this topic is continuation of the Migrate2rocky script not working - and severall so-called topics. The problem with migration FROM Centos 8 (8.5) - to Rocky8 OR Rocky9 - still gets stucked after (a bit stupid) communicate which says:
Getting system package names for CentOS Linux 8…Error: No Matches found
Can’t get package that provides /etc/yum.repos.d/CentOS-Linux-Base.repo.
[…]
Why a bit stupid ?
Because I have very well prepared (cleaned from all inconsistences) Centos 8 system here
Because It is synchronized to the newest (as with Centos8 poss.) version with dnf distro-sync and other commands ..
Because the rpm database is consistent and no errors (dependecies errors) were produced during any "dnf install | dnf remove | dnf update | rpm … " operations
Because Centos itself (by packet manager) knows exactly "what repository matches for Centos Linux Base , and others … and recognize it properly (but migrate2rocky script has problems with it …)
… and finally ..
Because I have managed to migrate over than 30 other machines from Centos 8 to Rocky 8 in the past - and always , everything went well …
So may I gently and kindly ask You - @pajamian - for providing some help with this :
to take a look at the migrate2rocky script (to improve its behaviour to make it workable again)
or to tell people who still have problems - HOW to DEAL with them manually ??
migrate2rocky can be used to migrate from CentOS 8 to Rocky Linux 8, migrate2rocky9 can be used to migrate from another EL9 distro to Rocky Linux 9. Migrate2rocky is not capable of migrating from CentOS 8 to Rocky Linux 9. If you need this then the only supported means is to do a fresh install of Rocky Linux 9 and then migrate your apps and data over yourself.
You can also use Elevate which is provided by Alma Linux to migrate from CentOS 8 to Rocky Linux 9 but this is not supported by Rocky Linux.
Is this CentOS 8 or CentOS 8 Stream? Since vanilla CentOS 8 (non-stream) repositories are in the vault now. So unless the CentOS-Linux-Base.repo is configured to look at vault then it won’t download anything.
Thank You @pajamian - for Your answer , but NO - I dont want to migrate DIRECTLY from Centos 8 to Roxky 9 … What I try to do (for 4 days) - is to DO proper migration - from Centos 8 to Rocky 8 …!
And after that - I’ll plan to migrate OS from Rocky 8 to Rocky 9 - manually in normal way - just using system normall procedures during Distrybution Upgrade …
What I put on my screens above - was only because some people suggested to try “migrate2rocky9.sh” instead “migrate2rocky.sh” in case of migration problems … (but it obviously refused to work as well)
So - I need Your, and I will ask You for (some) support - in only migrating to Rocky 8 ..
@iwalker - this is Centos 8 os_version - not the Centos 8 Stream . Additionally - You may take a look of my (proper) links to vault… locations of the repositories of the Centos 8 (which I use frequently in other Centos systems..) - on screens :
So you both suggest me to use Elevate mechanism (script) from Alma - instead of trying using a “normal Rocky native method ?”
My question is - why ?
Is it really impossibile to use “migrate2rocky.sh” at all ??
Or - in other words - do you suggest that everyone who wants to move their systems from Centos to RockyLinux - use only mechanisms from other distributions (which probably may work…) - so that it can be successful? Because using Rocky mechanism - there’s no possibility of doing that … ?
I personally have never used either method. Obviously whoever created the script would need to check to find out why it’s not working and fix it. Or you can take a look in the script finding the error output you encountered and figure out what the script is looking for and fix accordingly the /var/log/migrate2rocky.log should give hints as to what is failing as well, so you can take a look in there.
I prefer to reinstall my systems from scratch and use things like ansible to redeploy configurations, install required packages etc and restore data. Converted systems don’t always work as intended and have some side-effects/problems.
Also, it helps much more for you to stop providing screenshots, and copy/paste the text from the console and paste it into your post using the formatting tools to display it correctly. It makes it easier to read, and for us to copy/paste errors to search google or whatever for potential fixes. Screenshots are only useful when copy/pasting text is not possible. So when you check the migrate2rocky.log for errors and think about pasting it here, make sure it’s not a screenshot. Provide the log as text in full.
@iwalker :
In general, I agree with the colleague. But the problems lie - as usual - in the details…
the result that the log listing gives - it does not add anything extra to what is displayed on the screen. Anyway, let’s see:
migrate2rocky - Begin logging at Sat 21 Jun 2025 14:51:41 CEST.
Removing dnf cache
Preparing to migrate CentOS Linux 8 to Rocky Linux 8.
Determining repository names for CentOS Linux 8......
Found the following repositories which map from CentOS Linux 8 to Rocky Linux 8:
CentOS Linux 8 Rocky Linux 8
appstream appstream
baseos baseos
extras extras
Getting system package names for CentOS Linux 8........Error: No Matches found
Can't get package that provides /etc/yum.repos.d/CentOS-Linux-Base.repo.
An error occurred while we were attempting to convert your system to Rocky Linux. Your system may be unstable. Script will now exit to prevent possible damage.
So the log ( /var/log/migrate2rocky.log ) does not provide any important information that could help pinpoint the cause of the problem
On my part - of course I could (in theory) take care of fixing Rocky Linux migration scripts (because I also have such skills) - but due to really tight time and amount of work - (mainly as DevOps and SOC manager) I simply cannot afford to “dig into someone else’s code” at the moment… I simply have to migrate one machine that is important to me anyway - and that is why I am looking for help from the source
And nobody wants to become a complete “no-life” because of their work, everyone has a family life too…
(in any case, all my migrations from C->R - that I have done in the past have been successful - hence the hope that it should still be possible to do… )
I am considering the issue of installing a new version of Rocky on the old Centos system - of course - but due to many complex solutions (e.g. resources mounted from different sources etc.) it will not be a trivial operation…
Besides, nobody likes to force open doors - if some migration has CHANCE to succeed, so why grab a “pickaxe and shovel” you know
I’m sorry for providing only screenshots - but I just thought that they are a bit more credible and exhaustive for you - because anyone can post anything as formatted text…etc.
which according to the script comes from the function called final_message, which in turn is called from exit_message function. Searching exit_message in the script shows 38 results, and as such could be one of the below functions:
The Getting system package names... comes from the collect_system_info function.
The Can't get package that provides comes from either the provides_pkg or collect_system_info function. Usually the .repo files should be in centos-repos rpm package. Maybe it’s not installed/missing?
rpm -qa | grep -i centos | grep -i repos
Also:
rpm -qf /etc/yum.repos.d/CentOS-Linux-Base.repo
that above command will confirm if it belongs to a package or not, or has been created manually by someone else. The migrate2rocky script relies on standard repo files belonging to a package like centos-repos or centos-stream-repos package.
So narrows it down a bit at least as to where the problem lies. Could well be whatever repo file you have under /etc/yum.repos.d does not belong to a package, and because of this the migrate2rocky is bombing out. Suggest checking as above to see if centos-repos package is installed, and if the name of the repo file does not match what is in the rpm, change the name so that it does match, or remove the repo file.
I’ve never seen a repo file called CentOS-Linux-Base.repo on any of my CentOS installations, means that this is caused by your system not conforming to standard CentOS installs.
@iwalker - thank You for this little help (little but not unimportant !) which indicates the ranges (zones) where the problem should be looked for ..
Tonight I will sit down with this problem and try to work it out methodically look through all the suggested places and possibly determine what is happening
(tonight, because for now I am going out for a family walk)
Of course I will let you know - about the results of my “investigation”