Description of problem:

Looks like anaconda writes out files to /tmp in the chrooted installation environment, which doesn't use tmpfs on /tmp. So these files are in /tmp on disk, and then when systemd mounts tmpfs on /tmp it complains that it's not empty.


Version-Release number of selected component (if applicable):
anaconda 23.19.10-1

How reproducible:
Always


Steps to Reproduce:
1. Do an installation
2. Reboot
3.

Actual results:

Feb 11 19:48:10 f23s.localdomain systemd[1]: tmp.mount: Directory /tmp to mount over is not empty, mounting anyway.

Looking at actual /tmp without tmpfs mounted over it:
$ ls -l /mnt/tmp
total 4
drwxr-xr-x. 2 root root  19 Feb  6 04:59 hsperfdata_root
-rwx------. 1 root root 852 Feb 11 11:49 ks-script-bytw8xx2

And also:
$ grep ks-script /var/log/anaconda/program.log  
11:49:49,769 INFO program: Running... /bin/sh /tmp/ks-script-bytw8xx2
11:49:53,161 INFO program: Running... /bin/sh /tmp/ks-script-hy66_vxy
11:49:53,186 INFO program: Running... /bin/sh /tmp/ks-script-iqzupo_o



Expected results:

I don't know if these ks files have any long term importance, if they do they probably should get written to /var/log/anaconda rather than either live tmpfs on /tmp, or chrooted enviro's root fs /tmp. Or use /var/tmp I guess, but not /tmp.

Additional info: