understood what he meant, but I was simply writing out what my installation process was when I got the error. I thought, maybe, it’d help them in finding a solution to the bug.
And the reason why I did it that way (nginx/php before virtualmin), was because I wanted a certain version of those programs. And I didn’t want to have to do the whole uninstall, uninstall related dependencies, upgrade and install, and etc etc…
But although they suggest to install it on a fresh install, I think a bug that occurs should at the very least be let known. Whether they fix or not, well, that’s up to the dev team. If there is no fix, that’s fine.
On fresh CentOS 7 relase today July 16 near noon CET
Welcome to the Virtualmin GPL installer, version 6.2.0
////…
▣▣□ Phase 2 of 3: Installation
Installing EPEL release package [ ]
Installing yum-utils [ ]
Enabling extras repository [ ]
Installing scl-utils [ ]
Install Software Collections release package [ ]
Installing PHP7 [ ]
Marking ‘Virtualmin LAMP Stack’ for install [ ]
Marking ‘Virtualmin Core’ for install [ ]
Installing dependencies and system packages [ERROR] Failed with error: 1
[ ✘ ]
[ERROR] Something went wrong. Exiting.
[ERROR] The last few log entries were:
Spin pid is: 1741
Marking ‘Virtualmin Core’ for install: Success.
Spin pid is: 1977
Operating system is CentOS Linux
/var/tmp/rpm-tmp.Wb5F5L: line 5: /etc/webmin/stop: No such file or directory
Webmin install complete. You can now login to https://mynameserver:10000/
as root with your root password.
This Webmin module depends on the module virtual-server, which is
not installed on your system.
error: %pre(wbm-virtualmin-awstats-2:5.8-1.noarch) scriptlet failed, exit status 1
Error in PREIN scriptlet in rpm package 2:wbm-virtualmin-awstats-5.8-1.noarch
error: wbm-virtualmin-awstats-2:5.8-1.noarch: install failed
Installing dependencies and system packages: [2020-07-16 14:07:54 CEST] [ERROR] Failed with error: 1
[2020-07-16 14:07:54 CEST] [ERROR] Something went wrong. Exiting.
[2020-07-16 14:07:54 CEST] [ERROR] The last few log entries were:
(stops there) with prompt
Also, “https://mynameserver:10000/” is hostname and not fullyqualified domain name
I did checked /etc/hosts before and after installation
Before reboot, I could go to virtualmin, not after reboot