[clug] f14->f15 upgrade: now unclean boot (mdadm issue)

Thu Mar 8 01:14:38 MST 2012

An off-list suggestion was this may be the result of the switch from sysv init
to systemd. This is also my opinion as this looks like a dependency issue.

However, since RAID is rather common, and this is a simple setup (boot from
bare disk, RAID for data), I am surprised to see this case not handled
properly. This may be the result of the stuttering install process though.

Anyone knows the correct way to specify that RAID should be started before
/dev/md0 is mounted? Probably some xxx.wants somewhere to say
"fstab mounts depends on mdadm"?

TIA
	Eyal

On 03/07/12 22:05, Eyal Lebedinsky wrote:
> The upgrade (using preupgrade) was a nightmare (and I still have f15->f16 to do).
>
> I slowly picked up the pieces and now have one issue.
>
> I boot from sda and then have a RAID /dev/md0 where most of my data lives as /data1.
> The bootup attempts to mount /data1 before mdadm completed the preparation of /dev/md0.
>
> At least this is how it looks to me. Anyone understands what is happening and how to
> fix it (short of *not* auto-mounting /data1 and doing it in rc.local)?
>
> It seems like a common thing (if one trusts the talk on the web). Is this also the case
> in f16? If not then I will suffer quietly until the next upgrade (this weekend if one
> believed the weather forecast) (not anymore, clear weekend expected now...) since I
> do not reboot this server too often.
>
> Here is the log around that time:
>
> [ 14.980801] mdadm: sending ioctl 800c0910 to a partition!
> [ 14.981484] mdadm: sending ioctl 800c0910 to a partition!
> [ 14.981487] mdadm: sending ioctl 800c0910 to a partition!
> [ 14.981493] mdadm: sending ioctl 1261 to a partition!
> [ 14.981494] mdadm: sending ioctl 1261 to a partition!
> [ 14.981672] mdadm: sending ioctl 1261 to a partition!
> [ 14.981674] mdadm: sending ioctl 1261 to a partition!
> [ 14.987718] mdadm: sending ioctl 1261 to a partition!
> [ 14.987725] mdadm: sending ioctl 800c0910 to a partition!
> [ 14.987744] mdadm: sending ioctl 1261 to a partition!
> ...
> [ 23.750596] mdadm: sending ioctl 800c0910 to a partition!
> [ 23.750599] mdadm: sending ioctl 1261 to a partition!
> [ 23.750615] mdadm: sending ioctl 1261 to a partition!
> [ 23.750628] mdadm: sending ioctl 1261 to a partition!
> [ 23.750641] mdadm: sending ioctl 1261 to a partition!
> [ 23.750646] mdadm: sending ioctl 1261 to a partition!
> [ 23.750648] mdadm: sending ioctl 1261 to a partition!
> [ 23.750664] mdadm: sending ioctl 1261 to a partition!
> [ 23.750669] mdadm: sending ioctl 1261 to a partition!
> [ 23.750675] mdadm: sending ioctl 1261 to a partition!
> [ 23.782169] mtp-probe[748]: bus: 3, device: 2 was not an MTP device
> [ 23.783177] mtp-probe[658]: bus: 6, device: 2 was not an MTP device
> [ 23.784118] mtp-probe[659]: bus: 6, device: 3 was not an MTP device
> [ 23.785022] mtp-probe[749]: bus: 4, device: 2 was not an MTP device
> [ 23.796395] md: bind<sdg1>
> [ 23.800054] md: bind<sde1>
> [ 23.803426] md: bind<sdf1>
> [ 23.807301] md: bind<sdd1>
> [ 23.811194] md: bind<sdb1>
> [ 23.815536] md: bind<sdc1>
> [ 23.873241] EXT4-fs (md0): unable to read superblock
> [ 23.880180] mount[867]: mount: wrong fs type, bad option, bad superblock on /dev/md0,
> [ 23.880525] systemd[1]: data1.mount mount process exited, code=exited status=32
> [ 23.881648] mount[867]: missing codepage or helper program, or other error
