Should raidtools-19990309-0.90 manage a linear device without a SB ?
[ I can "mkraid" it, but once stopped, it can never be restarted ]
md8 fails, md7 is fine.
==============================================================================
raiddev /dev/md7
raid-level linear
chunk-size 32
persistent-superblock 1
nr-raid-disks 2
device /dev/hda10
raid-disk 0
device /dev/hda11
raid-disk 1
raiddev /dev/md8
raid-level linear
chunk-size 8
persistent-superblock 0
nr-raid-disks 2
device /dev/hda12
raid-disk 0
device /dev/hda13
raid-disk 1
==============================================================================
# raidstop /dev/md8
Mar 19 11:29:34 dean kernel: marking sb clean...
Mar 19 11:29:34 dean kernel: unbind<hda13,1>
Mar 19 11:29:34 dean kernel: export_rdev(hda13)
Mar 19 11:29:34 dean kernel: unbind<hda12,0>
Mar 19 11:29:34 dean kernel: export_rdev(hda12)
Mar 19 11:29:34 dean kernel: md8 stopped.
# raidstart /dev/md8
/dev/md8: Invalid argument
Mar 19 11:29:50 dean kernel: (read) hda12's sb offset: 3904 [events: ffffffff]
Mar 19 11:29:50 dean kernel: md: invalid raid superblock magic on hda12
Mar 19 11:29:50 dean kernel: md: hda12 has invalid sb, not importing!
Mar 19 11:29:50 dean kernel: could not import hda12!
Mar 19 11:29:50 dean kernel: autostart hda12 failed!
Mar 19 11:29:50 dean kernel: huh12?
# raidstart --version
raidstart v0.3d compiled for md raidtools-0.90
# raidstop /dev/md7
Mar 19 11:34:35 dean kernel: marking sb clean...
Mar 19 11:34:35 dean kernel: md: updating md7 RAID superblock on device
Mar 19 11:34:35 dean kernel: hda11 [events: 00000009](write) hda11's sb
offset: 3904
Mar 19 11:34:35 dean kernel: hda10 [events: 00000009](write) hda10's sb
offset: 3904
Mar 19 11:34:35 dean kernel: .
Mar 19 11:34:35 dean kernel: unbind<hda11,1>
Mar 19 11:34:35 dean kernel: export_rdev(hda11)
Mar 19 11:34:35 dean kernel: unbind<hda10,0>
Mar 19 11:34:35 dean kernel: export_rdev(hda10)
Mar 19 11:34:35 dean kernel: md7 stopped.
# raidstart /dev/md7
Mar 19 11:34:47 dean kernel: (read) hda10's sb offset: 3904 [events: 00000009]
Mar 19 11:34:47 dean kernel: (read) hda11's sb offset: 3904 [events: 00000009]
Mar 19 11:34:47 dean kernel: autorun ...
Mar 19 11:34:47 dean kernel: considering hda11 ...
Mar 19 11:34:47 dean kernel: adding hda11 ...
Mar 19 11:34:47 dean kernel: adding hda10 ...
Mar 19 11:34:47 dean kernel: created md7
Mar 19 11:34:47 dean kernel: bind<hda10,1>
Mar 19 11:34:47 dean kernel: md7: WARNING: hda11 appears to be on the same
physical disk as hda10. True
Mar 19 11:34:47 dean kernel: protection against single-disk failure might
be compromised.
Mar 19 11:34:47 dean kernel: bind<hda11,2>
Mar 19 11:34:47 dean kernel: running: <hda11><hda10>
Mar 19 11:34:47 dean kernel: now!
Mar 19 11:34:47 dean kernel: hda11's event counter: 00000009
Mar 19 11:34:47 dean kernel: hda10's event counter: 00000009
Mar 19 11:34:47 dean kernel: mask ffffffe0
Mar 19 11:34:47 dean kernel: rdev->size: 3904
Mar 19 11:34:47 dean kernel: masked rdev->size: 3904
Mar 19 11:34:47 dean kernel: new md_size: 3904
Mar 19 11:34:47 dean kernel: rdev->size: 3904
Mar 19 11:34:47 dean kernel: masked rdev->size: 3904
Mar 19 11:34:47 dean kernel: new md_size: 7808
Mar 19 11:34:47 dean kernel: md7: max total readahead window set to 128k
Mar 19 11:34:47 dean kernel: md7: 1 data-disks, max readahead per data-disk:
128k
Mar 19 11:34:47 dean kernel: md: updating md7 RAID superblock on device
Mar 19 11:34:47 dean kernel: hda11 [events: 0000000a](write) hda11's sb
offset: 3904
Mar 19 11:34:47 dean kernel: hda10 [events: 0000000a](write) hda10's sb
offset: 3904
Mar 19 11:34:47 dean kernel: .
Mar 19 11:34:47 dean kernel: ... autorun DONE.
