Degraded array not assembled via mdadm.conf?

Barry Tennison barry at ukph.org
Sat Nov 7 14:43:04 UTC 2009


I find mdadm a superb collection of tools - many thanks for all the work!

I'm posting just one thing that puzzles me.  For this issue, I'm running 
squeeze/testing on armv5tel (orion5) and have mdadm version 3.0-3.1

I have a degraded RAID1 array, with one active drive and one failed, removed 
drive.  This "array" assembles and runs fine with the manual command
   mdadm -A -u <correct UUID> /dev/md1
For the time being, I want to continue to use this array in degraded form, and 
have it auto-run at boot time.

However, when I put the any of the following lines in /etc/mdadm/mdadm.conf
  ARRAY /dev/md1 num-devices=2 UUID=<correct UUID>
  ARRAY /dev/md1 num-devices=1 UUID=<correct UUID>
  ARRAY /dev/md1 UUID=<correct UUID>
the array fails to assemble at boot.  The message in syslog is (always)
  mdadm[1336]: DegradedArray event detected on md device /dev/md1

I have got around this by putting the "manual" commands
  mdadm -A -u <correct UUID> /dev/md1
  mount /dev/md1
in rc.local, and now the array assembles, runs and is mounted fine.

I'm not sure whether this is a bug or a feature.  Is there a rationale for the 
difference between the results of attempting the same thing in mdadm.conf and 
"manually" in rc.local?

Thanks again for all the great work.

Barry



More information about the pkg-mdadm-devel mailing list