> [ 23.882328] mount[867]: (could this be the IDE device where you in fact use
> [ 23.882986] mount[867]: ide-scsi so that sr0 or sda or so is needed?)
> [ 23.883656] mount[867]: In some cases useful info is found in syslog - try
> [ 23.884342] mount[867]: dmesg | tail or so
> [ 23.894132] systemd[1]: Job fedora-autorelabel-mark.service/start failed with result 'dependency'.
> [ 23.894957] systemd[1]: Job fedora-autorelabel.service/start failed with result 'dependency'.
> [ 23.894963] systemd[1]: Job local-fs.target/start failed with result 'dependency'.
> [ 23.894968] systemd[1]: Triggering OnFailure= dependencies of local-fs.target.
> [ 23.897923] systemd[1]: Job boot.mount/start failed with result 'dependency'.
> [ 23.898738] systemd[1]: Job fsck at dev-disk-by\x2duuid-f9f345f0\x2d39d7\x2d4b2b\x2d8ed0\x2d57b25da62edf.service/start failed with result 'dependency'.
> [ 23.899556] systemd[1]: Unit data1.mount entered failed state.
>
> Above is the failure line leading to an emergency shell.
>
> [ 23.990638] async_tx: api initialized (async)
> [ 24.152290] xor: automatically using best checksumming function: generic_sse
> [ 24.158008] generic_sse: 9248.000 MB/sec
> [ 24.159064] xor: using function: generic_sse (9248.000 MB/sec)
> [ 24.374029] raid6: int64x1 1949 MB/s
> [ 24.392017] raid6: int64x2 2625 MB/s
> [ 24.410011] raid6: int64x4 1996 MB/s
> [ 24.428017] raid6: int64x8 1695 MB/s
> [ 24.446022] raid6: sse2x1 4449 MB/s
> [ 24.464020] raid6: sse2x2 6343 MB/s
> [ 24.482019] raid6: sse2x4 7300 MB/s
> [ 24.483084] raid6: using algorithm sse2x4 (7300 MB/s)
> [ 24.645247] md: raid6 personality registered for level 6
> [ 24.646319] md: raid5 personality registered for level 5
> [ 24.647381] md: raid4 personality registered for level 4
> [ 24.648613] bio: create slab <bio-1> at 1
> [ 24.649720] md/raid:md0: device sdc1 operational as raid disk 1
> [ 24.650813] md/raid:md0: device sdb1 operational as raid disk 0
> [ 24.651905] md/raid:md0: device sdd1 operational as raid disk 2
> [ 24.652997] md/raid:md0: device sdf1 operational as raid disk 4
> [ 24.654084] md/raid:md0: device sde1 operational as raid disk 3
> [ 24.655144] md/raid:md0: device sdg1 operational as raid disk 5
> [ 24.656670] md/raid:md0: allocated 6384kB
> [ 24.657769] md/raid:md0: raid level 6 active with 6 out of 6 devices, algorithm 2
> [ 24.658817] RAID conf printout:
> [ 24.658819] --- level:6 rd:6 wd:6
> [ 24.658820] disk 0, o:1, dev:sdb1
> [ 24.658822] disk 1, o:1, dev:sdc1
> [ 24.658824] disk 2, o:1, dev:sdd1
> [ 24.658825] disk 3, o:1, dev:sde1
> [ 24.658826] disk 4, o:1, dev:sdf1
> [ 24.658828] disk 5, o:1, dev:sdg1
> [ 24.658963] created bitmap (15 pages) for device md0
> [ 24.660378] md0: bitmap initialized from disk: read 1/1 pages, set 0 of 29809 bits
> [ 24.688848] md0: detected capacity change from 0 to 8001578598400
> [ 24.751155] systemd[1]: Startup finished in 2s 879ms 556us (kernel) + 15s 113ms 278us (initrd) + 6s 758ms 257us (userspace) = 24s 751ms 91us.
>
> At this point I log in, mount /data1 (no problem) and exit for the boot to complete. Note how the
> device completed coming up just above.
>
> My mount:
>
> [ 37.378259] md0: unknown partition table
> [ 37.601528] EXT4-fs (md0): mounted filesystem with ordered data mode. Opts: (null)
>
> exit, and now the bootup continues:
>
> [ 39.959976] Adding 16458588k swap on /dev/sda2. Priority:0 extents:1 across:16458588k
> ...
>

-- 
Eyal Lebedinsky	(eyal at eyal.emu.id.au)