i have been successfully able to install Webmin/Virtualmin on a new RL-10.0 by first doing the following:
echo ‘10.0’ > /etc/dnf/vars/releasever ; ## should this be removed afterward?
dnf -y install cyrus-sasl cyrus-sasl-devel ; ## webmin had trouble with this
dnf clean packages ; ## first time on 10.0 this was required for some reason
furthermore, i have successfully installed webmin/virtualmin on a new 10.1 by first doing the following:
dnf -y install cyrus-sasl cyrus-sasl-devel ; ## webmin had trouble with this
the webmin tech support suggested i stick with 10.0 for now. any opinions on this?
however, they suggested instead of manually creating a /etc/dnf/vars/releasever file, i should go in and edit the epel files.
i have never edited an epel file before. tech support suggested i edit a epel.repo file that does not exist.
it sounds as if they are suggesting i edit the rocky.epel file. i am guessing they are suggesting i change all 75 occurrences of $releasever to 10.0?
i am thinking its better just to create the /etc/dnf/vars/releasever file and then delete it afterwards.
please share any opinions on this with me since i have never done anything in this area before.
thank you. and please note that i need to stick with 10.0 for now.
what i was trying to do was to duplicate somehow whatever the shortcodes were doing with the hard-coded values. and i had no idea really how to convert ${releasever_minor:±z}except to just substitute it with 0-z
Sorry, I use a specified baseurl on my computer, not metalink. For metalink, it seems the URL “https://mirrors.fedoraproject.org/metalink?repo=epel-z-10.0&arch=x86_64“ is the only valid one if you need to stick with 10.0.
Much easier is to create a file /etc/dnf/vars/releasever and then in this file put 10.0 or 10.1. Then there is no need to edit the epel.repo file. However, it does mean all repo files that use the $releasever variable will use that value.
I presume that the mirrorlist uses same set of repo-ids as the metalink.
However, unknown id to mirrorlist gives a list of known ids.
The latter above – epel-10.0 – is not known. Curiously, the epel-10 is not on the list either – there must be some additional redirects as it nevertheless gives the epel-z-10.2
this certainly appears to be true. but for whatever reason, there appears to be a debate that has religious overtones for how strongly some people are against it.
rumor has it that whatever the epel problem was is now fixed so this post is now unnecessary.. i am testing it now.