While testing a hotfix'ed RHEL 8.4 kernel, we observed the following backtrace when running the metal4K iso-install scenario: ``` [ 11.581955] systemd[1]: Mounting /sysroot... [ 11.729982] SGI XFS with ACLs, security attributes, quota, no debug enabled [ 11.740353] XFS (dm-4): Mounting V5 Filesystem [ 11.748288] XFS (dm-4): Internal error !uuid_equal(&mp->m_sb.sb_uuid, &head->h_fs_uuid) at line 279 of file fs/xfs/xfs_log_recover.c. Caller xlog_header_check_mount+0x35/0xb0 [xfs] [ 11.751400] CPU: 0 PID: 960 Comm: mount Not tainted 4.18.0-305.39.1.el8_4.x86_64 #1 [ 11.752868] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015 [ 11.754380] Call Trace: [ 11.754923] dump_stack+0x5c/0x80 [ 11.755604] xfs_corruption_error+0x8b/0x90 [xfs] [ 11.756546] ? xlog_header_check_mount+0x35/0xb0 [xfs] [ 11.757559] ? xlog_do_io+0x8d/0x130 [xfs] [ 11.758376] xlog_header_check_mount+0x5f/0xb0 [xfs] [ 11.759362] ? xlog_header_check_mount+0x35/0xb0 [xfs] [ 11.760377] xlog_find_verify_log_record+0x115/0x230 [xfs] [ 11.761454] xlog_find_head+0x1c3/0x390 [xfs] [ 11.762319] xlog_find_tail+0x44/0x350 [xfs] [ 11.763171] ? try_to_wake_up+0x1cc/0x590 [ 11.763966] xlog_recover+0x2b/0x160 [xfs] [ 11.764785] ? xfs_trans_ail_init+0xbd/0xd0 [xfs] [ 11.765717] xfs_log_mount+0x280/0x2a0 [xfs] [ 11.766566] xfs_mountfs+0x44f/0x8c0 [xfs] [ 11.767393] xfs_fc_fill_super+0x318/0x560 [xfs] [ 11.768308] ? xfs_mount_free+0x30/0x30 [xfs] [ 11.769174] get_tree_bdev+0x186/0x260 [ 11.769911] vfs_get_tree+0x25/0xb0 [ 11.770609] do_mount+0x2e2/0x950 [ 11.771271] ? memdup_user+0x4b/0x70 [ 11.771966] ksys_mount+0xb6/0xd0 [ 11.772622] __x64_sys_mount+0x21/0x30 [ 11.773376] do_syscall_64+0x5b/0x1a0 [ 11.774110] entry_SYSCALL_64_after_hwframe+0x65/0xca [ 11.775097] RIP: 0033:0x7f9bcc5a292e [ 11.775801] Code: 48 8b 0d 5d 15 2c 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 90 f3 0f 1e fa 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 2a 15 2c 00 f7 d8 64 89 01 48 [ 11.779370] RSP: 002b:00007ffeaa990d78 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5 [ 11.780822] RAX: ffffffffffffffda RBX: 0000556c27db3a70 RCX: 00007f9bcc5a292e [ 11.782198] RDX: 0000556c27db5c80 RSI: 0000556c27db3c50 RDI: 0000556c27db4960 [ 11.783566] RBP: 00007f9bcd34f184 R08: 0000000000000000 R09: 0000000000000002 [ 11.784934] R10: 00000000c0ed0000 R11: 0000000000000246 R12: 0000000000000000 [ 11.786308] R13: 00000000c0ed0000 R14: 0000556c27db4960 R15: 0000556c27db5c80 [ 11.787685] XFS (dm-4): Corruption detected. Unmount and run xfs_repair [ 11.788963] XFS (dm-4): log has mismatched uuid - can't recover [ 11.790110] XFS (dm-4): failed to find log head [ 11.790992] XFS (dm-4): log mount/recovery failed: error -117 [ 11.792532] XFS (dm-4): log mount failed [ 11.795059] ignition-ostree-firstboot-uuid[952]: Clearing log and setting UUID [ 11.797491] ignition-ostree-firstboot-uuid[952]: writing all SBs [ 11.798745] ignition-ostree-firstboot-uuid[952]: new UUID = 642c68d7-863e-4f1a-ace9-373f424e5ba6 [FAILED] Failed to mount /sysroot. ``` Investigation into the trace showed previous evidence of it reported upstream at https://github.com/coreos/fedora-coreos-tracker/issues/619#issuecomment-683925234 It's believed that backporting https://github.com/coreos/fedora-coreos-config/pull/1357 to RHCOS 4.9 would fix the race condition.
This patch is already in 4.11.