Migrate data from Centos 7 to Ubuntu LTE from VMware
- Thread starter big_buddz
- Start date
In this case, it's probably easiest to rebuild from scratch.
If you were migrating to an RPM based distro such as Rocky, AlmaLinux, or Fedora, then you would have
a fairly straight-forward upgrade path.
But migrating to a Debian based distro is a big change. A lot of your package names and software will change.
Most of your RPMs won't work. There is a RPM tool for Ubuntu, but in practice most of the dependencies are
so different, that it often doesn't work. By default Ubuntu/Debian uses apt and .deb packages rather than
dnf/yum and .rpm packages. So any custom in house software you've created RPMs for, will likely need to be
re-created as .deb packages.
Certain commands are different, for example Ubuntu/Debian uses ufw instead of firewalld, and certain
software groups, like 'C Development Tools and Libraries' are called 'build essentials' in Ubuntu, and all
the packages that are in one group, are not in the other group.
Newer versions of CentOS/Alma/Rocky/Oracle/Redhat/Fedora support snap, appImage, and flatpak.
If you were nigrating to/from these it would be easier, as Ubuntu also supports these package formats,
but as far as I know, CentOS 6 and 7 do not.
The glibc libraries are different versions, the kernels are different versions, and even the default file-system
types are different. It really is easier to start over in most cases when you do something like this.
But even if you were to stay with RPM based distro's. There would be changes you have to make whether
you use Ubuntu or not. Most newer distro's use systemd, instead of sysV init. Also most use NetworkManager
instead of networkd and ifcfg files.
If you're using OS based clustering, like cman for example. Ubuntu doesn't come with that by default.
If you're using using NIC teaming or bonding, Ubuntu doesn't come with the modules to do this by default.
You can add kernel modules to do these things, but they aren't built-in by default.
If you were migrating to an RPM based distro such as Rocky, AlmaLinux, or Fedora, then you would have
a fairly straight-forward upgrade path.
But migrating to a Debian based distro is a big change. A lot of your package names and software will change.
Most of your RPMs won't work. There is a RPM tool for Ubuntu, but in practice most of the dependencies are
so different, that it often doesn't work. By default Ubuntu/Debian uses apt and .deb packages rather than
dnf/yum and .rpm packages. So any custom in house software you've created RPMs for, will likely need to be
re-created as .deb packages.
Certain commands are different, for example Ubuntu/Debian uses ufw instead of firewalld, and certain
software groups, like 'C Development Tools and Libraries' are called 'build essentials' in Ubuntu, and all
the packages that are in one group, are not in the other group.
Newer versions of CentOS/Alma/Rocky/Oracle/Redhat/Fedora support snap, appImage, and flatpak.
If you were nigrating to/from these it would be easier, as Ubuntu also supports these package formats,
but as far as I know, CentOS 6 and 7 do not.
The glibc libraries are different versions, the kernels are different versions, and even the default file-system
types are different. It really is easier to start over in most cases when you do something like this.
But even if you were to stay with RPM based distro's. There would be changes you have to make whether
you use Ubuntu or not. Most newer distro's use systemd, instead of sysV init. Also most use NetworkManager
instead of networkd and ifcfg files.
If you're using OS based clustering, like cman for example. Ubuntu doesn't come with that by default.
If you're using using NIC teaming or bonding, Ubuntu doesn't come with the modules to do this by default.
You can add kernel modules to do these things, but they aren't built-in by default.
Last edited:
