Re: [Evms-devel] "md: md0 still in use" just before reboot / poweroff

[Kevin C. <kev...@us...>]

Hi Christian,

On Thu June 29 2006 11:56 pm, Christian Pernegger wrote:
> I'm having a problem with an EVMS-managed RAID5 device -- it won't
> shutdown cleanly:
>
> [...]
> Will now halt.
> md: stopping all md devices.
> md: md0 still in use.
> Synchronizing SCSI cache for disk /dev/sdd:
> Synchronizing SCSI cache for disk /dev/sdc:
> Synchronizing SCSI cache for disk /dev/sdb:
> Synchronizing SCSI cache for disk /dev/sda:
> Shutdown: hde
> System halted.
>
> All disks are managed by EVMS but root is not on the RAID.
>
> EVMS 2.5.5-7, Linux 2.6.17-1-686 (both from Debian testing)
>
> The shutdown sequence above is always the same -- often the array
> comes up clean after that anyway, at least twice an automatic rebuild
> was started on next boot. Which might even be entirely unrelated.
>
> Nevertheless from the messages it looks that md0 couldn't be shutdown.
> Is this normal with EVMS?
>
> The box will be rebooted only to exchange failed drives and install
> kernel security updates -- having the md not shutdown cleanly when the
> array is degraded already doesn't sound like fun.

The messages you're seeing are from the kernel's MD/Software-RAID driver. 
Looking at the code in linux-2.6.17/drivers/md/md.c::do_md_stop(), it is 
checking the RAID device for an "active" value greater than 2. If so, it 
prints the above error message and refuses to stop that device.

Did this same problem occur with older kernels?

Can you describe your volume configuration in more detail (i.e., disks, 
segments, raid regions, lvm regions, volumes)?

Thanks,
-- 
Kevin Corry
kev...@us...
http://www.ibm.com/linux/
http://evms.sourceforge.net/