# cd /boot # ll init* -rw------- 1 root root 21577326 feb 21 20:22 initrd-5.10.16-desktop-1.mga8.img -rw------- 1 root root 12675779 apr 22 19:40 initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 21 20:27 initrd-desktop.img -> initrd-5.10.16-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 21 20:27 initrd.img -> initrd-5.10.16-desktop-1.mga8.img # rm initrd.img # rm initrd-desktop.img # ln -s initrd-5.10.30-desktop-1.mga8.img initrd-desktop.img # ln -s initrd-5.10.30-desktop-1.mga8.img initrd.img # ll init* -rw------- 1 root root 21577326 feb 21 20:22 initrd-5.10.16-desktop-1.mga8.img -rw------- 1 root root 12675779 apr 22 19:40 initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 22 20:20 initrd-desktop.img -> initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 22 20:16 initrd.img -> initrd-5.10.30-desktop-1.mga8.img Note also the huge difference in size. It's very strange for me that content of both initrd seems very different (not just the kernel version).
On my rpi4 ... $ ls -l /boot/*init* -rw------- 1 root root 14027069 Mar 31 07:27 /boot/initrd-5.10.27-desktop-1.mga8.img -rw------- 1 root root 14025220 Apr 10 00:41 /boot/initrd-5.10.28-desktop-1.mga8.img -rw------- 1 root root 14024672 Apr 10 13:19 /boot/initrd-5.10.29-desktop-1.mga8.img -rw------- 1 root root 14024347 Apr 15 13:06 /boot/initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 Apr 15 13:06 /boot/initrd-desktop.img -> initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 Apr 15 13:06 /boot/initrd.img -> initrd-5.10.30-desktop-1.mga8.img So it's working ok on mine. Did you create a separate partition for /boot? Also check ... [dave@rp4 ~]$ df -h |grep -v tmpfs Filesystem Size Used Avail Use% Mounted on /dev/mmcblk0p2 22G 8.1G 13G 40% / /dev/mmcblk0p1 1021M 24M 998M 3% /boot/EFI [dave@rp4 ~]$ df -hi |grep -v tmpfs Filesystem Inodes IUsed IFree IUse% Mounted on /dev/mmcblk0p2 1.4M 255K 1.2M 19% / /dev/mmcblk0p1 0 0 0 - /boot/EFI
CC: (none) => davidwhodgins
That's odd. Which image did you install from? I can try fresh dd on USB key and boot from there to double check if only noDE image is affected. $ LC_ALL=C df -h | grep -v tmpfs Filesystem Size Used Avail Use% Mounted on /dev/mmcblk0p2 20G 11G 8.6G 55% / /dev/sda1 877G 400G 433G 49% /mnt/SSD /dev/mmcblk0p1 128M 76M 53M 59% /boot/EFI $ LC_ALL=C df -hi | grep -v tmpfs Filesystem Inodes IUsed IFree IUse% Mounted on /dev/mmcblk0p2 1.3M 189K 1.1M 15% / /dev/sda1 56M 2.0M 54M 4% /mnt/SSD /dev/mmcblk0p1 0 0 0 - /boot/EFI
CC: (none) => filip.komar
(In reply to Filip Komar from comment #0) > # ll init* > -rw------- 1 root root 21577326 feb 21 20:22 initrd-5.10.16-desktop-1.mga8.img > -rw------- 1 root root 12675779 apr 22 19:40 initrd-5.10.30-desktop-1.mga8.img > Note also the huge difference in size. > It's very strange for me that content of both initrd seems very different > (not just the kernel version). Any comment on the content/size matter?
