Since Fedora-Rawhide-20190517.n.0, when kernel-5.2.0-0.rc0.git6.1.fc31 appeared, openQA ext3 install tests are all failing. Every time, it seems to fail because the ext3 root partition is remounted read-only shortly after it is initially mounted; this obviously breaks the install process. These are the relevant messages from the system journal: 22:05:56,542 INFO kernel:EXT4-fs (vda3): mounting ext3 file system using the ext4 subsystem 22:05:56,549 INFO kernel:EXT4-fs (vda3): mounted filesystem with ordered data mode. Opts: (null) ... 22:06:37,040 CRIT kernel:EXT4-fs error (device vda3): ext4_get_branch:171: inode #8: block 7165: comm jbd2/vda3-8: invalid block 22:06:37,041 ALERT kernel:jbd2_journal_bmap: journal block not found at offset 6156 on vda3-8 22:06:37,041 ERR kernel:Aborting journal on device vda3-8. 22:06:37,046 CRIT kernel:EXT4-fs error (device vda3): ext4_journal_check_start:61: Detected aborted journal 22:06:37,046 CRIT kernel:EXT4-fs (vda3): Remounting filesystem read-only Proposing as a Final blocker per criterion "The installer must be able to create and install to any workable partition layout using any file system and/or container format combination offered in a default installer configuration" - https://fedoraproject.org/wiki/Fedora_30_Final_Release_Criteria#Disk_layouts jforbes says this is known upstream, but I want to have a Fedora bug for blocker and tracking purposes.
Looks like this got fixed in the 20190524.n.1 compose; the test hasn't failed since then. There was a new kernel in that compose, so that checks out, looks like we just got a fix from upstream.