Description of problem: Version-Release number of selected component (if applicable): 5.10.45 Steps to Reproduce: Attempt to update from 5.10.45
That can happen either due to corruption during the download from the mirror, or during the download from the main Mageia repository to the mirror. Please run (as root) urpmi --clean and then try installing the update again. If it still fails please post the output of urpmq --list-url|grep "Core Updates" so we know which mirror to check.
CC: (none) => davidwhodgins
I second Dave's comment. I happen to have just done a system update, including the kernel in question 'kernel-desktop-5.10.46-1.mga8-1-1.mga8', without problem. So it is not inherent in the package.
CC: (none) => lewyssmith
> Please run (as root) urpmi --clean and then try installing the update again. > That worked, thanks. It appears we had a Heisenbug, though it happened on several attempts. I consider this closed.
Status comment: (none) => Fixed
Just fyi, the reason it failed on multiple attempts is that the initially corrupted download was still in /var/cache/urpmi/rpms/ Repeating the attempt to download/get the update finds the file already there, doesn't download it, and just repeats trying to install the corrupt copy. The urpmi command with the --clean option deletes the previously downloaded copy. That's not done by default so that a partially downloaded file can resume downloading from where it left off. That's needed for those who have a flaky connection, or who pay by the byte (MB/GB, etc.) for downloads in which case redoing the whole download costs them actual money. Linux isn't M$. :-)
I should add that resuming from where it left off doesn't work if the whole file was transferred, but with a bit incorrect, which is likely the case here. Closing as invalid, since it's not a Mageia bug.
Resolution: (none) => INVALIDStatus: NEW => RESOLVEDStatus comment: Fixed => (none)
(In reply to Dave Hodgins from comment #5) Thanks for the assistance, and the very clear explanation. I appreciate it. > Closing as invalid, since it's not a Mageia bug. Despite which, Mageia produces a solution. :-)*