(In reply to Filip Komar from comment #3) > (In reply to Filip Komar from comment #0) > > # ll init* > > -rw------- 1 root root 21577326 feb 21 20:22 initrd-5.10.16-desktop-1.mga8.img > > -rw------- 1 root root 12675779 apr 22 19:40 initrd-5.10.30-desktop-1.mga8.img > > > Note also the huge difference in size. > > It's very strange for me that content of both initrd seems very different > > (not just the kernel version). > > Any comment on the content/size matter? I assume the first one is a install-time "include-all", whereas the second one only contains stuff that dracut detect is needed to boot
From rpm -q --scripts kernel-desktop-5.10.30-1.mga8-1-1.mga8, it has the lines ... pushd /boot > /dev/null if [ -e initrd-5.10.30-desktop-1.mga8.img ]; then ln -sf vmlinuz-5.10.30-desktop-1.mga8 vmlinuz-desktop ln -sf vmlinuz-5.10.30-desktop-1.mga8 vmlinuz ln -sf initrd-5.10.30-desktop-1.mga8.img initrd-desktop.img ln -sf initrd-5.10.30-desktop-1.mga8.img initrd.img fi popd > /dev/null Please save the lines from the pushd to the popd, inclusive to a text file, in /root, then run (as root) "bash -x -v ./thattextfile". Does it create the symlinks?
Thanks to Dave & tmb for your comments. Leaving with bugsquad for now.
Status: NEW => NEEDINFOCC: (none) => lewyssmith
I tried fresh dd on USB key and boot from there with no other change to double check if only noDE image is affected and the results were very different. Besides constant very annoying and useless (due to USB boot so w/o SD) emmc timeout errors I noticed that the initrd-5.10.16 was rebuild during update. I don't know if that is intentionally but I guess it might be useful in some cases. The sizes are comparable and the symlinks are fine. Unfortunately I forget to use tee before urpmi.update -a && urpmi --auto-select but the end result is here. cat /run/media/user/e45a161d-aa49-4c6c-b8b7-af67a32a9f80/root/ll.txt -rw------- 1 root root 12018119 Apr 23 22:43 /boot/initrd-5.10.16-desktop-1.mga8.img -rw------- 1 root root 12015024 Apr 23 22:56 /boot/initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 Apr 23 22:57 /boot/initrd-desktop.img -> initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 Apr 23 22:57 /boot/initrd.img -> initrd-5.10.30-desktop-1.mga8.img Forgot to add that I didn't add or remove any partition on original SD but I did resize root one in gparted before first boot. AFAIK that should not matter. (In reply to Thomas Backlund from comment #4) > (In reply to Filip Komar from comment #3) > > Any comment on the content/size matter? > > I assume the first one is a install-time "include-all", whereas the second > one only contains stuff that dracut detect is needed to boot Makes sense. Thx to all. Strange. Closing as noise ;). I really hope that SD card is not acting up or even dying. Quick Gparted test didn't find any errors. Need to test more.
Status: NEEDINFO => RESOLVEDResolution: (none) => INVALID
(In reply to Dave Hodgins from comment #5) > Does it create the symlinks? Yes. I'm unsure but I reopened as I reproduced the bug with old kernel install and it's kinda funny as symlink were not from the latest installed nor bigger version but last one in urpmi line. $ sudo urpmi kernel-desktop-5.10.27-1.mga8 kernel-desktop-5.10.25-1.mga8 $ ls -l /boot/*init* -rw------- 1 root root 21577326 feb 21 20:22 /boot/initrd-5.10.16-desktop-1.mga8.img -rw------- 1 root root 13024789 apr 24 01:58 /boot/initrd-5.10.25-desktop-1.mga8.img -rw------- 1 root root 13024995 apr 24 01:57 /boot/initrd-5.10.27-desktop-1.mga8.img -rw------- 1 root root 12675779 apr 22 19:40 /boot/initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 24 01:58 /boot/initrd-desktop.img -> initrd-5.10.25-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 24 01:58 /boot/initrd.img -> initrd-5.10.25-desktop-1.mga8.img # chmod u+x bug_test.sh # bash -x -v ./bug_test.sh pushd /boot > /dev/null + pushd /boot if [ -e initrd-5.10.30-desktop-1.mga8.img ]; then ln -sf vmlinuz-5.10.30-desktop-1.mga8 vmlinuz-desktop ln -sf vmlinuz-5.10.30-desktop-1.mga8 vmlinuz ln -sf initrd-5.10.30-desktop-1.mga8.img initrd-desktop.img ln -sf initrd-5.10.30-desktop-1.mga8.img initrd.img fi + '[' -e initrd-5.10.30-desktop-1.mga8.img ']' + ln -sf vmlinuz-5.10.30-desktop-1.mga8 vmlinuz-desktop + ln -sf vmlinuz-5.10.30-desktop-1.mga8 vmlinuz + ln -sf initrd-5.10.30-desktop-1.mga8.img initrd-desktop.img + ln -sf initrd-5.10.30-desktop-1.mga8.img initrd.img popd > /dev/null + popd # ls -l /boot/*init* -rw------- 1 root root 21577326 feb 21 20:22 /boot/initrd-5.10.16-desktop-1.mga8.img -rw------- 1 root root 13024789 apr 24 01:58 /boot/initrd-5.10.25-desktop-1.mga8.img -rw------- 1 root root 13024995 apr 24 01:57 /boot/initrd-5.10.27-desktop-1.mga8.img -rw------- 1 root root 12675779 apr 22 19:40 /boot/initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 24 02:07 /boot/initrd-desktop.img -> initrd-5.10.30-desktop-1.mga8.img lrwxrwxrwx 1 root root 33 apr 24 02:07 /boot/initrd.img -> initrd-5.10.30-desktop-1.mga8.img
Status: RESOLVED => REOPENEDResolution: INVALID => (none)
(In reply to Filip Komar from comment #8) > (In reply to Dave Hodgins from comment #5) > > Does it create the symlinks? > Yes. > > I'm unsure but I reopened as I reproduced the bug with old kernel install > and it's kinda funny as symlink were not from the latest installed nor > bigger version but last one in urpmi line. please check "rpm -qa --last |less" for the install order
Thanks reporting this. Assigning.
Status: REOPENED => NEWAssignee: bugsquad => pkg-bugsComponent: Release (media or process) => RPM PackagesCC: sysadmin-bugs => j.biernacki+mga, ouaurelien, pterjan
(In reply to Thomas Backlund from comment #9) > (In reply to Filip Komar from comment #8) > > (In reply to Dave Hodgins from comment #5) > > > Does it create the symlinks? > > Yes. > > > > I'm unsure but I reopened as I reproduced the bug with old kernel install > > and it's kinda funny as symlink were not from the latest installed nor > > bigger version but last one in urpmi line. > > > please check "rpm -qa --last |less" for the install order # rpm -qa --last |head -3 kernel-desktop-5.10.25-1.mga8-1-1.mga8.aarch64 sob 24 apr 2021 01:56:05 kernel-desktop-5.10.27-1.mga8-1-1.mga8.aarch64 sob 24 apr 2021 01:56:01 hdparm-9.60-1.mga8.aarch64 čet 22 apr 2021 22:21:46
Yes the size difference is expected, as Thomas explained in comment #4. The image is not using hostonly as it is not generated on that hardware (and the image could in theory be used on several types of storage).
If the symlinks are pointing to the latest kernel installed (not necessarily the one with the highest version number), then it's working as designed.
(In reply to Dave Hodgins from comment #13) > If the symlinks are pointing to the latest kernel installed (not necessarily > the one with the highest version number), then it's working as designed. Ok. Implications are small I guess. And with simple workaround. I'll wait for another kernel update. Or another test report. Bug 24084 is way more critical anyway.
Status: NEW => UNCONFIRMEDEver confirmed: 1 => 0
Seems fine with 5.10.30 and 33.
Resolution: (none) => FIXEDStatus: UNCONFIRMED => RESOLVED