Bug 25420

Description of problem:
The XFS filesystem on a raw disk crashes after updating packages on the root filesystem which is on a different disk.

Version-Release number of selected component (if applicable):
drakxtools-backend 18.21-2.mga8

How reproducible:

Steps to Reproduce:
1. Create a XFS filesystem on /dev/sdX (not on a partition), mount it, write some files to it
2. update a bunch of packages on the system [my guess: install a new kernel]
3. wait a little while

Jul 22 10:35:15 zem kernel: [2208802.716374] XFS (sdb): Metadata CRC error detected at xfs_inobt_read_verify+0x15/0xc0 [xfs], xfs_inobt block 0x38 
Jul 22 10:35:15 zem kernel: [2208802.721897] XFS (sdb): Unmount and run xfs_repair
Jul 22 10:35:15 zem kernel: [2208802.727379] XFS (sdb): First 128 bytes of corrupted metadata buffer:
Jul 22 10:35:15 zem kernel: [2208802.732798] 00000000: 49 41 42 33 00 00 00 f6 ff ff ff ff 00 00 00 14  IAB3............
Jul 22 10:35:15 zem kernel: [2208802.732800] 00000010: 00 00 00 00 00 00 00 38 00 00 00 60 00 03 d1 ce  .......8...`....
Jul 22 10:35:15 zem kernel: [2208802.732801] 00000020: f2 93 cc 5d e2 66 45 c5 bf e4 e1 25 ad 3d 2f 11  ...].fE....%.=/.
Jul 22 10:35:15 zem kernel: [2208802.732803] 00000030: 00 00 00 00 65 7d 3a 6b 00 00 00 60 00 00 00 00  ....e}:k...`....
Jul 22 10:35:15 zem kernel: [2208802.732804] 00000040: 00 00 00 00 00 00 00 00 00 00 09 40 00 00 00 00  ...........@....
Jul 22 10:35:15 zem kernel: [2208802.732805] 00000050: 00 00 00 00 00 00 00 00 00 00 09 80 00 00 00 00  ................
Jul 22 10:35:15 zem kernel: [2208802.732806] 00000060: 00 00 00 00 00 00 00 00 00 00 09 c0 00 00 00 00  ................
Jul 22 10:35:15 zem kernel: [2208802.732807] 00000070: 00 00 00 00 00 00 00 00 00 01 63 20 00 00 00 00  ..........c ....
Jul 22 10:35:15 zem kernel: [2208802.733199] XFS (sdb): metadata I/O error in "xfs_trans_read_buf_map" at daddr 0x38 len 8 error 74
Jul 22 10:35:15 zem kernel: [2208802.929576] XFS (sdb): xfs_do_force_shutdown(0x1) called from line 328 of file fs/xfs/xfs_trans_buf.c. Return address = 00000000136a611d
Jul 22 10:35:15 zem kernel: [2208802.936027] XFS (sdb): I/O Error Detected. Shutting down filesystem
Jul 22 10:35:15 zem kernel: [2208802.942330] XFS (sdb): Please unmount the filesystem and rectify the problem(s)
Jul 22 10:35:15 zem kernel: [2208802.948709] XFS (sdb): xfs_difree_inobt: xfs_inobt_lookup() returned error -117.
Jul 22 10:35:15 zem kernel: [2208802.950922] XFS (sdb): xfs_imap_to_bp: xfs_trans_read_buf() returned error -5.
Jul 22 10:35:15 zem kernel: [2208802.976538] XFS (sdb): xfs_imap_to_bp: xfs_trans_read_buf() returned error -5.

I'm filing this bug on drakxtools-backend because I guess it's caused by bootloader-config, maybe by it calling test_for_bad_drives, which writes to sector 62, matching corruption in 0x38-0x40