"array md0 already has disks" with kernel 2.6.29

I've been running EVMS 2.5.5 for years.  I have a hand-built initrd that calls
evms_activate (I know, old school).  I have two MD/RAID1 regions: md0 mounted on
/boot, and md1 mounted on /.

This setup has worked fine up through kernel 2.6.26.  Now when I try to upgrade
to kernel 2.6.29, boot halts after running evms_activate.  The console shows a
bunch of the following messages:

md: array md0 already has disks!
md: array md1 already has disks!

I'll try to get a more detailed log and comparison-- unfortunately I have to
write it all down by hand.

Google shows a lot of confusion about the cause of this problem.  Some people
report that it occurs randomly due to a change in the order of naming the disk
devices.  Others report it when the MD array is degraded.  Neither of these is
true in my case:  it occurs every time with kernel 2.6.29 and never with 2.6.26,
and when I do boot into 2.6.26, EVMS shows that the MD array is operating
normally.

It does seem likely that a change in boot device order is to blame here, but I'm
not sure how to diagnose it when I can't boot.  The /linuxrc script in my initrd
tries to drop me into a shell when it can't mount the root FS, but with 2.6.29
the shell is unresponsive-- no input or output.

Anyone else have this problem?  Suggestions?

Thanks,
Andrew.


------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and 
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today. 
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Evms-devel mailing list
[email protected]
To subscribe/unsubscribe, please visit:
https://lists.sourceforge.net/lists/listinfo/evms-devel